Threshold photoelectron photoion coincidence spectroscopy is used to study the dissociation of energy-selected X(CH3)3 + ions (X ) As, Sb, Bi) by methyl loss, the only process observed up to 2 eV above the ionization energy. The ion time-of-flight distributions and the breakdown diagrams are analyzed in terms of the statistical RRKM theory to obtain accurate ionic dissociation energies. These experiments complement previous studies on analogous trimethyl compounds of the N group where X ) N and P. However, trimethylamine was observed to lose only an H atom, whereas trimethylphosphine was shown to lose methyl radical, H atom, and, to a lesser extent, methane in parallel dissociation reactions. Both kinetic and thermodynamic arguments are needed to explain these trends. The methyl radical loss has two channels: either a H transfer to the central atom, followed by CH3 loss, or a direct homolytic bond cleavage. However, the H transfer channel is blocked in trimethylamine by an H loss channel with an earlier onset, and, thus, the methyl loss is not observed. Bond energies are defined based on ab initio reaction energies and show that the main thermodynamic reason behind the trends in the energetics is the significantly weakening CdX double bond in the ion in the N f As direction. The first adiabatic ionization energies of Sb(CH3)3 and Bi(CH3)3 have also been measured by ultraviolet photoelectron spectroscopy to be 8.02 ( 0.05 and 8.08 ( 0.05 eV, respectively. Introduction Although the chemical properties of homologous compounds are often similar, small changes in the electronic structure can result in large differences in chemical properties. We discovered one such case during the investigation of the dissociation patte s of trimethylamine and trimethylphosphine ions.1,2 While energy-selected trimethylamine ions lose exclusively an H atom, forming H2CdN(CH3)2 + ions (isoelectronic with isobutene) over a large inte al energy range, the trimethylphosphine ions dissociate not only by H atom loss but also via methyl and methane loss, the methyl loss being the lowest energy and dominant channel. The methyl loss channel can involve a simple X-C bond break to form X(CH3)2 + or can take place via a prior H atom transfer step to produce the pentavalent intermediate, (CH3)2X(H)dCH2 +, followed by methyl loss to finally produce the tetravalent (CH3)(H)XdCH2 + ion. An overview of these processes is presented in Figure 1. The difference in the products for the trimethylamine and phosphine cannot be explained by the thermodynamics of the product species, as the calculated stability differences between the H atom loss and an H transfer followed by CH3 loss (the preferred path for methyl loss in these species) suggest a strong preference for the latter in trimethylamine, the opposite of what is observed. Therefore, explaining the trends in the dissociation patte involves not only knowing how bond energies change from the second to the third row central atom but also kinetic arguments. In order to shed more light on this trend, we have embarked on a study of three higher mass homologues: trimethylarsane, -stibane, and -bismane, which completes the series X(CH3)3, X ) N, P, As, Sb, Bi. As in the previous study, ions are produced in the gas phase, and energy selected by threshold photoelectron photoion coincidence (TPEPICO) using a monochromatic photon beam in the vacuum UV range from 7 to 14 eV.3-10 The ion inte al energy in a photoionization process is given by Eion ) hν + Eth - IE - Eel, where the quantities are the photon energy, the neutral molecule’s thermal energy, the adiabatic ionization energy, and the electron kinetic energy, respectively. By measuring the ions in coincidence with initially * Author to whom correspondence should be addressed. E-mail: bsztaray@ pacific.edu. † Eo¨tvo¨s University. ‡ Paul Scherrer Institut. § Current address: Department of Chemistry, Stanford University, Stanford, CA 94305. | University of North Carolina. ⊥ University of the Pacific. Figure 1. Possible dissociation reaction schemes for X(CH3)3 + (X ) N, P, As, Sb, Bi) at low inte al energies. J. Phys. Chem. A XXXX, xxx, 000 A 10.1021/jp900920r CCC: $40.75 XXXX American Chemical Society Downloaded by AUSTRIA CONSORTIA on July 6, 2009 Published on June 11, 2009 on http://pubs.acs.org | doi: 10.1021/jp900920r zero-energy electrons (threshold electrons), the ion is effectively energy selected. The time-of-flight (TOF) mass spectra as a function of the photon energy provide information about the ion dissociation energies. The data are then analyzed with the aid of ab initio calculations of the various reaction paths. Experimental Section Trimethylarsane was purchased from Sigma-Aldrich and used without further purification. Trimethylstibane was synthesized according to Fleming:11 antimony trichloride was purified by vacuum sublimation; 25.4 g (0.111 mol) of the halide was dissolved in 55 cm3 of diethyl ether and added to a 190 cm3 ether solution of methylmagnesium iodide (0.334 mol). The reaction mixture was subjected to fractional distillations in order to obtain pure trimethylstibane (16.5 g, 88% yield). Trimethylbismane was prepared similarly via the method described by Amberger:12 a solution made of 33.1 g (0.105 mol) of dry bismuth(III) chloride and 250 cm3 of diethyl ether was added dropwise to a methylmagnesium iodide solution (0.320 mol in 190 cm3 diethyl ether). The reaction product was purified with fractional distillation to produce 18.1 g of trimethylbismane (68% yield). Photoelectron Spectroscopy. In order to overcome the significant uncertainty conce ing the adiabatic IEs, photoelectron spectra were recorded for Sb(CH3)3 and Bi(CH3)3, using He(I) radiation with the ATOMKI ESA-32 UV photoelectron spectrometer, equipped with a Leybold-Heraeus UVS 10/35 high-intensity gas discharge photon source and a hemispherical analyzer.13 The energy resolution was better than 30 meV, and the sample was introduced through a gas inlet system at room temperature. Nitrogen was used as an inte al standard. Threshold Photoelectron Photoion Coincidence (TPEPICO). The TPEPICO apparatus at the University of North Carolina is only briefly presented, as it has been described in detail elsewhere.14,15 The sample is introduced in the ionization chamber through an inlet needle. The sample gas was in thermal equilibrium with the inlet system; therefore, room temperature was assumed in the data analysis. Vacuum UV light from a hydrogen discharge lamp dispersed by a 1 m normal incidence monochromator is used to ionize the gas-phase thermal sample. The wavelength resolution of 1 Å translates to 8 meV (0.8 kJ mol-1) at the onset energies obtained in this study (approximately 10 eV). The photon energy was calibrated using the intense Lyman-R emission line of the light source at 1215.67 Å. Electrons and ions are accelerated out of the ionization region in a homogeneous 20 V cm-1 field. Threshold electrons are velocity focused along the extraction axis onto a 1.5 mm aperture located 12 cm from the ionization region and detected by a Channeltron electron multiplier. Because energetic electrons with initial velocity vectors directed parallel to the extraction axis are also focused onto this aperture, the threshold electron signal is contaminated with “hot” electrons. They are accounted for by collecting a second set of electrons, whose trajectory ends in a 2 mm × 5 mm rectangular aperture close to the central aperture where a second Channeltron detects them. Electrons collected by the off-axis rectangular aperture have been found to be a good representation of the hot electrons at the central collector,16 so that the off-axis signal can be used to correct for hot electron contamination to yield a pure threshold electron signal. Ions are accelerated in the constant 5 cm long 20 V cm-1 field to -100 V, and then within 5 mm to -260 V as they enter the main drift region of the linear time-of-flight analyzer. The ions travel through a 26 cm long field-free drift region and are detected by tandem Burle multichannel plates. The two electron signals are fed into two Ortec 467 time-to-pulse height converters and used as start signals for the ion TOF measurement, and the ion TOF distributions are obtained for both the central and the off-axis electron signals with Ortec TRUMP- 8K multichannel analyzers. The purpose of the 5 cm long acceleration region is to disperse slowly dissociation ions into asymmetric fragment ion peaks, the shape of which provides information about the ion dissociation rates in the 104-107 s-1 range. In order to extend the rate measurements to 103 s-1, it is possible to decelerate ions some 20 cm into the drift region.17 Without this deceleration stage, daughter ions formed in the field-free region have the same time-of-flight as their parents. But because their mass is smaller, they can be decelerated more than the parent ions and thus distinguished from them. This extends rate information to nearly the full parent ion TOF. The experimental results consist of the ion TOF distributions at various photon energies. The fractional abundances of the various parent and fragment ions were corrected for the hot electron contribution and plotted as a function of the photon energy to produce the breakdown diagram. In some cases, the TOF distribution of the fragment ion exhibited an asymmetry, which is characteristic of slow dissociations. The peak shapes are modeled to obtain the ion dissociation rate constants as a function of the ion inte al energy.
