555
J. Phys. Chem. 1991, 95, 555-558
I1
!I I
3320
This illustrates the utility of FDMR in obtaining structural information which can be used to identify the recombining cation. Diphenylamine was the only compound studied for which such decomposition was important.
I
3340 9360 3380 Magmtlc Fkld (Gaur)
Figure 4. FDMR spectra from carbazole and diphenylamine. (a) Simulated spectrum of diphenylamineradical cation using coupling constants from ref 34 (N, 9.03 G H,10.98 G; 4 H,3.46 G 4 H,1.31 G; 2 H, 4.86 G; 1-G line width (fwhm)). (b) FDMR spectrum from recirculated 5 X IO-' M diphenylamine in 2-propanol (MW(0,200)BC(200,250)ns, microwave magnetic field 0.2 G, nine scans of 30 laser pulses per point, with background subtraction). (c) FDMR spectrum from 1 X lo-* M carbazole in 2-propano1, settings as in (d). (d) Simulated spectrum of carbazole radical cation using coupling constants for 3,bdimethylcarbazole from ref 35 replacing the methyl coupling constant with 3.30 G (N, 6.89 G; H, 8.09 G; 2 H, 3.84 G; 4 H,0.80 G; 2 H, 3.30 G; 1.0-G line width (fwhm)).
the EPR spectrum reported for diphenylamine catiod4 but were similar to the FDMR spectrum we found for carbazole (Figure 4). The photocyclization reactions of diphenylamine and its derivatives are well-known, and indeed analysis by gas chromatography/mass spectroscopy (GCMS) showed almost total conversion of the diphenylamine in our sample into carbazole after long laser irradiation. (34) Neugebauer, F. A,; Bamberger, S.;Groh, W. R. Chem. Ber. 1975, 108,2406. (35) Bamkrger, S.; Hellwinkel, D.;Neugekuer, F. A. Chem. Ber. 1975,
108,2416. (36) Lewis, I. C.; Singer, L. S.J. Chem. Phys. 1%5,13, 2712.
Conclusions In this study we have shown the utility of the FDMR technique in establishing the spin multiplicity of the photoionization precursor state. Our results for diphenylamine illustrate how this technique is able to monitor the identity of the recombining species as a cross-check on our results and demonstrate unequivocally photoionization occurs predominantly from the singlet state. The agreement of the experimental and simulated power dependence for anthracene establishes the mechanism of photoionization as sequential, two-photon absorption, occurring predominantly through the first excited singlet state. Observation of a similar power dependence for TMPD and carbazole and of exclusively singlet FDMR signals for all the aromatic compounds studied suggests that, for the experimental conditions used, this is the predominant photoionization mechanism for such compounds. Because of the biphotonic nature of such processes, it is important to allow for differences in pulse shape, pulse length, and power when comparing results from different studies. Our studies show that the high peak powers available with lasers are likely to favor ionization through singlet rather than triplet states. The observation of FDMR signals from a wide variety of aromatic compounds shows that this technique should be generally applicable. Acknowledgmenr. The authors thank Dr. D. W. Werst, Dr. M. C. Sauer, Jr., and Dr. C. D. Jonah for many helpful discussions. Dr. K. H. Schmidt is gratefully acknowledged for the use of his "Photion" and "Intdep" programs to simulate the laser power dependence. Our thanks also go to Dr.K. Anderson for running the GCMS analysis of diphenylamine and Mr. J. Gregar for crafting our flow cells.
A Coefficients of the alA States of NF and NH Gregory R. Bradburn* and Harvey V. Lilenfeld McDonnell Douglas Research Laboratories, P.O. Box 516, St. Louis, Missouri 63166 (Receiued: May 4, 1990)
The Einstein A coefficients of the paramagnetic species NF(a'A) and NH(a'A) have been measured in the gas phase. The excited statm were detected by both optical emission spectroscopy and electron paramagnetic resonance (EPR) spectroscopy. The optical detection systems were calibrated by using luminescence from the emission of either the air afterglow reaction or 02(alA). The Einstein A coeficient of NF(aIA) was determined to be 0.15 (f0.04) S-I, in good agreement with a previous measurement. This work is believed to be the first experimental determination of the gas-phase Einstein A coefficient of NH(a'A). Our value for this A coefficient is 3.7 (f0.6) s-I. In addition, we believe that this is the first report of the detection of NH(alA) in the gas phase using EPR.
Introduction In recent years there has been much interest in gas-phase reactions with long-lived highly excited products.'-8 One highly ( I ) Arnold, S.J.; Finlayson, N.; Ogryzlo, E. A. J . Chem. Phys. 1966,14,
2529.
(2) Derwent, R. G.; Thrush, B. A. Trans. Faraday Soc. 1971.67, 2036. (3) Clyne, M.A. A.; Connor, J. J . Chem. Soc., Faraday Trans. 2 1972, 68, 1220. (4) Curran, A. H.; MacDonald, R. G.; Stone, A. J.; Thrush, B. A. Proc. R. Soc. London A 1973,332,355. (5) Herbelin. J. M.Chem. Phys. Len. 1976, 12, 367. (6) McDermott, W. E.;Pchelkin, N.R.; Benard, D.J.; Bousek, R. R. Appl. Phys. Len. 1978, 32. 469. (7) Hays, G. N.; Fisk, G. A. Appl. Phys. Len. 1983,12,3. (8) Hovis, F. E.;Whitefield, P.D.;Lilenfeld, H. V.; Bradburn, G. R. J. Phys. Chem. 1988. 92. 5133.
0022-3654/91/2095-0555$02.50/0
excited molecule that can be produced chemically is N2(A3Z:). This molecule is metastable and could potentially be used as a high-energy (6.2 eV) storage molecule for a transfer laser. The H NF2 reaction is expected to produce N2(A3Z:) under conditions of excess H atom^.^,^ In this system production of N2(A3X:) is thought to occur by the following sequence of reactions: H + NF2 HF(u) + NF(X3Z-) (la)
+
-- ++ + +
HF(u)
H ~~
HF(u) NF(a'A)
NF(aIA)
(lb)
NF(b'Z+) H F N(2D)
(IC)
+
(2)
~~
(9) Cheah, C. T.; Clyne, M. A. A.; Whitefield, P. D. J . Chem. Soc., Faraday Trans. 2 1980, 76, 7 1 1.
0 1991 American Chemical Society
Bradburn and Lilenfeld
556 The Journal of Physical Chemistry, Vol. 95, No. 2, 1991
N(2D)
- +
+ NF(a'A) N2(B)
N2(B,W) + F
N,(A)
hu
(4) Early work on this system indicated that the branching ratio of reaction 1 is greater than 90% to produce NF(a1A).5*'0s" This excited state is metastable and, with an appropriate transfer partner, could be used as an energy storage molecule in a transfer laser. Workers at this laboratorye noted that a much lower yield of NF(a'A) from reaction 1 was one way to explain the results of their experiments. The initial reports of large branching ratios to NF(a'A) were based on the inability to detect significant amounts of the other electronic states of N F in the reaction products. In particular, until recently no one had been successful in detecting the ground state, NF(X3Z-), directly; Malins and Setser'l implied its production in small amounts from the presence of HF(o=4). In addition, reaction 1 proceeds through a vibrationally excited intermediate, HNF$,ll*I* and collisional deactivation of the intermediate could produce small amounts of ground-state N F without production of HF(u=4). Heidner and co-workers have recently demonstrated the use of laser-induced fluorescence to monitor NF(X3Z) produced by KrF laser photolysis of NF2" They were also unable to detect NF(X3Z-) in the H + NF2 reaction and concluded that the branching ratio to NF(a'A) > 0.9.14 In the previous work done a t this laboratory the NF(alA) Einstein A coefficient of 0.18 s-I as measured by Setser et al." was used to determine the NF(alA) population. Because our results indicated that we were obtaining NF(a'A) concentrations much lower than those expected on the basis of the branching ratio of reaction 1, it seemed prudent to verify the determination of the lifetime of NF(aIA). We independently measured the Einstein A coefficient of NF(a'A) using a technique (described below) that is well-suited for measuring lifetimes of metastable molecules. Additionally, it has been proposed that an analogous reaction, F N H 2 HF(o) + N H , might yield high densities of the NH(a'A) excited state which could also be a useful energy storage molecule. This reaction would have an advantage over the H + NF2 reaction system in that H atoms would not be present to deactivate the excited state. Because the value of a molecule for energy storage depends on the Einstein A coefficient and because of the similarity of the experimental requirements, we measured the A coefficients of the alA states of both N F and N H .
+
-
Theory The spontaneous emission of radiation is a first-order process as shown in the equations B* B hu (5)
-
+
where B* is an excited state of the species B which emits at frequency u, A is the Einstein A coefficient for B*, 7 is the zerepressure lifetime of B+, I is the emission intensity in arbitrary units, and a is the instrument calibration factor which converts the emission intensity to an absolute photon flux. For species with short radiative lifetimes the Einstein A coefficients can be determined by monitoring the decay in the emission intensity as a function of time. Since the emission intensity decay is dependent on all mechanisms for deactivation of the excited state, several measurements are needed under different conditions to extract the decay rate due to the radiative process a10ne.l~ However, for long-lived excited species this technique does not work well ~~
~~~
~~
~~
~
~
(10) Cheah, C. T.; Clyne, M. A. A. J. Chem. Soc., Faraday Trans. 2 1980,
76, 1543.
R. J.; Setser, D. W. J . fhys. Chem. 1981,85, 1342. (12) Herbelin. J. M.;Cohen, N.Chem. fhys. Lea. 1973,20, 605. (13) Heidner, R. F., 111; Helvajian, H.; Holloway, J. S.; Koffend, J. B. J. fhys. Chem. 1989,93, 7813. (14) Heidner, R. F.,111; Helvajian, H.; Holloway, J. S.; Koffend, J. B.J. fhys. Chem. 1989,93, 7818. (IS) Bradburn, G. R.;Armstrong. R.A.; Davis, S. J. Opt. Eng. 1980, 19, ( 1 I ) Malins,
66.
H, in He
(3)
cavity J no. 2 port Figure 1. Schematic diagram of experimental apparatus for determining Einstein A coefficients of long-lived species. NP4Me
n==wk
"am
Figure 2. Schematic diagram of heated injector for dissociation of N2F4.
because of experimental limitations due to low signal-to-noise ratios, long-term signal instabilities, and quenching rates that are large compared to the radiative rates. We have determined the Einstein A coefficients of NF(a'A) and NH(a'A) by the more direct method of measuring the absolute emission rate, using a calibrated optical detection system, and the absolute concentration of the excited radical using an EPR spectrometer. These values are substituted into eq 6 to obtain the Einstein A coefficient. Because we are able to simultaneously measure the emission rate and the excited-state concentration, this technique does not require correction for quenching processes which may be occurring. Apparatus
A schematic of the experimental apparatus is shown in Figure 1. The apparatus is a fast-flow reaction chamber with optical emission spectroscopy (OB)and electron paramagnetic resonance spectroscopy (EPR) detection equipment. The reaction chamber consists of a Suprasil quartz flow tube (2.25-cm i.d.) with a movable heated quartz injector, a quartz side arm which passes through a microwave discharge cavity, a gas handling system, and a pressure port. The flow tube passes through the center of the EPR cavity and is coated with halocarbon wax up to the point at which it enters the EPR cavity to reduce wall recombination reactions. The chamber is evacuated by a 150 L s-I Stokes Model 212-221 Roots style blower backed by a 140 L s~I Stokes Model 412 Microvac mechanical pump. The pressure in the reaction chamber is adjusted by throttling the flow at a gate valve downstream of the observation ports. A schematic of the heated injector is shown in Figure 2. The heating element is a Glo-Quartz Electric Heater Co., Inc., Model LHP400S quartz heater rod. It is inserted, via an O-ring seal, into a 1.4-cm-0.d. quartz tube with a gas injection port. The other end of this sleeve is closed with six radially symmetric 0.05-cmdiameter holes. The temperature of the tip of the heater was monitored by a thermocouple attached to the heater with a single wrapping of Teflon tape. Because the O-ring seal fails a t high temperature, we restricted the movement of the injector to the 18-cm region closest to the gas input port. This section of the sleeve was coated with Fisher Scientific Co. Fluorolube Grease GR-90 to provide lubrication and to prevent recombination of atomic species on the wall of the injector. To prevent atom recombination on the hotter portions of the injector, the 30-cm nearest the tip was wrapped with Teflon tape. Power was supplied to the heater element through a Variac for temperature control.
A Coefficients of the alA States of N F and N H
The Journal of Physical Chemistry, Vol. 95, No. 2, 1991 557
Atomic (H, F, or 0) and electronically excited molecular (02(a'A)) reactants are produced by passing the appropriate source gases (H2, F2, or 02),in a He carrier, through the microwave discharge plasma generated in an Evenson type cavity connected to an Opthos Instruments, Inc., Model MPG-4M microwave generator (2.45 GHz). The injection port for the effluent of the microwave discharge is located about 55 cm upstream of the center of the EPR cavity. The gas handling system consists of a remotely controlled mixing system for diluting the N2F4with He, Tylan Model FC-260 flow controllers, and an MKS Baratron capacitance manometer for monitoring the gas pressure. The NZF4 was typically diluted to 50% with He at a total pressure of about 50 KPa in 40-L stainless steel accumulators. The mixtures were allowed to stand for at least 2 h before any experiments were performed to allow for complete mixing. The OES system consisted of two separate detection systems. The first system, used for detecting NF(a'A), NH(a'A), HF(3-0), and NO2* emission, consisted of a Jarrell-Ash 0.25-m monochromator with an 1180 groove mm-' grating blazed at 600 nm. Typically we used 250-pm slits with this system which fully resolved the individual rotational lines of the P branch and partially resolved the R branch of the HF(3-0) emission. The detector was a Hamamatsu R636 photomultiplier tube, which is sensitive out to 930 nm, operated at -900 Vdc. An optical fiber was used to collect the signal and bring it to the entrance slit of the monochromator so that optical emission spectra could easily be recorded both upstream and downstream of the EPR activity by using the same optical system. The optical signal was chopped and amplified by means of a lock-in amplifier. The monochromator scan was controlled by a Digital Equipment Corp. Minc 11/23 microcomputer, which also recorded the spectra and performed initial analysis of the data. The second OES detection system, used for HF(3-1) and 02(a1A)emission, was essentially the same as described above with the following modifications: the grating had 590 groove mm-' and was blazed at 1000 nm; the detector was a North Coast EO-8 17 intrinsic Ge detector cooled to liquid nitrogen temperatures and operated at -300 Vdc; the slits were set to 500 pm; the optical train included a filter holder between the fiber optic and the entrance slit so that a neutral density filter could be used to match system sensitivities of the visible and IR systems; the scan was controlled by an internal scanning stepper motor; and the spectra were recorded on a strip chart recorder and then digitized and stored on the computer for analysis. The EPR system consists of a Varian Model 917500-14 E-109 IO-GHz EPR spectrometer with a Varian Model 909807-03 V7400 38-cm magnet assembly capable of obtaining 1.9-T fields with a 5.4-cm gap. The data were recorded with the same microcomputer used to acquire the optical data.
by using the molecular constants of HF as reported in ref 16 and the H F transition moments reported in ref 17. To get the relative sensitivity of the intrinsic Ge optical detection system in the HF(3-1) and 02(a'A) emission bands, we used the relative system response to the output of a General Electric 30A/T24/13 lamp calibrated against NBS reference standards by the Eppley Laboratory, Inc., from 250 to 2600 nm. For the NH(a'A) measurements we used the 0 + N O reaction as our calibration source. This reaction results in broad-band emission which is relatively intense in the region in which the NH(a'A) emits. It has been calibrated by several investigators including workers at this We used a rate constant of 3.8 (f0.9) X lo-'* cm-3 s-l, as determined at this laboratory for the range 778.5-813.5 nmeZ2 Before performing the calibration, it was necessary to show that the excited-state species were present in sufficient concentrations for detection and measurement by EPR. Figure 3 shows an EPR spectrum taken at 9.1 105 GHz of the products of a reaction of D atoms, from a microwave discharge of D2 in He, mixed with NF2, from thermally dissociated NzF4, in He. Above the EPR spectrum are shown the calculated positions of the NF(a'A) EPR lines. A similar spectrum was obtained by using H atoms from a microwave discharge of H2 in He. Our EPR spectrum of NF(a'A) matches the spectrum reported by Curran et al.4 A similar process was carried out with the NH(a'A) Einstein A coefficient. Since the EPR spectrum of this molecule has not been observed before, we calculated the positions of the 24 expected lines. The calculated NH(a'A) spectrum is shown in Figure 4 for a cavity frequency of 9.0935 GHz. Also shown in Figure 4 is the measured EPR spectrum from an F NH3 flame. The 0 N O reaction emission is broad-band and is still relatively intense in the range 778.5-813.5 nm, where NH(a'A) emits. The absolute emission rate for the reaction is proportional to the product [O][NO]. We produced oxygen atoms by a discharge in a mixture of O2in He. The effluent of the discharge was mixed with NO, and the optical intensity was recorded over the region in which the NH(alA) emits. The concentration of 0 atoms was measured by EPR, and the concentration of N O was calculated from the total pressure and partial flows of the gases used. Using these values and the published emission rate constant, we calculated the absolute sensitivity factor for the optical detection system. Next we acquired optical spectra of NH(alA) upstream and downstream of the EPR cavity while recording the NH(alA) EPR spectrum. By calibrating the EPR with known concentrations of the paramagnetic species, NO, and calculating the relative sensitivities of the EPR for N O and NH(aIA) using the method outlined in ref 23, we determined the concentration of NH(aIA). Using this concentration, the average integrated intensity of NH(aIA), and the absolute sensitivity of the optical detection system, we determined the Einstein A coefficient for NH(a'A).
Materials The gases used and their purities were as follows: He, Union Carbide, 99.995%; H2, Matheson UHP, 99.999%; 02,Matheson UHP, 99.98%; NO, Matheson CP, 99.0%; N2F4, Air Products, 95% NzF4 and 5% N2.
Calculation The EPR spectra of N F and N H (ala) were calculated by using an analysis similar to that developed by Curran et al.4 The Hamiltonian consisted of three contributions including a rotation, hyperfine, and Zeeman term. The energies of the M, levels of the J = 2 ('A) ground state for both N F and N H were calculated as a function of field strength in the range from 0.9 to 1.O T, using a Hamiltonian consisting of rotational, hyperfine, and Zeeman interactions. For these calculations we used the matrix elements of ref 4 and Clebsch43ordon coefficients of ref 24. The coupling
Experiment The instrument calibration factor, a in eq 6, is determined by referring to a gas-phase emission with a known emission rate. It is a system-dependent variable, and so great care must be taken to make all measurements with precisely the same optical system in place. For the NF(alA) measurements we used the emission of 02(a1A)as our calibration reference. Since 02(a'A) emits at 1270 nm and NF(a'A) emits around 870 nm, a correction for the relative sensitivities of the detection systems at these two wavelengths was necessary. The u'= 3 to d'= 0 and V" = 1 transitions of H F occur at about 865-920 and 1310-1495 nm, respectively. The HF(3-0) transition has good overlap with the NF(a'A) emission, and the HF(3-1) emission is in a relatively high sensitivity region of the intrinsic Ge detector. The relative emission intensities of the HF(3-0) and the HF(3-1) lines were calculated
+
+
(16) Mann, D. E.; Thrush, B. A,; Lide, D. R., Jr.; Ball, J. J.; Aquista, N. J. Chem. Phys. 1961, 34,420. (17) Oba, D.; Agrawalla, B. S.;Setser, D. W. J. Quant. Specrrosc. Radiat. TransJer 1985, 34, 283. (18) Fontijn, A.; Meyer, C. B.; Schiff, H. I. J . Chem. Phys. 1964, 40, 64. (19) Vanpee, M.; Hill, K. D.; Kineyko, W. R. AIAA J . 1971, 9, 135. (20) Woolsey, G. A.; Lee, P. H.; Slafer, W. D. J. Chem. Phys. 1977, 67, 1220. (21) Sutoh, M.; Morioka, Y.;Nakamura, M.J . Chem. Phys. 1980,72,20. (22) Bradburn, G. R.; Lilenfeld. H. V. J. Phys. Chem. 1988, 92, 5266. (23) Westenberg, A. A. frog. Reacf. Kinef. 1973, 73, 23. (24) Condon, E. U.; Shortley, G. H. The Theory of Aromfc Spectra; Cambridge University Press: London, 1964.
558 The Journal of Physical Chemistry, Vol. 95, No. 2, 1991
1
Bradburn and Lilenfeld
1
I
,
I
0.92
0.94
0.96
0.98
I
1.oo
1
I
1.02
1.04
J
Fielm
Figure 3. Calculated (a) and experimental (b) EPR spectra of NF(a'A) at 9.1 105 GHz.
..r_.._..__.....l
0.958 0.962 0.966 0.970 0.974 0.978 0.982 0.986
Field/T Figure 4. Calculated (a) and experimental (b) EPR spectra of N H (a'AJ-2) at 9.0935 GHz.
constants for N F were obtained from ref 4, and those for N H were obtained from ref 25. Good agreement was obtained when the calculated spectra were compared with the experimental results, thus confirming the identity of the EPR spectra. (25) Leopold, K. R.; Evenson, K. M. J . Chem. Phys. 1986, 85, 324.
Conclusions The A coefficient of NF(a'A) was previously determined by Setser" to be 0.18 s-l. Our measurement of 0.15 (f0.04)s-l is the same within experimental error. The stated uncertainty is the 95% confidence limit on the statistical error in the measurements. Systematic errors could add an additional 40% to the uncertainty. Therefore, we conclude that our NF(a'A) yields8 were far below the measured branching ratio for eq 1. It is possible that this is a result of dissociating the NF2 since we saw some evidence of fluorine atoms in the reaction zone (Le., the H2 did not have to be dissociated to get a reaction when mixed with the NF2 flow). An alternate explanation might involve rapid deactivation of NF(a'A) under the conditions used in ref 8. This work is believed to be the first experimental determination of the A coefficient of NH(a'A). The A coefficient was determined to be 3.7 (f0.6)s-I where the uncertainty is the 95% confidence level on the statistical error. Systematic errors could add an additional 40% to the error estimate.
Acknowledgment. This research was conducted under the McDonnell Douglas Independent Research and Development program. Registry NO. NF, 13967-06.,1;NH,13774-92-0 H, 12385-13-6; NF2, F, 14762-94-8; NH2, 13774-92-0.
3744-07-8;