State-Specific Reactions of Cu+(1S, 3D) with ... - ACS Publications

Apr 17, 2012 - William S. Taylor*, Micah L. Abrams, Cullen C. Matthews, Seth Byers, Scott Musial, and Charles M. Nichols. Department of Chemistry ...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JPCA

State-Specific Reactions of Cu+(1S, 3D) with CH3X and CF3X (X = Cl, Br, I): Exploring the Influence of Dipole Orientation on Association and C−X Bond Activation William S. Taylor,* Micah L. Abrams, Cullen C. Matthews,† Seth Byers,‡ Scott Musial, and Charles M. Nichols§ Department of Chemistry, University of Central Arkansas, Conway, Arkansas 72035, United States ABSTRACT: The reactions of gas-phase Cu+(1S) and Cu+(3D) with CF3X and CH3X (X = Cl, Br, and I) have been examined experimentally using the drift cell technique at 3.5 Torr in He at room temperature. State-specific product channels and overall bimolecular rate constants for depletion of the two Cu+ states were determined using electronic state chromatography. The results showed that Cu+(1S) participates exclusively in association with all of these neutrals, whereas, depending on the neutral, Cu+(3D) initiates up to three bimolecular processes, resulting in the formation of CuX+, CuC(H/F)3+, and C(H/F)3X+. Possible structures for the singlet association products were explored using density functional methods. These calculations indicated that Cu+ preferentially associates with the labile halogen (Cl, Br, I) with all neutrals except CF3Cl, for which a “backside” geometry occurs in which Cu+(1S) is weakly bound to the −CF3 end of the molecule. All products observed on the triplet reaction surface can be understood in terms of either known or calculated thermochemical requirements. Product distributions and overall reaction efficiencies for C−X bond activation (X = Br, I) through Cu+(3D) suggest that the orientation of the neutral dipole has little or no effect in controlling access to specific product channels. Likewise, second-order rate constants for reactions with X = Br and I indicate efficient depletion of Cu+(3D) and do not exhibit the dramatic variations in reaction efficiency previously observed with CH3Cl and CF3Cl. These results suggest that C−X bond activation proceeds through a bond-insertion mechanism as opposed to direct abstraction.



of gas-phase metal ion chemistry, the reactions of Cu+ have generated considerable interest and are the subject of recent reviews.14,15 In addition, state-specific comparisons have highlighted differences in the chemistry of Cu+(1S) and Cu+(3D).16,17 Activation of the sigma bond in CH4 represents the prototypical example of C−H bond activation and is the first step in the conversion of this abundant resource into more useful molecules. Numerous gas-phase studies of methane activation by transition-metal ions have revealed a wealth of mechanistic information regarding this important process.18 However, C−H bond activation in CH4 is endoergic from most transition-metal-ion ground states, including Cu+(1S). Proton abstraction through Cu+(3D) is likewise energetically unfavorable, and C−H activation by Cu+(3D) to yield CuCH2+ is spinforbidden. Alternatively, activation of the weaker C−X bond in many halogenated methane analogues is energetically accessible under thermalized conditions. Furthermore, the polarity of the C−X bond might contribute to other interesting reactive features, as discussed further below. Reactions of bare metal

INTRODUCTION The ability to activate sigma bonds is one of the most interesting characteristics of transition metals. Arising as a consequence of the energetically accessible manifold of d orbitals, this capability lies at the core of their use as catalysts for processes in which typically unreactive molecules are functionalized to yield useful products. The study of gas-phase metal ions can provide intrinsic information regarding bondactivation mechanisms, while at the same time avoiding the complicating effects arising from the presence of a solvent. As a result, the behavior of gas-phase metal ions has been the subject of numerous experimental and theoretical examinations spanning more than two decades.1−13 A continuing focus of research in our laboratory has been the determination of various fundamental factors influencing the outcomes of such reactions in an effort to shed light on the mechanisms by which they occur. One such determining factor is the electronic state of the metal. A wide variety of studies have illustrated that the outcomes of these reactions can be dramatically influenced by the electronic state of the metal through thermochemical, kinetic, molecular orbital, and quantum-mechanical (i.e., spin) restrictions.1−10,13 Following this line of inquiry, we have been interested in the reactions of the 1S and 3D states of Cu+ with a number of small organic molecules. Within the broader context © 2012 American Chemical Society

Received: January 17, 2012 Revised: March 20, 2012 Published: April 17, 2012 3979

dx.doi.org/10.1021/jp300587y | J. Phys. Chem. A 2012, 116, 3979−3988

The Journal of Physical Chemistry A

Article

ions (including Cu+) with methyl halides have been the focus of a number of previous studies by others.19−27 In addition, we have previously described the results of the reactions of Cu+(1S, 3 D) with CH3Br in which the observed bimolecular chemistry can be understood within the context of known reaction energetics and quantum mechanical requirements with regard to each state; however, state-specific kinetic determinations were not carried out.28 In a later study, we reported the results of the reactions of Cu+(1S, 3D) with CH3Cl, CH2FCl, CHF2Cl, and CF3Cl.29 In that work, Cu+(3D) was observed to statespecifically react with all four neutrals to produce CuCl+. Most notably, depletion of Cu+(3D) by CF3Cl occurred at only 7% of the collision limit as predicted by the average dipole orientation (ADO) model.30 Both reaction efficiency and product formation in the triplet reactions was rationalized in these systems by postulating a direct abstraction mechanism in which Cl must be oriented toward the metal for a productive encounter to occur. Within this mechanistic scenario, the inefficiency of the Cu+(3D)/CF3Cl system can be explained by noting that the dipole orientation in CF3Cl results in an initial interaction with Cu+(3D) in which the reactive site on the neutral (the Cl atom) is directed away from the metal. In the work described here, we further test this idea by comparing our earlier CF3Cl and CH3Cl results with those obtained using CF3Br, CH3Br, CF3I, and CH3I. These neutrals have dipole orientations similar to that of the CH3Cl/CF3Cl pair; however, the weaker C−Br and C−I bonds should result in more favorable overall reaction energetics with respect to C−X bond activation. As in our previous work, Cu+ is well-suited for this study for several reasons. It has a limited number of low-lying excited states compared with earlier transition-metal ions; therefore, in the absence of the ability to selectively populate a given state, this reduces the number of states sampled for subsequent reaction and simplifies the reaction analysis. Further, the 1S and 3 D states have different electronic configurations (3d10 and 3d94s1, respectively). This allows them to be distinguished on the basis of differences in their mobilities in He using electronic state chromatography.31,32

the reactant ion beam was focused on the entrance aperture of a 4.0-cm drift cell that was described in detail elsewhere.33 In typical operation, the drift cell was charged with 3−5 Torr of He and a small partial pressure of the desired reactant neutral. For this work, the mole fraction of the reactant gas in He was on the order of 10−4 at a total pressure of 3.5 Torr. Under the conditions employed here, this concentration was sufficiently greater than the ion number density that pseudo-first-order conditions existed with respect to depletion of the reactant ion. Kinetic determinations were carried out by measuring reactant ion depletion at a constant reaction time as a function of the number density of the neutral reactant. In this apparatus, reactant ions are drawn through the drift cell by means of a small electric field maintained by a set of seven guard rings, during which time reaction can occur. This electric field imparts an additional amount of translational heating to the reactants that is typically characterized by the ratio of the electric field strength to the number density within the drift cell, or E/N. In general, translational energies are regarded as thermal (i.e., defined by the Maxwellian temperature within the drift cell) when E/N values are less than ∼6.4 Td (1 Td = 1 × 10−17 cm2·V). The experiments described herein were carried out with E/N values in the approximate range of 7−9 Td and are thus best characterized as “near-thermal”. Nonetheless, these reaction conditions were such that only exothermic or thermoneutral reactions were typically observed, and we concluded that little translational heating occurred. Under these reaction conditions, residence times for reactant ions within the drift cell were on the order of 100 μs. Temperature control of the drift cell was accomplished through a copper shroud through which heated or cooled gases can be circulated. All of the reactions described in this work were carried out at room temperature. Temperatures within the drift cell were monitored using a platinum RTD (resistance temperature device). Ions exiting the drift cell were mass-analyzed with a second quadrupole and detected using an electron multiplier operated in pulse-counting mode. Metal Ion Source. Cu+ ions for use in this work were produced by a glow discharge source that has been described previously.34 Briefly, this ion source produces metal ions through a sputter bombardment process in which ions of a working gas (Ne in this work) are accelerated to a cathode made from the desired metal. Metal atoms are sputtered from the surface of the cathode and diffuse into an intensely luminous region of the discharge known as the negative glow. The metal atoms are then ionized either by Penning ionization through metastables of the working gas or by ionization through fast electrons being accelerated from the cathode. Metal ions are sampled directly from the discharge plasma. We previously demonstrated that this ion source is capable of producing metal ions in excited states as well as in their ground states. Further, excited-state populations can be controlled to some extent by manipulation of discharge parameters.35 In this work, Cu+(3D)/Cu+(1S) ratios were increased or decreased by decreasing or increasing, respectively, the distance between the sampling aperture and the cathode. Determination of Cu+ State Distribution. For the work described here, specific configurations of Cu+ ions produced in the glow discharge were identified within the drift cell using a specific application of ion mobility mass spectrometry known as electronic state chromatography (ESC), which characterizes ions on the basis of their mobilities in He.31 ESC is most effective in distinguishing between electronic configurations



EXPERIMENTAL METHODS Instrument Description. Experiments were carried out using a selected-ion drift cell apparatus that was described previously.29 Cu+ ions used in this work were produced using a dc glow discharge ion source, as discussed in the next section. The instrument also incorporated an electron ionization source that was used to produce several of the molecular ions examined here. Ions from either source were directed through a quadrupole deflector (turning lens) and then to a quadrupole mass filter for mass selection. In this work, the selection quadrupole was operated in two modes. For reactions in which Cu+ was the reactant ion, we used Q1 as a high-pass filter by operating it in radio-frequency-only mode and setting the cutoff mass sufficiently high to reject unwanted ions produced by the discharge (which consisted mainly of lower-mass ions produced from residual atmospheric species and the discharge gas). This mode of operation produced a reactant ion beam composed exclusively of Cu+. In examinations of several of the follow-on reactions initiated by Cu+, Q1 was operated in the normal mass-resolving mode, and the required reactant ion in each case (CF3+, CF2X+, etc.) was generated by electron ionization of a suitable precursor species and mass-selected. In either mode, 3980

dx.doi.org/10.1021/jp300587y | J. Phys. Chem. A 2012, 116, 3979−3988

The Journal of Physical Chemistry A

Article

Figure 1. Cu+ ATDs and comparison spectra for the reaction of Cu+(1S, 3D) with CF3I in which the ion source has been adjusted to either enhance (red) or suppress (blue) production of Cu+(3D). T = 306 K, Ptot = 3.5 Torr, XCF3I = 1.7 × 10−4. Spectra have been normalized to the Cu+ intensity.

some amount of Cu+(1D) might be present in the 3d94s1 (highmobility) feature, it is likely that the major contributors to this configuration are the energetically more accessible 3D3,2,1 states. Further, the higher multiplicity of the 3D state relative to the 1D state (15 microstates versus 5) also favors population of the triplet excited state. Previous tests for the presence of Cu+(1D) using charge-transfer bracketing reactions revealed no evidence of this state, suggesting that the 3d94s1 Cu+ configuration that we observed contained exclusively Cu+(3D). Determination of State-Specific Products. State specificity with respect to product formation was determined using two methods: (1) acquiring product mass spectra while manipulating reactant ion state distributions and (2) correlating reactant and product ATDs. In the first method, the discharge conditions were first adjusted such that production of Cu+(3D) was minimized. A continuous beam of the reactant ion was then injected into the drift cell containing halocarbon/He mixtures as described above. Mass spectra accumulated under these conditions were then compared to those obtained when the discharge was adjusted to produce more of the 3D state. Definitive assignment of reactant ion precursor states for each product was obtained by correlating product ATDs to those of the two Cu+ states using a variation of the ESC experiment that was outlined previously.36 The Cu+ ATD is first collected under the same drift cell conditions as the reaction of interest minus the reactant neutral. A small amount of the reactant neutral is then admitted into the drift cell, and the mass filter is tuned to the desired product ion mass while the reactant ion beam is pulsed. The ATDs for both species are then corrected to account for differences in flight times through the analysis quadrupole and then overlaid. A product ion formed near the entrance of the drift cell will exhibit a residence time in the drift cell characteristic of the mobility of that product. If the product mobility is lower than that of the reactant ion (as was the case here), this “early-converted” product will ultimately arrive at the detector later than its reactant ion precursor would have. Alternatively, a product ion formed near the exit of the drift cell will exhibit a residence time in the drift cell representative of

differing significantly in size, such as those that differ by either the presence or absence of an s electron. The larger size of the s orbital results in a less attractive interaction between the ion and the He bath gas, which reduces the number of capture collisions. In terms of the first-row ions, this means that ions with 3dn−14s1 configurations have higher mobilities in the bath gas than those with 3dn configurations. As a consequence, a pulse containing a given metal ion in both configurations will be separated within the drift cell such that the higher-mobility configuration arrives at the detector first. Thus, configurations of sufficiently different mobilities appear as different peaks in an arrival time distribution (ATD). Ion mobilities used in assigning configurations were obtained from the ATDs by measuring the flight times of the different configurations as a function of the reciprocal of the drift voltage as described previously.31 ESC analysis of Cu+ extracted from the discharge indicated the presence of two configurations. Mobility measurements indicated that these correspond to the 3d10 and 3d94s1 configurations. A previous analysis of excited states formed in the glow discharge suggested that this ion source is capable of producing excited metal ion states with energies up to approximately 11.8 eV above the atom ground state.35 For Cu+, this energy range includes the 1S(3d10) ground state, as well as the 3D(3d94s1) and 1D(3d94s1) excited states, which lie 2.842 eV (averaged over J levels) and 3.257 eV, respectively, above the 1S state. The low-mobility feature in our Cu+ ATDs is undoubtedly the 1S ground state, but because the first and second excited states are indistinguishable on the basis of their mobilities, we cannot rule out the presence of both within the high-mobility feature on the basis of ESC alone. Our previous examination of ionization/excitation within the discharge suggested that excited atomic ions extracted from the discharge are produced primarily by an electron ionization process and that , in the absence of any preferential depletion mechanism(s), the relative populations of the excited metal ions will depend on the electron energy distribution function at the point of sampling within the discharge. Thus, although 3981

dx.doi.org/10.1021/jp300587y | J. Phys. Chem. A 2012, 116, 3979−3988

The Journal of Physical Chemistry A

Article

the mobility of the reactant ion. The arrival times of these “lateconverted” products will be the shortest and will correlate directly with their reactant ion precursors. Thus, the ATD of a product ion with a lower mobility than the reactant will originate at the same time as that of the reactant ion producing it. In this way, state-specific reaction sequences were identified for the reactions of both Cu+ states with CF3Br, CH3I, and CF3I. Copper cathodes used as sputter targets were fashioned from used oxygen-free copper gaskets into 5.0-mm-diameter rods. Neon (the discharge gas) and CH3Br were both obtained from Matheson Tri-Gas Inc. with purities of 99.999% and 99%, respectively. CH3I purchased from Spectrum Chemicals was 99.5% pure. Helium used as the buffer gas in the drift cell was obtained from Air Products Inc. with a purity of 99.9999%. CF3Br and CF3I were obtained from Synquest Gases Inc. with purities of 99% and 97%, respectively. Computational Methods. Singlet association product structures were determined with the Gaussian 03 suite of programs using density functional methods.37 All calculated energies were corrected for zero-point energy contributions. These calculations were carried out using the B3LYP functional38 in conjunction with the following basis sets obtained from the EMSL Basis Set Exchange:39,40 for Cu+, C, F, Cl, and Br, aug-cc-pVTZ;41−43 for I, aug-cc-pVTZ-PP;44 and for H, cc-pVTZ.42 All of these are correlation-consistent basis sets that include polarization functions with triple-ζ splitvalence and diffuse functions in the augmented versions. Augmented basis sets were used to better model potentially weak interactions between Cu+ and these neutral species. In addition, the basis set used for I includes a small-core pseudopotential that has been shown to enhance calculational efficiency for larger atoms. Calculations on open-shell species were spin-unrestricted, whereas those for singlet species were carried out in restricted mode. In singlet structures for which two stable geometries were found, transition states between the two stationary points were located using the synchronous transit-guided quasi-Newton (STQN) method developed by Peng et al.45

Figure 2. Cu+, CuBr+, and Cu+·CF3Br ATDs for the reaction of Cu+(1S, 3D) with CF3Br. Reactant ATDs have been fit to Gaussians. T = 306 K, Ptot = 3.5 Torr, XCF3Br = 1.7 × 10−4.

Similarly to its previously reported behavior with CH3Cl, CF3Cl, and CH3Br, Cu+(1S) exhibits association exclusively with CF3Br, CH3I, and CF3I under these experimental conditions.28,29 No significant association was observed to arise from Cu+(3D) within the 0.5-μs time resolution of the correlation experiment. Likewise, no evidence of quenching of Cu+(3D) to Cu+(1S) was noted (which would have been observed as bridging between the two Cu+ ATD features). Analogously to the behavior of Cu+(3D) with CH3Cl, CH3Br, and CF3Cl reported in our previous work, this Cu+ state reacts to form CuX+ (X = Br, I) with CF3Br, CH3I, and CF3I as the major bimolecular product with each neutral. In addition, charge transfer was observed from this Cu+ state with CH3Br, CH3I, and CF3I. Formation of CuCH3+ as a minor product through Cu+(3D) occurred with all of the methylated neutrals, as well as CuCF3+ from CF3I. Formation of CuCH3+ was not reported in our previous examination of the reactions of Cu+(3D) with CH3Cl and CH3Br because of limitations in the sensitivity of our apparatus at the time of those studies.28,29 In addition to these primary products, C(H/F)3+ and C(H/F)2X+ (X = Br, I) were observed as higher-order products at larger extents of reaction. All product channels initiated by Cu+(3D) are summarized in the general reaction sequence given in Figure 3. With regard to the X-abstraction product channel, the behavior of the brominated and iodinated neutrals is consistent with a sequence of reactions that we previously proposed for CH3Cl and CF3Cl in which CuX+ participates in secondary X− abstraction to yield either CH3+ or CF3+.29 C(H/F)3+ then abstracts H− (or F−) to form C(H/F)2X+ as the terminal bimolecular product. CH3+ and CF3+ were previously shown to produce CH2X+ (with CH3X) and CF2X+ (with CF3X), respectively.46,47 For the methylated neutrals, independent confirmation of this terminal reaction was obtained in our apparatus by forming CH3+ through electron ionization of CH3Cl and subsequently injecting it into the drift cell containing either CH3Br or CH3I. In the case of CH3I, these tests also revealed a second reaction pathway for the production of CH3I+ through charge transfer with CH3+. Primary product branching ratios were measured at low extents of reaction to minimize contributions from higher-order charge transfer to the



RESULTS AND DISCUSSION Representative examples of product comparison spectra and product ATD correlations for the Cu+/CF3I and Cu+/CF3Br systems are given in Figures 1 and 2, respectively. In Figure 1, it is clear that the amounts of CuI+, CF3I+, and CF2I+ formed were all enhanced when the proportion of Cu+(3D) was increased; however, this fact does not provide explicit information as to the sequence of product formation. Further insights into the immediate precursors for each observed product can be gained from correlations such as that given in Figure 2 for the Cu+/CF3Br system. Here, one can see that the ATDs for CuBr+ and Cu+·CF3Br exhibit good correlations with the excited and ground states of Cu+, respectively. Secondary and tertiary products (observed in all systems) can be likewise identified by correlating their product ATDs with their respective precursor ions. In similar fashion, state-specific product sequences were determined for all of the reactions examined here. Primary product channels observed in these systems are listed in Table 1, along with our previous results for CF3Cl, which have been included for comparison. We have also included product data for the CH3Cl and CH3Br systems that have not been previously reported. 3982

dx.doi.org/10.1021/jp300587y | J. Phys. Chem. A 2012, 116, 3979−3988

The Journal of Physical Chemistry A

Article

Table 1. Primary Products and Overall Rate Constants for the Reactions of Cu+ Precursor States Cu+(1S)

a

Cu+(3D) −9 a

neutral reactant

ionic product

kobs (×10 )

kobs/kADO

CH3Clc

Cu+·CH3Cl

0.71 ± 0.07

0.40

CF3Clc CH3Br

Cu+·CF3Cl Cu+·CH3Br

0.31 ± 0.01 0.81 ± 0.05

0.32 0.57

CF3Br CH3I

Cu+·CF3Br Cu+·CH3I

0.48 ± 0.03 0.51 ± 0.07

0.50 0.37

CF3I

Cu+·CF3I

0.90 ± 0.05

0.79

b

ionic product

product distribution (%)

CuCl+ CuCH3+ CuCl+ CuBr+ CuCH3+ CH3Br+ CuBr+ CuI+ CuCH3+ CH3I+ CuI+ CuCF3+ CF3I+

96 ± 1 4±1 100 72 ± 9 13 ± 6 15 ± 8 100 >14 ± 4 3±2 −130,b > −128,b > −204,b > −216,b

−41 > −1c −97 > −70 −152 > −139, ≤0c

67 196 −2 81 −99 −28

≤0 ≤0 ≤0 ≤0

a In units of kJ/mol. bUses calculated value for Cu+−X bond strength (see text) cUses calculated value for Cu+−CF3 bond strength (see text)

energetically possible from Cu+ only when the sum of the Cu ionization energy and the Cu+(3D) state energy exceeds the ionization energy of the reactant neutral. Thus determined, the energetic changes for these three product channels are given in Table 5. Where available, the experimentally determined bond strengths given in Table 2 were used in calculating the energetic changes listed in Table 5. To our knowledge, the binding energies for the three CuX+ species have not been measured directly; however, the Cu+−Cl binding energy can be obtained indirectly from related experimental parameters (see Table 2) and was used to calculate the energetics associated with the formation of CuCl+. Reaction energetics for the production of CuX+ through the brominated and iodinated neutrals were determined using the calculated binding energies of 150 and 169 kJ/mol for Cu+−Br and Cu+−I, respectively. These values were determined using the same level of theory as discussed above and were corrected for thermal contributions at 298 K. Given the calculated overbinding already noted for CuCl+, we anticipate that the calculated bond strengths for CuBr+ and CuI+ (also open-shell) are likely to be overestimated as well. If so, this will correspondingly be manifested as an overestimate in the magnitude of the predicted exothermicity for CuBr+ and CuI+ formation listed in Table 5. Therefore, the tabulated exothermicities for production of these two species are listed as lower limits. Having said this, we note that, even if the calculated Cu+−Br and Cu+−I bond strengths are too high by 60% (the approximate error in the Cu+−Cl binding), this does

not alter the predicted exothermicity of the reactions yielding CuBr+ and CuI+. Overall reaction energetics for the formation of CuCF3+ are also less certain, as we are aware of no measured or calculated values for the Cu+−CF3 binding energy. If we assume that formation of CuCF3+ is dictated solely by thermochemistry, the fact that CuCF3+ does not occur in the Cu+(3D)/CF3Br system implies that the Cu−CF3+ binding energy is