2975
J . Phys. Chem. 1991,95. 2975-2982
in the fluorescence lifetime is ascribed to the retardation of the rate of intersystem crossing to the nearby triplet m a n i f ~ l d , ' ~ * ~ ~ i.e., a decrease in k,,(P) with pressure. In order to calculate the true pressure dependence of k,,(P) the change in 7, with pressure must be taken into account by using the correction given by eq 5, since, as was discussed in the previous section, T , is not constant but depends on pressure as a result of changes in the refractive index. Equation 7 can then be rewritten as follows:
19
knr(P) = 1 / ~ f i P-) 1/Tr(P)
18
= 1/~1(P)- n2(P)/n2(P=0.1)T,(P=O.I)
J
(8)
The values for k,,(P) which were calculated without correcting for the dependence of T,(P)on pressure is denoted k::(P), whereas the values corrected for changes in the refractive index are denoted k z ( P ) . These values are plotted in Figure 8. The two curves deviate significantly from each other at higher pressures and, consequently, the apparent activation volumes, AV',,, calculated by using eq 9, Le., the slopes of the two curves given in Figure
E
4
17
RT 16
0
200
400
600
Pressure (MPa) Fipre 8. Pressure dependence of the nonradiative rate constant of MEA in MCH at 25 O C (a) after and (b) before correcting for the change in the refractive index with pressure by using eq 8. The solid lines represent the best-fit quadratic polynomials.26
7b are much broader than those of DMEA, though both compounds exhibit similarly structured sharp absorption spectra as is seen from a comparison of Figure 5a and Figure 7a. An Important Consequence ofthe n2 Dependence of 7,. Since the fluorescence lifetime is more often affected by nonradiative processes whose overall rate constant is given by k,,, 7fiP) is more generally written as follows for many other anthracene derivatives: 7dP) = 1 / [ 1 / ~ r+ knr(P)) (7) In many cases the fluorescence lifetime increases as the pressure is increased. For example, as has been reported previou~ly,~~ the fluorescence lifetime of 9-methylanthracene (MEA) is 5.0 ns at 0.1 MPa and it increases gradually with increasing pressure and appears to level off at a value of 10.2 ns at 700 MPa. This increase
a In k,,/aP = -AVnr
(9)
8, also differ significantly. The corrected and uncorrected activation volumes calculated at 300 MPa are 9.1 f 0.5 and 6.4 f 0.7 cm3 mol-', respectively. This correction for changes in refractive index given by eq 8 would also be necessary when activation energies for radiationless processes are calculated from Tf measured a t different temperatures, since the refractive index of a medium also changes with temperature.20 In conclusion, it has been shown clearly in this work that the natural radiative lifetime is influenced by changes in the refractive index of the medium in the form of n2,but hardly by the factor given by eq 6 . Since the n2 dependence given in eq 1 can be used to predict satisfactorily the radiative lifetime, this relation should be of great use in many cases where experimental values for fluorescence quantum yields are difficult to obtain. Acknowledgment. This work is partly supported by Grantsin-Aid 01470010 and 02245102 to S.H. and 6254033 to M.O. Registry No. DCNA, 1217-45-4; DMEA, 781-43-1; DMEDA, 2395-97-3; DNA, 1210-12-4; DPHA, 1499-10-1; MEA, 779-02-2. ( 2 5 ) Tanaka, F.; Okamoto, M.; Yamashita, S.; Teranishi, H. Chem. Phys. Lett. 1986, 123, 295. (261 Isaacs. N. S . Liauid Phase Hiah Pressure Chemistry. _ . Wilev: . ChiChester, U.K., 1981; p 183.
-
Collision-Induced Intersystem Crossing in NH(a'A)
-
NH(X3Z-)
J. Stephen A d a d and Louise Pasternack* Chemistry Division, Code 61 10, Naval Research Laboratory, Washington, D.C. 20375-5000 (Received: August 24, 1990; I n Final Form: November 9, 1990) Laser-induced fluorescence (LIF) of individual rovibronic states of gas-phase NH(a'A) and NH(X31T) is used to study intersystem crossing (isc) induced by collisions with Nz, CO, Xe, and Oz. Hydrazoic acid, HN3, is photolyzed at 266 nm to produce NH(a'A). The time-dependent behavior of NH(a'A) in its ground and excited vibrational levels is probed by LIF. Appearance rates for NH(X3Z-) in several vibrational levels are also measured. The vibrational level distribution of NH(X%) is measured and compared to prior distribution calculations in order to predict the position of the potential energy curve crossing for the collision-induced isc. The results of these measurements on the mechanism of NH(a'A) decay are discussed in light of recent theoretical calculations and by comparison to isc occurring in the isoelectronic O(ID).
Introduction The study of electronically excited species is of considerable interest to both the experimentalist and theoretician. Various electronic relaxation processes have been extensively studied Over *NRL/NRC Postdoctoral Research Associate.
the past few years, particularly the electronic relaxation of small molecules in which spin-orbit coupling is h " a n t . When these molecules are in the collision-free regime, decay may occur via a radiative transition; however when the molecular system is perturbed by collisions, nonradiative processes may occur. Our interest is in the study of collision-induced intersystem crossing
0022-3654/91/2095-2975%02.50/0 0 1991 American Chemical Society
2976
The Journal of Physical Chemistry, Vol. 95, No. 8,1991
(isc) where collisions alter the spin multiplicity of the molecule. Production of NH(a'A) and its subsequent reactions and collision-induced intersystem crossing have been the subject of many experimental and theoretical studies. In particular, much attention has been focused on the processes that can lead to ~ s c ' - ~ of NH(a'A) to NH(X3Z-). Although early studies found little electronic quenching of NH(a'A),12 with rates less than cm3 s-I for most quencher gases, more recent studies have provided examples of collisions with N2,+ in which the isc of NH(a'A) is on the order of 8 X cm3 s-' and collisions with Xe4q5and C03*597 in which the isc of NH(a'A) is on the order of one-tenth gas kinetic. For NH(a'A) O2quenching, recent studiesp5 give cm3s-I. rates on the order of 6 X Although early studies reported that the NH(a'A) produced from 248- and 266-nm laser photodissociation of HN3 was predominantly in u = 0,4,8-'0Kenner et al." have measured a u = 1 to u = 0 ratio of 0.95 f 0.30 following 248-nm photolysis of HN3. Both Nelson and McDonaldI2 and Hack', have also measured high ratios of NH(alA) u = 1, 2, and 3 to u = 0 following 266-nm photolysis of HN,. The isc of NH(a'A,u=O) by N2 has been the subject of detailed experimental and theoretical studies in the past two years. This quenching takes place via a long-lived HN, complex. The singlet NH(a)-N2 potential surface is crossed by the energetically lower triplet NH(X)-N2 surface. CASSCF calculations by Alexander et al.I4 predict a barrier on the singlet exit channel for HN, NH(alA) + N,. Experimental measurements using both IR multiphoton dissociation of DN3I5and the temperature dependence of the rate constant for NH(a'A) + N2 quenching5 have confirmed the presence of a barrier. The quenching of NH(alA,u=O) by CO has been measured by several investigators to be quite rapid,,Jv7 with rates about 200 times faster than quenching by N2, which is isoelectronic. In addition, no barrier was ~bserved.~JAlthough the qualitative picture is similar to NH(a'A) isc by N2, the greater energetic stability of HNCO compared to H N 3 is believed to reduce the long-range barrier.5 Detailed theoretical calculations have recently been performed on this system and predict a small barrier (-450 cm-').I6 Recently, Hack and Rathmann7 have measured NCO as a reactive product of NH(aIA) + CO. They estimated that the reactive pathway is dominant; however, they also observed a significant yield of the NH(X3Z-) isc product. For NH(a'A,u=O) Xe, there are no reactive channels; consequently, only decay through spin-forbidden isc is energetically allowed. The rate of NH(a'A,u=O) depletion induced by collisions with Xe has been measured to be 1.1 X lo-" cm3 This is more than 3 orders of magnitude faster than isc of NH(alA) by Kr and more than 4 orders of magnitude faster than isc by Ar or He.2 Several studies of the NH(a'A) + O2 system have been in disagreement. Early e x p e r i m e r ~ t son ~ , ~the ~ rate of NH(X3Z-)
+
-
+
(1) Rohrer, F.; Stuhl, F. Chem. Phys. Lett. 1984, 111, 234. (2) COX,J. W.; Nelson, H. H.; McDonald, J. R. Chem. Phys. 1985, 96, 175. (3) Freitag, F.; Rohrer, F.; Stuhl, F. J . Phys. Chem. 1989, 93, 3170. (4) Hack, W.; Wilms, A. J . Phys. Chem. 1989, 93, 3540. (5) Nelson, H. H.; McDonald, J. R.: Alexander, M. H. J . Phys. Chem. 1990, 91, 3291. (6) Bower, R. D.; Jacoby, M. T.; Blauer, J. A. J . Chem. Phys. 1987,86, 1954. (7) Hack, W.; Rathmann, K. J . Phys. Chem. 1990, 94, 3636. (8) McDonald, J. R.; Miller, R. J.; Baronavski, A. P. Chem. Phys. Left. 1977, 51, 57. (9) Baronavski, A. P.; Miller, R. G.; McDonald, J. R. Chem. Phys. 1978, 30, 119. (IO) DeKoven, B. M.; Baronavski, A. P. Chem. Phys. Leu. 1982,86,393. (1 1) Kenner, R. D.; Rohrer, F.: Stuhl, F. J . Chem. Phys. 1987,86,2036. (12) Nelson, H. H.; McDonald, J. R. J . Chem. Phys. 1990, 93, 8777. (13) Hack, W.; Mill, Th. J . Phys. Chem., in press. (14) Alexander, M. H.; Werner, H.-J.; Dagdigian, P. J. J . Chem. Phys. 1988, 89, 1388. (15) Stephenson, J . C.; Casassa, M. P.; King, D. S.J . Chem. Phys. 1988, 89, 1378. (16) Alexander, M., unpublished results.
Adams and Pasternack production using techniques similar to those described in this paper reported rates of 1.6 X lo-" 6331, s-I; however, a study of the phosphorescence by Freitag et aL3 quenching of NH(a'A-X%) reported a rate of 6.2 X cm3 s-'. In the latter study, the total removal rate of NH(a'A,u=O) and NH(a'A,u=l) was measured, although no difference in rate was observed when only the NH(alA,u=O) was monitored. Hack and Wilms4 measured the NH(alA,u=O) disappearance rate using laser-induced fluorescence (LIF) on the NH(cIn-a'A) 0-0band and measured a rate of depletion of NH(a'A,u=O) of 4.5 X cm3 s-I. They also monitored the NH(X3B-,u=O) production and observed the same rate as for NH(a'A,u=O) depletion. Nelson et ale5have also measured the rate of NH(a'A,u=O) depletion using LIF and measured a rate of 5.7 X cm3 s-I. In this paper, we report measurements of NH(alA,u= 1,2) disappearance rates and NH(X3Z-,u=0,1,2) appearance rates that confirm and extend to higher vibrational levels the slower rate constant measurements, although we observe an increase in rate constants for the higher vibrational levels for both the disappearance of the alA state and the appearance of the X3Z- state. We are unable to determine the cause of the discrepancy with the rate constants reported by Drozdoski et but can only conclude that their results are in error.I8 An incorrect interpretation of insufficient data on NH(a'A) O2 quenching reported by Cox et aL2 most likely explains their erroneous quenching rate. However, the remainder of the results reported by Cox et aLz on the temperature dependence of reactions of NH(alA) with HN,, H2, and saturated and unsaturated hydrocarbons were done correctly and are in good agreement with subsequent experiments in other l a b o r a t ~ r i e s . ~ ~ ~ ~ ' ~ There is also some uncertainty in the literature as to the products of NH(alA) quenching by 02.Although Drozdoski et al.I7 suggested the spin-allowed channel to NH(X3Z-) + 02(a1$) products as most likely due to energy resonances, Freitag et al., observed 02(b1B+J as a major product using phosphorescence studies. Although exothermic reaction channels exist, Hack and Wilms4 measure the depletion of NH(alA) due to quenching to be the major channel. In this paper, we report the vibrational level dependence of the rates of both NH(a'A) depletion and NH(X3Z-) appearance and disappearance for isc by collisions with NZ,CO, Xe, and 02.We also measure the vibrational level distribution of the NH(X3Z-) products. The implications of these results on the mechanism of NH(alA) isc are discussed both in terms of the recent theoretical calculations14 on NH(a'A) N 2 and by comparison to the isoelectronic O(lD) and CH2(%'Al)isc.
+
+
Experimental Section Our studies involve the quenching of the N H radical from its lowest singlet electronic state, a' A, to its triplet ground state, X3Z-, following collisions with 02,N2, CO, and Xe gases. The NH(a'A) radicals are generated by photolysis of hydrazoic acid (HN,) using 266-nm pulses from a frequency-quadrupled Nd:YAG laser (Quantel Model YG581C) operating at a repetition rate of 10 Hz. The fourth harmonic light was separated from the YAG fundamental and second harmonic by means of dichroic mirrors. An iris diaphragm apertured the beam to 6-mm diameter, with a pulse energy ranging from 10 to 40 mJ. Following each excitation pulse, an excimer-pumped dye laser system was fired after a computer-controlled delay in order to generate laser-induced fluorescence signals from the NH(c'II+a'A) 1-1 and 1-2 transitions and the NH(A311+X3Z-) 0-0,1-1, and 2-2 transitions. The excimer-pumped dye laser system consisted of a Lumonics HyperEx (420 Series) operating at 308 nm (XeCI), which pumped a Lambda Physik dye laser (FL 2002E). The dye laser was operated in its eighth order, having a bandwidth of ca. 0.2 cm-I. p-Terphenyl dissolved in dioxane was the laser dye employed for much of this experiment, permitting the efficient probing of the 0, 1-1, 2-2 transitions of the X3Z-state and the 1-1 transitions (17) Drozdoski, W. S.; Baronavski, A. P.; McDonald, J. R. Chem. Phys. L e r r . 1979, 64, 421.
(18) Nelson, H. H.; McDonald, J. R., private communication.
Intersystem Crossing in NH(a'A)
+
NH(X3Z-)
of the a'A state. The u = 2 level of the alA state was studied by using the 1-2 transition for which a BBQ dye solution was used. Dye laser pulse energy ranged from 1 to 10 pJ for the spectral scans (to avoid saturation of rotational transitions) and 10 to IO00 pJ for the kinetics scans. The dye beam diameter was approximately 2 mm. The photolysis and probe lasers collinearly counterpropagate through a 5-cm-diameter stainless steel cross equipped with long glass sidearms. A toluene solution filter was placed within the dye laser in order to prevent dye solution degradation from the 266-nm photolysis laser. The fluorescence emission from N H is collected at right angles to the laser beam axis. This uncollimated emission is separated from scattered laser light by appropriate filters and detected by an RCA 31000 photomultiplier tube (PMT). The PMT output is processed by a gated integrator (SRS Model 235), digitized, and stored by a laboratory microcomputer. For spectral scans, a digital delay generator (SRS Model DG535) is used to fix a constant delay between the firing of the photolysis and probe lasers. The resulting fluorescence intensity is divided shot-by-shot by the probe laser intensity (as monitored by an RCA IP28 PMT) in order to correct for fluctuations in the probe laser energy. For the kinetics scans, a signal background is taken 100 ps before the firing of the photolysis laser. The kinetic profiles are collected after a set minimum delay (ranging from 0.25 to 5 ps) after the firing of the photolysis laser. The delay before data acquisition ensures that scattered photolysis light and prompt emission of photoproducts will have insignificant effects on the kinetic data. Under these conditions, the NH(a'A) and NH(X3Z-) radicals are rotationally relaxed prior to the kinetics scans. A typical N H kinetic profile contains 500 points, each of which represent an average of 5-20 laser shots. Pseudo-first-order conditions were obtained by maintaining a substantial pressure difference between the N H precursor and the collision partner, typically at least an order of magnitude. NH(aiA) has a decay that is well fit by a single-exponential function in the presence of an added reagent or in its absence. The appearance and decay of N H in its triplet ground state (X3F) was well fit by the difference of two exponentials in most cases; however, some of the slower N H ( X 3 Z ) decays did require the sum of two exponential functions to achieve an accurate fit. Decays were typically fit over 4-6 reaction lifetimes. The N H precursor, HN3, was synthesized from sodium azide and molten stearic acid according to a standard literature procedure.I9 The product H N 3 was stored in a 10-L glass bulb at 200 Torr of total pressure. A 1:lOO mixture of HN, and He was made in a high-pressure steel tank for use in this experiment. The HN3/He mixture was further diluted with additional He and a reactant gas, allowed to mix, and then slowly flowed into the reaction cell. Typical experiments were run at 20 Torr of total pressure. Gas flows were measured by Tylan gas flow meters or controllers. Typical flow rates ranged from 100 to 200 m.Most experiments were performed at a total pressure of 20 Torr, as measured by a capacitance manometer. H e (Air Products industrial grade, 99.998%, N, (Air Products industrial grade, 99.995%), O2 (Matheson ultrahigh purity, 99.98%) and Xe (Spectra Gases research grade, 99.999%) are used as received; however, the C O (Matheson CP, 99.5%) was passed through a liquid N 2 trap prior to use. All experiments were performed a t 296 f 2 K.
Results A . Rate Constant Measurements. We have measured the rate constants for formation of NH(X3Z-) in u = 0, 1, and 2 from collision-induced intersystem crossing (isc) of NH(a'A) by N2, CO, Xe, and 02.We have also obtained the disappearance rates of NH(a'A,u=l) with these partners and the disappearance rate of NH(aiA,u=2) with 02.First-order rates for the appearance and disappearance of NH(X3Z-) are measured over a range of (19) Krakow, B.;Lord, R. C.; Neely, G. 0.J. Mol. Specirosc. 1968,27, 148.
The Journal of Physical Chemistry, Vol. 95, No. 8, 1991 2977
250.0
0.0
4000.0
time (ccd Figure 1. Plot of the rise and decay of the NH(X'E-,u=O) fluorescence intensity due to collisions with O2 P ( 0 , ) = 2.8 Torr, P(HN3) = 0.005 Torr, and P(He) = 17.2 Torr. The solid line is the nonlinear least-squares fit to the equation Y(r) = A(exp(-kjt) - exp(-k?)) where k)and k', are the pseudo-first-order rate constants for the decay and the rise of the signal, respectively. Measurements are made at 293 K.
25
-
- 20 -z -
CI
I
I
I
1
I
1
,
a v=2 A V=o
I
v1 m
c
15-
10
Y
5-
0
- )
I
2978 The Journal of Physical Chemistry, Vol. 95, No. 8. 1991 12
Adams and Pasternack
11: Observed Ratios (flu) of NH(X3C) in = &1:2 Dw to 2 TABLE Collision-Induced Intersystem Crossing of NH(alA) by Ob Nh CO, I
and Xe U
collision partner N2
co
O) compared to the rate of decay of NH(alA,u=O) induced by collisions with CO. For the ground triplet state, the disappearance rate of the u = 2 level, 4.8 X cm3 s-I, was measured to be at least 2 orders of magnitude faster than the disappearance of the u = 0 or u = 1 levels, implying a slow rate of vibrational energy transfer. In contrast to the isc of NH(alA) by N2, isc by CO generates significant vibrational excitation in the N H ( X 3 Z ) product. We determine an initial excess energy of 6000 f 2000 cm-' using the method described for NH(a'A) isc by collisions with N2, assuming the initial population ratio in the vibrational levels of NH(a'A)I2 is unchanged by vibrational relaxation. The large error limit is primarily due to the uncertainty in the initial population ratio of NH(a1A).12 This result suggests that the singlet-triplet curve crossing is considerably lower with respect to the NH(a'A) + CO asymptote than for isc by N2. Our results are in qualitative agreement with recent theoretical calculations by Alexander16 that predict the crossing to be 8500 cm-' below the NH(alA) C O asymptote. We also observe similar rates for the formation of NH(X3Z) u = 0, 1, and 2, which is consistent with complex formation followed by a statistical redistribution of energy. D. NH(alA) + Xe. Rapid isc of NH(alA) by Xe has also recently been ~ b s e r v e d . ~No - ~activation energy was observed for this process.5 Hack and Wilms4 suggest that an NH-Xe exciplex is formed, comparable to the isoelectronic isc of O('D) induced by collisions with Xe. We observe an identical rate for N H (a' A,u>O) disappearance compared to NH(a' A,o=O) disappearance due to isc by collisions with Xe, consistent with vibrational energy-transfer rates being significantly less than isc rates. These
+
+
+
+
+
+
(55) Yarkony, D. R. J . Chem. Phys. 1990, 92, 320. (56) Spiglanin, T. A,; Chandler, D. W. Chem. P h p . Lett. 1987, 141, 428.
Intersystem Crossing in NH(alA)
-
The Journal of Physical Chemistry, Vol. 95, No. 8, 1991 2981
NH(X3Z-)
results are analogous to the results for NH(a'A) isc by CO. Additionally, no deviations from exponential decay of NH(a'A,u=O) have been ob~erved.~' We also observe no vibrational energy transfer in the ground triplet state. We observe identical rates for formation of NH(X3Z-) in vibrational levels u = 0, 1, and 2 (see Table I), which may be interpreted as evidence for a mechanism involving formation of a complex. The vibrational distribution in NH(X3Z-), as shown in Table 11, implies an excess initial energy of 3000 f 1500 cm-I. This energy is equivalent to the position of the curve crossing with respect to the NH(a'A) Xe asymptote using the model described in this paper for the isc of NH(a'A) by collisions with N2, and assuming no vibrational relaxation N H (a' A). E . NH(alA) 02.An activation barrier for NH(alA) 0, quenching has previously been measured by Nelson et ai? This barrier is believed to be in the entrance channel for complex formation, analogous to the barrier in the isc of NH(a'A) by N2 collisions. Freitag et aL3 have observed 02(b'Z+J as a major product of this reaction with a reaction yield estimated at near unity, They measured a rise of 02(b1Z+J that was faster than the fall of N H ( a ' A p 0 ) by approximately a factor of 3. We have measured a disappearance rate for NH(a'A,u>O) that is significantly faster than the disappearance rate for NH(a' A,v=O) (see Table I), which may account for this discrepancy, particularly since relative large populations for NH(a'A) in higher vibrational levels have been measured.12J3 The spin-allowed isc channel for NH(a'A,u=O) 02(X3Z-,) NH(X3Z-,u=O) + 02(b'Z+,,u=O) is endothermic by 430 cm-I. The more rapidly quenched higher vibrational levels of NH(a'A) would be required to populate the higher vibrational levels of NH(X3Z-), consistent with our observations. This increase in rate of decay for NH(a'A,u>O) compared to the decay of NH(a'A,u=O) induced by collisions with 02,could also be attributed to vibrational relaxation similar to the case of collisions of NH(a'A) with N,; however, this possibility seems to be less likely since no rise was observed in the signal of NH(a'A,u=O) in the presence of OPS1We also observe an increase in the rate of NH(X3Z-) depletion with increasing vibrational energy due to collisions with 02. Since exothermic reaction pathways are present in this system, we cannot separate the effects of vibrational energy transfer from enhanced reactivity of vibrationally excited N H . A second spin-allowed isc channel for NH(alA) with 02(X3Z-,) to form NH(X3Z-,) and 02(a'$) is possible, although there have been no studies of this reaction. In the isoelectronic O('D) quenching by 02,production of 02(a'A ) has been found to be much less important than production of O2(b1Z+,).4* For NH(a'A) + 02,several exothermic reactive channels are possible to yield N O + OH, NO2 H, and H N O 0. Hack and Wilms4 measured the formation of O H and set an upper limit of 6% for its yield. The other possible reactive channels have not been measured. Additionally, Hack and Wilms4 determined that the ratio of isc of NH(alA) to total depletion of NH(alA) is 10.6. They obtained this result by comparison of the intensity of the NH(X3Z-) produced by O2 to that produced by Xe isc where reactive channels are unavailable. Although there are also reactive channels available to NH(a'A) O2collisions, our data can best be interpreted by a mechanism in which the major quenching pathway for NH(a'A) depletion is via a long-lived complex that dissociates to O2(b'Z+,) as Freitag et aL3suggest. From a prior distribution calculation using a value for the excess energy determined from the difference between the NH(a'A) 02(X3Z3)and NH(X3Z-) 02(b'2+.J asymptotes, weighted by the relative vibrational level populations of NH(a'A) determined by Nelson and McDonaldi2 assuming vibrational relaxation of the NH(a'A) is unimportant, we calculate a vibrational population ratio for NH(X3Z-) that agrees with our experimental ratio to within experimental error. In addition, the different rates of appearance of the NH(X3Z-) in the different vibrational levels can qualitatively be explained by faster disappearance rates of the higher vibrational levels of NH(alA), which are energetically required to populate the higher vibrational levels of NH(X3Z-). If the more exothermic channel to 02(a'A,)
+
+
+
+
-
+
+
+
+
+
formation was the major pathway, we would expect NH(X3Z-) to be produced with significantly greater vibrational excitation than we observe, However, due to the uncertainties in the experimental vibrational distribution measurements, we cannot eliminate the possibility of some contribution to the vibrational level distribution of NH(X3Z-) from quenching by O2to produce 02(a1Ag). On the basis of our vibrational population data and with assumption of the validity of this model, we estimate the contribution of this channel to the total quenching of NH(alA) at