From Iron Pentacarbonyl to the Iron Ion by Imaging Photoelectron

Article pubs.acs.org/JPCA

From Iron Pentacarbonyl to the Iron Ion by Imaging Photoelectron Photoion Coincidence Eileen M. Russell,† Elvis Cudjoe,†,∥ Michael E. Mastromatteo,†,⊥ James P. Kercher,*,† Bálint Sztáray,‡ and Andras Bodi§ †

Department of Chemistry, Hiram College, Hiram, Ohio 44234, United States Department of Chemistry, University of the Pacific, Stockton, California 95211, United States § Paul Scherrer Institut, Villigen 5232, Switzerland ‡

ABSTRACT: The dissociation dynamics of internal energy selected iron pentacarbonyl cations, Fe(CO)5+, have been investigated using the imaging photoelectron photoion coincidence (iPEPICO) spectrometer at the Swiss Light Source. The molecular ion loses all five carbonyl ligands in sequential dissociations in the 8.5−20 eV photon energy range. The Fe(CO)5+ parent ion is metastable at the onset of the first dissociation reaction on the time scale of the experiment. The slightly asymmetric and broad daughter ion time-of-flight distributions indicate parent ion lifetimes in the microsecond range, and are used to obtain an experimental dissociation rate curve. Further carbonyl losses were found to be fast at threshold. The fractional parent and daughter ion abundances as a function of the photon energy, that is, breakdown diagram, as well as the dissociation rates for the first CO loss were modeled using the statistical Rice−Ramsperger−Kassel−Marcus (RRKM) and statistical adiabatic channel model (SSACM) theories. The excess energy redistribution in the products was also taken into account in a statistical framework. The 0 K dissociative photoionization thresholds for the five carbonyl-loss channels were found to be 9.015 ± 0.024 eV, 10.199 ± 0.027 eV, 10.949 ± 0.033 eV, 12.282 ± 0.39 eV, and 13.821 ± 0.045 eV for the processes leading to Fe(CO)4+, Fe(CO)3+, Fe(CO)2+, Fe(CO)+, and Fe+, respectively. The iron cation thermochemistry is well-known, and these onsets connect the bare metal ion to the other fragment ions as well as to the gas phase neutral Fe(CO)5.

■

INTRODUCTION

To derive neutral thermochemistry by the positive ion cycle in organometallic complexes, the dissociative photoionization threshold for the MLn + hν → M+ + n L + e− process is perhaps the most desired, because it connects the transition metal complex to the well characterized free transition metal ion and the uncoordinated ligands. This often remains elusive using laboratory based tunable vacuum ultraviolet (VUV) light sources because of the limited scanning range. In such cases, threshold photoionization experiments can be combined with collision-induced dissociation16 results to yield heats of formation.17 In the present, alternative approach, the photon energy range is increased by using VUV synchrotron radiation.18 In threshold photoelectron-photoion coincidence (TPEPICO) experiments, internal energy selected photoions are prepared in the gas phase by threshold single photon photoionization with tunable VUV radiation. The dissociative photoionization products are mass analyzed by time-of-flight (TOF) mass spectrometry. Their threshold energy is determined by scanning the photon energy and analyzing fractional ion abundances plotted in the breakdown diagram as well as modeling slow dissociation rate constants to account for

Iron pentacarbonyl is used extensively as an intermediate to produce high purity iron,1 as well as a synthetic and catalytic reagent.2 As one of the few binary metal carbonyls, it is an intriguing model system to understand coordination and bonding in organometallic complexes that are relevant in catalytic processes.3 Bomb calorimetric measurements on transition metal compounds are notoriously difficult to interpret because of the multitude of oxidation states of the transition metal in the products.4 On the other hand, using the positive ion cycle has been proven to be a robust pathway to obtaining neutral thermochemical data, and accurate bond dissociation energies.5 The ion cycle has, for some time, been established as an alternative route to neutral bond energies, too.4,6 The measured (ionic or neutral) bond energies can then be interpreted in terms of the electronic structure of the molecule,7 steric (e.g., trans) effects,8 and, in transition metal coordination complexes, in terms of the Dewar−Chatt− Duncanson model.9−11 In favorable cases, bond dissociation energies, enthalpies of formation, and other thermochemical properties can be derived with close to or sub-kJ mol−1 accuracy.12,13 Dissociative photoionization thresholds are less accurate for mid- to large-sized molecules, where the parent ion is metastable on the experimental time-scale, and the resulting kinetic shift14 must also be included in the modeling.15 © XXXX American Chemical Society

Received: March 10, 2013 Revised: April 30, 2013

A

dx.doi.org/10.1021/jp402443e | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

kinetic shifts.15 The latter are manifested as broad, quasiexponentional TOF peaks, which become symmetric as the photon energy increases and the dissociation becomes fast. The appearance of the breakdown diagram can help identify complex dissociation mechanisms even by cursory inspection.19,20 Parallel processes are characterized by breakdown curves with gentle slopes, because the relative abundance of the products is determined by the slowly changing dissociation rate constant ratio or, in other words, the branching ratio. In sequential processes, on the other hand, the product ion abundance depends on the fraction of parent ion internal energy distribution above the dissociation barrier, which changes rapidly with increasing photon energy. Therefore, the corresponding breakdown curves will be steep, and get progressively broader in each dissociation step because of internal energy distribution broadening due to excess energy redistribution between the fragment ion, the neutral fragment, and the translational degrees of freedom.21 In organometallics with little reorganization and numerous sequential ligand losses, the main difficulty in determining accurate dissociative photoionization threshold lies in extrapolating the measured unimolecular dissociation rates to the threshold for the slow first ligand loss(es) and in modeling the excess energy redistribution, that is, the internal energy of the intermediate fragment ions at higher energies, when several ligands are lost.8 The broadening effect of the product energy redistribution on the internal energy distribution of the dissociating sample is also evident in the iron pentacarbonyl TPEPICO study of Fieber-Erdmann et al.22 They found that the sample temperature has a major effect on the low-energy breakdown diagram, but little to no effect on the high-energy part. This is because the internal energy distribution of the Fe(CO)2+ and FeCO+ intermediates is dominated by the excess energy partitioning in the dissociation steps leading up to these daughter ions and hardly influenced by their share of room temperature internal energy of the neutral. From the successive broadening of the internal energy distributions, it is also clear that a canonical description of the product energy distributions is insufficient as the dissociation processes cannot be thought of as cooling of the energized precursor ion. In the early 1990s, a coincidence photoionization work by Norwood et al. and a threshold collision-induced dissociation (TCID) guided ion-beam mass spectrometry (GIB-MS) study by the Armentrout group were published on Fe(CO)5, in quick succession.23,24 Since both relied and mostly confirmed earlier photoionization results by Distefano,25 as later corrected by Halle et al.,26 this line of research spans more than 42 years. Guided ion beam tandem mass spectrometry for bimolecular and CID processes has since proved to be a versatile tool to measure thermodynamic data.27 Threshold photoelectron photoion coincidence (TPEPICO) experiments later benefited greatly by pulsed field extraction in the study of small molecules28 and velocity map imaging for large electron collection efficiencies at high resolution for photoion internal energy selection.5,29 TPEPICO and TCID experiments claim to measure the same bond energies by different, complementary approaches. It is thus crucially important to clear up the significant discrepancies between Norwood et al.23 and Schultz et al.24 works, in particular the 50 kJ mol−1 difference in the Fe+ production onset, as first pointed out by Schultz et al. Both studies discussed adiabatic and concerted dissociation processes at length. This is in contrast with later photoionization works, in which diabatic, sequential processes were found to determine

the dissociative photoionization dynamics in transition metal complexes even when several intersystem crossings were required during dissociation.30 Although intersystem crossing rates have yet to be addressed quantitatively in dissociative photoionization, the singlet−triplet conversion rate in neutral Fe(CO)4 is nevertheless predicted to be almost 109 s−1, even at room temperature.31 Our goal in this study is to determine accurate dissociative photoionization thresholds for the iron carbonyl fragment ion series, Fe(CO)n+, n = 5−1 by applying a statistical model to the experimental data, which includes dissociation rates as well as statistical excess energy redistribution in each carbonyl-loss process. In addition to new, more accurate Fe(CO)5 + hν → Fe(CO)5−n+ + n CO + e−, n = 1−5, reaction energies, we will also show that the TCID and TPEPICO experiments can be reconciled if the kinetic shifts and the initial parent ion internal energy distribution as well as its broadening caused by excess energy redistribution are explicitly taken into account.

■

EXPERIMENTAL AND THEORETICAL APPROACH The imaging photoelectron photoion coincidence mass spectrometer (iPEPICO) has been described in detail elsewhere,32 and only a summary is presented below. The iPEPICO experiment is located at the VUV beamline33 at the Swiss Light Source of the Paul Scherrer Institut in Villigen, Switzerland. Synchrotron VUV light is dispersed by a grazing incidence monochromator with higher order radiation suppressed by a compact gas filter. The photon energy, with a resolution of ≈3 meV, is calibrated using first and second order Ar and Ne autoionization lines. The room temperature sample of Fe(CO)5 (Sigma−Aldrich) is introduced effusively into the ionization region through a Teflon tube, and is intersected and ionized by the incident radiation with a 2 mm × 2 mm cross section. Photoelectrons and photoions are extracted in opposite directions by a 40 V cm−1 static electric field. The electrons are velocity map imaged onto a Roentdek delay line detector with a kinetic energy resolution better than 1 meV at threshold. The central, threshold electron signal is contaminated by energetic, “hot” electrons with near-zero transverse momentum. This hot electron contamination was subtracted based on a small ring area around the central spot,29 yielding an overall energy resolution of ≈3.5 meV. Photoions were accelerated in the first 5 cm to −130 V, and then further accelerated in a short region to −600 V before entering the 55 cm long flight tube, where they were space focused and detected by a 40 cm diameter Jordan MCP detector. Synchrotron light is quasi-continuous on the experimental time scale,34 but since electrons have a 100 ns flight time with only a few ns spread, they can be used as start signal for the time-of-flight analysis of photoions, which have a TOF of several tens of μs and a thermal peak width of about 100 ns. A master clock is used to record the arrival time and position of all electrons as well as ion arrival times. Electrons in the total image or in parts of it can then be correlated with all ions within a suitable TOF window. This multiple-start/multiplestop data acquisition scheme has been shown to deliver the best performance in continuous two-particle coincidence experiments.35 Photoions spend several μs in the first, 40 V cm −1 acceleration region. If they dissociate during this time, the resulting ion TOF will be between that of the fragment ion and the parent ion, resulting in a quasi-exponential shape of the daughter ion time-of-flight peak, from which the dissociation B



Article

rate constants are measured in the 103 s−1 < k < 107 s−1 range. Only the parent ion, Fe(CO)5+, was found to be slightly metastable on the experimental time scale; the sequential carbonyl-loss dissociation rates are fast at threshold. The parent ion dissociation rates and the fragment ion fractional abundances, that is, the breakdown curves, were modeled simultaneously to extract the 0 K dissociative photoionization onsets. The model (described below) takes into account the room temperature internal energy distribution of the neutral, the ionization energy, and the ion optics parameters.15 The thermal energy distribution, the ion density, and transition state number of states as well as the density of states of the leaving CO, all required in calculating the product energy distributions, are based on rotational constants and harmonic vibrational frequencies from quantum chemical calculations carried out with the Gaussian 09 software package.36 The ground state geometries of neutral and ionic species were optimized using density functional theory (DFT) with the B3LYP functional, and the 6-311++G(d,p) basis set. The vibrational frequencies were used without scaling.

Figure 2. Select time-of-flight (TOF) distributions showing the first CO-loss channel from Fe(CO)5+ producing the Fe(CO)4+ fragment ion. The blue points are the experimental TOF distributions, and the solid red line is the fit. The asymmetric peak shape of Fe(CO)4+ in the 28.0−28.80 μs TOF range arises from slowly dissociating ions in the acceleration region.

■

EXPERIMENTAL RESULTS AND ANALYSIS Iron pentacarbonyl ions dissociate via five sequential carbonyl loss reactions in the 8.5−20 eV photon energy range yielding the Fe+ cation and five CO neutral fragments. Three photoionization experiments yielded very similar adiabatic ionization energies (IEad), namely, 7.95 ± 0.03,37 7.96 ± 0.02,38 and 7.98 ± 0.01 eV.25 In contrast, the Norwood et al. ionization energy at 7.88 eV23 appears to be an outlier, whereas Angeli et al. later reported the complete valence TPES of Fe(CO)5 at a moderate energy resolution of 30 meV without proposing an updated adiabatic ionization energy.39 In calculating the internal energy distribution of the photoions, we thus use the Distefano IEad of 7.98 eV. The breakdown diagram of Fe(CO)5 is given in Figure 1, and select TOF distributions showing the first CO-loss dissociation are shown in Figure 2. At low photon energies, hν < 8.5 eV,

only the molecular ion, Fe(CO)5+, is present. At 8.5 eV, the first carbonyl-loss product appears, corresponding to the iron tetracarbonyl cation, Fe(CO)4+. As the time-of-flight distributions in Figure 2 show, Fe(CO)5+ is observed as a sharp, symmetric peak centered at 30.45 μs, while the Fe(CO)4+ peak is broad and stretches over the entire 28.0−28.80 μs TOF region at low photon energies because of slowly dissociating parent ions. In modeling PEPICO experimental data, two statistical rate theories have recently been in use to account for kinetic and competitive shifts.15 The statistical adiabatic channel model (SSACM) 4 0 and Rice−Ramsperger−Kassel−Marcus (RRKM)41 rate equations both have the same form, k(E) =

σN ⧧(E − E0) hρ(E)

(1)

where E and E0 are the ion internal energy and the activation energy measured from the ground state of the ion, respectively, σ is the reaction symmetry, ρ(E) is the ion density of states, and N⧧(E − E0) is the transition state sum of states. The two rate theories differ in how the transition state number of states is calculated. Rigid-activated-complex (RAC−) RRKM assumes an energy-independent transition state as the bottleneck between the reactant and products. In bond breaking reactions without a saddle point along the reaction energy curve, a constrained geometry optimization and vibrational analysis are carried out at a constant, elongated bond length R ≈ 4−5 Å.5 The transitional frequencies are then scaled by an optimized factor to reproduce the experimental rate curve using eq 3. Thus, the transition state sum of states is calculated using calculated harmonic frequencies of the conserved vibrational modes along with scaled frequencies for the transitional modes. In SSACM,40 the transition state is approximated to be at the dissociation products at low excess energies and the rigidity factor is used to account for its shift to a tighter transition state

Figure 1. Breakdown curve of Fe(CO)5 showing the five sequential carbonyl losses leading to the Fe+ ion. The open points are the experimental breakdown curve, and the solid lines are the statistical fit. C



Article

Table 1. Dissociative Photoionization Onsets for the Sequential CO-Loss Channels of Fe(CO)5 E0/eV Fe(CO)5 Fe(CO)5 Fe(CO)5 Fe(CO)5 Fe(CO)5

→ → → → →

Fe(CO)4+ + CO Fe(CO)3+ + 2 CO Fe(CO)2+ + 3 CO FeCO+ + 4 CO Fe+ + 5 CO

iPEPICO E0a

Distefanob

Norwood et al.c

Schultz et al.d

9.015 ± 0.024 10.199 ± 0.027 10.949 ± 0.033 12.282 ± 0.039 13.821 ± 0.045

8.77 ± 0.10 9.87 ± 0.10 10.68 ± 0.10 12.43 ± 0.10 14.03 ± 0.10

8.67 ± 0.03 9.76 ± 0.04 10.88 ± 0.05 12.67 ± 0.05 14.38 ± 0.07

9.10 ± 0.04 10.17 ± 0.07 10.86 ± 0.09 12.42 ± 0.11 14.01 ± 0.13 13.78 ± 0.13e

a

This work. bDistefano25 as adjusted by Halle et al.26 cNorwood et al.23 dSchultz et al.24 the sum of the bond dissociation energies were converted to dissociative photoionization onsets using the reported IE = 7.933 ± 0.044 eV. eaAssuming the 4F9/2 Fe+ exit channel in the TCID experiments.

as the energy is increased. In the present work, we use the rigidity factor expression frigid (E) = exp[−(E − E0)/c]

P(E I , Eexc) =

(2)

ρI (E I) ∫

Eexc − E I

0

Eexc

∫0 ρI (y)(∫0

where the parameter c is adjusted in the fit to reproduce the experimental rate curve.15,40 In this framework, the transition state sum of states is calculated by scaling the product rotation contribution by f rigid before convolution with the conserved vibrational modes. Thus, there are two fitting parameters in the statistical model, the 0 K dissociation onset, E0, and either a scaling factor for the transitional modes (RAC-RRKM) or a rigidity constant c (SSACM). For fast dissociations without a kinetic shift, the only adjustable model parameter is E0. In the RRKM model, the four transitional frequencies were scaled to reproduce the TOF peak shapes together with the breakdown diagram. The dissociation onset was thus determined as E0 = 8.987 ± 0.024 eV, indicating a kinetic shift of about 70 meV, with an activation entropy of ΔS⧧600K = 46 J mol−1 K−1. The SSACM model yielded E0 = 9.015 ± 0.024 eV, corresponding to a kinetic shift of about 40 meV, with a rigidity factor of c = 350 cm−1. The uncertainties were determined by scanning the assumed 0 K onset energy and determining the low and high energy limits at which the fit turned out to be unacceptable. As SSACM has been shown to extrapolate to the threshold more reliably in halobenzenes,40 we will use the SSACM appearance energy henceforth. It is worth noting that even though the two rate theories yield a significantly different kinetic shift, its small magnitude means that the first onset energy lies close to the disappearance energy of the parent ion, and the dissociative photoionization onset can be determined accurately. As illustrated in the breakdown diagram in Figure 1, the second carbonyl-loss step producing Fe(CO)3+ appears at photon energies greater than 9.5 eV. The fragment ion TOF peaks are symmetric and narrow even at the lowest energies, indicating an absence of a kinetic shift. In each of the sequential CO losses from Fe(CO)5+, only part of the excess energy is deposited as internal energy in the fragment ion, while some of the energy is released as kinetic energy, and some is lost to further dissociations as ro-vibrational internal energy of the leaving neutral CO. The internal energy distribution of the intermediate fragment ions depends on the excess energy and its statistical redistribution in each CO-loss step. In one such step at excess energy Eexc, the CO-loss fragment ion internal energy distribution, P(EI, Eexc) will be given by

Eexc − y

ρCO (x) ρtr (Eexc − EI − x) dx ρCO (x) ρtr (Eexc − y − x) dx) dy (3)

where ρtr is the translational density of states, ρI and ρCO are the ro-vibrational densities of states of the fragment ion and the CO neutral fragment.41 In sequential dissociation reactions, the internal energy distribution of each intermediate ion is obtained according to eq 3. Aside from the number of translational degrees of freedom (three in this analysis),21,41 eq 3 provides a rigid prediction for the internal energy distribution of the fragment ions without adjustable parameters. In modeling the Fe(CO)5 breakdown diagram, we have found that the fragment ion breakdown curves in the onset and crossover regions were modeled very accurately using the respective E0 values as the only fitting parameters. There were, however, slight deviations from the model curves in the form of stretched tails starting with the second CO loss product, indicating a nonstatistical probability of high excess energy loss to translations or CO vibrations and rotations. Ion imaging experiments could decide whether this excess energy is lost to translation.34 In the current model, we accounted for higher than statistical probability of high kinetic energy release by increasing the CO density of states. Since the 0 K onset energies are primarily determined by the onset of the fragment ion signal, this should not affect the experimental E0 determination. The dissociation onset for the second CO loss reaction producing the Fe(CO)3+ cation was determined to be 10.199 ± 0.027 eV. Each of the following sequential reactions is modeled as described above, with the dissociation onset as the only adjustable parameter in the model. This is optimized to 10.949 ± 0.033, 12.282 ± 0.039, and 13.821 ± 0.045 eV for the production of Fe(CO)2+, Fe(CO)+, and Fe+, respectively and can be used to determine the heats of formation of the fragment ions as described below. The 0 K dissociation onsets are also shown in Figure 1 and summarized in Table 1. The first dissociative photoionization threshold lies close to the disappearance energy of the parent ion. As some of the excess energy is lost to the dissociating ion, the 0 K onset energy moves closer to the actual appearance energy of the fragment in subsequent dissociation steps. In the last one yielding Fe+, it is even below the photon energy at which the Fe+ signal first rises above the noise in the TOF spectrum. The standard deviation of the differences between our dissociative photoionization onsets and the corresponding values reported by Distefano,25 Norwood et al.23 as well as the ones derived from the bond dissociation energies reported D



Article

the iron vaporization enthalpy,46 which yields ΔfHθ0K[Fe(g)] = 413.13 ± 1.25 kJ mol−1 (Table 1). This, in turn, leads to ΔfHθ0K[Fe+(g)] = 1175.60 ± 1.25 kJ mol−1,47 which can be compared with 1175.54 ± 0.25 kJ mol−1 used by Schultz et al. and 1173.2 ± 1.7 kJ mol−1 by Norwood et al. The 0 K iron ion enthalpy of formation corresponds to the most stable 6D9/2 state, while there are four more spin−orbit states (6D7/2−6D1/2) in the 384−977 cm−1 energy range. The first excited electronic state is the quartet 4F, with spin−orbit states in the 1873−3117 cm−1 energy range (0.232−0.386 eV, corresponding to 4 F9/2−4F3/2).48 The heat of formation of Fe(CO)5 vapor, ΔfHθ298K[Fe(CO)5(g)] is less accurately known experimentally. Cotton et al.49 arrived at ΔfHθ298K[Fe(CO)5(l)] = −764.0 ± 7.1 kJ mol−1, based on which they derived the 298 K reaction enthalpy for Fe(CO)5(l) → Fe(g) + 5 CO as 615.88 ± 7.1 kJ mol−1. The auxiliary thermochemical values in Table 1 yield 626.8 kJ mol−1 for this quantity, indicating a significant discrepancy. By discounting earlier combustion calorimetry determinations because of incomplete product characterization, the Cotton value is updated to ΔfHθ298K[Fe(CO)5(l)] = −766.09 ± 7.1 kJ mol−1 in the NIST-JANAF compendium.50 The 298 K vaporization enthalpies of liquid Fe(CO)5 scatter between 38.1 ± 0.4 and 41.17 kJ mol−1, possibly because of partial decomposition of the sample.51−53 We use the average value of ΔvapHθ298K[Fe(CO)5(l)] = 39.6 ± 2 kJ mol−1 in this work and, from the NIST-JANAF condensed phase heat of formation, derive the gas-phase room-temperature value as ΔfHθ298K[Fe(CO)5(g)] = −726.5 ± 7.4 kJ mol−1. Based on the thermal enthalpies listed in Table 1, this converts to ΔfHθ0K[Fe(CO)5(g)] = −728.6 ± 7.4 kJ mol−1, which can be compared with the NIST-JANAF value of −729.5 ± 7.1 kJ mol−1. Using the reaction thermochemistry of Fe(CO)5 → Fe+ + 5 CO, we determine the 0 K heat of formation of Fe(CO)5 to be −726.9 ± 4.4 kJ mol−1 in this study. The Schultz et al. roomtemperature value of −727.8 ± 7.1 kJ mol−1 was incorrectly converted to 0 K because of a sign error, and their conversion should have yielded ΔfHθ0K[Fe(CO)5(g)] = −731.2 ± 7.1 kJ mol−1. The 0 K heat of formation of Fe(CO)5 was reported by Norwood et al. as ΔfHθ0K[Fe(CO)5(g)] = −782.8 ± 7.1 kJ mol−1, which they mistakenly assumed to be in agreement with the 298 K literature data available at the time, because of an incomplete conversion to 298 K, in which the elemental thermal enthalpies were disregarded, leading to an about 50 kJ mol−1 error. Schultz et al. pointed out that the Norwood result disagreed significantly with the literature heat of formation and raised the possibility of 4F9/2 Fe+ production as well as a kinetic shift to explain the discrepancy. TPEPICO yields experimental dissociation rates and only the first CO loss is affected by a kinetic shift on a μs experimental time scale. The issue of the exit channel is more intriguing. Albeit a real possibility in CID experiments, we think it is unlikely that this channel plays a role in threshold photoionization. The existence and dynamics of higher lying exit channels do not affect photoionization measurements at threshold. Thus, if a statistical model predicts the product ion to rise at a lower energy or with a steeper slope than is observed, it can only be due to nonstatistical energy release, that is, an impulsive dissociation, a dissociation over a reverse barrier,54 or a fast competing relaxation process, such as fluorescence.55 The existence of a higher lying exit channel merely offers another route to the same products at higher energies, but cannot per se block the lower lying one. The first

by Shultz et al.24 assuming the 4F9/2 Fe+ exit channel are 24, 43, and 9 kJ mol−1, respectively. Out of these, only the TCID− iPEPICO standard deviation of 9 kJ mol−1 (100 meV) is commensurate with the assumed error bars. The newly established agreement between the collision induced dissociation and the photoionization data reaffirms that, barring differences in the mechanism, both techniques measure the same barrier heights and can be used to derive energetics data. The bond dissociation energies can be determined with the same accuracy for each of the mass-selected iron carbonyl ions in TCID, and the sum of the five BDEs leading to the Fe+ ion will suffer from error accumulation. With recent advances of the CRUNCH modeling program used in the TCID data analysis, the TCID experimental data could be reinterpreted including consecutive dissociations, presumably decreasing the uncertainties.42,43 In the PEPICO experiment, the 0 K appearance energies are determined directly from the experimental data, the accuracy of which is limited by the slower rising breakdown curves because of the ever broader internal energy distribution of the daughter ions. At the same time, kinetic shifts and broadened internal energy distributions in sequential dissociations mean that only careful statistical modeling of the experimental data can yield accurate 0 K appearance energies.

■

THERMOCHEMISTRY The 0 K enthalpies of formation for the iron carbonyl ion series were calculated using the experimental iPEPICO dissociative photoionization energies (Table 1) and the 0 K heats of formation for Fe+ and CO, as listed in Table 2. The 0 K enthalpy of formation of carbon monoxide is well-known as Δ f H θ 0K (CO) = −113.81 ± 0.03 kJ mol−1 . 44,45 Desai recommended ΔsubHθ298K(Fe) = 415.47 ± 1.25 kJ mol−1 as Table 2. Derived and Literature Thermochemistry on Fe(CO)5, Its Dissociative Photoionization Products, and Ancillary Species

Fe(CO)5(g)

Fe(CO)4+ Fe(CO)3+ Fe(CO)2+ FeCO+ Fe(s) Fe(g) Fe+ CO C O2

(H298K − H0K) /kJ mol−1

Δf Hθ0K/kJ mol−1

ΔfHθ298K/kJ mol−1

−728.6 ± 7.4a −726.9 ± 4.4c −731.2 ± 7.1d −729.5 ± 7.1e 256.7 ± 5.1c 260.1d 484.7 ± 5.5c 477.3d 670.9 ± 5.3c 657.6d 913.3 ± 5.9c 922.5d

−726.5 ± 7.4a −724.8 ± 4.4c

33.53b

264.2 ± 5.1c

33.60b

497.5 ± 5.5c

26.27b

689.1 ± 5.3c

21.05b

936.9 ± 5.9c

11.54b

413.13 ± 1.25 1175.60 ± 1.25i −113.81 ± 0.03g

415.47 ± 1.25

f

g

4.507e 6.850h

−110.53 ± 0.03g 1.05f 8.683f

a

Derived herein based on a critical review of previously published data. B3LYP/6-311+G(d) results based on the ion (stationary electron) convention. cThis study, based on the 0 K dissociative photoionization appearance energies and thermal enthalpies. dSchultz et al.24 eChase, M.W. (NIST-JANAF).50 fDesai.46 gActive Thermochemical Tables.45 h Sugar and Corliss.48 iUsing IE = 7.9025 eV from Schoenfeld et al.47 b

E



Article ⊥

electronically excited state of CO lies 6 eV above the ground state,56 and there is no indication of enhanced kinetic energy release in Fe(CO)5 dissociative photoionization. Indeed, concerted CO-loss from Fe(CO)2+ as proposed by Norwood et al., in line with the Cr(CO)2+ dissociation mechanism by Das et al.,57 would lead to lower kinetic energy release and a faster rising Fe+ signal as a function of photon energy. Therefore, the involvement of a higher lying exit channel in dissociative photoionization to explain the discrepancy between CID and photoionization experiments is not verisimilar. It is more likely that Norwood et al. linearly interpolated the breakdown curves of the fragments to derive their appearance energies, and based the thermochemical derivation on these results. This means that the discrepancies are probably a consequence of the incomplete data analysis and not of the measurement. Nonetheless, the iron pentacarbonyl 0 K heat of formation determined in the present study, −726.9 ± 4.4 kJ mol−1, is in excellent agreement with the previously reported values of −728.6 ± 7.4 kJ mol−1 obtained through the heat of vaporization of Fe(CO)5(l), −731.2 ± 7.1 kJ mol−1 obtained by Schultz et al., and −729.5 ± 7.4 kJ mol−1 listed in the NISTJANAF compilation. The enthalpy of formation of Fe(CO)4+ is determined from the first CO loss reaction; Fe(CO)5 → Fe(CO)4+ + CO. The measured dissociation onset of 9.015 eV, coupled with the known ancillary thermochemistry yields ΔfHθ0K[Fe(CO)4+] = 256.7 ± 10.1 kJ mol−1. This compares nicely with ΔfHθ0K[Fe(CO)4+] = 260.2 ± 10.1 kJ mol−1 from Schultz et al. Similarly, the rest of the photoionization onsets can be used to derive the heats of formation of the fragment ions as listed in Table 2. Generally, our results agree well with the TCID study as opposed to the previous photoionization works. This underlines the importance of quantitatively including the kinetic shift, the thermal energy distribution at the experimental temperature as well as the product energy distribution in sequential processes in the data analysis.

Department of Chemistry, University of MassachusettsAmherst, Amherst, MA 01003, U.S.A. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The experimental work report here was carried out at the VUV beamline of the Swiss Light Source at the Paul Scherrer Institut. J.P.K. and B.S. would both like to thank the American Chemical Society Petroleum Research Fund for financial support through an Undergraduate New Investigator Award (PRF# 51896UNI4) and a New Directions Award (PRF# 49930-ND6), respectively.

■

■

CONCLUSIONS We have determined the dissociation energies of the iron carbonyl ion series by imaging photoelectron photoion coincidence spectroscopy. The first CO loss reaction is modeled using the SSACM and RRKM statistical theories of unimolecular dissociation rates, and the further sequential CO losses are modeled by statistical theory to calculate the product energy distributions. Dissociative photoionization onsets are thus derived, which, by way of the ionization potential of iron, connect the neutral Fe(CO)5 to the gas phase iron atom. Using this connection, we update the thermochemistry the iron carbonyl cationic series and that of iron pentacarbonyl, the latter to ΔfHθ298K[Fe(CO)5(g)] = −724.8 ± 4.4 kJ mol−1. We have shown that the determined bond dissociation energies measured by TCID and PEPICO experiments are in agreement when the kinetic shifts and the initial parent ion internal energy distribution as well as its broadening caused by excess energy redistribution are explicitly taken into account.

■

REFERENCES

(1) Wildermuth, E.; Stark, H.; Friedrich, G.; Ebenhöch, F. L.; Kühborth, B.; Silver, J.; Rituper, R. Iron Compounds: ULLMANN’S Encyclopedia of Industrial Chemistry. Wiley: New York, 2000. (2) Samson, S.; Stephenson, G. R. Pentacarbonyliron. In Encyclopedia of Reagents for Organic Synthesis; Paquette, L., Ed.; J. Wiley & Sons: New York, 2013. (3) Astruc, D. Organometallic Chemistry And Catalysis; Springer: Berlin, Germany, 2007. (4) Chipperfield, J. R.; Sneyd, J. C. R.; Webster, D. E. The Thermochemistry of the Di-n5-Cyclopentadienyl Derivatives of the First Transition Series and their Unipositive Ions. J. Organomet. Chem. 1979, 178, 177−189. (5) Baer, T.; Sztáray, B.; Kercher, J. P.; Lago, A. F.; Bodi, A.; Scull, C.; Palathinkal, D. Threshold Photoelectron Photoion Coincidence Studies of Parallel and Sequential Dissociation Reactions. Phys. Chem. Chem. Phys. 2005, 7, 1507−1513. (6) Sztáray, B.; Szepes, L.; Baer, T. Neutral Cobalt−Carbonyl Bond Energy By Combined Threshold Photoelectron Photoion Coincidence and He(I) Photoelectron Spectroscopy. J. Phys. Chem. A 2003, 107, 9486−9490. (7) Hemberger, P.; Bodi, A.; Schon, C.; Steinbauer, M.; Fischer, K. H.; Kaiser, C.; Fischer, I. A Pass Too Far: Dissociation of Internal Energy Selected Paracyclophane Cations, Theory and Experiment. Phys. Chem. Chem. Phys. 2012, 14, 11920−11929. (8) Fischer, K. H.; Schneider, M.; Fischer, I.; Pfaffinger, B.; Braunschweig, H.; Sztáray, B.; Bodi, A. Bonding in a Borylene Complex Investigated by Photoionization and Dissociative Photoionization. Chem.Eur. J. 2012, 18, 4533−4540. (9) Chatt, J.; Duncanson, L. A. Olefin Co-Ordination Compounds. Part III. Infra-Red Spectra and Structure: Attempted Preparation of Acetylene Complexes. J. Chem. Soc. 1953, 2939−2947. (10) Dewar, M. J. A Review of π Complex Theory. Bull. Soc. Chim. Fr. 1951, 18, 3−4. (11) Pongor, C. I.; Szepes, L.; Basi, R.; Bodi, A.; Sztáray, B. MetalCarbonyl Bond Energies in Phosphine Analogue Complexes of Co(CO)3NO by Photoelectron Photoion Coincidence Spectroscopy. Organometallics 2012, 31, 3620−3627. (12) Borkar, S.; Ooka, L.; Bodi, A.; Gerber, T.; Sztáray, B. Dissociative Photoionization of Sulfur Chlorides and Oxochlorides: Thermochemistry and Bond Energies Based on Accurate Appearance Energies. J. Phys. Chem. A 2010, 114, 9115−9123. (13) Borkar, S.; Sztáray, B.; Bodi, A. Dissociative Photoionization Mechanism of Methanol Isotopologues (CH3OH, CD3OH, CH3OD and CD3OD) by iPEPICO: Energetics, Statistical and Non-statistical Kinetics and Isotope Effects. Phys. Chem. Chem. Phys. 2011, 13, 13009−13020. (14) Lifshitz, C. Time-resolved Appearance Energies, Breakdown Graphs, and Mass Spectra: The Elusive Kinetic Shift. Mass Spectrom. Rev. 1982, 1, 309−348. (15) Sztáray, B.; Bodi, A.; Baer, T. Modeling Unimolecular Reactions in Photoelectron Photoion Coincidence Experiments. J. Mass Spectrom. 2010, 45, 1233−1245.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Addresses ∥

Macromolecular Science and Engineering, Case Western Reserve University, Cleveland, OH 44106, U.S.A. F



Article

(35) Bodi, A.; Sztáray, B.; Baer, T.; Johnson, M.; Gerber, T. Data Acquisition Schemes for Continuous Two-Particle Time-of-Flight Concidence Experiments. Rev. Sci. Instrum. 2007, 78, 084102-1− 084102-7. (36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Revision B.01; Gaussian, Inc.: Wallingford, CT, 2009. (37) Vilesov, F. I.; Kurbatov, B. L. Photoionization of Esters and Metal Carbonyls in the Gaseous Phase. Proc. Acad. Sci. USSR, Phys. Chem. Sect. 1961, 140, 792. (38) Lloyd, D. R.; Schlag, E. W. Photoionization Studies of Metal Carbonyls. I. Ionization Potentials and the Bonding in Group VI Metal Hexacarbonyls and in Mononuclear Carbonyls and Nitrosyl Carbonyls of Iron, Cobalt, and Nickel. Inorg. Chem. 1969, 8, 2544−2555. (39) Angeli, C.; Berthier, G.; Rolando, C.; Sablier, M.; Alcaraz, C.; Dutuit, O. Threshold Photoionization Study of Fe(CO)5 versus ab initio Calculations. J. Phys. Chem. A 1997, 101, 7907−7913. (40) Stevens, W.; Sztáray, B.; Shuman, N.; Baer, T.; Troe, J. Specific Rate Constants k(E) of the Dissociation of the Halobenzene Ions: Analysis by Statistical Unimolecular Rate Theories. J. Phys. Chem. A 2008, 113, 573−582. (41) Baer, T.; Hase, W. L. Unimolecular Reaction Dynamics: Theory and Experiments; Oxford University Press: New York, 1996. (42) Rodgers, M. T.; Ervin, K. M.; Armentrout, P. B. Statistical Modeling of Collision-Induced Dissociation Thresholds. J. Chem. Phys. 1997, 106, 4499−4508. (43) Carl, D. R.; Chatterjee, B. K.; Armentrout, P. B. Threshold Collision-Induced Dissociation of Sr2+(H2O)x complexes (x = 1–6): An Experimental and Theoretical Investigation of the Complete Inner Shell Hydration Energies of Sr2+. J. Chem. Phys. 2010, 132, 044303− 044312. (44) Klopper, W.; Ruscic, B.; Tew, D. P.; Bischoff, F. A.; Wolfsegger, S. Atomization Energies from Coupled-Cluster Calculations Augmented with Explicitly-Correlated Perturbation Theory. Chem. Phys. 2009, 356, 14−24. (45) Ruscic, B. Active Thermochemical Tables, ver. 1.110; Argonne National Laboratory: Lemont, IL, 2010. (46) Desai, P. D. Thermodynamic Properties of Iron and Silicon. J. Phys. Chem. Ref. Data 1986, 15, 967−983. (47) Schoenfeld, W. G.; Chang, E. S.; Geller, M.; Johansson, S.; Nave, G.; Sauval, A. J. High Excitation Rydberg Levels of Fe I from the ATMOS Solar Spectrum at 2.5 and 7 μm. Astron. Astrophys. 2013, 301, 593−601. (48) Sugar, J.; Corliss, C. Atomic-Energy Levels of the Iron-Period Elements - Potassium through Nickel. J. Phys. Chem. Ref. Data 1985, 14, 1−664. (49) Cotton, F. A.; Fischer, A. K.; Wilkinson, G. Heats of Combustion and Formation of Metal Carbonyls. III. Iron Pentacarbonyl; The Nature of the Bonding in Metal Carbonyls. J. Am. Chem. Soc. 1959, 81, 800−803. (50) Chase, M. W. NIST-JANAF Thermochemical Tables, 4th ed.; American Institute of Physics: New York, 1998. (51) Leadbetter, A. J.; Spice, J. E. The Third Law Entropy and Structure of Iron Pentacarbonyl. Can. J. Chem. 1959, 37, 1923−1929. (52) Behrens, G. Thermodynamics of Transition Metal Carbonyls I. Fe(CO)5, Ru(CO)5, Os(CO)5. J. Less-Common Met. 1977, 56, 55−68. (53) Acree, J.; Chickos, J. S. Phase Transition Enthalpy Measurements of Organic and Organometallic Compounds. Sublimation, Vaporization and Fusion Enthalpies From 1880 to 2010. J. Phys. Chem. Ref. Data 2010, 39, 043101−043942. (54) Harvey, J.; Bodi, A.; Tuckett, R. P.; Sztáray, B. Dissociation Dynamics of Fluorinated Ethene Cations: from Time Bombs on a Molecular Level to Double-Regime Dissociators. Phys. Chem. Chem. Phys. 2012, 14, 3935−3948. (55) Nenner, I.; Guyon, P. M.; Baer, T.; Govers, T. R. A Threshold Photoelectron Photoion Coincidence Study of the N2O+ Dissociation between 15 and 20.5 eV. J. Chem. Phys. 1980, 72, 6587−6592.

(16) Ervin, K. M.; Armentrout, P. B. Translational Energy Dependence of Ar+ + XY -> ArX+ + Y (XY = H2 ,D2, HD) from Thermal to 30 eV c.m. J. Chem. Phys. 1985, 83, 166−189. (17) Gengeliczki, Z.; Sztáray, B.; Baer, T.; Iceman, C.; Armentrout, P. B. Heats of Formation of Co(CO)2NOPR3, R = CH3 and C2H5, and its Ionic Fragments. J. Am. Chem. Soc. 2005, 127, 9393−9402. (18) Leone, S. R.; Ahmed, M.; Wilson, K. R. Chemical Dynamics, Molecular Energetics, and Kinetics at the Synchrotron. Phys. Chem. Chem. Phys. 2010, 12, 6564−6578. (19) Bodi, A.; Stevens, W. R.; Baer, T. Understanding the Complex Dissociation Dynamics of Energy Selected Dichloroethylene Ions: Neutral Isomerization Energies and Heats of Formation by Imaging Photoelectron Photoion Coincidence. J. Phys. Chem. A 2011, 115, 726−734. (20) Hemberger, P.; Bodi, A.; Gerber, T.; Würtemberger, M.; Radius, U. Unimolecular Reaction Mechanism of an Imidazolin-2-ylidene: An iPEPICO Study on the Complex Dissociation of an Arduengo-Type Carbene. Chem.Eur. J. 2013, DOI: 10.1002/chem.201204465. (21) Baer, T.; Walker, S. H.; Shuman, N. S.; Bodi, A. One- and TwoDimensional Translational Energy Distributions in the Iodine-Loss Dissociation of 1,2-C2H4I2+ and 1,3-C3H6I2+: What Does This Mean? J. Phys. Chem. A 2012, 116, 2833−2844. (22) Fieber-Erdmann, M.; Holub-Krappe, E.; Bröker, G.; Dujardin, G.; Ding, A. Fragmentation Spectroscopy of Photoionized Fe(CO)5: A Molecular Model for a Heterogeneous Cluster. Int. J. Mass Spectrom. Ion. Proc. 1995, 149−150, 513−520. (23) Norwood, K.; Ali, A.; Flesch, G. D.; Ng, C. Y. A Photoelectron Photoion Coincidence Study of Iron Pentacarbonyl. J. Am. Chem. Soc. 1990, 112, 7502−7508. (24) Schultz, R. H.; Crellin, K. C.; Armentrout, P. B. Sequential Bond Energies of Iron Carbonyl Fe(CO)x+ (x = 1−5): Systematic Effects on Collision-Induced Dissociation Measurements. J. Am. Chem. Soc. 1991, 113, 8590−8601. (25) Distefano, G. Photoionization Study of Fe(CO)5 and Ni(CO)4. J. Res. Natl. Bur. Stand. (U. S.) 1969, 74A, 233−239. (26) Halle, L. F.; Armentrout, P. B.; Beauchamp, J. L. Ion Beam Studies of the Reactions of Group VIII Metal Ions with Alkanes: Correlation of Thermochemical Properties and Reactivity. Organometallics 1982, 1, 963−968. (27) Armentrout, P. B. Mass Spectrometry not just a Structural Tool: The Use of Guided Ion Beam Tandem Mass Spectrometry to Determine Thermochemistry. J. Am. Soc. Mass Spectrom. 2002, 13, 419−434. (28) Weitzel, K. M.; Malow, M.; Jarvis, G. K.; Baer, T.; Song, Y.; Ng, C. Y. High-Resolution Pulsed Field Ionization Photoelectron Photoion Coincidence Study of CH4: Accurate 0 K Dissociation Threshold for CH3+. J. Chem. Phys. 1999, 111, 8267−8270. (29) Sztáray, B.; Baer, T. The Suppression of Hot Electrons in Threshold Photoelectron Photoion Coincidence Spectroscopy Using Velocity Focusing Optics. Rev. Sci. Instrum. 2003, 74, 3763−3768. (30) Li, Y.; Sztáray, B.; Baer, T. The Dissociation Kinetics of EnergySelected CpMn(CO)3+ Ions Studied by Threshold Photoelectron Photoion Coincidence Spectroscopy. J. Am. Chem. Soc. 2001, 123, 9388−9396. (31) Besora, M.; Carreon-Macedo, J. L.; Cowan, A. J.; George, M. W.; Harvey, J. N.; Portius, P.; Ronayne, K. L.; Sun, X. Z.; Towrie, M. A Combined Theoretical and Experimental Study on the Role of Spin States in the Chemistry of Fe(CO)5 Photoproducts. J. Am. Chem. Soc. 2009, 131, 3583−3592. (32) Bodi, A.; Johnson, M.; Gerber, T.; Gengeliczki, Z.; Sztáray, B.; Baer, T. Imaging Photoelectron Photoion Coincidence Spectroscopy with Velocity Focusing Electron Optics. Rev. Sci. Instrum. 2009, 80, 034101−034107. (33) Johnson, M.; Bodi, A.; Schulz, L.; Gerber, T. Vacuum Ultraviolet Beamline at the Swiss Light Source for Chemical Dynamics Studies. Nucl. Instrum. Methods Phys. Res., Sect. A 2009, 610, 597−603. (34) Bodi, A.; Hemberger, P.; Gerber, T.; Sztáray, B. A New Double Imaging Velocity Focusing Coincidence Experiment: i2PEPICO. Rev. Sci. Instrum. 2012, 83, 083105−083108. G



Article

(56) Shi, D. H.; Li, W. T.; Sun, J. F.; Zhu, Z. L. Theoretical Study of Spectroscopic and Molecular Properties of Several Low-Lying Electronic States of CO Molecule. Int. J. Quantum Chem. 2012, 113, 934−942. (57) Das, P. R.; Nishimura, T.; Meisels, G. G. Fragmentation of Energy-Selected Hexacarbonylchromium Ion. J. Phys. Chem. 1985, 89, 2808−2812.

H


From Iron Pentacarbonyl to the Iron Ion by Imaging Photoelectron

Recommend Documents