J . Phys. Chem. 1986, 90, 1806-1811
1806
Hot Hydrogen Atom Reactions Moderated by H, and He S. Aronowitz,* Fairchild Research Center, 4001 Miranda Avenue, Palo Alto, California 94304
T. Scattergood,+ Department of Earth and Space Science, S.U.N.Y. at Stony Brook, Stony Brook, New York 11794
J. Flores, and S. Chang Planetary Biology Branch, NASA-Ames Research Center, Moffett Field, California 94035 (Received: September 11, 1985)
Photolysis experiments were performed on the H2-CD4-NH3and He-CD4-NH3 systems. The photolysis (I849 A) involved only NH3. Mixtures of H2:CD,:NH3 included all combinations of the ratios (200,400,800):( 10,20,40):4. Two He:CD4:NH, mixtures were examined where the ratios equalled the combinations 100:(10,20):4. Abstraction of a D from CD4 by the photolytically produced hot hydrogen from ammonia was monitored by mass spectrometric determination of HD. Both experiment and semiempirical hot-atom theory show that H2 is a very poor thermalizer of hot hydrogens with excess kinetic energy of about 2 eV. Applications of the hard-sphere collision model to the H2-CD4-NH3 system resulted in predicted ratios of net HD production to NH3 decomposition that were two orders of magnitude smaller than the experimental ratios. On the other hand, helium is found to be a very efficient thermalizer; here, the classical model yields reasonable agreement with experiments. Application of a semiempirical hot-atom program gave quantitative agreement with experiment for either system.
1. Introduction The ability to describe quantitatively reactions initiated by atoms or molecular fragments possessing kinetic energy in excess of their thermal environment has direct impact in delineating an extensive set of phenomena. The range of that set covers plasma etching of silicon] to secondary reactions associated with photolysis of ammonia or other photolytically active gases in the overwhelmingly hydrogen-rich atmospheres of Jupiter and other giant planets. In this context, a particularly sensitive application is the determination of the efficiency of molecular hydrogen in thermalizing radicals, principally hydrogen atoms that possess a large excess of kinetic energy. Classically,2 moderately hot (several electronvolts) hydrogen atoms are projected to be thermalized after three or four collisions with molecular hydrogen. However, the classical picture of a system of hydrogen interacting with its sisters might be substantially incorrect due to quantum effects; as an elementary example, there will always exist a region in any collision between the hot hydrogen atom (H) and H, where it is impossible to discern conceptually which hydrogen of the triplet is really the hot one. In turn, because quantum effects are expected to be important, this system is a critical test of the usefulness of the semiempirical hot-atom approach, detailed in a previous comm ~ n i c a t i o nthat , ~ melded hot-atom reactions with ordinary kinetics. Experimentally, the problem is how to monitor the thermalization of hot H in excess H,. A suitable abstraction process involving the hot H appears to be the best monitor, though it is only an indirect measure of the competing hot H thermalization process. Our desire to minimize introduction of isotopic species and restrict photolysis to NH,, in order to make our experimental simulations as realistic as feasible, render tritium radioassay gas chromatographic technique^^*^ inappropriate. Unfortunately, analyses relying on conventional gas chromatographic techniques are found to be inadequate. An ingenious alternative analytical approach was developed some time ago by Martin and Willard.6 In their work, the quantity of H D formed by photolyzing hydrogen bromide in mixtures of (HBr CD4) and (DBr CH4) was assayed by mass spectrometry. Since the HD, in their experiments, could be formed only by the abstraction of a hydrogen from methane, their results were a direct, sensitive measure of successful
+
+
‘Present address: NASA-Amcs Research Center, Planetary Biology Branch, Moffett Field, CA 94035.
0022-3654/86/2090-1806$01.50/0
abstraction. Application of semiempirical hot-atom theory to Martin and Willard’s experiments resulted in quantitative agreement with their quantum yields and also demonstrated that even in this simple two-component system a hard-sphere elastic collision modeling of the system would not yield agreement with e~periment.~ One process that is relatively clean and straightforward while offering insight into questions of secondary reactions attributable to hot-atom chemistry in planetary atmospheres is the photolysis (at 1849 A) of ammonia with methane present. The hot hydrogen produced has an excess kinetic energy (quantity above 3 / 2 k T )of about 2 eV. In our experiments NH, was chosen as the source of hot hydrogens with CD4 providing the abstraction vehicle. The experimental design and precautions initiated to ensure that only ammonia is photolyzed are described in the following section. Since ammonia is the sole source of hot H, then any €ID produced must be from abstraction of CD4, exclusively. Since extremely sensitive measurements of H D can be made, the approach permits examination of abstraction processes occurring in the presence of a quantity of molecular hydrogen that exceeds all other species by two orders of magnitude. The same experimental protocol is also used in experiments where helium replaces hydrogen. The differences between helium and hydrogen as potential moderators are extensive. Helium is spherically symmetric and has no vibrational or rotational motion. Moreover, it is distinguishable from a hydrogen atom at any separation on an atomic scale. Therefore, helium might be expected to behave more as a classical thermalizer for a hot H atom than would molecular hydrogen. Furthermore, the helium system serves as another test of the flexibility of the semiempirical hot-atom theory. 11. Experimental Section Initial experiments showed that reasonable photolytic conversion ( E 20%) of N H 3 resulted in less than 1% of the initial CH, to (1) Winters, H. F.; Coburn, J. W.; Chuang, T. J., J . C’ac. Sci Technol. B 1983, I , 469. (2) Libby, W. F. J . A m . Chem. Sor. 1947, 69, 2523. (3) Aronowitz, S.; Chang, S.; Scattergood, T. J . Phys. Chem. 1981, 85, 360. (4) Wolfgang, R.; Rowland, F. S . Anal. Chem. 1958, 30, 903. (5) Lee, J. K.; Lee, E. K. C.; Musgrave, B.; Tang, Y.-N.; Root, J. W.; Rowland, F. S . Anal. Chem. 1962, 34, 741. (6) Martin, R . M.; Willard, J . E . J . Chem. Phys. 1964, 40, 3007.
0 1986 American Chemical Society
The Journal of Physical Chemistry, Vol. 90, No. 9, 1986 1807
Hot Hydrogen Atom Reactions
A -Sample Bulb B -Liquid N2 Tmpr, C -Toeppler Rlmp D -3-Woy Stopcock E -Molecular Sieve F -Gloss Bead Trap$ G -Collsclion Bulb @ stopcock^
TC -Thermocouple Pressure Gouger --Go$ Flow
Figure 1. System for the separation and collection of hydrogen gases. be decomposed. This rendered accurate gas chromatographic determination of the change in CH, very difficult, sometimes unmeasurable, despite the use of a high-precision fixed-loop sampling system. Also problems were encountered with an unknown hydrocarbon contamination that added to the uncertainty of the measured changes in methane. These problems could be circumvented by the use of deuterium-labeled methane, CD,. Reaction of CD, with a hot H from photolysis of NH, yields the labeled product HD. This species can be measured very accurately in small amounts by stable isotope ratio mass spectrometry. Interference from system sources of hydrogen thus was eliminated. Photolyses of various mixtures of Hz/CD4/NH3 were carried out in a straightforward manner using Spectmsil quartz photolysis cells. The cells (sample and control) were cylinders 10 cm long by 2 cm in diameter affixed with greaseless O-ring stopcocks. The various mixtures were prepared in a mercury free vacuum system which contained a bulb of known volume for each component to be included. This system was also used in preparing He/CD4/ NH, mixtures. After filling each hulh to a predetermined pressure of the respective component, we allowed the gases to mix wemight, resulting in mixtures good to *2% in mole fraction for each of the components. Both photolysis cells were then filled simultaneously to provide an unphotolyzed control and a photolyzed sample for each experiment. Tests using pure hydrogen containing a known amount of HD showed no isotope fractionation during the cell-filling procedure. The gases used in this study were Mathewn research grade H,, 99.999% pure He, electronic grade NH,, and Merke, Sharp and Dohme 99 atom % pure CD,. The photolyses were carried out using a microwavedriven (2.45 GHz) low-pressure mercury discharge lamp. Initial experiments showed that some photolysis of CD, took place. Examination of the spectral output of the lamp revealed the presence of significant flux at 1650 A even though a spare piece of the quartz used for the cell windows was used as a filter. To eliminate this line, and to give clean blanks with pure methane, we placed a 2 cm long oxygen filter filled to about 1 atm between the lamp and the photolysis cell. The sample mixtures were then exposed to 1849-R light for various times to give about 15% decomposition of the ammonia. The output of the lamp at 1849 & . was monitored by decomposition of pure NH, at 4 mmHg and was found to average (1.2 i 0.2) X loi5hvls over this study. All measurements of the ammonia were made using a Cary I 4 spectrophotometer. After each photolysis was completed the control and test samples were each divided into four aliquots hy simultaneous expansion into evacuated bulbs. In those experiments involving helium, a measured quantity of hydrogen (of known D/H ratio) was introduad. The following procedure was then used to separate the hydrogen (H, + HD) or hydrogen and helium (H, He HD) to be analyzed from other product and reactant gases. The separation system is shown schematically in Figure 1. Each of the sample bulbs [labeled A in Figure I ] was attached to the system and the gas mixture was allowed to expand into two U-traps (B), cooled with liquid nitrogen, to freeze out the ammonia
+ +
and other condensable gases. The noncondensable gases, primarily H , He, HD, and C D , were then transferred via a Toeppler pump (C) through a three-way stopcock (D) into a section containing a W r a p (E) filled with 250 mg of molecular sieve 13X. Protection of the molecular sieve from mercury poisoning and other condensable contaminants was accomplished by placing traps (F) containing glass beads on both sides of the molecular sieve trap. These traps were cooled in liquid nitrogen before opening the U-trap containing molecular sieve (E) to the system. The traps (F) were maintained in liquid nitrogen throughout the separation. After the U-trap (E) was cooled in liquid nitrogen, the sample was cycled through the sieve to remove the CD,. Upon removal of the methane, the stopcock (D) was turned to the collection bulb (G)and the hydrogen was collected via the Toeppler pump (C). All transfers were monitored by the use of thermocouple pressure gauges [labeled T C in Figure I]. Suitable tests with samples of H, of known HD composition with and without added CD, confirmed the validity of the procedure; for our purposes the isotope fractionation was within tolerable limits (these limits are quantified in the next paragraph). The ratios of HD ( m l e 3) to Hz ( m / e 2) for each of the eight samples (four controls and four samples) were determined by stable hydrogen isotope ratio mass spectrometry. Calibration of the Varian MAT 6D-150 mass spectrometer was accomplished by measurements of an internal standard that was checked frequently against two reference standards, SMOW (standard mean Ocean water) and GISP (Greenland ice sheet precipitation). The raw values for HD/H, for the samples are referenced to SMOW, for which the HD/H, value is well-known. Using the customary notation, we can express the measured values for HD/H, as 6 %, (parts per thousand) where
The ratio' (D/HJsMow = %(HD/Hz)s~ow= 1.58 X lo4 The net HD/H2, and hence the extent of the hot H atom abstraction, is the difference between the HD/H, ratios for the photolyzed sample and unphotolyzed control. This difference A6
A(HD/Hz) = 2(Lmpie- ~,.,,,)(D/H)sMow/~OOO
Testson samples of known HD/H, ratios that have been processed according to the procedure described in this section yield an overall precision of 2 ~ 8 %for the technique. Due to the nature of the experiments and the novelty of the technique used to monitor the hot hydrogen abstraction, numerous controls were run to elevuate possible system effects. Mixtures doped with oxygen (from residual air) and water vapor (from the walls of the vacuum system and cells) were photolyzed and showed, respectively, no effect and a slight inhibitory effect on the production of H D as compared to similar experiments without the added contaminants. The effect of the presence of walls (surface effects) on the experimental results was investigated by a series of photolyses in cells (the same as used previously) which contained fired quartz wool with a surface area 23 times that of the cell itself. No changes in the production of HD greater than 20% relative to the appropriate experiments with the empty cell were found. Thus we conclude that any surface effects introduced by the presence of the cell wall were negligible for the series of experiments examined in this article. Finally, replicate experiments were done for three of the HJCD,/NH, mixtures in order to ensure reproducibility of the results. The largest disparity in A6 found for the three sets was 7.9% [which is within our overall analytical error @%)I. 111. Resulk A. H2-CD4-NH3 System. The experimental results for the Hz-CD4-NH, system are presented in Table I. Hydrogen (H,) and methane (CD,) partial pressures were stepwise varied by
1808
The Journal of Physical Chemistry, Vol. 90, No. 9, I986
TABLE I: Experimental Results with the H,-CD,-NH,
Aronowitz et al.
System
200/10/4 400/10/4 800/ l 0 / 4
13.5 18.6 14.6
33.9 [13.3; 5.81 29.8 [13.7: 7.4]* 15.1 [12.9; 9.0]*
3.93 5.06 7.12
200/20/4 400/20/4 800/20/4
15.2 14.7 15.2
88.0 [14.3; 6.01 56.3 [ f 3 . 1 ; 9.6]* 54.3 [12.4; 4.91
9.1 1 12.6 22.6
200/40/4 400/40/4 800/40/4
19.3 16.6 14.9
163.1 [ f 1 . 4 , 2.41 104.9 [ & 6 . 5 ; 12.91 89.4 [12.5; 4.61
13.7 20.0 37.9
Partial pressures in mmHg. 'Percent decomposition of NH,. Results marked by * are averages of replicate experiments whose results were within 1 7 % of the listed data. dStandard deviation of four determinations of H/D ratio. PRatio of net HD formed to net NH, decomposed.
, 400
O200
factors of two from initial pressures of 200 and 10 mmHg, respectively. Ammonia (NH,) was held constant at 4 mmHg; this pressure was a lower bound in satisfying the requirements for negligible errors in pressure and optical measurement while maintaining the highest possible ratio of H, or CD4 to NH,. Restricting the decomposition of ammonia to about 15% ensured a high precision in the optical measurements while limiting the contributions from photolysis of secondary products. The fourth column in Table I lists the experimentally determined AHD:ANH, ratios; it shows that increasing H, pressure yields increasing quantities of HD produced per NH, loss. This result would be paradoxical if molecular hydrogen were as efficient a thermalizer of hot hydrogen (H*) radicals as predicted by classical estimates; if molecular hydrogen behaved classically then an increase in the H2 ratio would result in a decrease in the quantity of the HD produced. This observed increase is found for all partial pressures of CD4. When the ratio, AHD/ANH3, is corrected for the photolysis time, however, the new function decreases with increasing H 2 pressure [see Figure 21. That is, less ammonia is lost per unit time as the H2 pressure is increased. The low-pressure mercury lamp used for the photolysis was monitored over the complete set of experiments; its flux varied less than 20% over the entire project and varied considerably less (
