View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Classical Trajectories and RRKM Modeling of Collisional Excitation and Dissociation of Benzylammonium and tert-Butyl Benzylammonium Ions in a Quadrupole-Hexapole-Quadrupole Tandem Mass Spectrometer Vadim D. Knyazeva,b and Stephen E. Steina a National Institute of Standards and Technology, Physical and Chemical Properties Division, Gaithersburg, Maryland, USA b Research Center for Chemical Kinetics, Department of Chemistry, The Catholic University of America, Washington, District of Columbia, USA Collision-induced dissociation of the benzylammonium and the 4-tert-butyl benzylammonium ions was studied experimentally in an electrospray ionization quadrupole-hexapole-quadrupole tandem mass spectrometer. Ion fragmentation efficiencies were determined as functions of the kinetic energy of ions and the collider gas (argon) pressure. A theoretical Monte Carlo model of ion collisional excitation, scattering, and decomposition was developed. The model includes simulation of the trajectories of the parent and the product ions flight through the hexapole collision cell, quasiclassical trajectory modeling of collisional activation and scatter- ing of ions, and Rice-Ramsperger-Kassel-Marcus (RRKM) modeling of the parent ion decom- position. The results of modeling demonstrate a general agreement between calculations and experiment. Calculated values of ion fragmentation efficiency are sensitive to initial vibrational excitation of ions, scattering of product ions from the collision cell, and distribution of initial ion velocities orthogonal to the axis of the collision cell. Three critical parameters of the model were adjusted to reproduce the experimental data on the dissociation of the benzylammonium ion: reaction enthalpy and initial internal and translational temperatures of the ions. Subse- quent application of the model to decomposition of the t-butyl benzylammonium ion required adjustment of the internal ion temperature only. Energy distribution functions obtained in modeling depend on the average numbers of collisions between the ion and the atoms of the collider gas and, in general, have non-Boltzmann shapes. (J Am Soc Mass Spectrom 2010, 21, 425–439) © 2010 American Society for Mass Spectrometry ragmentation of gaseous ions has long been an quantitative prediction of CID mass spectra. It is gen- important source of information on the properties erally understood that the successful modeling of colli- Fof these ions and the corresponding parent spe- sional ion dissociation would require correct quantita- cies. Studies of collision-induced dissociation (CID) tive description of three factors: (1) the chemical mass spectra of many types of ions provide important mechanism of fragmentation (i.e., the sequence of in- information on their structure, thermochemistry, and tramolecular rearrangements and decomposition and other properties; recently, the main focus of such stud- the potential energy surfaces of these processes), (2) ies shifted to the analysis of large biomolecules, such as rovibrational excitation of the parent ion as a result of proteins and peptides (e.g., [1–11]. collisions with the inert collider gas, and (3) kinetics of As a result of many studies, the many features of the dissociation of ions in competition with deactivating collisional ion dissociation have been elucidated, in- collisions and removal of ions (via scattering or detec- cluding those of fragmentation of the ions of large tion) from the collision cell of the mass spectrometer. A biomolecules (e.g., [9, 12–17] and references therein). number of earlier studies concentrated on modeling one However, the factors that determine absolute and rela- or two of these factors; reviews can be found in [9, 18]. tive abundances of possible fragments are not well Among these three factors, the collisional excitation of understood, certainly not to the point that would enable ions is probably the one least well understood. Usually, the results of modeling of experimental ion fragmenta- tion data are presented in a form of effective tempera- Address reprint requests to Professor V. D. Knyazev, Research Center for Chemical Kinetics, Department of Chemistry, The Catholic University of tures, with an implicit assumption that single or multi- America, Washington, DC 20064, USA). E-mail: [email protected] ple collisions of ions with the inert collider gas yield Published online December 4, 2009 © 2010 American Society for Mass Spectrometry. Published by Elsevier Inc. Received May 26, 2009 1044-0305/10/$32.00 Revised November 24, 2009 doi:10.1016/j.jasms.2009.11.007 Accepted November 24, 2009 426 KNYAZEV AND STEIN J Am Soc Mass Spectrom 2010, 21, 425–439 energy distributions that are similar to the Boltzmann lider mass were studied. Rather broad P(E, E=, ECOLL) distribution in their shapes, although evidence of non- distribution functions were obtained, with considerable Boltzmann behavior exists in the literature (e.g., [10, contributions at both small and large ⌬E ϭ E=ϪE values. 19–22]. However, the goal of achieving the ability to Average energy-transfer efficiency was shown to be predict these energy distributions is far from being large, in the range of ϳ40–80%. Energy-transfer was realized. The most important characteristic of collisional found to moderately increase with the size of the energy-transfer needed to be understood is the per- peptide and to be somewhat more efficient for folded collision activation function, P(E, E=,ECOLL), the prob- structures as compared to extended conformations. It ability of energy transfer from energy E= to energy E was demonstrated that the energy-transfer efficiency is upon a collision with a collider gas atom or molecule sensitive to the repulsive part of the intermolecular with the relative kinetic energy ECOLL. potential and to the collider mass: heavy atoms such as In a number of earlier works, experimental results of Kr and Xe are more efficient activators than a light CID experiments were used to derive various features collider like He. Meroueh and Hase also analyzed of energy distributions resulting from gas-phase colli- collisional rotational activation; the results demon- sional activation of ions [23–30]. In a series of studies strated that between 1% and 8% of the relative kinetic ([31–33] and references therein), the “survival yield” energy of colliders can be deposited into degrees of method was used to evaluate the shapes of energy freedom of peptide corresponding to its overall rota- distributions. In this method, the survival yield (frac- tions. Efficiency of rotational excitation decreased with tion of nonfragmented ions) was studied as a function the increasing relative kinetic energy. In a recent work, of the ion fragmentation energy barrier for a series of Martinez-Nunez et al. [47] extended the method of similar ions with different substituents (such as ben- Meroueh and Hase by including a Morse function term zylpyridinium salts). The resultant dependences were in the force field describing the potential energy surface analyzed using Rice-Ramsperger-Kassel-Marcus (RRKM) of the ion and thus enabling modeling of ion dissocia- modeling to derive the energy distribution functions. tion. These authors applied the method to modeling ϩ Vekey and coworkers performed a series of RRKM and CID of Cr(CO)6 in collisions with Xe. The results of Master Equation modeling studies of collisional energy- modeling were in general agreement with the experi- transfer and CID, with the “survival yield” and other mental data of Muntean and Armentrout [20]. types of experiments supplying empirical data (e.g., ref- In the current study, we performed a first-principles erences [30, 32–37]. Laskin and coworkers performed a modeling of collisional activation and dissociation of series of studies directed at characterization of colli- two small ions (benzylammonium and 4-tert-butyl ben- sional energy-transfer in multicollisional CID MS and zylammonium) in a quadrupole-hexapole-quadrupole analyzing fragmentation energetics of large ions, in- tandem CID mass spectrometer. cluding peptide ions [18, 19, 38–42]. These authors used ϩ ¡ ϩ ϩ RRKM/Master Equation modeling of CID in FT-ICR C6H5CH2NH3 C6H5CH2 NH3 (1) mass spectrometer and proposed an analytical form of ϩ ¡ ϩ ϩ the activation function [19]. An informative review of C(CH3)3C6H4CH2NH3 C(CH3)3C6H4CH2 NH3 (2) studies directed at determining energy distributions resulting from collisional ion activation is given in [18]. The potential energy surfaces of the reactions were In a series of studies, Armentrout and coworkers inves- studied using quantum chemical methods. Collisional tigated collision-induced dissociation of small poly- activation was described by performing quasiclassical atomic ions under the conditions of low pressures, trajectory calculations using the method similar to that where most ions experience only single collisions with of Meroueh and Hase. Kinetics of ion dissociation was the inert collider gas (e.g., references [11, 20, 43, 44]. modeled using the RRKM method. The trajectories of These authors concentrated on determination of reac- ion flight through the collision cell of mass spectrometer tion threshold energies, and performed RRKM model- were modeled using classical mechanics in a Monte ing of the experimental data using a functional form of Carlo approach; processes of collisional activation, de- the internal energy distribution derived