J. Phys. Chem. A 2010, 114, 11285–11291
11285
Dissociation Dynamics of Energy Selected, Propane, and i-C3H7X+ Ions by iPEPICO: Accurate Heats of Formation of i-C3H7+, i-C3H7Cl, i-C3H7Br, and i-C3H7I† William R. Stevens,‡ Andras Bodi,§ and Tomas Baer*,‡ Department of Chemistry, UniVersity of North Carolina, Chapel Hill, North Carolina 27599-3290, and Molecular Dynamics Group, Paul Scherrer Institut, Villigen 5232 Switzerland ReceiVed: May 8, 2010; ReVised Manuscript ReceiVed: June 22, 2010
The dissociation dynamics of energy selected i-C3H7X (X ) H, Cl, Br, and I) ions have been investigated by imaging photoelectron-photoion coincidence (iPEPICO) spectroscopy using synchrotron radiation from the X04DB VUV beamline in the Swiss Light Source of the Paul Scherrer Institut. The 0 K dissociation energy (E0) for i-C3H8 was determined to be 11.624 ( 0.002 eV. This leads to a 298 K isopropyl ion heat of formation of 805.9 ( 0.5 kJ mol-1. The ∆fH298K°(i-C3H7+) combined with the measured 0 K onsets for i-C3H7+ formation from isopropyl chloride (11.065 ( 0.004 eV), isopropyl bromide (10.454 ( 0.008 eV), and isopropyl iodide (9.812 ( 0.008 eV) yields the 298 K isopropyl chloride, bromide, and iodide heats of formation of -145.7 ( 0.7, -95.6 ( 0.9, and -38.5 ( 0.9 kJ mol-1, respectively. These values provide a significant correction to literature values and reduce the error limits. Finally, the new i-C3H7+ heat of formation leads to a predicted adiabatic ionization energy for the isopropyl radical of 7.430 ( 0.012 eV and a 298 K proton affinity for propene of 744.1 ( 0.8 kJ mol-1. Introduction The isopropyl ion is a common product in the mass spectrometry of alkanes. However, a reliable high accuracy value of its heat of formation has yet to be determined. One route to measuring the ∆fH°(i-C3H7+) is via dissociative photoionization
i-C3H7X + hV f i-C3H7+ + X + e- (X ) H, Cl, Br, I) (1) It is known that these systems have stable parent ions that rapidly dissociate to form i-C3H7+, and that the X loss reaction is the lowest energy dissociation path, which means that their photoionization onsets are relatively easily measured. However, only the propane molecule has a well established heat of formation1 so that it is effectively the only route to establishing the propyl ion heat of formation by dissociative photoionization. Despite this, numerous studies of the alkyl halides dissociative photoionization2-7 have been used to establish the propyl ion heat of formation. However, these values have ranged from 799 to 808 kJ mol-1 with no clear trend in sight. Clearly, it would make more sense to establish the propyl ion heat of formation on the basis of the H loss from C3H8+ and then use this value to establish heats of formation of the i-propyl halides. The 298 K dissociative photoionization of C3H8 to form i-C3H7+ was first observed by Steiner et al. in 19618 and measured again by Chupka and Berkowitz9 in 1967 using photoionization mass spectrometry (PIMS). The 298 K appearance energy (AE298K) of 11.53 ( 0.01 eV reported by Steiner et al.8 appears to be quite reasonable; however, its extrapolation to 0 K (11.59 ( 0.03 eV) was based on a poorly characterized temperature dependence, as reflected by the higher uncertainty. †
Part of the “Klaus Mu¨ller-Dethlefs Festschrift”. * Corresponding author. E-mail:
[email protected]. ‡ University of North Carolina, Chapel Hill. § Paul Scherrer Institut.
Chupka and Berkowitz9 reported a similar 11.590 ( 0.010 eV 0 K onset apparently using the Steiner et al. extrapolation. The currently accepted extrapolation of such 298 K onsets involves the addition of the internal thermal energy of the molecule to the 298 K onset. In the case of propane, the sum of the rotational and vibrational energy at room temperature is 0.085 eV, which would suggest that these previous photoionization onsets should extrapolate to 11.615 eV at 0 K. Propane and its deuterated analogues has also been investigated by threshold photoelectron-photoion coincidence (TPEPICO) from 11 to 17 eV by Stockbauer and Inghram.10 However, this study focused more on the mechanism and H-atom scrambling than on the extraction of an accurate dissociation limit. The dissociative photoionization of the isopropyl halides has been reported several times, but the reported onsets have varied by as much as 85 meV, far beyond the usually claimed error limits of 0.01 eV. (For comparison with our new data, see Table 1 in the Results section.)2-6 The wide scatter in the reported onset energies, along with the uncertainty in the propyl halide heats of formation, has contributed to the current uncertainty in the i-propyl ion heat of formation. Traeger2 reported photoionization onsets in 1980 for all three i-propyl halides. In 1982, Rosenstock et al.4 used the more sophisticated technique of TPEPICO spectroscopy to report onsets for Br and I atom loss from the corresponding ion that are 50 and 70 meV, respectively, lower than the Traeger values.4 In that study, Rosenstock et al.4 reported that their experiment suggested a slow dissociation via metastable parent ions, which would shift their onset to higher energies. However, these metastable ions have not been observed in subsequent studies, so that accounting for a nonexistent kinetic shift might well account for the lower reported onset. In 2000, the 0 K dissociation onsets (E0) of isopropyl ion from isopropyl chloride, isopropyl bromide, and isopropyl iodide were measured using the much higher resolution pulsed field ionization-photoelectron-photoion coincidence spectroscopy (PFI-PEPICO) with molecular beam cooled samples by Baer et al.5 to be 11.085 ( 0.005, 10.505 ( 0.008,
10.1021/jp104200h 2010 American Chemical Society Published on Web 07/09/2010
11286
J. Phys. Chem. A, Vol. 114, No. 42, 2010
Stevens et al.
TABLE 1: Comparison of E0(i-C3H7+) in Electronvolts with Previous Values E0 C3 H8 i-C3H7Cl
i-C3H7Br
i-C3H7I
AE0 - AE298K
AE298K
11.61 ( 0.03 11.61 ( 0.01 11.624 ( 0.002d 11.03 ( 0.02a 11.085 ( 0.005f 11.036 ( 0.010g 11.065 ( 0.004d 10.44 ( 0.02a 10.42 ( 0.01h 10.42 ( 0.01g 10.505 ( 0.020f 10.454 ( 0.008d 9.81 ( 0.02a 9.77 ( 0.02h 9.851 ( 0.025f 9.818 ( 0.010g 9.8180 ( 0.0037i 9.812 ( 0.008d a
11.53 ( 0.01 11.530 ( 0.01c
0.085
10.92e
0.105
10.33e
0.110
b
in PA(C3H6) is therefore due to ∆fH298K°(i-C3H7+). As a result of the scatter in ∆fH298K°(i-C3H7+), PA(C3H6) has oscillated between 746 and 743 kJ mol-1. Finally, calculations can also be used to determine the ∆fH0K°(i-C3H7+). Lau and Ng16 have recently determined values of 90.0 and 806.4 kJ mol-1 for the isopropyl radical and ion 298 K heats of formation at the CCSD(T)/CBS level. These values appear to be within the scatter of previously reported values. The calculated 0 K values suggest an ionization energy of 7.437 eV, which is significantly higher than the experimentally observed IE of 7.36 eV.13,14 Experimental Approach
9.70e
0.113
a These values were converted from an AE298K by adding AE0 AE298K ) + . Steiner’s extrapolation, which ignored the rotations, yielded a 0 K onset of 11.59 ( 0.03 eV. b Steiner et al. (1961).8 c Chupka and Berkowitz (1967).9 d This work. e Traeger (1980).2 f Baer et al. (2000).5 g Brooks et al. (2004).6 h Rosenstock et al. (1982).4 i Park et al. (2001).7
and 9.851 ( 0.025 eV, all of which are at the upper end of the previously reported onsets.2-6 However, as recently pointed out by Bodi et al.,11 the PFI-PEPICO molecular beam experiments were plagued by dimers that caused the reported onsets to be too high. In 2004, Brooks et al.6 reported 0 K TPEPICO onsets for Cl and Br loss from their respective molecules of 11.036 ( 0.010 and 10.420 ( 0.010 eV. This Cl loss onset, which is at the lower extreme of the published values, was obtained with a room-temperature sample, thus eliminating complications due to dimer formation. However, it was recently determined that the photon energies of the isopropyl chloride data were not properly calibrated and so that the reported 0 K dissociation onset (and therefore ∆fH°(i-C3H7+)) was too low by 0.030 meV. Finally, Park et al.7 reported on a mass-analyzed threshold ion (MATI) study of the i-propyl iodide using VUV laser, in which they reported a 0 K onset of 9.8180 ( 0.0037 eV, which is by far the most accurate value to date. Another possible route to ∆fH°(i-C3H7+) is the measurement of the ionization energy of the isopropyl radical. Although the radical has a relatively well established 0 K heat of formation of 106.2 ( 1 kJ mol-1,12 determining an accurate adiabatic ionization energy is no trivial matter. In 1979, Houle and Beauchamp measured the IE(i-C3H7) to be 7.36 ( 0.02 eV,13 a value later confirmed by Dyke et al.14 Beauchamp, relying upon the then accepted value for ∆fH298K°(i-C3H7 · ) of 74 kJ/mol, reported a value of 784 ( 4.6 kJ mol-1 for ∆fH298K°(i-C3H7+). Using today’s 0 K value for the radical heat of formation (106.2 ( 1 kJ mol-1) would yield ∆fH298Ko(i-C3H7+) ) 799 kJ mol-1. However, the photoelectron spectrum (PES) of such alkyl radicals is very broad and structureless so that the assignment of the adiabatic IE is not much more than a guess. Nevertheless, the Houle and Beauchamp value, corrected for the new radical energy, is within the scatter of other propyl ion heats of formation, which are based on the dissociative photoionization of the i-propyl halides. The proton affinity (PA) of propene plays an important role in calibrating the proton affinity scale. The heats of formation of C3H6 and H+ are very well known.1,15 Most of the uncertainty
The data presented here were obtained using the imaging photoelectron-photoion coincidence (iPEPICO) spectrometer located at the X04DB VUV beamline at the Swiss Light Source Synchrotron of the Paul Scherrer Institut.17 The apparatus has been described in detail18,19 and is only briefly reviewed here. The pure sample is effusively introduced at room temperature to the ionization region through a Teflon tube. In addition, a 5% propane sample in Ar was expanded in a continuous molecular beam through a 50 µm nozzle to investigate a cold sample. The source chamber was evacuated by a 5000 L s-1 cryo pump as well as a 1250 L s-1 turbo pump. Synchrotron radiation from a bending magnet, dispersed by a grazing incidence monochromator and with higher orders suppressed by a gas filter, was used to ionize the sample in a 2 × 4 mm interaction region. The photon energy, with a resolution of ∼2 meV, was calibrated using the well-known Ar 11s′-14s′ and Ne 13s′, 14s′, 12d′, and 13d′ autoionization lines. Upon ionization, an 80 V cm-1 electric field accelerated the electrons and ions in opposite directions for the case of the propyl halides. The H loss from propane was done with both 80 and 20 V cm-1 extraction fields. Velocity map imaging was used to focus the electrons onto a DLD40 Roentdek position sensitive delay-line detector with a kinetic energy resolution better than 1 meV at threshold. The ions were extracted from the same region by a two-stage Wiley-McLaren20 space-focused time-of-flight (TOF) mass spectrometer with a 6 cm long first acceleration region and a 50 cm drift region, after which they were detected by a Jordan TOF C-726 microchannel plate detector. We corrected for the contamination of the threshold electron signal from energetic electrons by determining the energetic electron background from a small ring around the threshold region of the electron image and subtracting this from the threshold signal. Using this setup, the energy resolution was ultimately limited by the photon energy resolution of ∼2.5 meV. Electron hit times and positions and ion hits were recorded in the triggerless mode of an HPTDC time to digital converter card, and electrons and ions were correlated on the fly to obtain TOF distributions without dead time. This multiple-start/ multiple-stop data acquisition scheme18 enables data acquisition at high ionization rates, which is beneficial at a high intensity light source, such as the synchrotron. The experimental data were analyzed and plotted in several ways: the threshold electron signal as a function of the photon energy yields a threshold photoelectron spectrum (TPES); the threshold electron signal detected in coincidence with an ion in a particular TOF range yields a mass-selected TPES; and the fractional ion abundances as a function of the photon energy yield the breakdown diagram, which has the significant benefit of being independent of volatile ambient parameters such as sample pressure, photon intensity, and Franck-Condon factors. Molecular parameters such as vibrational frequencies and rotational constants and geometries used in this article were
Dissociation Dynamics Investigated by iPEPICO
Figure 1. Room-temperature i-C3H7Cl threshold photoelectron spectrum (TPES) obtained with 2 meV resolution.
Figure 2. Room-temperature i-C3H7Br threshold photoelectron spectrum (TPES) obtained with 2 meV resolution.
J. Phys. Chem. A, Vol. 114, No. 42, 2010 11287
Figure 4. Breakdown diagram for i-C3H7Cl+. The open circles indicate the percent i-C3H7Cl+ abundance and the solid squares represent i-C3H7+ abundance. The solid lines through the data points indicate the best fit, whereas the dashed lines show the fits with the (4 meV variations in the onset energies. The TPES (red line) was smooth and unstructured between 10.7 and 11.2 eV.
to Br, to I. The spin orbit splitting, barely visible in the C3H7Cl TPES, becomes progressively more obvious in the bromide (0.30 eV) and the iodide (0.55 eV) compounds. The breakdown diagrams for the isopropyl halides, which are the fractional abundances of the parent and fragment ions as a function of the photon energy, are shown in Figures 4-6. The only fragment ion formed over the energy range of interest is i-C3H7+. The i-C3H7+ peaks in the TOF spectra were symmetric at low energies for all three compounds, indicating that the dissociation rates for i-C3H7+ formation are faster than the time scale of our apparatus (k(E) > 107 s-1). When the dissociation rate is fast, all ions, whose total internal energy exceeds the dissociation limit, will fragment. The total ion internal energy, relative to the ground state of the ion, is given by Eion(hν) ) hν - IE + Eth, where IE is the adiabatic ionization energy, hν is the photon energy, and Eth is the neutral thermal energy. We assume that in the photoionization process, the neutral thermal energy distribution, P(E), is transposed to the ion manifold. The relative abundance of the parent ion, BDP(hν), is determined by the portion of the ion internal energy distribution that lies below the 0 K dissociation threshold E0(i-C3H7+)
BDP(hν) )
∫0E -hν P(E) dE 0
for
hν < E0
(2)
The relative fragment ion abundance is then Figure 3. Room-temperature i-C3H7I threshold photoelectron spectrum (TPES) obtained with 2 meV resolution.
determined using the Gaussian 03 computer software.21 Geometry optimizations and normal-mode analyses, unless otherwise specified, were carried out using the B3LYP hybrid functional with the 6-311++G** basis set for C and H and 6-311G** basis set for iodine. Results Room-Temperature Data for i-Propyl Halides. The roomtemperature threshold photoelectron spectra for i-C3H7Cl, i-C3H7Br, and i-C3H7I are shown in Figures 1-3. These TPES show progressively more vibrational structure going from Cl,
BDF(hν) ) 1 - BDP(hν)
(3)
When the photon energy equals or exceeds the E0, the parent ion disappears so that the fractional parent and daughter ion signals remain 0 and 1, respectively. We use the C3H7X frequencies and rotational constants to determine the ion internal energy distribution. Although the molecule’s thermal energy distribution is often faithfully transposed to the ionic manifold upon ionization,22,23 it is not always the case, and the temperature is therefore used as a fitting parameter along with the 0 K dissociation energy, E0. For i-C3H7Cl, i-C3H7Br, and i-C3H7I, the best fit onsets were 11.064 ( 0.004, 10.454 ( 0.008, and 9.812 ( 0.008 eV with best-fit temperatures of 305 ( 10, 298 ( 20, and 320 ( 15 K, respectively. The reported E0 for i-C3H7I is within the experimental error of the previous TPEPICO
11288
J. Phys. Chem. A, Vol. 114, No. 42, 2010
Figure 5. Breakdown diagram for i-C3H7Br+. The open circles indicate the percent i-C3H7Br+ abundance, and the solid squares represent i-C3H7+ abundance. The solid lines through the data points indicate the best fit, whereas the dashed lines show the fits with the (8 meV variations in the onset energies. The red line shows the threshold photoelectron spectrum over the indicated energy range.
Figure 6. Breakdown diagram for i-C3H7I+. The open circles indicate the i-C3H7I+ abundance and the solid squares represent the i-C3H7+ abundance. The solid lines through the data points indicate the best fit (between 9.70 and 9.84 eV), whereas the dashed lines show the fits with the (8 meV variations in the onset energies. The red line shows the threshold photoelectron spectrum (arbitrary units).
measurement by Brooks et al.6 (9.818 ( 0.010 eV) and the laserbased MATI value of 9.818 ( 0.004 eV.24 Table 1 summarizes the current and previously measured onset energies. The improved energy resolution of 2.5 meV of the iPEPICO apparatus reveals some structure in the breakdown diagrams for i-C3H7Br and i-C3H7I. Similar features have been observed in the case of CH3I.11 It is evident in Figures 5 and 6, where the TPES is plotted along with the breakdown diagram, that the i-C3H7Br+ and i-C3H7I+ abundance is lower than expected in the region of the Franck-Condon gaps in the TPES. The production of threshold electrons in Franck-Condon gaps is not well understood. Guyon et al.25 and Chupka et al.26 proposed a mechanism that relies on the existence of long-lived neutral Rydberg states. Because there is a quasicontinuum of these Rydberg states converging to various ion states, it is possible for the neutral molecule to absorb a photon and access these neutral Rydberg states in regions where the probability of direct ionization is low (i.e., in a Franck-Condon gap). From these states, the neutral molecule can cross over to a neutral dissociative surface. As the molecule dissociates, it may cross over to a high n Rydberg state converging to the ground electronic ion state, from which it can vibrationally autoionize, generating a threshold electron. Because of energy conservation, the kinetic energy of the dissociating neutral fragments becomes
Stevens et al.
Figure 7. Room-temperature C3H8 threshold photoelectron spectrum (TPES) obtained with 2 meV resolution.
vibrational energy in the ion. It was suggested in the study of CH3I that the warmer ion internal energy distribution observed as an increase in fragment ion abundance, is a result of this process being aided by rotational excitation.11 A similar effect appears operative in the isopropyl halide ion dissociation. For instance, we observe peaks in the i-C3H7Br+ abundance (Figure 6) that coincide perfectly with peaks in the TPES, the most notable among which is the peak at 10.35 eV. Alternatively, dips in the TPES are associated with increased fragment ion abundance. To summarize: ions are produced by both direct ionization (peaks in the TPES) and autoionization in the FC gap regions. If the latter process is enhanced by rotational energy, then the autoionization route tends to produce hotter ions that preferentially dissociate. It is important to mention that because the dissociation is fast, this structure in the breakdown diagram does not shift the location of the E0. Thermal and Molecular Beam Data for Propane. The propane TPES, shown in Figure 7, is structureless from its IE of 10.94 ( 0.05 eV27 to well past the H-loss onset of 11.64 eV. The sharp dips are due to Kr absorption lines in the gas filter, which could be used as calibration lines. The breakdown diagrams for the H-loss channel from the propane ion were measured several times. The first data sets at 80 V cm-1 extraction field suggested onsets of 11.639 and 11.624 eV, respectively, for the room-temperature and molecular beam onsets, a discrepancy of 15 meV. In addition, the TOF distributions for the H-loss fragment were slightly asymmetric, which indicates a slow reaction. These breakdown diagrams, Figures S1 and S2, are available as Supporting Information. When these experiments were repeated (Figures 8 and 9) at an extraction field of 20 V cm-1, the molecular beam onset remained the same, whereas the room-temperature onset was reduced to 11.630 eV, leaving a discrepancy of just 6 meV. An RRKM calculation at the onset as well as modeling the daughter ion peak positions28 yielded comparable 2 × 106 and 3 × 106 s-1 dissociation rates, respectively. The residence time of the parent ion in the acceleration region using an 80 V cm-1 extraction field is 2.5 µs so that >99.3% of the parent ions should have time to dissociate in this region, even when the rate is only 2 × 106 s-1. Therefore, we should not see a kinetic shift in the onset. Our explanation of these observed kinetic shifts in the onsets is that H loss from the propane ion involves a tight transition state (vibrational frequencies are higher than those of the ion) so that a centrifugal barrier could shift the onset to higher
Dissociation Dynamics Investigated by iPEPICO
Figure 8. Breakdown diagram for the room-temperature effusive C3H8+ sample with an extraction field of 20 V cm-1. The open circles indicate the relative C3H8+ abundances and the solid squares represent relative fragment ion abundances. The fragments are predominantly the i-C3H7+ ion with the C2H4+ ions contributing
