Near-Thermal Reactions of Au+(1S,3D) with CH3X - American

Jan 18, 2012 - Department of Chemistry, University of Central Arkansas, Conway, Arkansas 72035, United States. 'INTRODUCTION. Studies exploring variou...
0 downloads 0 Views 1010KB Size
ARTICLE pubs.acs.org/JPCA

Near-Thermal Reactions of Au+(1S,3D) with CH3X (X = F,Cl) William S. Taylor,* Cullen C. Matthews, Ashley J. Hicks, Kendall G. Fancher, and Li Chen Chen Department of Chemistry, University of Central Arkansas, Conway, Arkansas 72035, United States ABSTRACT: Reactions of Au+(1S) and Au+(3D) with CH3F and CH3Cl have been carried out in a drift cell in He at a pressure of 3.5 Torr at both room temperature and reduced temperatures in order to explore the influence of the electronic state of the metal on reaction outcomes. State-specific product channels and overall two-body rate constants were identified using electronic state chromatography. These results indicate that Au+(1S) reacts to yield an association product in addition to AuCH2+ in parallel steps with both neutrals. Product distributions for association vs HX elimination were determined to be 79% association/21% HX elimination for X = F and 50% association/50% HX elimination when X = Cl. Reaction of Au+(3D) with CH3F also results in HF elimination, which in this case is thought to produce 3AuCH2+. With CH3Cl, Au+(3D) reacts to form AuCH3+ and CH3Cl+ in parallel steps. An additional product channel initiated by Au+(3D) is also observed with both methyl halides, which yields CH2X+ as a higher-order product. Kinetic measurements indicate that the reaction efficiency for both Au+ states is significantly greater with CH3Cl than with CH3F. The observed two-body rate constant for depletion of Au+(1S) by CH3F represents less than 5% of the limiting rate constant predicted by the average dipole orientation model (ADO) at room temperature and 226 K, whereas CH3Cl reacts with Au+(1S) at the ADO limit at both room temperature and 218 K. Rate constants for depletion of Au+(3D) by CH3F and CH3Cl were measured at 226 and 218 K respectively, and indicate that Au+(3D) is consumed at approximately 2% of the ADO limit by CH3F and 69% of the ADO limit by CH3Cl. Product formation and overall efficiency for all four reactions are consistent with previous experimental results and available theoretical models.

’ INTRODUCTION Studies exploring various aspects of gas phase metal ion chemistry have been summarized in a number of reviews spanning more than two decades.1 13 The volume and breadth of work described in these surveys amply demonstrate the intense interest that metal ion chemistry has experienced and attest to its continuing relevance as an area of scientific inquiry. Of particular interest has been the ability of metal ions to activate sigma-bonds in reactions that often have parallels within condensed-phase catalytic mechanisms. Gas phase studies offer the advantage of providing insights into intrinsic reactivity while avoiding complicating effects arising from the presence of a solvent. An ongoing focus of research in our laboratory has been the determination of various fundamental factors influencing the outcomes of reactions of metal ions with small molecules in order 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 thoroughly 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 More correctly, access to specific product channels is frequently governed by the interaction between these various requirements.10 Clearly, the promise with respect to catalysis is that these mechanistic controls might somehow be exploited in order to selectively form desired products. The catalytic activity of dispersed gold nanoparticles has generated considerable interest, particularly with regard to their r 2012 American Chemical Society

ability to oxidize CO at room temperature.14 16 Studies of dispersed gold on MgO substrates have implicated cationic gold as a necessary participant in this process.17,18 The activity of gold has been shown to extend to the gas phase as well in experiments in which CO is oxidized by ionic gold clusters.19,20 More relevant to the work described here is the fact that gold nanostructures have been shown to catalytically oxidize hydrocarbons.15,21 Methane represents the prototypical model for hydrocarbon activation and as such, reactions of CH4 with gas phase metal ions have also been the subject of a number of studies.13 Activation of the C H bond in methane is the first step in the conversion of this abundant resource into molecules that are more useful as fuels and feedstocks. However, with few exceptions, C H bond activation in CH4 is endoergic from transition metal ground states, including Au+(1S).22 25 As a consequence, we have primarily focused our attention on reactions of metal ions with a number of methyl halides. For our use, these compounds offer the advantage of the weaker C X bond, thus making metalinduced molecular rearrangements energetically possible at low interaction energies. Previous work by others has examined the low-energy reactions of Au+(1S) with CH3X (X = F, Cl, Br, I).26 28 In addition, we have previously reported on the state-specific reactions of Au+(1S) and Au+(3D) with CH3Br.29 Dependent on the available energetics and electron spin Received: September 28, 2011 Revised: November 22, 2011 Published: January 18, 2012 943

dx.doi.org/10.1021/jp2093912 | J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

requirements, these systems have been shown to exhibit a variety of bimolecular product formation dependent on the Au+ electronic state, including HX elimination, X atom abstraction, CH3 abstraction, and association. In the case of CH3Br, state-specific product formation suggests that the reaction surfaces for Au+(1S) and Au+(3D) are well-behaved with respect to conservation of spin. The state-specific reactions of Au+(3D, 1S) with CH3X (X = F, Cl, Br) have also been investigated in several computational studies in which possible mechanisms leading to different products are explored.30 32 In this article, we describe the state-specific reactions of Au+(3D, 1S) with CH3F and CH3Cl at low interaction energy. The results summarized here offer useful comparisons to the largely ground state chemistry reported previously for this metal in addition to providing a direct experimental test of reaction pathways predicted by theory.

Temperature control of the drift cell is accomplished via a copper shroud through which heated or cooled gases can be circulated. The reduced reaction temperatures utilized in this work were achieved using liquid nitrogen as the cryogen. Temperatures within the drift cell are monitored using a Pt-RTD (resistance temperature device). Ions exiting the drift cell are mass-analyzed with the use of a second quadrupole and detected using an electron multiplier operated in pulse-counting mode. Metal Ion Source. Au+ ions for use in this work were produced by a glow discharge source, which has been described previously.35 Briefly, this ion source produces metal ions via a sputter bombardment process in which ions of a working gas 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 by either Penning ionization via metastables of the working gas or by electron impact ionization via fast electrons being accelerated from the cathode. Metal ions are sampled directly from the discharge plasma. We have 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.36 In this work, Au+(3D)/Au+(1S) ratios were controlled either by changing the identity of the discharge gas or by adjusting the distance between the gold cathode and the sampling orifice. The working gases used in these experiments were Ne and Xe (both 99.999%). Gold cathodes utilized here were obtained in the form of 3 mm rods with a purity of 99.9985%. Determination of Au+ State Distribution. For the work described here, specific configurations of Au+ ions produced in the glow discharge were identified within the drift cell using the electronic state chromatography technique (ESC), which characterizes them on the basis of the their mobilities in He.37 ESC is most effective in distinguishing between electronic configurations differing significantly in size, such as those which 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 capturecollisions. With respect to third-row transition metal ions, this means that those with 5dn 16s1 configurations have higher mobilities in the bath gas than those with 5dn 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). ESC analysis of Au+ extracted from a Ne discharge indicates the presence of two configurations. Previous mobility measurements indicate that these correspond to 5d10 and 5d96s1.36 ESC resolution can be enhanced by reducing the temperature of the He bath gas and was necessary in this work in order to distinguish between the two Au+ configurations. As such, all state-specific measurements (kinetic determinations and product ion correlations) utilizing ESC were carried out at temperatures ranging from 166 to 230 K. We have also previously shown that the glow discharge is capable of producing excited metal ion states with energies up to approximately 11.8 eV above the atom ground state.36 For Au+, this energy range includes the 1S(5d10) ground state, as well as the 3D3 and 3D2 excited spin orbit states, which lie 1.86 and 2.19 eV above the 1S state and both possess the 5d96s1 configuration. Given these energetic constraints,

’ EXPERIMENTAL METHODS Instrumental Description. Experiments were carried out using the selected ion drift cell apparatus, which has been described previously.33 Au+ ions used in this work were produced using a dc glow discharge ion source, which is discussed below. The instrument also incorporates an electron impact ion source, which was used to investigate higher-order reactions initiated by several of the molecular ions observed in the reactions described here. Ions from either source are 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 Au+ was the reactant ion, Q1 was used as a high-pass filter by operating it in rfonly mode and setting the cutoff mass sufficiently high as to reject unwanted ions produced by the discharge (which consist mainly of lower mass ions produced from residual atmospheric species). This mode of operation produced a reactant ion beam composed exclusively of Au+. In examinations of the reactions of secondary and tertiary molecular ions, Q1 was operated in the normal mass-resolving mode and the required reactant ion in each case (CH3+, CH2X+, etc.) was generated by electron impact on a suitable precursor species and mass-selected. In either mode, the reactant ion beam is focused onto the entrance aperture of a 4.0 cm drift cell, which has been described in detail elsewhere.34 In typical operation, the drift cell is charged with 3 5 Torr of He and a small partial pressure of the desired reactant neutral. In this work, He was obtained with a purity of 99.9999%. Stock reactant gas mixtures of CH3Cl (99.5%) and CH3F (99%) were initially prepared in He at mole fractions on the order of 10 3, then subsequently diluted upstream from the drift cell such that the mole fraction during reaction was on the order of 10 4 at a total pressure of 3.5 Torr. Under the conditions employed here, this concentration is sufficiently greater than the ion number density that pseudo first-order conditions exist with respect to the depletion of the reactant ion. Kinetic determinations were carried out by measuring reactant ion depletion at constant reaction time as a function of the number density of the neutral reactant. Reactant ions are drawn through the drift cell by means of a small electric field maintained by a set of seven guard rings. E/N values for the experiments described here were on the order of 8.5 Td (1 Td = 1  10 17 cm2 3 V) at room temperature, and ∼6 Td at approximately 220 K. Residence times for reactant ions under these conditions were on the order of 100 μs. These reaction conditions are such that little translational heating occurs and only exothermic or thermoneutral reactions are observed. 944

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Product spectra for Au+ + CH3Cl under conditions in which Au+(3D) is either enhanced (Ne discharge) or suppressed (Xe discharge). T = 307 K; E/N = 8.4 Td; [CH3Cl] = 5.3  1012 molec/cm3. Spectra are normalized to Au+.

Figure 2. Au+, AuCH2+, and Au+•CH3Cl ATD’s for the reaction of Au+(1S,3D) with CH3Cl. Au+ ATD’s are fit to Gaussians. T = 166 K; E/N = 5.2 Td; XCH3Cl = 5.6  10 5.

we assign the low mobility feature of our Au+ ATD as 1S, and the high mobility feature as a mixture of the 3D3 and 3D2 spin orbit states. As we will show below, the observation of limited chargetransfer with CH3Cl suggests that our high mobility Au+ feature is composed primarily of 3D3. Arrival time distributions of Au+ ions extracted from a Xe discharge indicate little or no production of the 5d96s1 configuration. We have previously hypothesized that excited state metal ion production in the glow discharge occurs primarily via electron impact involving high energy electrons.36 The absence of excited state Au+ configurations suggests the proportion of electrons with energies capable of producing excited Au+ is substantially reduced in the Xe discharge.

desired product ion mass, while the reactant ion beam is pulsed. The ATD’s for both species are then corrected to account for differences in flight times through the analysis quadrupole and then overlaid. The reactant ion can be converted into the product at any point within the drift cell, but a product ion formed near the exit of the drift cell will exhibit a flight time characteristic of the reactant ion producing it. Thus, the ATD of a product ion with a lower mobility than the reactant (as was the case for all product ions discussed here due to the incorporation of the ligand) will originate at the same time as the reactant ion producing it. A representative example of this technique is shown in Figure 2 (again for the Au+/CH3Cl system), where we see that both AuCH2+ (indicative of HCl elimination) and Au+•CH3Cl correlate directly to the 1S state. Note that the slight displacement of the association product to longer arrival times in this set of ATD’s is an indication of its lower mobility. Using these methods, products arising from both Au+ states were identified and are summarized in Table 1. In general, the products observed in these reactions result from several processes with respect to the neutral. These include association (arising exclusively from Au+(1S)) as well as a number of bimolecular processes including HX elimination, charge-transfer, and both hydride and halide abstraction. In addition, the reaction environment represented by the drift cell is one in which higher-order reactions are almost always observed. Accordingly, a number of higher-order products are formed in these reactions. In many cases, these follow-on products are obvious, such as when the product is the result of secondary or tertiary associations with the neutral reactant. However, the immediate ionic precursor for a product resulting from hydride or halide abstraction is not always apparent. The methods described above for determining state-specificity are useful in identifying the Au+ state ultimately responsible for a particular product but do not explicitly indicate its direct precursor. Additional insights into the formation of possible primary products can be ascertained by examining the thermochemical and quantum mechanical requirements for reaction. Thermochemical changes for possible primary reactions that would originate from each Au+ state can be calculated from known bond strengths and ionization energies and are listed in Table 2

’ RESULTS AND DISCUSSION State-specific product channels in these reactions were identified using two methods. Products arising from Au+(1S) were unambiguously determined using a Xe discharge to produce sputtered gold ions. Low-temperature ESC tests under these ionization conditions indicate that Au+(1S) is produced exclusively when Xe is used as the working gas in the discharge; thus, any products formed in the drift cell must arise from this state. Identification products formed from Au+(3D) were determined using a Ne discharge as described above. Comparison of mass spectra of the reaction mixture acquired using the two discharge gases provides a direct indication of products arising from both Au+ states. This is illustrated in Figure 1 for the Au+/CH3Cl system where we see that, consistent with the known endothermicity for hydride abstraction relative to the energetic content of Au+(1S), CH2Cl+ is formed only in the presence of Au+(3D). In addition to comparing spectra under different ionization conditions, state-specificity when both Au+(1S) and Au+(3D) were injected into the drift cell was confirmed via ESC by correlating product ion arrival time distributions to their reactant ion precursor states.38 In this technique, the reactant ion ATD is first collected under the same drift cell conditions as the reaction of interest. A small amount of the reactant neutral is then admitted into the drift cell and the analysis quadrupole is tuned to the 945

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

Table 1. Ionic Products Arising from Au+ Precursor States and State-Specific Overall Rate Constants for Au+ Depletion Au+(1S) neutral reactant

ionic products

kobs (10 9)a

Au+(3D) kobs/kADO

kobs (10 9)a

307 K +

CH3F

Au CH3F

0.045 ( 0.008

kobs/kADO

226 K 0.028

0.051 ( 0.007

0.93

1.6 ( 0.1

0.032

0.036 ( 0.003

1

AuCH3+

1.0 ( 0.1

Au+(CH3F)2 AuC2H4+ AuC2H4+(CH3F) 307 K Au+CH3Cl

1.3 ( 0.1

kobs/kADO

226 K AuCH2+ CH2F+ AuC2H4+ CH2F+(CH3F) CH3+(CH3F)

AuCH2+

CH3Cl

kobsa (10 9)

ionic products

218 K

0.023

218 K

AuCH2+

CH3Cl+

Au+(CH3Cl)2

CH3+

AuC2H4+ AuC2H4+(CH3Cl)

CH2Cl+

0.69

CH2Cl+(CH3Cl) CH3+(CH3Cl)

a

3

Units of cm molec

1

1

s .

Table 2. Primary Product Channel Requirements neutral reactant CH3F

CH3Cl

metal ion

primary product channel

Au+(1S)

1

AuCH2+ + HFa

44

Au+(3D)

3

AuCH2+ + HFa,b

97,

Au+(1S)

CH2F+ + AuHc 1 AuCH2+ + HCla

Au+(3D)

AuCH3+ + Cld CH3Cl+ + Au

a

ΔΣ

thermochemistry (kJ/mol) 0 128,

66 ( 22, 15 39.1, 16.40,

250 98 ( 22,

219 ( 22

primary? Y

0

Y

(1 0

N Y Y

70.2,

191.3

0, ( 1, ( 2

14.75,

135.80

0, ( 1, ( 2

Y

(1

N

(1

N

CH2Cl+ + AuHc

105 ( 3,

CH3+ + AuCl

50.5,

137 ( 3,

81.6,

202.7

258 ( 3

Thermochemistry calculated using Au CH2 binding energy of 393 kJ/mol from ref 24. b Thermochemistry calculated using 126.4 kJ/mol singlet triplet splitting for AuCH2+ from ref 25. c Uncertainties due to range of values reported for CH2X ionization energy. d Thermochemistry calculated using Au+ CH3 binding energy of 209.4 kJ/mol from ref 25. +

along with possible spin changes. Energetics for excited state reactions are listed separately for each Au+(3D) spin orbit state. The near-thermal reaction environment, which exists within the drift cell, places the requirement of exothermicity on processes that can be observed. As such, only reactions that are energetically accessible from at least one Au+ spin orbit state have been included in Table 2. Further, as is discussed below, no conclusive evidence of coupling between the singlet and triplet surfaces was observed in this work, and we thus assume that all reactions formally conserve spin. Primary product channels that can be further excluded on this basis have been identified accordingly in Table 2 and, therefore, cannot be the means by which those particular product ions are formed. Table 1 also summarizes the bimolecular rate constants for depletion of the two Au+ states. State-specific kinetic determinations were carried out using a method we have described previously in which Au+ arrival time distributions (ATD’s) were acquired at fixed reaction times with different concentrations of the neutral reactant.33 Rate constants were calculated from the pseudo first-order decay of the ATD’s as a function of the neutral number density. Rate constant determinations were carried out at reduced temperatures in order to resolve the two Au+ states via ESC. Examples of kinetic plots for all four reactions are given in

Figure 3. These plots clearly illustrate the significant difference in reaction efficiency exhibited by both Au+ states with the two neutrals examined here. Indeed, the exceedingly low rates for the CH3F reactions restricted the observable kinetic dynamic range as a result of instrumental limitations. Ground state rate constants were also determined at room temperature using a Xe discharge to ensure that Au+(3D) was not present during the reaction. Also provided is a comparison of these measured rate constants to the theoretical limiting rate predicted by the parametrized average dipole orientation (ADO) model.39 Singlet Reactions. Analysis of the products of the reactions of Au+(1S) with CH3F and CH3Cl can be summarized in the generalized reaction sequence given in Scheme 1. There is strong evidence provided by our results and those of others that primary association occurs in parallel with HX elimination resulting in the formation of AuCH2+ as is described in Scheme 1. Product correlations, such as that given in Figure 2, indicate that both AuCH2+ and Au+•CH3X arise directly from the ground state for both neutrals. This is in agreement with a previously reported result for the reaction of singlet Au+ with CH3F under similar experimental conditions in which association was observed to compete with HF elimination.28 Similarly, AuCH2+ has been previously shown to occur as a primary product in the Au+/CH3Cl system.26 946

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

Our room temperature rate constant of 4.5  10 11 is approximately five times larger than the value of 8.9  10 12 reported by Zhao et al.28 This is not unexpected given the difference in pressures in the two experiments (3.5 Torr in the drift cell as opposed to 0.35 Torr in the flowtube) since the association product channel is dependent on the frequency of relaxing collisions with He. The overall inefficiency of this reaction implies that the initial Au+(1S)/CH3F interaction is weak and that unimolecular decay of the orbiting complex back to reactants competes effectively with collisional stabilization by He and with access to the HF elimination channel. Li et al. have calculated the exothermicity of HF elimination via Au+(1S) to be as large as 45 kJ/mol.31 This in good agreement with the value of 44 kJ/mol predicted by bond-additivity using a previously calculated Au+ CH2 binding energy of 393 kJ/mol (see Table 2).24 We thus conclude that the inefficiency in HF elimination cannot be the result of unfavorable overall reaction energetics but is instead due to some combination of a weakly bound encounter complex and a possible kinetic barrier in the exit channel. Although the same product channels are exhibited by the Au+(1S)/CH3Cl system as in the Au+(1S)/CH3F system, there are distinct differences in both the distribution of the products and in the efficiency of their formation. Under the conditions extant in the drift cell, formation of AuCH2+ competes effectively with association such that the product distribution is 50/50, thus indicating that there is no kinetic preference for either. Pressuredependent studies were not carried out here; however, we would expect this product distribution to be influenced by pressure since the association product is dependent on the number of stabilizing collisions with He. These two primary products have been reported previously for this reaction when carried out in an ICR cell.26 In that study however, HCl elimination to form AuCH2+ dominated the product distribution due to the low pressure reaction environment presented by the ICR apparatus. Reaction of Au+(1S) is also significantly more efficient here than with CH3F. Rate constants measured at 307 and 218 K for depletion of Au+(1S) by CH3Cl indicate that this reaction is proceeding at the ADO limit at both temperatures. Taken together with the fact that HCl elimination competes effectively with association, this suggests that there are no kinetic barriers regulating access to the HCl elimination channel in excess of the reactant energies. The experimental results described above are not inconsistent with recent computational models proposed for these singlet reactions.30,31 These models, which utilize a combination of density functional and coupled-cluster methods, predict that the most stable species with either neutral reactant is one in which Au+(1S) is associated directly with the halogen atom (frontal association). For both methyl halides, the binding is predicted to be largely electrostatic, which in the case of CH3Cl, agrees with previous experimental structural determinations based on ligand exchange reactions.26 The theoretical models also predict that the Au+(1S)•CH3Cl interaction is enhanced through a donor acceptor interaction in which p electrons on the Cl are donated to the vacant 6s orbital of Au+. As a consequence, the Au+(1S)•CH3Cl binding energy is calculated to be substantially larger than that of Au+(1S)•CH3F (196.1 kJ/mol vs 111.4 kJ/ mol). We would expect the increased binding in Au+(1S)•CH3Cl to contribute to the more rapid association rate we observe when X = Cl by increasing the likelihood that the initial encounter complex will be collisionally stabilized. In addition, Li et al.

Figure 3. Pseudo first-order kinetic decay plots for the reaction of CH3F at 226 K (squares) and CH3Cl at 218 K (circles) with Au+(1S) (closed symbols) and Au+(3D) (open symbols). E/N = 6.3 Td (CH3F) and 6.1 Td (CH3Cl).

Scheme 1. Generalized Reaction Sequence for Au+(1S) + CH3X

Parallel product formation of this nature is consistent with a process in which the initial encounter complex can either unimolecularly decay into the HX elimination channel or be relaxed via collisions with the He bath gas and subsequently detected. In addition to the two primary products, higher-order reactions are exhibited by one or both methyl halides consisting of secondary HX elimination and secondary and tertiary associations. In the case of CH3F, association is the major product, representing approximately 79% of the total product signal when the discharge is adjusted to yield Au+(1S) exclusively. Given the very low efficiency of this reaction, we believe that this product distribution is in reasonable agreement with the 88%/12% distribution measured previously in a flowtube at 295 K and a pressure of 0.35 Torr.28 In that work, Au+ ions produced in an ICP torch were thought to be largely thermalized via collisional deactivation prior to reaction. Our results suggest that this is a reasonable assertion, given that no evidence of CH2F+ was reported in the flowtube experiments. As we will discuss later, we have observed this to be a product arising exclusively from Au+(3D) in this reaction. Depletion of Au+(1S) with this neutral is inefficient with an apparent two-body rate constant representing ∼3% of the limiting rate predicted by the ADO model. 947

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

Scheme 2. Au+(3D) + CH3F Reaction Sequence

predict that access to the HX elimination channel via the frontal association complex is kinetically prohibited for both neutrals.30,31 Thus, we conclude that the frontal geometry is representative of the collisionally stabilized association species that we observe. These theoretical models also indicate that association can occur via the interaction of Au+(1S) with one of the methyl hydrogens; however, the binding energy for this backside interaction (62.4 kJ/mol for CH3F and 65.9 kJ/mol for CH3Cl) is predicted to be significantly weaker than that for frontal association. It is likely that both the frontal and backside structures are sampled by the energized encounter complex; however, HX elimination with both neutrals is predicted to be kinetically possible only from the backside geometry, which subsequently leads to the oxidative addition of Au+ to the C H bond. This is then followed by reductive elimination of HX, which proceeds via a four-center transition state representing the highest energetic barrier between reactants and products for either neutral reactant. In the case of CH3Cl, the energy of this transition state is calculated to be 30.4 kJ/mol below the reactant energy and thus represents little or no kinetic restriction to the formation of AuCH2+. Conversely, the hypothesized four-center transition state on the Au+(1S)/CH3F surface is calculated to lie 5.4 kJ/mol above the reactant asymptote, thus limiting access to the HF elimination channel. This being the case, we would expect formation of AuCH2+ to exhibit a positive temperature dependence; however, we note no significant difference in the depletion rates of Au+(1S) at 307 and 226 K. This could simply be an indication that this temperature differential is too small or that enhancements in the bimolecular channel due to thermal activation are offset by concomitant increases in the rate of dissociation back to reactants. The apparent lack of temperature dependence notwithstanding, we nonetheless conclude that the kinetic behavior that we observe with respect to the formation of AuCH2+ is consistent with the relative magnitudes of the activation barriers predicted by Li et al. for HX elimination in the reactions of Au+(1S) with these two molecules.30,31 Table 1 also lists several higher-order products observed in both singlet reactions at higher extents of reaction. These include formation of Au+•(CH3X)2 as well as secondary HX elimination via AuCH2+ to yield AuC2H4+. Secondary association was reported in the previously noted flowtube study of the Au+(1S)/CH3F system, whereas AuC2H4+ was not.28 In the work reported here, the product sequence that includes AuC2H4+ represents only a small portion of the total product ion signal with both methyl halides (∼16% for CH3F and ∼2% for CH3Cl). Formation of AuC2H4+ in the Au+(1S)/CH3Cl system was observed in an earlier study in an ICR cell; however, no structural determination was reported.26 Structural determinations are not possible in our apparatus; however, if we entertain the possibility that the AuC2H4+ species is a Au+ ethene structure, this would indicate that carbon carbon bond formation has been facilitated by the metal center. Given a previously reported Au+ ethene binding energy of ∼272 kJ/mol,40 bond-additivity calculations predict that formation of AuC2H4+ via AuCH2+ is exothermic by approximately 250 and 222 kJ/mol for CH3F and CH3Cl, respectively, and thus cannot be excluded on the basis of energetics. With CH3F, AuC2H4+ is itself consumed in a tertiary association process to yield AuC2H4+•(CH3F); however, this tertiary association was not observed with CH3Cl. Triplet Reactions. As seen in Table 1, the behavior exhibited by Au+(3D) is quite different for the two methyl halides examined here. A proposed sequence of reactions summarizing

Figure 4. Au+ and AuCH2+ ATD’s for the reaction of Au+(1S,3D) with CH3F. All ATD’s are fit to Gaussians. T = 170 K; E/N = 4.9 Td; XCH3F = 2.7  10 4.

product formation for the Au+(3D)/CH3F system is given in Scheme 2. Interestingly, AuCH2+ is again identified as a primary product. Evidence for this assignment is seen in the product correlation ATD given in Figure 4, which clearly indicates that some portion of this product originates from Au+(3D). Elimination of HF to yield singlet AuCH2+ via Au+(3D) represents a formal violation of conservation of spin. This requires that the two reaction surfaces become coupled through some intermediate species or by a reversal of the relative energies of the two spin states at some point along the reaction coordinate. Numerous examples of such surface couplings have been discussed, although these occur more frequently in endothermic processes than in exothermic ones.10,13,41 However, given the large spin orbit effects in third-row ions, the possibility of interaction between the singlet and triplet surfaces cannot be disregarded out of hand. An obvious possibility that must be considered is the quenching of Au+(3D) to Au+(1S) induced by CH3F. If occurring, the nascent Au+(1S) could go on to yield AuCH2+, while still correlating to the Au+(3D) arrival time feature. By necessity, such a process must occur early on the reaction coordinate (presumably in the initial encounter complex) and would yield some portion of the Au+(1S) population, which is not consumed. These relaxed Au+(1S) ions would exhibit intermediate flight 948

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

times in the drift cell indicated by bridging in the Au+ ATD. We observe no evidence of such bridging and thus conclude that quenching of Au+(3D) by CH3F does not take place to any significant extent. There are likewise reasonable arguments to suggest that the reaction surfaces are not coupled later in the HF elimination mechanism. Not the least of these is the fact that the ordering of the Au+ states does not satisfy the requirement for two-state reactivity that the high-spin state be lower in energy than the low-spin state.41 Not only is the triplet surface energetically higher than the singlet with respect to the reactants, recent calculations predict at least two triplet states for AuCH2+, which lie 1.31 and 3.42 eV higher than the singlet for this species.25 Thus, the two reaction surfaces do not cross as a result of a reversal of the ordering of the reactant and product states. Further, the widely accepted mechanism for formation of AuCH2+ from methane and methane analogues is one in which Au+ forms an insertion intermediate with the substrate molecule. We would expect that the triplet state of such an insertion intermediate should lie higher in energy than its singlet counterpart as a result of unpairing two electrons. Finally, previous examinations of the reactions of Au+(1S,3D) with CH425 and CH3Br29 have suggested that the singlet and triplet reaction surfaces associated with H2 and HBr elimination remain isolated from one another. It therefore seems unlikely that the two reaction surfaces become coupled through some intermediate. We therefore conclude that the AuCH2+ formed via Au+(3D) is in an excited triplet state in a process that is formally spin-allowed. Excited state carbene production has been previously reported in the reaction of Au+(3D) with CH4 in which 3AuCH2+ is formed in a 3B1 state at an energy of 1.31 eV above the ground state .25 If we assume that 3AuCH2+ is likewise formed in the in 3B1 state in our work, the process is exothermic with respect to all three Au+(3D) spin orbit states (see Table 2) and thus meets the energetic requirements imposed by our experiment. The only other product in the Au+(3D)/CH3F system formed in significant amounts is CH2F+. Energetically, this product ion is only possible from the excited state. This is in fact confirmed when product spectra are compared in which Au+ is produced in Xe and Ne discharges. As indicated in Table 2, direct formation of CH2F+ via hydride abstraction by Au+(3D) is formally spinforbidden. In light of the arguments made above in support of isolated reaction surfaces, it seems more likely that CH2F+ is formed in a later spin-allowed step from some intermediate. However, no immediate precursors to this ion could be identified by direct observation. We do note the presence of CH3+•CH3F, which indirectly indicates that CH3+ is formed at some point in the reaction sequence. On the basis of available bond strengths and ionization energies, the reaction thermochemistry for hydride abstraction by CH3+ is 106 to 63 kJ/mol. (The uncertainty in the energetics for this reaction is due to a range of reported values for the ionization energy of CH2F.) Indeed, CH3+ has been demonstrated previously to abstract hydride from CH3F under low energy conditions.42 In addition, direct production of CH2F+ by CH3+ was independently verified in our own apparatus via experiments in which CH3+ was produced by electron impact prior to injection into the drift cell. Thus, we infer that CH3+ is the immediate precursor ion yielding CH2F+ as a higher-order product. The source of CH3+ is, however, itself unclear. As with CH2F+, formation of this ion is thermochemically possible only from the 3D state of Au+; however, direct fluoride abstraction is again spin-forbidden, suggesting the possibility that

Scheme 3. Au+(3D) + CH3Cl Reaction Sequence

some other (as yet unidentified) short-lived primary product ion is the precursor to CH3+. Kinetics for the Au+(3D) /CH3F system were possible only at 226 K in order to independently measure the decay of the two Au+ states via ESC. As was the case for Au+(1S), depletion of Au+(3D) by CH3F is inefficient, occurring at approximately 2% of the ADO limit. The fact that we do not observe association as a significant product of Au+(3D) even at the reduced temperature is a clear indication that the interaction between Au+(3D) and CH3F in the initial encounter complex is very weak. This undoubtedly contributes to the low rates we observe by facilitating rapid unimolecular decay back to reactants. However, kinetic barriers in the bimolecular product channels shown in Scheme 2 may also play a role in reducing the overall reaction efficiency. Temperature-dependent kinetic determinations would shed further light on the energies of these activation barriers relative to the reactants; however, such studies were not carried out here due to instrumental limitations. Because the process ultimately leading to CH2F+ is not clear, it is difficult to speculate regarding the mechanistic origin giving rise to an activation barrier in that product channel. However, it seems reasonable to postulate that a kinetic barrier to HF elimination exists on the triplet surface in a manner similar to that which has been proposed for H2 elimination by Au+(3D) from CH3F,31 CH4,25 and CH3Cl.30 Overall reaction energetics notwithstanding (formation of AuCHF+ is endothermic), a three-center transition state representing hydrogen atom migration to the Au center is calculated to be the rate determining structure in these reactions. One could envisage similar transition states for HF elimination on the triplet surface, which arise from initial activation of a either a C H bond (followed by F migration) or the C F bond (followed by H migration). Products in the Au+(3D)/CH3Cl system are summarized by the reaction sequence shown in Scheme 3 where three parallel product channels are observed. Of these, charge-transfer and methyl abstraction cumulatively represent less than 4% of the total products, while the bulk of the reactant ions initiate a series of steps yielding CH2Cl+ as the terminal bimolecular product. The observation of CH3Cl+ indicates the presence of Au+ excited states at least as high as 3D2, which is consistent with the energetic limits we have previously observed with the glow discharge.36 Having said this, the small amounts of chargetransfer observed here suggest that the majority of the excited Au+ is in the lower 3D3 spin orbit state. We have also previously reported formation of AuCH3+ under similar experimental 949

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A conditions for the reaction of Au+(3D) with CH3Br.29 In that reaction, AuBr+ was also observed as a parallel product; however, the analogous reaction to yield AuCl+ is absent here. Brown et al. have calculated the Au+ Cl bond dissociation energy to be 130.5 kJ/mol using ab initio calculations carried out at the CCSD(T) level of theory.27 Using this value, we see that formation of AuCl+ from CH3Cl is endothermic from both excited Au+ spin orbit states, which are believed to be present, and is therefore inaccessible in our instrument on that basis. We note that the lack of production of AuCl+ in our apparatus is consistent with previous results reported by others for this reaction carried out under similar low-energy conditions.26,27 The formation of AuCHCl+ is also notably absent in the reactions initiated by Au+(3D). This product is predicted computationally to arise from an overall exothermic process in which a C H bond is initially activated, followed by H atom migration to the metal center and subsequent elimination of H2.30 No activation barriers in excess of the reactants are indicated in the computational study. Thus, our failure to observe this product represents an apparent inconsistency with this theoretical description. As noted above, the major reaction sequence initiated by Au+(3D) with CH3Cl to yield CH2Cl+ is analogous to the behavior observed with CH3F. CH2Cl+ can only arise via Au+(3D) under the energetic requirement of exothermicity. This fact is clearly demonstrated in Figure 1. Having said this, we believe that direct hydride abstraction by Au+(3D) is unlikely. As with the analogous process with CH3F, direct formation of CH2Cl+ in this reaction is spin-forbidden. We also note that the correlation of CH2Cl+ to the Au+(3D) ATD feature is poor, giving further weight to the idea that it is formed in a secondary or tertiary step via some intermediate species. We believe that CH3+ is again the most likely candidate for the immediate precursor to CH2Cl+ and (unlike the CH3F reaction) is in fact observed directly in product spectra for this reaction. Further, hydride transfer from CH3Cl to CH3+ has been previously demonstrated33,42 and is exothermic by approximately 120 kJ/mol.33 Here again, the source of CH3+ is uncertain since direct chloride abstraction by Au+(3D) is spinforbidden if both CH3+ and AuCl are formed in their singlet ground states. Formal conservation of spin would be satisfied if one of these products is formed in an excited triplet state; however, this possibility is precluded due to unfavorableQenergetics. On the basis of spectroscopic measurements, the 3 first excited state of AuCl has been determined to lie 228.2 kJ/mol above the ground state. This excitation energy is large enough to render chloride abstraction endothermic from all three Au+(3D) spin orbit states.43 Likewise, the 364 kJ/mol singlet f triplet excitation energy for CH3+ also results in unfavorable chloride abstraction energetics.44 This suggests that some other unobserved species is the immediate precursor to CH3+ in the sequence of steps initiated by Au+(3D). Kinetic measurements at 218 K indicate a reaction rate representing 69% of the ADO limit, indicating no significant energetic impediment to product formation overall. As we have stated, we believe that the small amount of charge-transfer observed here reflects a limited Au+(3D2) population and thus does not necessarily inform us as to the relative kinetic favorability of this product channel. Conversely, since formation of AuCH3+ and the process resulting in CH2Cl+ are both exothermic from all Au+(3D) spin orbit states, the relative product distribution between the two must reflect a kinetic preference for the hydride abstraction product to the extent that it dominates product formation. As such, the efficiency

ARTICLE

in the formation of CH2Cl+ is the dictating factor in the efficiency of the overall reaction.

’ CONCLUSIONS Reactions of Au+(1S) and Au+(3D) with CH3F and CH3Cl have been carried out in a drift cell at low interaction energy at both room temperature and reduced temperatures. With both methyl halides, two parallel product channels are observed to arise from Au+(1S) corresponding to the association and elimination of HX. In this regard, the behavior of the singlet reactions is entirely consistent with that which has been reported previously by us and others for these and related reactions under similar energetic conditions. The reactions of Au+(3D) with these methyl halides results in a variety of bimolecular processes dependent on the neutral reactant. Most notably, HX elimination is observed to occur in the reaction of Au+(3D) with CH3F. We believe this occurs entirely on the excited surface resulting in the formation of 3AuCH2+. In addition, Au+(3D) initiates a second reaction sequence with CH3F, which ultimately results in the formation of CH2F+. A similar process occurs in the Au+(3D)/ CH3Cl system to yield CH2Cl+. Rather than arising directly from Au+(3D), we believe that these two hydride abstraction products are the result of higher-order reactions. AuCH3+ and CH3Cl+ are also observed to occur as minor products arising from the reaction of Au+(3D) with CH3Cl. All products that could be unambiguously identified as arising directly from one of the Au+ states could be rationalized on the basis of the available energetics and conservation of spin. State-specific kinetic determinations at reduced temperatures reveal that the depletion of both Au+ states by CH3F occurs at less than 5% of the ADO limit. With respect to Au+(1S), this inefficiency is consistent with the idea that the initial interaction complex is weakly bound and that some form of kinetic barrier to reaction occurs in the exit channel for HF elimination. HF elimination on the triplet surface is likewise kinetically inhibited. Reaction of both Au+ states with CH3Cl occurs at or near the ADO limit, indicating no significant energetic barriers in excess of the reactant energies on either surface. These results yield no evidence of coupling between the singlet and triplet reaction surfaces with either neutral. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT Support for this research was provided by the National Science Foundation under Grant Nos. CHE-0078771 and CHE-0956393. ’ REFERENCES (1) Armentrout, P. B.; Beauchamp, J. L. Acc. Chem. Res. 1989, 22, 315–321. (2) Armentrout, P. B. Science 1991, 251, 175–179. (3) Eller, K. E.; Schwarz, H. Chem. Rev. 1991, 91, 1121–1177. (4) Armentrout, P. B. Int. J. Mass Spectrom. Ion Processes 2003, 227, 289–302. (5) Armentrout, P. B. Acc. Chem. Res. 1995, 28, 430–436. (6) Armentrout, P. B. Gas Phase Inorg. Chem 1989, 1–42. (7) Weisshaar, J. C. Acc. Chem. Res. 1993, 26, 213–219. (8) Armentrout, P. B. Annu. Rev. Phys. Chem. 1990, 41, 313–344. (9) Weisshaar, J. C. Adv. Chem. Phys. 1992, 82 (Pt. 1), 213–262. 950

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951

The Journal of Physical Chemistry A

ARTICLE

(10) Schwarz, H. Int. J. Mass Spectrom. 2004, 237, 75–105. (11) Schwarz, H. Angew. Chem., Int. Ed. 2003, 42, 4442–4454. (12) Metz, R. B. Int. Rev. Phys. Chem. 2004, 23, 79–108. (13) Roithova, J.; Schr€oder, D. Chem. Rev. 2010, 110, 1170–1211. (14) Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. J. Catal. 1989, 115, 301–309. (15) Haruta, M.; Date, M. Appl. Catal., A 2001, 222, 427–437. (16) Laoufi, I.; Saint-Lager, M.-C.; Lazzari, R.; Jupille, J.; Robach, O.; Garaudee, S.; Cabailh, G.; Dolle, P.; Cruguel, H.; Bailly, A. J. Phys. Chem. C 2011, 115, 4673–4679. (17) Guzman, J.; Gates, B. C. J. Am. Chem. Soc. 2004, 126, 2672– 2673. (18) Guzman, J.; Gates, B. C. J. Phys. Chem. B 2002, 106, 7659–7665. (19) Johnson, G. E.; Reilly, N. M.; Tyo, E. C.; Castleman, A. W., Jr. J. Phys. Chem. C 2008, 112, 9730–9736. (20) B€urgel, C.; Reilly, N. M.; Johnson, G. E.; Mitric, R.; Kimble, M. L.; Castleman, A. W., Jr.; Bonacic-Koutecky , V. J. Am. Chem. Soc. 2008, 130, 1694–1698. (21) Hughes, M. D.; Xu, Y.-J.; Jenkins, P.; McMorn, P.; Landon, P.; Enache, D. I.; Carley, A. F.; Attard, G. A.; Hutchings, G. J.; King, F.; Stitt, E. H.; Johnston, P.; Griffin, K.; Kiely, C. J. Nature 2005, 437, 1132–1135. (22) Irikura, K. K.; Beauchamp, J. L. J. Am. Chem. Soc. 1991, 113, 2769–2770. (23) Irikura, K. K.; Beauchamp, J. L. J. Phys. Chem. 1991, 95, 8344–8351. (24) Irikura, K. K.; Goddard, W. A., III. J. Am. Chem. Soc. 1994, 116, 8733–8740. (25) Li, F.-X.; Armentrout, P. B. J. Chem. Phys. 2006, 125, 133114–133126. (26) Chowdhury, A. K.; Wilkins, C. L. J. Am. Chem. Soc. 1987, 109, 5336–5343. (27) Brown, J. R.; Schwerdtfeger, P.; Schr€oder, D.; Schwarz, H. J. Am. Soc. Mass Spectrom. 2002, 13, 485–492. (28) Zhao, X.; Koyanagi, K.; Bohme, D. K. J. Phys. Chem. A 2006, 110, 10607–10618. (29) Taylor, W. S.; May, J. C.; Lasater, A. S. J. Phys. Chem. A 2003, 107, 2209–2215. (30) Li, T. H.; Liu, X. Y.; Yu, S. W.; Zhao, N.; Liu, S. Q.; Ao, X. Y.; Xie, X. G. J. Mol. Struct. 2009, 899, 18–24. (31) Li, T. H.; Wang, C. M.; Yu, S. W.; Liu, X. Y.; Li, X. H.; Xie, X. G. Chem. Phys. Lett. 2008, 463, 334–339. (32) Li, T. H.; Wang, C. M.; Liu, X. Y.; Xie, X. G. Chin. Chem. Lett. 2008, 19, 881–884. (33) Taylor, W. S.; Matthews, C. C.; Parkhill, K. S. J. Phys. Chem. A 2005, 109, 356–365. (34) Taylor, W. S.; Campbell, A. S.; Barnas, D. F.; Babcock, L. M.; Linder, C. B. J. Phys. Chem. A 1997, 101, 2654–2661. (35) Taylor, W. S.; Everett, W. R.; Babcock, L. M.; McNeal, T. L. Int. J. Mass Spectrom. Ion Processes 1993, 125, 45–54. (36) Taylor, W. S.; Spicer, E. M.; Barnas, D. F. J. Phys. Chem. A 1999, 103, 643–650. (37) Kemper, P. R.; Bowers, M. T. J. Phys. Chem. 1991, 95, 5134–5146. (38) van Koppen, P. A. M.; Kemper, P. R.; Bowers, M. T. In Organometallic Ion Chemistry; Freiser, B. S., Ed.; Kluwer Academic Publishers: Boston, MA, 1996. (39) Su, T.; Bowers, M. T. Int. J. Mass Spectrom. Ion Phys. 1973, 12, 347–356. (40) Sch€oder, D.; Schwarz, H.; Hrusak, J.; Pyykk€o, P. Inorg. Chem. 1998, 37, 624–632. (41) Sch€oder, D.; Shaik, S.; Schwarz, H. Acc. Chem. Res. 2000, 33, 139–145. (42) Henis, J. M. S.; Loberg, M. D.; Welch, M. J. J. Am. Chem. Soc. 1974, 96, 1665–1671. (43) O’Brien, L. C.; Elliot, A. L.; Dulick, M. J. Mol. Spectrosc. 1999, 194, 124–127. (44) Ignatyev, I. S.; Schaefer, H. F.; Schleyer, P. V. Chem. Phys. Lett. 2001, 337, 158–168. 951

dx.doi.org/10.1021/jp2093912 |J. Phys. Chem. A 2012, 116, 943–951