Oxidation of Atomic Gold Ions: Thermochemistry for the Activation of

Mar 22, 2010 - molecular oxygen adsorption.17-20 Oxidation of small gold clusters by ... generated in a direct current discharge flow tube source desc...
0 downloads 0 Views 1018KB Size
J. Phys. Chem. A 2010, 114, 11043–11052

11043

Oxidation of Atomic Gold Ions: Thermochemistry for the Activation of O2 and N2O by Au+ (1S0 and 3D)† Feng-Xia Li,‡ Katrine Gorham,§ and P. B. Armentrout* Department of Chemistry, UniVersity of Utah, 315 S. 1400 E. Rm 2020, Salt Lake City, Utah 84112-0850 ReceiVed: January 20, 2010; ReVised Manuscript ReceiVed: March 4, 2010

Reaction of Au+ (1S0 and 3D) with O2 and N2O is studied as a function of kinetic energy using guided ion beam tandem mass spectrometry. A flow tube ion source produces Au+ primarily in its 1S0 (5d10) electronic ground state level but with some 3D and perhaps higher lying excited states. The distribution of states can be altered by adding N2O, which completely quenches the excited states, or CH4 to the flow gases. Cross sections as a function of kinetic energy are measured for both neutral reagents and both ground and excited states of Au+. Formation of AuO+ is common to both systems with the N2O system also exhibiting AuN2+ and AuNO+ formation. All reactions of Au+ (1S0) are observed to be endothermic, whereas the excitation energy available to the 3D state allows some reactions to be exothermic. Because of the closed shell character of ground state Au+ (1S0, 5d10), the reactivity of these systems is low and has cross sections with onsets and peaks at higher energies than expected from the known thermochemistry but lower than energies expected from impulsive processes. Analyses of the endothermic reaction cross sections yield the 0 K bond dissociation energy (BDE) in eV of D0(Au+-O) ) 1.12 ( 0.08, D0(Au+-N2) g 0.30 ( 0.04, and D0(Au+-NO) ) 0.89 ( 0.17, values that are all speculative because of the unusual experimental behavior. Combining the AuO+ BDE measured here with literature data also yields the ionization energy of AuO as 10.38 ( 0.23 eV. Quantum chemical calculations show reasonable agreement with the experimental bond energies and provide the electronic structures of these species. 1. Introduction A key virtue of metallic gold is its low reactivity, particularly with regard to oxidation, allowing even gold leaf to retain its luster for years. The low reactivity of gold is evidenced even at the atomic level. For instance, we have demonstrated in our laboratory that C-H bond activation by Au+ requires more energy than any other third-row transition metal cation,1 consistent with previous studies at thermal energies.2-4 However, recent studies show that the reactivity of gold becomes enhanced for small clusters (about 10 atoms or so), for instance in the activation of H25 or oxidation of CO,6-9 a phenomenon also explored theoretically.10-12 Castleman and co-workers have explored the oxidation of CO by gold oxide cations and anions, finding that the anions are more reactive and that cations with an odd number of gold atoms are more reactive than those with even numbers.13-16 In reactions with O2, neutral and cationic gold clusters are generally unreactive (except for Au10+), whereas anionic clusters with an even number of atoms exhibit molecular oxygen adsorption.17-20 Oxidation of small gold clusters by molecular oxygen has been studied theoretically using density functional theory (DFT),21 as well as more advanced theoretical methods.22 The latter study specifically comments on the fact that DFT approaches generally produce only qualitative agreement with experiment, whereas CCSD(T) methods perform more accurately. Finally, gold nanoparticles have been shown to be effective at activating molecular oxygen and then oxidizing hydrocarbons.23 Insight into the interaction of gold metal with O2 can be obtained by examining relevant reactions in the gas phase using †

Part of the “Klaus Mu¨ller-Dethlefs Festschrift”. Present address: Covance Laboratories Inc., Madison, WI 53704. § Present address: Polar Field Services, Littleton, CO 80127. ‡

guided ion beam tandem mass spectrometry (GIBMS). The gas phase is an ideal arena for detailed study of the energetics of bond-making and bond-breaking processes at a molecular level. Because metal supports and interactions are absent, quantitative thermodynamic and intrinsic mechanistic information for various bond activation processes can be obtained. Such insight may be useful in better understanding under what circumstances gold is reactive. Previously we have used GIBMS to systematically study the oxidation of atomic metal cations of the first-row,24-32 second-row,26,30,33-38 and third-row39,40 transition series, and other metals.30,41-43 In the present work, we extend these studies to the third-row transition metal ion, Au+. Here, we examine the kinetic energy dependence of the oxidation reaction with both O2 and N2O of this atomic ion in two electronic states. Analyses of such data provide experimental BDEs that can be used as benchmarks for comparison with theoretical models of the structure and bonding of AuO+. The present work provides thermodynamic, dynamic, and mechanistic information for the activation of O2 and N2O by Au+. In addition, this study is part of ongoing efforts in our laboratory to understand the periodic trends in the BDEs of metal oxides.44-46 2. Experimental Section 2.1. General Procedures. The guided ion beam tandem mass spectrometer on which these experiments were performed has been described in detail previously.47,48 Briefly, Au+ ions are generated in a direct current discharge flow tube source described below, extracted from the source, accelerated, and focused into a magnetic sector momentum analyzer for mass selection of primary ions. The mass-selected ions are then decelerated to a desired kinetic energy and focused into an octopole ion beam guide that uses radio frequency electric fields to trap the ions in the radial direction and ensure complete

10.1021/jp100566t  2010 American Chemical Society Published on Web 03/22/2010

11044

J. Phys. Chem. A, Vol. 114, No. 42, 2010

Li et al.

collection of reactant and product ions.49,50 The octopole passes through a static gas cell with an effective length of 8.26 cm that contains the reaction partner at a low pressure (usually less than ∼0.3 mTorr) so that multiple ion-molecule collisions are improbable. All results reported here result from single bimolecular encounters, as verified by pressure dependence studies. The unreacted parent and product ions are confined radially in the guide until they drift to the end of the octopole, where they are extracted, focused, and passed through a quadrupole mass filter for mass analysis of products. Ions are subsequently detected with a secondary electron scintillation ion detector using standard pulse counting techniques. Reaction cross sections are calculated from product ion intensities relative to reactant ion intensities after correcting for background signals.51 Uncertainties in the absolute cross sections are estimated to be (20%. The kinetic energy of the ions is varied in the laboratory frame by scanning the dc bias on the octopole rods with respect to the potential of the ion source region. Laboratory (lab) ion energies are converted to energies in the center-of-mass frame (CM) by using the formula ECM ) Elabm/(m + M), where m and M are the neutral and ionic reactant masses, respectively. Two effects broaden the cross section data: the kinetic energy distribution of the reactant ion and thermal motion of the neutral reactant gas (Doppler broadening).52 The absolute zero and the full width at half-maximum (fwhm) of the kinetic energy distribution of the reactant ions are determined using the octopole beam guide as a retarding potential analyzer, as described previously.51 The distributions of ion energies, which are independent of energy, are nearly Gaussian and have typical fwhm between 0.6 and 1.0 eV (lab) in these studies. Uncertainties in the absolute energy scale are (0.05 eV (lab). 2.2. Ion Source. Au+ ions are produced in a direct current discharge flow tube (DC/FT) source,48 consisting of a cathode held at high negative voltage (0.7-1.5 kV) over which a flow of approximately 90% He and 10% Ar passes at a total pressure of 0.3-0.4 Torr and ambient temperature. Ar+ ions created in the discharge are accelerated toward a gold-plated aluminum cathode, thereby sputtering Au+ ions, which are swept down a one-meter long flow tube. The ions undergo ∼105 thermalizing collisions with He and ∼104 collisions with Ar before entering the guided ion beam apparatus. Excited states of Au+ are observed to survive these flow conditions, as found in previous work on the reactions of Au+ with CH4,1 and in the present study. For Au+, the ground state is 1S0 (5d10) with a first and second excited state of 3D (6s15d9) and 1D (6s15d9). Spin-orbit splitting energies are appreciable such that the 3D3, 3D2, 3D1, and 1D2 levels are higher in energy than the ground state by 1.865, 2.187, 3.443, and 3.673 eV, respectively.53 In our previous study,1 we found that excited species having excitation energies above about 2.5 eV (i.e., all but 3D3 and 3D2) are removed by introducing CH4 to the flow tube about 15 cm downstream of the discharge zone at a pressure of ∼100 mTorr, although populations of the remaining excited states varied appreciably with different amounts of CH4, as demonstrated below. Complete quenching of all excited states can be achieved by the addition of N2O as a cooling gas. This relies on the fact that all states of Au+ except the 1S0 ground state react exothermically with N2O to form AuO+ + N2, as confirmed below. 2.3. Data Analysis. To determine the energy threshold for product formation at 0 K, E0, endothermic reaction cross sections are modeled using eq 1,54-57

σ(E) ) σ0

∑ gi(E + Eel + Ei - E0)n/E

(1)

where σ0 is an energy-independent scaling factor, E is the relative kinetic energy of the reactants, Eel is the electronic energy of the Au+ reactant ion, and n is an adjustable parameter that characterizes the energy dependence of the process.58 The sum considers contributions from rovibrational states of the reactants at 300 K, denoted by i, having energies Ei and populations gi, where Σgi ) 1. The various sets of vibrational frequencies and rotational constants used to determine Ei in this work are taken from the literature for O259 and N2O.60 Before comparison with the experimental data, eq 1 is convoluted with the kinetic energy distributions of the reactant ions and neutral reactants at 300 K. The σ0, n, and E0 parameters are then optimized using a nonlinear least-squares analysis to give the best reproduction of the data.51 Uncertainties in E0 are calculated from the range of threshold values for different data sets over a range of acceptable n values combined with the absolute uncertainties in the kinetic energy scale and internal energies of reactant ions. 2.4. Theoretical Calculations. To establish the electronic character of the product ions formed in this work and to examine the potential energy surface of the AuO2+ system, quantum chemistry calculations were computed with the B3LYP hybrid density functional method61,62 and performed with the GAUSSIAN 03 suite of programs.63 We generally start with the B3LYP functional, which is based on the hybrid gradient-corrected exchange functional proposed by Becke49 combined with the gradient-corrected correlation functional of Lee, Yang, and Parr.50 In most cases, a 6-311+G(3df) basis set is used for oxygen and nitrogen. Calculated bond energies are corrected for zero point energies with unscaled vibrational frequencies. As a point of comparison, the 0 K single point bond energies for O-O, N2-O, and N-NO calculated at the B3LYP/6311+G(3df) level are 5.277, 1.887, and 5.060 eV compared to the experimental values of 5.116 ( 0.002,59 1.672 ( 0.004, and 4.924 ( 0.005 eV.60 Three basis sets for Au were considered. Ohanessian et al.64 describe a basis set that is based on the small core (60 electron) relativistic effective core potential (ECP) and valence basis set of Hay-Wadt (HW),65 equivalent to the Los Alamos ECP (LANL2DZ) basis set. Whereas the HW-ECP is optimized for neutral atoms, the altered basis set of Ohanessian et al. (HW+) accounts for differential contraction of the s orbitals compared to the d orbitals induced by the positive charge. Three alternative basis sets for Au were also considered: the Stuttgart-Dresden relativistic small core ECP basis set of Dolg et al.66 (SDD); the SBKJC VDZ ECP basis set (SBKJC, also known as CEP121G);67 and the Def2TZVPP basis set68 (used for Au, O, and N), a balanced set of triple-ζ quality + polarization functions that also uses the SDD ECP for Au. The SBKJC basis set was augmented with one f polarization (exponent ) 0.89) and one s and one p diffuse function (both exponents ) 0.01) as per Varganov et al.22 and will be called SBKJC*. All basis sets on gold retain 19 explicit electrons. Evaluation of these basis sets is achieved by comparison to five experimental quantities. (1) The 1S-3D excitation energy of Au+, Eex(Au+) ) 2.288 eV for the statistical average of the experimental energies for the 3DJ levels.53 (2) The ionization energy of Au, IE(Au) ) 9.22553 eV.69 (3) The 0 K bond energy of AuO, D0(Au-O) ) 2.27 ( 0.22 eV.70,71 (4) The 0 K bond energy of AuH+, D0(Au+-H) ) 2.17 ( 0.08 eV.1 (5) The 0 K bond energy of AuCH2+, D0(Au+-CH2) ) 3.70 ( 0.07 eV.1 Table 1 compares these five experimental quantities with those calculated at various levels of theory. For the B3LYP functional, the SBKJC* and Def2TZVPP basis sets yield the

Oxidation of Atomic Gold Ions

J. Phys. Chem. A, Vol. 114, No. 42, 2010 11045

TABLE 1: Comparison of Experimental and Theoretical Properties of Gold (in eV) level of theory

Eex(Au+)

IE(Au)

D0(AuO)

D0(AuH+)

D0(AuCH2+)

MADa

Exp B3LYP/HW+ B3LYP/SDD B3LYP/SBKJC* B3LYP/Def2TZVPP CCSD(T,full)/HW+ CCSD(T,full)/SDD CCSD(T,full)/SBKJC* CCSD(T,full)/Def2TZVPP BLYP/SBKJC* BP86/SBKJC*

2.288 2.361 2.307 2.366 2.271 2.473 2.411 2.411 2.411 2.265 2.286

9.22553 9.390 9.443 9.271 9.413 8.090 8.404 8.573 8.942 9.358 9.550

2.27 ( 0.22 1.94 1.75 2.03 2.03 2.16 1.40 2.17 1.88 2.50 2.64

2.17 ( 0.08 2.08 1.99 2.00 2.08 1.66 1.26 1.84 1.91 2.22 2.38

3.70 ( 0.07 3.69 3.62 3.62 3.75 3.23 2.78 3.62 3.63 4.06 4.19

0.13 0.20 0.12 0.12 0.48 0.73 0.26 0.23 0.16 0.28

a

Mean absolute deviation from experimental values.

best results, followed by the HW+, with the SDD yielding the worst agreement with experiment. To ensure that this judgment is not limited to applications with the B3LYP functional, calculations at the CCSD(T,full) level were also performed for these properties with all three basis sets. CCSD(T) calculations of AuCH2+ utilize B3LYP geometries. Here, the Def2TZVPP basis set performs the best for all five properties, although it is notable that the CCSD(T) calculations do not calculate the ionization energy nearly as accurately as the B3LYP functional. Calculations at the MP2(full) and MP4(SDTQ,full) levels show trends similar to those of the CCSD(T,full) level. Wu72 has tested eight different functionals for predicting the properties of various charge states of the gold chalcogenide diatoms, using the SBKJC basis set on Au. He found that the bond dissociation energies are highly dependent on the functional used. BLYP and BP86 were concluded to yield the best agreement with available experimental values. For example, his calculations yield D0(AuO) of 1.83, 2.29, and 2.41 eV for the B3LYP, BLYP, and BP86 functionals, respectively (where we have used a theoretical vibrational frequency of 551 cm-1 to convert from De). However, if we repeat these calculations using the SBKJC* basis set (both calculations use 6-311+G(3df) on O), we find bond energies of 2.03, 2.51, and 2.64 eV, respectively. Thus, with a polarization function and diffuse functions added to the basis set, the B3LYP functional performs better with the BLYP and BP86 functionals overbinding. The same is true for the AuH+ and AuCH2+ bond energies. Calculation of Eex(Au+) and IE(Au) using the BLYP and BP86 functionals and SBKJC* basis set (Table 1) give excitation energies that improve slightly compared to B3LYP/SBKJC* values, but the ionization energies disagree with experiment by more. Overall, the B3LYP/SBKJC* and B3LYP/Def2TZVPP combinations appear to provide the best reproduction of the calibration data, with the CCSD(T)/Def2TZVPP combination being only slightly worse and the best CCSD(T) approach. 3. Experimental Results 3.1. Reaction of Au+ with O2. Figure 1 shows cross sections for the reaction of Au+ with O2 as a function of kinetic energy under different conditions in the flow tube ion source, with no quenching gas present and N2O or CH4 added. In all cases, the only product observed is AuO+, as formed in reaction 2.

Au+ + O2 f AuO+ + O

(2)

When there is no quenching gas present, Figure 1a, the cross section exhibits three features: below ∼0.5 eV is a small feature that declines with increasing energy indicating an exothermic

reaction; starting around 2 eV, there is a shoulder in the cross section; and above 4 eV, the cross section rises rapidly reaching a maximum at about 7 eV before declining. When N2O is admitted to the flow tube source, Figure 1b, the two low energy cross section features disappear, leaving only the major feature rising from an apparent threshold of ∼4 eV (which is lower than the 0 K threshold extracted from modeling described below because of the energy distributions involved). Previous work1 has shown that N2O efficiently quenches all the excited states of Au+, leaving only the 1S0 (5d10) ground state. Thus, the major cross section feature is identified as reactivity of the ground electronic state, and the two low energy features must correspond to reactions of excited electronic states. The observation that the apparent threshold in the low-energy shoulder is shifted to energies lower than that for the ground state cross section by about 2 eV is consistent with the excitation energies of the 3 D3 and 3D2 states of Au+, 1.865 and 2.187 eV, respectively. Compared to the Langevin-Gioumousis-Stevenson (LGS) collision cross section,73 which has an E-0.5 energy dependence, the exothermic reactivity observed here matches the predicted kinetic energy dependence but has a magnitude between 1000 and 4000 times smaller. This feature can therefore be attributed to excited state species comprising as little as 0.10-0.025% of the ion beam. Figure 1c shows results for reaction 2 where Au+ is formed with varying pressures of CH4 used as a quenching gas in the flow tube. As noted above, previous work indicated that methane quenches all but the lowest excited electronic states, 3D3 and 3 D2. This is consistent with the failure to observe the exothermic reactivity found in Figure 1a. However, the magnitude and shape of the cross section for reaction 2 is now very dependent on the CH4 pressure used (an observation not previously made in our laboratory for any other metal). At the lowest CH4 flow, there are two distinct endothermic features that begin at about 2 and 4 eV. This cross section is very similar in both energy dependence and magnitude to that shown in Figure 1a with no quenching gas, except that the exothermic feature has disappeared. As the amount of CH4 admitted to the source increases, the low energy shoulder increases in magnitude and the ground state feature peaking near 7 eV decreases. If we compare the magnitudes of the AuO+ cross sections for the four conditions shown in Figure 1c, we find that the amount of excited states present in the Au+ ion beam increases with CH4 pressure from about 6 to 45 to 70 to 100%, with a commensurate decrease in the amount of ground state present. We believe that the enhanced production of excited states at high methane pressures occurs because the methane begins to take an active part in the discharge, altering the conditions for ion formation, as opposed to simply reacting with the gold ions downstream of the discharge.

11046

J. Phys. Chem. A, Vol. 114, No. 42, 2010

Li et al.

Au+ + O2 f AuO+ + O f Au+ + O + O

(3)

which has a threshold equal to D0(O2). The observation that the peak of the AuO+ cross section occurs at higher energies indicates that the reaction dynamics of Au+ (1S0) with O2 favor placing excess energy in translation rather than in internal energy of the diatomic product. Similar observations have been made for the other group 10 metal cations, Cu+ and Ag+, reacting with O2.32,33 As discussed in these studies, one means by which this can happen is an impulsive process, where the energy relevant to the reaction is no longer the center-of-mass (CM) energy but a pairwise interaction energy between the incoming ion and one atom of the neutral reactant. The pairwise and CM energies are related for the generic reaction A + BC f AB + C by eq 4,

E(pair) ) E(CM)(A + B + C)B/[(B + C)(A + B)]

(4) where A, B, and C represent the masses of the corresponding species. For reaction 3, E(3,pair) ) 0.538 E(3,CM) such that the peak position in the AuO+ cross section is predicted to lie at 5.116/0.538 ) 9.51 eV in the CM frame for an impulsive pairwise interaction. Clearly the experimental data do not conform to this prediction, Figure 1b, but instead the peak position lies about midway between the thermodynamic limit and this impulsive pairwise limit. 3.3. Reaction of Au+ with N2O. Figure 2 shows cross sections for the reaction of Au+ with N2O as a function of kinetic energy at various source conditions. The reaction yields three products, AuO+, AuN2+, and AuNO+, formed in reactions 5-7.

Au+ + N2O f AuO+ + N2

(5)

f AuN2+ + O

(6)

f AuNO+ + N

(7)

Results for the reaction of Au+ with N2O obtained when the ions are quenched with N2O in the flow tube are shown in Figure 2a. Because the apparent threshold for reaction 5 is near 1 eV, this reaction is exothermic for all excited states of Au+, explaining why reaction of N2O at thermal energies leaves only the 1S0 ground state. The dominant reaction pathways observed, processes 5 and 6, result from cleavage of the weak N2-O bond, D0 ) 1.672 ( 0.004 eV,60 followed by binding one of the fragments to Au+. The peaks in the cross sections for reaction 5 and lower energy feature for reaction 6 occur at about 2 eV, somewhat higher than the thermodynamic thresholds for the overall reactions 8 and 9, equal to D0(N2-O). Figure 1. Cross sections for reaction of Au+ with O2 as a function of kinetic energy in the center-of-mass frame (lower axis) and laboratory frame (upper axis). Results are shown for Au+ produced with no quenching gases (a), N2O (b), and different pressures of CH4 (c) added to the flow tube source. Expansions of the data are shown in parts a (×10) and b (×20 and offset from zero by 0.2 Å2).

3.2. Reaction of Au+(1S0) with O2. In reaction of the Au+(1S0) ground state, the cross section peaks at ∼7 eV, Figure 1b, which is higher than the dissociation energy of O2, D0(O2) ) 5.116 eV.59 This behavior is unusual because the decline in the cross section at high energies must result from dissociation of AuO+ in the overall reaction 3,

Au+ + N2O f AuO+ + N2 f Au+ + O + N2

(8)

f AuN2+ + O f Au+ + O + N2

(9)

As for the O2 reaction, this observation may imply some impulsive character in these reactions, although the impulsive pairwise limit for these reactions are again higher, 4.06 eV for reaction 8 where E(8,pair) ) 0.411 E(8,CM) and 2.45 eV for reaction 9 where E(9,pair) ) 0.682 E(9,CM). The second feature in the cross section for reaction 6 starting about 3 eV is probably associated with dissociation of the N2-O bond along a singlet surface to form excited state O

Oxidation of Atomic Gold Ions

J. Phys. Chem. A, Vol. 114, No. 42, 2010 11047

Au+(1S0) + N2O(1Σ+) f AuN2+(1Σ+) + O(3P)

(6a)

f AuN2+(1Σ+) + O(1D)

(6b)

We also considered whether the low-energy feature for the AuN2+ product cross section might be a result of the secondary reaction 10.

AuO+ + N2O f AuN2+ + O2

(10)

Studies of the pressure dependence of this reaction do not support this possibility and verify that this cross section is the result of a single collision between Au+ and N2O. The third product observed, AuNO+, involves cleavage of the N-N bond of N2O, D0(N-NO) ) 4.924 ( 0.005 eV, and occurs at elevated kinetic energies in reaction 7. Note that the apparent threshold for this reaction in Figure 2a, >5 eV, lies aboVe this bond energy, indicating that the observed reactivity cannot be occurring at its thermodynamic threshold. Similar to the O2 system, the peak for this cross section lies at higher energies than expected for reaction 11,

Au+ + N2O f AuNO+ + N f Au+ + NO + N

(11)

Figure 2. Cross sections for reaction of Au+ with N2O as a function of kinetic energy in the center-of-mass frame (lower axis) and laboratory frame (upper axis). Results are shown for Au+ produced with N2O (a), no quenching gases (b), and CH4 (c) added to the flow tube source.

(1D), 1.967 eV above O (3P). In this regard, we note that reaction 6a is spin-forbidden, whereas reaction 6b is spinallowed, consistent with the larger magnitude of the higher energy feature.

which has a thermodynamic threshold equal to D0(N-NO). For this reaction, the peak position lies close to that predicted for an impulsive pairwise reaction, 6.80 eV as given by E(11,pair) ) 0.724 E(11,CM). The related AuN+ + NO channel was not observed in this system, despite a careful search. Figure 2b shows results for the same reaction but with Au+ formed with no quenching gas added in the flow tube. For reactions 5 and 6, the cross sections remain similar in magnitude and energy dependence to those of Figure 2a, with the addition of exothermic features in both cross sections. This is consistent with formation of electronically excited states, which all have energies exceeding the apparent thresholds of both of these reactions. For reaction 7, the ground state feature of Figure 2a remains with a similar magnitude but a distinct endothermic feature is now evident beginning at about 1.5 eV. This new feature must correspond to reactions of excited states. Note that the magnitudes of the major cross sections features in Figure 2a,b are essentially the same, consistent with the same behavior observed in Figure 1. In both cases, this indicates that the amounts of excited states present in this ion beam with no quenching gases are small. Figure 2c shows results for the reaction of Au+ with N2O when CH4 is added as a quenching gas to the flow tube. Clearly, the exothermic features in reactions 5 and 6 have grown appreciably, consistent with the increased presence of electronically excited states of Au+, whereas the endothermic cross sections have magnitudes comparable to those in Figure 2a. The low energy endothermic feature in the AuNO+ cross section has increased appreciably, again consistent with it being attributable to excited states of Au+. 4. Thermochemical and Theoretical Results for AuL+ The endothermic cross sections are analyzed in detail using eq 1 as described above and the optimum values of the fitting parameters are listed in Table 2. Because the rotational, vibrational, and translational energy distributions of reactants are explicitly included in the modeling, the E0 thresholds determined using eq 1 correspond to 0 K. From the thresholds

11048

J. Phys. Chem. A, Vol. 114, No. 42, 2010

Li et al.

TABLE 2: Parameters of Eq 1 Used in Modeling Various Reaction Systems reactants + 1

products

Au ( S0) + O2 Au+ (1S0) + N2Oa a

Au+ (3D2) + O2b ( 3 D 3) Au+ (3D2) + N2Ob ( 3 D 3)

+

n

σ0

E0 (eV)

D0(Au+-L) (eV) >0.42 ( 0.20 >0.55 ( 0.07 >0.30 ( 0.04 >–0.76 0.80 ( 0.08 1.12 ( 0.08 0.57 ( 0.17 0.89 ( 0.17

AuO + O AuO+ + N2 AuN2+ + O AuNO+ + N AuO+ + O

1.85 ( 0.42 0.56 ( 0.16 0.03 ( 0.01 0.06 ( 0.02 1.27 ( 0.21

1.6 ( 0.3 1.2 ( 0.1 1.2 (fixed) 1.0 (fixed) 1.1 ( 0.2

4.70 ( 0.20 1.12 ( 0.07 1.37 ( 0.04 5.68 ( 0.14 2.13 ( 0.08

AuNO+ + N

0.22 ( 0.19

1.1 ( 0.2

2.17 ( 0.17

a Au+ (1S0) data as quenched with N2O. b Data for Au+ quenched with CH4 or not quenched. Bond energies derived assume that the threshold corresponds to reaction of Au+ (3D2) or (3D3).

measured, the BDEs at 0 K for the different product ions can be calculated using eq 12,

D0(Au+-L) ) D0(L-R) - E0

(12)

where the D0(L-R) values are given above. This equation assumes that there are no activation barriers in excess of the endothermicity of a given reaction, an assumption that is often true for ion-molecule reactions because of the long-range attractive forces.51,56 In the present systems, however, the reaction thresholds measured may not correspond directly to their thermodynamic values as indicated by the high energies observed in the peaks of the cross sections. 4.1. Au+-O: Experiment. AuO+ is observed in the reaction of Au+ (1S0) with O2 and N2O. The thresholds obtained from analysis of these data using eq 1 are given in Table 2 and can be converted to bond dissociation energies for D0(Au+-O) of 0.42 ( 0.20 and 0.55 ( 0.07 eV, respectively, using eq 12. Although these values do agree with one another within experimental uncertainty, both reactions exhibit signs of impulsive reactivity that preferentially places energy in translational degrees of freedom, such that the thresholds are shifted to energies higher than the thermodynamic limits. Therefore, these values are best viewed as lower limits. Upper limits to the true thermodynamic values can probably be obtained by assuming that the reactions are completely impulsive, which the peak positions suggest is not quantitatively accurate. In the impulsive limit, the measured thresholds are adjusted to true thermodynamic values using the scaling factor of eq 4, 0.538 and 0.411, for reactions 2 and 5, respectively, to give adjusted thresholds of 2.53 ( 0.11 and 0.46 ( 0.03 eV, respectively. These impulsive thresholds lead to AuO+ BDEs of