JOURNAL OF CHEMICAL PHYSICS VOLUME 111, NUMBER 6 8 AUGUST 1999 Reaction of state-selected ammonia ions with methane Michael A. Everest, John C. Poutsma, Jonathan E. Flad, and Richard N. Zarea) Department of Chemistry, Stanford University, Stanford, California 94305 ͑Received 19 April 1999; accepted 18 May 1999͒ We have investigated the reaction of ammonia ions with methane molecules (CD4) over the ␯ collision energy range of 0.5–10.0 eV and for ammonia ion vibrational states ranging from 2 ϭ ϩ ϩ 1 – 10. Under these conditions, the two main product channels are NH3D and CD3 . The cross ϩ section for formation of NH3D is enhanced with increasing internal energy at collision energies below 6.0 eV, and independent of internal energy at higher collision energies. The enhancement is greater for forward-scattered products indicating that ammonia-ion vibrational energy enhances ϩ reactivity at large impact parameters. The mechanism for formation of CD3 involves ͑ ϩ͒ collision-induced dissociation of CD4 or NH3 which leads to the formation of a short-lived ϩ ϩ ͓NH3CD3͔ (͓NH2CD4͔ ) complex, which then decays to products. This reaction is found not to be vibrationally mode selective, which is consistent with the hypothesis that mode selectivity in reactions of ammonia ions is driven by the Franck–Condon overlap whenever charge transfer is involved. © 1999 American Institute of Physics. ͓S0021-9606͑99͒02130-3͔ I. INTRODUCTION pect they were products resulting from multiple collisions, as their experiments were done at much higher pressures than We have studied the effects of different types of energy the present work. on the reactivity of ammonia ions with methane molecules to Equation ͑1͒ shows the energetics of selected product understand better the dynamics of ion-molecule reactions. channels Information regarding the microscopic dynamics of the reac- ϩ Ϫ tion is deduced by investigating the effects of collision en- NH4 CH3 0.84 eV ϩϩ ϩ ergy, ion internal energy, and ion vibrational state on reac- CH3 NH3 H 4.24 eV tion efficiency, product branching, and product scattering. ϩ ϩ NH ϩCH CH NH ϩH Ϫ0.31 eV . ͑1͒ The ammonia ion/methane system is of interest for sev- 3 4˜ 3 3 ϩ Ά CH NH ϩH ϩH 1.08 eV eral reasons. First, the reaction system has only two heavy 2 2 2 atoms, so that it serves to test the best contemporary theories ϩϩ CH4 NH3 2.54 eV of ion-molecule reaction dynamics. Second, the ammonia ion ⌬ 7 is one of the few that can be prepared in a wide variety of The energies reported are Hrxn at 298 K. Note that these specific internal states owing to the nature of the ionization reactions and energetics involve only hydrogen atoms, scheme employed.1 Finally, methane is isoelectronic both whereas the experiments discussed below involve hydroge- with neutral ammonia, previously observed to react in a nated ammonia and deuterated methane. mode-selective manner,2 and with neutral water, previously observed not to react mode selectively.3 The ammonia ion/ II. EXPERIMENTAL METHOD methane system therefore is a test case for understanding the necessary and sufficient criteria of mode selectivity in ion- The experimental method employed is described more molecule reactions. fully elsewhere;8,9 only a brief description is given here. Am- This system has not been investigated in detail in the monia ions are prepared in selected vibrational states with past. The thermal rate constant for the reaction of (2ϩ1) resonance-enhanced multiphoton ionization1 ϩϩ ϩϩ NH3 CD4 NH3D CD3 was measured by Huntress ͑REMPI͒ of ammonia in a pulsed molecular beam. This 4,5 ˜ Ϫ Ϫ Ϫ et al., and was found to be 3.9ϫ10 10 cm3 mol 1 s 1. scheme is used to prepare ammonia ions with 0 or 1 quanta ␯ ͑ ͒ They observed no other product channels from this reaction. in the 1 symmetric stretching ‘‘breathing’’ mode and ␯ The only state-selected study to date is that of Conaway, 0–10 quanta of vibrational energy in the 2 symmetric bend- 6 Ebata, and Zare. They measured the reactivity of ing ͑‘‘umbrella’’͒ mode.10,11 A particular vibrational prepa- ϩ ϩ m n ␯ NH3 CH4 as a function of both collision energy and am- ration is described as 1 2 for m quanta in the 1 mode and ␯ n monia 2 bending. The observed ionic products include the n quanta in the ␯ mode. If mϭ0 only 2 may be specified. ϩ ϩ 2 following: NH4 from direct atom abstraction, CH3 , The ions are extracted orthogonally from the molecular beam ϩ ϩ CH3NH3 , and CH2NH2 . While they claim the latter conden- and injected into a quadrupole mass filter that passes only sation products to be direct products of the reaction, we sus- ions with a mass-to-charge ratio of the desired reactant ion. The reactant ion beam then passes through a collimating lens 12 a͒Author to whom correspondence should be addressed; electronic mail: and into an octopole ion guide in which the reactive colli- [email protected] sions occur. The presence of the ion guide ensures the col- 0021-9606/99/111(6)/2507/6/$15.002507 © 1999 American Institute of Physics Downloaded 20 Oct 2011 to 171.64.124.19. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions 2508 J. Chem. Phys., Vol. 111, No. 6, 8 August 1999 Everest et al. lection of all products, regardless of scattering angle.8 The neutral collision gas ͑methane-d4, CIL, 99% D, lot 94-1͒ is introduced through a leak valve into a short collision cell that is concentric with the octopole ion guide. The reaction collision energy is experimentally con- trolled by varying the dc float voltage of the octopole ion guide. The voltage of the field-free ionization region is kept constant at 6.00 V. Changing the octopole dc float voltage from 5.08 to Ϫ12.50 V corresponds to changing the center- of-mass collision energy range from 0.50 to 10.0 eV. The product ions and unreacted ammonia ions drift through the octopole and are mass selected with a second quadrupole. Time-of-flight profiles for reactants and products are then recorded simultaneously with a multichannel scaler and a current transient recorder. FIG. 1. Relative reaction cross sections ͑arbitrary units͒ for the reaction We report relative integral cross sections that are calcu- ϩ͑ 0 5͒ϩ NH3 1 2 CD4 as a function of center-of-mass frame collision energy lated by normalizing the counted signal for the product in ϩ ϩ (Ecom /eV) to form the two observed products: NH3D and CD3 . The lines question to the integrated current of the reactant ion. Here represent the mean of the four data sets included in the figure, and the error ‘‘relative’’ refers to the fact that we are unable to measure bars represent twice the standard deviation (Ϯ2s). absolute cross sections because we know neither the exact pressure in the collision cell nor the effective path length. bon species ͑e.g., pump oil fragments͒ in the chamber to Although the units are arbitrary, all of the cross sections ϩ form NH . reported in the present work may be compared with each 4 Because the experiment is initiated with a short laser other, as the pressure in the collision cell was kept nearly pulse ͑Ϸ10 ns͒, and the detector is time sensitive (␶ constant ͑vide infra͒ for all the experiments. We also report Ͻ5 ns), the time of flight of the products can also be mea- product branching ratios that are defined as the intensity of a sured, yielding some indication of the differential reaction particular product normalized by the total product intensity cross section. The instrument function is estimated by simu- under the same experimental conditions. Although branching lating the guided-ion beam to determine the correlation be- ratios contain less information than relative cross sections tween ion flight time and ion velocity in the octopole. The ͑i.e., a change in branching ratio does not indicate which reactant and product time-of-flight profiles are then trans- product became more intense or which lost intensity͒, they formed into the velocity domain. Product signal can be plot- have the advantage of being invariant to any experimental ted as an histogram versus the projection of the product ve- error that has an equal multiplicative effect on all products. locity onto the octopole axis ͑i.e., w cos(␪) where w is the Examples of such errors may include the integral of the col- p p product velocity in the center-of-mass frame and ␪ is the lision cell pressure over the cell path length, or any nonlinear scattering angle͒. The instrument axis is coincident with the process in the estimation of reactant intensity. initial velocity of the ammonia-ion reactant, and therefore, The estimated pressure in the collision cell ranged be- the average velocity of the center of mass. Although the tween 65 and 85 ␮Torr with most experiments being done complete differential scattering cannot be deduced from with a collision cell pressure of 75 ␮Torr. ͑These values are these histograms ͓i.e., separation of the w and cos(␪)͔, in- highly uncertain.͒ To ensure single-collision conditions, the p formation regarding the degree of forward or backward scat- branching ratios for the various products were measured over ͑ ␮ ͒ tering is obtained. a broad collision-cell pressure range roughly 16–810 Torr The raw transformed data are binned into 100-m/s-wide and were found to be independent of pressure. bins—much smaller than the instrumental blurring. A The possibility of signal arising from sources other than sharply increasing feature toward low velocity ͑zero in the reactions in the collision cell was checked at all collision laboratory frame, to be precise͒ in the velocity profiles arises energies by repeating the experiment without introducing from random dark current at very long times or very slow CD4 into the reaction chamber.