J . Phys. Chem. 1989, 93, 7818-7823
7818
a disproportionation rate constant, k , , of 5 X cm3/(molecule-s) lies considerably below the experimental data while the model result with kl = 2 X lo-', cm3/(molecule.s) is in reasonable agreement. Consideration of the data in Figure I O leads us to a conservative estimate that the NF(X3Z-) disproportionation rate coefficient is 0.90 i 0.03, in accord with previous results. The measured NF(a'A) quantum yield from KrF photolysis of NF,, @'NF(a) < 0.04 f 0.02, also agrees with prior work. The NF, photolysis quantum yield is observed to be unity, based on results from an HCI titration of the F atoms from the 249-nm photolysis of NF,.
Introduction The rapid reaction of hydrogen atoms with the NF2 radical H + NF, NF(X,a,b) HF(u54) k298= 1.3 X IO-" cm3/(molecule.s) (ref 5 and 6) ( I )
-
+
is among the very few that produce electronically excited products in high yield.'-7 The production of the NF(alA) radical has been ( I ) Herbelin. J . M.; Cohen, N. Chem. Phys. Lett. 1973, 20, 605. (2) Clyne, M. A. A.; White, I . F. Chem. Phys. Lett. 1970, 6, 465. (3) King, D. L.; Setser, D. W. Annu. Reu. Phys. Chem. 1976, 27, 407.
0022-3654/89/2093-7818$01 .50/0
cited as the dominant channeL6v7 It has been ~ u g g e s t e d lthat -~ this reaction proceeds via an addition-elimination mechanism forming HNF, as the intermediate. Since the triplet surface leading to the entrance channel is expected to be repulsive, the reaction occurs via the HNFt 'A surface, and spin considerations (4) Herbelin, J. M. Chem. Phys. Lett. 1976, 42, 367. (5) Cheah, C. T.: Clyne, M. A. A,; Whitefield, P. D. J . Chem. Soc., Faraday Trans. 2 1980, 76, 71 1. ( 6 ) Cheah, C. T.; Clyne, M. A. A. J . Chem. SOC., Faraday Trans. 2 1980, 76, 1543. (7) Malins, R. J.; Setser, D. W. J . Phys. Chem. 1981, 85, 1342.
0 1989 American Chemical Society
Determination of the H
+ NF2 Branching Ratio
The Journal of Physical Chemistry, Vol. 93, No. 23, 1989 7819 TEFLON MIX TANK
COMPUTER e-
HBr TRANSFER LINE
XCARICOMPUTER Ch 3
CELL PRESSURE AND/% THERMOCOUPLE (inlernal & ex1 temp)
Ch 1 NF(b) SIGNAL Ch 2 NF(b1 BACKGROUND
LENS STACK AND
Ch 3 Nd YAG
-
REF DIODE
Ch 5 EXCIMER ENERGY
CELL PRESSURE
TO TRANSIENT OIGITIZERICOMPUTER Ch 4
Figure I . Diagram of the
experimental apparatus.
preclude formation of triplet NF. The abstraction reaction, however, favors the production of NF(X3Z-).4 Several indirect measurements of the NF(a' A) branching ratio have been performed. Cheah et a1.6 monitored time profiles of N atoms produced from the H / N F 2 system and, using spin conservation arguments, conclude that the NF(a' A) branching fraction is in excess of 90%. Their results stem from the fact that N(4S),produced primarily from the reaction of H and NF(X32-), was not observed at early times during the reaction. Malins and Setser' also report a large NF(a'A) branching ratio of 91%. Their results are based upon relative emission intensity measurements of HF(v) and NF(a). They also examined the vibrational distributions of the N F and H F products, and their RRKM analysis supports an addition-elimination reaction mechanism. More recently,* however, absolute photometry measurements were performed on the H NF, flame in a flow tube reactor. They determined that the NF(a'A) yield, under hydrogen-rich conditions, is only 15-30% of the initial NF, density. We have performed a direct determination of the NF(alA) branching fraction using a technique that does not depend upon absolute detectivity calibrations. Relative amplitudes of both NF(a'A) and NF(X3Z-) produced from the 249-nm photolysis of NF, and from the H NF2 reaction have been measured by use of emission spectroscopy and laser-induced fluorescence (LIF), respectively. The H NF, reaction was studied by the ArF photolysis of HBr to create H atoms in the presence of NF,. A knowledge of the NF, absorption cross section at 249 nm and the HBr cross section at 193 nm allows the extraction of the absolute NF(a'A) yield from the experimental data, since the experimental geometry was held fixed throughout the experiment. In addition, any loss of NF2 during its transport through the gas-handling system to the observation region does not affect our results. Such a loss of NF,, possibly due to wall assisted NF, disproportionation, may explain the low NF(a'A) yield reported in ref 8.
+
+
+
Experimental Section NF2 and HBr were photolyzed by use of a Lumonics 400 Hyperex excimer laser operating with KrF (249 nm) or ArF (193 (8) Hovis, F. E.; Whitefield, P. D.; Lilenfeld, H. V.; Bradburn, G. R. J . Phys. Chem. 1988, 92, 5133.
Ch 6 Nd YAG
MASS FLOW METERS
nm), respectively. The excimer beam was directed into the photolysis cell through a series of beam-shaping lenses. Two slit apertures were placed in the beam path to minimize scattered excimer light. A Quantel Datachrome 581C Nd:YAG pumped dye laser system operating with Coumarin 500 dye provided tunable radiation to probe the N F b-X transition near 530 nm. Baffles on the dye laser entrance arm of the cell helped reduce scattered dye laser light. Uncoated Suprasil windows were used to provide small fractions of the dye and excimer beams for relative laser energy measurements, obtained from a silicon photodiode and a Laser Precision RJP-734 energy meter, respectively. The relative response of the energy meter for the KrF and ArF wavelengths was performed in situ using a large area (5-in. diameter) calorimetric power meter that had been previously calibrated. The exit window of the photolysis cell was removed, and the large power meter sampled the entire excimer beam for these measurements. The energy meter was absolutely calibrated in the same manner for the HCI titration experiments described below. A Lasertechnics Model lOOF Fizeau wave meter served to determined the dye laser wavelength. A diagram of the experimental arrangement is shown in Figure 1. The gases in these experiments used without further purification were CO, (Matheson 99.99%), Ar (Matheson 99.99%), SF6 (MG Gases 99.9%), and N,F, (Hercules 96%). HCI (Matheson 99.0%) and HBr (Matheson 99.8%) were first cooled to liquid nitrogen temperatures and pumped to remove any free H,. They were then vacuum distilled to remove any Br2 or C1,. The purified HBr or HCI was mixed with Ar and stored at 1000 Torr in a Teflon-coated reservoir. Mixtures of N2F4 in argon were prepared and stored at 100 psia in a stainless steel reservoir. Reagents were flowed into the photolysis cell through calibrated Tylan flowmeters. The cell, constructed of stainless steel and internally Teflon coated, was wrapped in heating tape and maintained at 145 OC. At this temperature, the N2F4is >95% dissociated into NFl9-I3 for the N F 2 densities used in these experiments. Cell pressure was measured with an MKS Baratron capacitance manometer. (9) Johnson, F. A.; Colburn, C . B. J . Am. Chem. SOC.1961, 83, 3043. (10) Modica, A. P.; Hornig, D. F. J . Chem. Phys. 1968, 49, 629. (11) Foner, S. N.; Hudson, R. L. J . Chem. Phys. 1973, 58, 581. (12) Evans, P. J.; Tschuikow-Roux, E. J . Chem. Phys. 1976, 65, 4202. (13) Evans, P. J.; Tschuikow-Roux, E. J . Phys. Chem. 1978, 82, 182.
7820 The Journal of Physical Chemistry, Vol. 93, No. 23, 1989
NF(a) was monitored with a GaAs photomultiplier tube viewing the N F a-X emission through a I nm fwhm interference filter centered at 874.3 nm. For the F atom titration experiments, emission of the H F 2-0 overtone was observed by using a LN, cooled intrinsic Ge detector through a 1270-nm interference filter (30 nm fwhm). The GaAs or Ge detector output was amplified and processed with a LeCroy Model 2264 transient digitizer whose output was averaged and stored in a DEC 11/73 computer. The N F b-X LIF was detected with a gatable EM1 9816QB photomultiplier tube mounted on one arm of the photolysis cell. A 1 nm fwhm filter centered at 528.9 nm was used to isolate part of the N F b-X (0,O) band near the strong QP/QR head. Since the strong N F b-X vibronic bands are diagonal in vibrational quantum n ~ m b e r ’ ~(900 - ’ ~ = 0.96),14 it was essential to gate the photomultiplier off during the dye laser probe pulse to eliminate scattered dye laser radiation. We observed a signal from the N F b-X phototube in the absence of the dye laser. This background is primarily due to cell window fluorescence caused by scattered excimer light in addition to N F b-X chemiluminescence from the NF2 photolysisI7and from ~ b-X signal is ultimately traceable the H NF, r e a ~ t i o n .The to NF(alA) HF(v) E-V energy pooling. The excimer laser was operated at a repetition rate of I O Hz while the dye laser was pulsed at 5 Hz. The dye laser was synchronized to fire upon every other excimer laser shot, and dual channel detection was employed in order to perform background subtraction. The PMT output was sent into a Tektronix AM-502 amplifier and processed by using two channels of an S R S SR250 boxcar integrator. One channel is triggered at 5 Hz, in phase with the dye laser, and acquires the N F b-X LIF signal. The second boxcar, also triggered at 5 Hz but out of phase with the dye laser, obtains the signal due to the excimer laser alone. The two boxcar outputs were digitized and recorded with a DEC 11/73 laboratory computer. The relative dye laser and excimer laser energy measurements were also acquired by using the computer. A third boxcar was used for the dye laser while the analog output from the energy meter was used directly. NF(X3Z-) time profiles were generated by scanning the delay between the excimer laser photolysis and the dye laser probe pulses while the NF(alh) and HF(o=2) time behavior was obtained directly by using the transient digitizer. Throughout this study, N F b-X LIF was excited by tuning the dye laser to the QP(9)/QR(9) line of the (0,O) band at 18 909.29 cm-I. The pulsed wave meter was used to adjust the laser’s frequency to line center. I n all cases, final adjustment was made by maximizing the N F b-X LIF signal in real time. I n the case of the ArF photolysis of HBr/NFz mixtures where no LIF was detected, the ArF laser was blocked and 266-nm radiation was directed into the cell to create NF(X3Z-) from NF2 photolysis for peaking up the LIF signal prior to each run. The 266-nm radiation was produced by frequency doubling the excess 532-nm light from the YAG laser system. The NF(X3Z-) quantum yield is nearly 99% at 266 nm18 and near the peak of the absorption band at 260 nm7.38-20Extreme care was taken to ensure that the detection geometry was held constant for the experiments which investigated relative NF(a’A) and NF(X3Z-) yields from the KrF photolysis of NF, and those which probed the branching ratio from the ArF photolysis of HBr/NF, mixtures.
+
+
Results and Discussion A. NF2 Photolysis Quantum Yield. The first absorption band Under 0.25-cm-’ of NF, at 260 nm exhibits diffuse (14) Tennyson, P. H.; Fontijn, A.; Clyne. M. A. A. Chem. Phys. 1981.62,
171.
( 1 5 ) Douglas, A . E.; Jones, W. E. Can. J . Phys. 1966, 4 4 , 2251. (16) Vervloet. M.; Watson, J . K. G . Can. J . Phys. 1986, 64, 1529. (17) Koffend, J . B.; Gardner, C. E.; Heidner, R . F. J . Chem. Phys. 1985, 83, 2904. (18) Heidner, R. F.; Helvajian, H.: Koffend, J. B. J . Chem. Phys. 1987, 87, 1520. (19) Goodfriend, P. L.; Woods, H. P. J . Mol. Spectrosc. 1964, 13, 63. (20) Kuznetsova, L. A,; Kuzyakov, Y . Y.; Tatevskii, V. M. Opt. Speclrosc. 1964, 16, 295.
Heidner et al. 30
1 1
I
I
!o
4
’
I [
I
I
I I I
6’
I
1
I
lb
112
i 4!
moleculesrcm3i
[HCI]
Figure 2. Results of HCI titration of F atoms from KrF laser photolysis of NF,.‘The NF, density was held constant a t 5.0 X lOI5 molecules/cm3, and the NF, photolysis fraction was 6 f 1%. The vertical lines indicate the calculated titration point of 5.7 X loi4molecules/cm3 using an NF(X) photolysis quantum yield of 90% (see text).
resolutionI8 no additional structure is observed and the features are probably indicative of the excited NF2 electronic surface and dissociation process, analogous to O3 photolysis in the Hartley band.*I Photolysis of NF2 in its 260-nm absorption band leads primarily to NF(X).17,18The NF(alA) photolysis quantum yield has been reported to be 0.10 f 0.0517 for the KrF laser photolysis of NF,. The wavelength dependence of the NF(a) quantum yield has also been investigated,18and the yield is observed to decrease monotonically from a value of 0.17 at 244 nm to 0.01 at 267 nm. Energetic considerations dictate that the only other possible photolysis product is NF(X).’2*’5 Previous work in this l a b ~ r a t o r y ”has ~ ~shown ~ that following U V NF, photolysis, the appearance of the NF(a’A) fragment occurs on a time scale of -80 p s . In contrast, the NF(X3Z-) photofragment is observed to be produced promptly,22 behavior more in accord with continuum dissociation. In light of these data, however, an argument could be made for a short-lived NF2 excited state playing a role in its dissociation and the question is raised whether the NF2 photolysis yield is unity, an assumption inherent in the data analysis presented below. In order to measure the photolysis yield, an HCI titration of the F atoms from the KrF photolysis of NF, was undertaken. The following reactions must be considered for the titration: NF, hu(249 nm) @photo(%F(a)NF(a) + ( 1 - %,,,))NF(X) + F) (2)
-
+
F k298 = 1.6
+ HC1 X
-
HF(v) + CI IO-” cm3/(molecule.s) (ref 23)
NF(X3Z-)
-
+ NF(X3Z-)
NF(X3Z-) + NF,
Nz
+ 2F
NzF2 + F
(3) (4) (5)
where is the NF, photolysis yield and @.NF(a) is the fraction of dissociation events yielding NF(a). We see from (4) and (5) that any NF(X3Z-) formed results in an addtional F atom created either from bimolecular disproporti~nation~-~ or by reaction with NF,.,, Any direct reaction of NF(alA) or electronic quenching resulting in F atom production can be neglected under our experimental condition^.'^,^^ Hence, each photon absored by NF, results in the formation of aPhoto( 1 ( 1 - @NF(a))) fluorine atoms. In a series of experiments, NF,/HCI/Ar mixtures were photolyzed by using a KrF laser. The total pressure and NF, density were held constant and the HC1 partial pressure was varied while time profiles of HF(u=2) were recorded using the H F (2-0) overtone emission near 1270 nm. The HF(u=2) time profiles were fit to a rising and falling double exponential to obtain amplitudes.
+
(21) Imre, D. G.; Kinsey, .J. L.; Field, R. W.; Kayatama, D. H. 7th International Symposium on Gas Kinetics, Goltingen, Springer Verlag: Berlin,
1982. (22) Heidner, R. F.; Helvajian, H.; Holloway, J. S.; Koffend. J . B. J . Phys. Chem.. preceding paper in this issue. (23) Clyne, M. A . A,; Nip, W. S. Int. J . Chem. Kinet. 1978, IO, 367. ( 2 4 ) Cha, H.; Setser, D. W. J . Phys. Chem. 1987, 91, 3658.
Determination of the H
+ NF2 Branching Ratio
The Journal of Physical Chemistry, Vol. 93, No. 23, 1989 7821
Figure 2 shows a set of HF(u=2) emission amplitudes plotted as a function of HCI density. Also shown is the calculated titration point with @ hD = 1 and the NF, absorption cross section reported in ref 18. h n c e reaction 3 produces H F having u 5 2?5 the titration is not complicated by cascading from higher H F vibrational levels. Titration curves were generated from HF(uz2) amplitudes so that the effect of increased HF(u=2) quenching as the HCI density is increased is included. The major source of the estimated error in the calculation is due to the uncertainty in determination of the KrF beam area. However, there is good agreement between the experiment and the calculation. We repeated the titration using a different NF, density with similar results. We conclude, within an estimated error of 20%, that NF, dissociates with unit quantum efficiency at 249 nm. B. NF(a) Branching Fraction. The reaction of H atoms with NF2 yields an N F molecule with an H F m ~ l e c u l e . ~The -~ N F product is formed in either the X3Z-, alh, or the b'Z state. Malins and Setser7 report NF(X):NF(a):NF(b) branching fractions of 0.07:0.9 I :0.02. Their value of 2% for the NF(b) state yield is based upon the 200-ms N F b-X radiative lifetime reported by Clyne and White., Subsequent work hasZbZ*showed that the b-X radiative lifetime is a factor of 10 shorter. Thus, the NF(b) branching fraction is lowered to 0.2% and it can be effectively neglected. Our ability to probe both the NF(X) and NF(a) states enables us to perform a direct determination of the H NF, reaction branching ratio without the need for absolute calibration. The N F b-X LIF and N F a-X emission signal amplitudes, normalized for KrF laser flux, from the photolysis of NF2 are given by Sa-x = 1249Fa-~@a6249[NFzl (6)
KIF ENERGY (arb units) 141
I
1
I
I
I
I
I
1
+
and (7) = 1249Fb-X(1 - @a)a249[NF21 where Fa-xand Fbx are constants containing detection efficiency, geometries, and appropriate conversion factors, @a is the NF(a) photolysis quantum yield, 6249 is the NF2 absorption cross section at 249 nm, 1249 is the KrF photolysis flux in photons/cm2, and [NF,] is the NF2 number density. Similar expressions can be written for the signal amplitudes from ArF photolysis of HBr in the presence of NF,: Sb-X
S'a-x = 1193F'a-~Bab193[HBrl
(8)
S',X = 1193F'b-X(1 - Ba)r193[HBrl
(9)
and where F\-x and PbX are similar to the constants Fa-xand Fbx appearing in eq 6 and 7, B, is the NF(a) branching ratio, u193is the HBr absorption cross section at 193 nm, is the ArF photolysis flux in photons/cm2, and [HBr] is the HBr number density. Equations 6 and 7 are valid under the assumption that the NF, photolysis quantum yield is unity, discussed in the previous section. Equations 8 and 9 assume that each H atom reacts with N F2 and that the HBr photolysis quantum yield is unity. At our typical operating temperatures of 420 K, the rate constant for the cm3/(molereaction of H atoms with HBr (k420 = 5.9 X c u l e . ~ ) ) ~is~ more * ~ ' than a factor of 2 slower than that for H + NF27( k = 1.30 X lo-" ~ m ~ / ( m o l e c u l e . s ) )Furthermore, .~~~ for the work reported here, the HBr/NF2 ratios were kept at values so that at least 80% of the H atoms react with NF,. This point is discussed further below. Equations 6-9 state that the observed signals are proportional to the photolyis laser flux. In order to ensure that this indeed is (25) Kirsch, L. J.; Polyanyi, J . C. J . Chem. Phys. 1972, 57, 4498. (26) Kwok, M. A.; Herbelin, J. M.; Cohen, N. EIectronic Transition Lasers I; Steinfeld, J., Ed.; MIT Press: Cambridge, MA, 1976. (27) Tennyson, P. H.; Fontijn, A.; Clyne, M. A. A. Chem. Phys. 1981.62, 171. (28) Lin, D.; Setser, D. W. J . Phys. Chem. 1985, 89, 1561. (29) Cha, H.; Setser, D. W. J . Phys. Chem. 1989, 93, 235. (30) Endo, H.; Glass, G. P. J . Phys. Chem. 1976, 80, 1519. 0 1 ) Baulch, D. L.; Duxbury, J.; Grant, S. J.; Montague, D. C. J . Phys. Chem. Re/. Dura 1981, IO.
DYE LASER ENERGY (arb
units)
Figure 3. Plots of the N F b-X LIF intensity as a function of laser flux. The dye laser was tuned to excite the QP(9)/QR(9) line of the (0,O) band, and the delay between the KrF photolysis laser and dye laser was fixed at 100 ps. The upper figure shows the dependence of the LIF intensity upon the KrF laser flux while the lower plot shows the variation with dye laser flux.
the case, NF(a) emission and N F b-X LIF were recorded as a function of laser flux in separate experiments. Figure 3 displays a plot of the N F b-X LIF intensity as a function of KrF flux. Linear behavior is observed and the plot extrapolates through the origin. The same result was obtained for the NF(a) emission for both the KrF and ArF cases. Since we used N F b-X LIF as an NF(X) probe, we also investigated the LIF signal as a function of dye laser energy at constant [NF(X)], to demonstrate linearity. Figure 3 also contains a plot of the N F b-X LIF intensity versus dye laser energy. Again, passage through the origin and linear behavior is observed. Emission signals of N F b-X LIF and N F a-X emission from KrF laser photolysis of NF, and A r F photolysis of HBr/NF2 mixtures were recorded as a function of [NF,] and [HBr], respectively. The b-X (0,O) QP(9)/QR(9)line was probed with the dye laser. The addition of 5 Torr of SF6 ensured that the NF(X) state vibrational and rotational distributions were relaxed.,, In addition, 5 Torr of CO, was also present to quench vibrationally * ~ ~effectively excited H F produced in the H NF2 r e a ~ t i o n . ~This prevents any energy pooling'J8 between HF(v) NF(a) to produce NF(b) which would contribute to the background signal. The NF(a) emission and N F b-X LIF time profiles were normalized for excimer laser flux. In addition, the b-X LIF signal was also normalized for the dye laser energy to correct for dye laser power fluctuations during the course of the measurements. The experimental NF(a) and N F b-X time profiles were fit to single-exponential decays, and amplitudes were obtained from an extrapolation to time t = 0. Plots for NF(a) amplitudes are shown in Figure 4. Figure 5 contains a plot of the N F b-X LIF amplitude as a function of NF, density. In all cases the plots are linear in accord with eq 6-9 above. We note further that the NF(a) amplitude plot from the HBr photolysis passes through the origin within experimental error, indicating that the source of N F is via reaction of H atoms with NF, and that NF2 is not appreciably photolyzed by 193-nm radiation. However, we were not able to observe any N F b-X LIF signal from the ArF pho-
+
(32) Bott, J . F. J . Chem. Phys. 1976, 65, 4239.
+
7822
Heidner et al.
The Journal of Physical Chemistry, Vol. 93, No. 23, I989
x
,
Oo2t -0 02 5287 8
[NF,]
04
02
[NF,]
06
08
52882
52684
52686
52888
5289
IVAVELENGTH [Vacuum Angslroms)
m~lecules~cm~i
Figure 4. Upper figure: Plot of NF(a) amplitude from the 193-nm photolysis of HBr in the presence of NF,. The NF2 concentration was held constant at 3.9 X I O i 5 mo~ecu~es/cm3 with s F 6 = 1.7 X I O i 7 molecules/cm3, Ar = 2.0 X and CO, = 1.1 X I O l 7 molecules/cm3. Also shown is a linear fit to the data. Lower figure: Plot of NF(a) amplitude from the 249-nm photolysis of NF,. Buffer gases densities were SF6 = 1.7 X l O I 7 molecules/cm3, Ar = 2.0 X and CO, = 1 . 1 X I O l 7 molecules/cm3. Also shown is a linear fit to the data.
"2
5286
10
molecules cm31
Figure 5. Plot of N F b-X LIF amplitude as a function of [NF,] from the KrF laser photolysis of NF,. Other species densities were SF, = 1.8 X I O l 7 molecules/cm3, Ar = 2.0 X and C 0 2 = 1 . 1 X lo1' molecules/cm3. A linear fit to the data is indicated.
tolysis of H B r / N F, mixtures. Two different methods were used to search for an N F b-X signal, and the results are displayed in Figure 6. Also shown for comparison is the LIF observed from the KrF laser photolysis of NF, alone. In one experiment the dye laser was tuned to excite the (0,O) QP(9)/QR(9)line and the delay between the ArF and dye lasers was scanned, while in another the time delay between the two lasers was held fixed and the wavelength of the dye laser was varied. No b-X LIF could be detected in either case. If no change is made in the excitation and detection geometries when signals are measured for both KrF and ArF laser photolysis, then F'&, = FbX and Fa-,= Fa-xin eq 6-9 and one obtains
Figure 6. Upper trace: Time profiles of N F b-X LIF intensity obtained by tuning the dye laser to the QP(9)/QR(9)line of the (0,O) band and varying the delay between the excimer laser and dye laser pulses. The large signal is from the KrF laser photolysis of NF2 while the other trace is from the ArF photolysis of an HBr/NF2 mixture. Experimental conditions for the KrF case had NF, = 1.2 X I O t 6 molecules/cm3, SF6 = 1.0 X 10I7 molecules/cm3, Ar = 6.2 X 10l6,and CO, = 1.3 X l O I 7 molecules/cm3. The ArF data had HBr = 3.3 X I O l 5 molecules/cm3, NF, = 7.7 X 1OI6 molecules/cm3, SF, = 8.2 X 10I6molecules/cm3, Ar = 8.0 X IOi6, and C 0 2 = 1.0 X IOI7 molecules/cm3. Lower trace: N F b-X excitation spectra obtained by fixing the delay between the excimer and dye lasers to 15 p s and scanning the dye laser wavelength. The spectrum is a portion of the b-X (0,O) band and the strong QP/QR branch is labeled with the N"rotationa1 quantum number. The larger signal is from the KrF laser photolysis of NF2 while the other is from the ArF photolysis of an HBr/NF, mixture. Conditions for the KrF case were NF, = 2.8 X molecules/cm3, SF6 = 1.8 X lOI7 molecules/cm3, Ar = 8.7 X 1 0I6,and CO, = 1.2 X IOI7 molecules/cm3. The ArF data had HBr = 1.9 X 10'' molecules/cm3, NF, = 3.9 X l o i 5molecules/cm3, SF6 = I .7 X IO" molecules/cm3, Ar = 3.4 X I 0l6, and C 0 2 = 1 . 1 X IO1' molecules/cm3.
(S&X/I24~)/(S'~X/~,~3), respectively. Solving for B, and obtains Ba = (1 @a
-
Rbp)/b(Ra
= ( 1 - RbP)/(I
-
one
Rb)]
(12)
Rb/Ra)
(13)
-
where p = u193/u249is the ratio of the HBr absorption cross section at 193 nm and the NF, absorption cross sections at 249 nm. Equations 12 and 13 allow the extraction of the N F product branching fractions and photolyis quantum yields from experimental measurements of NF(a) and N F b-X signal amplitudes, excimer laser flux, and absorption cross sections. Since the strong LIF occurs at the same wavelength as the dye laser, it was necessary to use a gated detector (see Experimental Section), and a fixed delay, At = 50 ps, is introduced between the time the dye laser excites the LIF and the time the LIF is detected. Therefore, the NF(b) will decay somewhat during this time interval between its production by laser pumping and its detection. The magnitude of this decay is given by ek'*', where k'is the NF(b) removal rate. One would expect that this factor is very close to unity for 249-nm photolysis of NF, and for the ArF laser photolysis of HBr and NF,, since NF(b) quenching rate constants are very small for the species p r e ~ e n t . ~ *This l ~ ~ expectation was confirmed in the KrF laser experiments. However, under our operating conditions during the ArF laser studies, the
Determination of the H
+ NF,
Branching Ratio
The Journal of Physical Chemistry, Vol. 93, No. 23, 1989 1823
TABLE I: Experimental N F b-X LIF and a-X Emission Amplitudes versus Density"
process NF, + hu(249 nm) NF(a) + F NF, + hu(249 nm) NF(X) + F NF, NF,
+H +H
--
-
-
+ FC NF(X) + FC NF(a)
NF(a) branching fraction' NF(a) 249-nm photolysis quantum yieldC
1
0
50
I
I
I
experimental valuesb run no. 2 9.5 (1.8) x 10-32
run no. 1 2.1 (0.4) x 10-31
6.2 (1.2) x 10-33
3.8 (0.8) x 10-33
8.6 (1.7) X 0.90 f 0.03 0.91 f 0.03
