Thermal Decomposition of 2- and 4-Iodobenzyl Iodide Yields Fulvenallene and Ethynylcyclopentadienes: A Joint Threshold Photoelectron and Matrix Isolation Spectroscopic Study

The thermal decomposition of 2- and 4-iodobenzyl iodide at high temperatures was investigated by mass-selective threshold photoelectron spectroscopy (ms-TPES) in the gas phase, as well as by matrix isolation infrared spectroscopy in cryogenic matrices. Scission of the benzylic C–I bond in the precursors at 850 K affords 2- and 4-iodobenzyl radicals (ortho- and para-IC6H4CH2•), respectively, in high yields. The adiabatic ionization energies of ortho-IC6H4CH2• to the X̃+(1A′) and ã+(3A′) cation states were determined to be 7.31 ± 0.01 and 8.78 ± 0.01 eV, whereas those of para-IC6H4CH2• were measured to be 7.17 ± 0.01 eV for X̃+(1A1) and 8.98 ± 0.01 eV for ã+(3A1). Vibrational frequencies of the ring breathing mode were measured to be 560 ± 80 and 240 ± 80 cm–1 for the X̃+(1A′) and ã+(3A′) cation states of ortho-IC6H4CH2•, respectively. At higher temperatures, subsequent aryl C–I cleavage takes place to form α,2- and α,4-didehydrotoluene diradicals, which rapidly undergo ring contraction to a stable product, fulvenallene. Nevertheless, the most intense vibrational bands of the elusive α,2- and α,4-didehydrotoluene diradicals were observed in the Ar matrices. In addition, high-energy and astrochemically relevant C7H6 isomers 1-, 2-, and 5-ethynylcyclopentadiene are observed at even higher pyrolysis temperatures along with fulvenallene. Complementary quantum chemical computations on the C7H6 potential energy surface predict a feasible reaction cascade at high temperatures from the diradicals to fulvenallene, supporting the experimental observations in both the gas phase and cryogenic matrices.


Velocity Map Imaging
In velocity map imaging (VMI), ions are focused onto concentric rings according to their momentum in the image plane.The ion image is centered in the background (BG), corresponding to molecules that are rethermalized upon collisions with the wall, and hence exhibit a room temperature (RT) velocity distribution.In the absence of external factors, the x and y momentum components of the BG should be similar, so it appears as a circle in the center of the ion image.On the other hand, the molecular beam (MB) possesses a high velocity due to the adiabatic expansion into vacuum and is thus offset from the BG.Due to skimming the MB exhibits a narrow velocity distribution perpendicular to the molecular beam expansion direction (Figures S3-4).In the VMI image (Figure S3, left), the yellow-red narrow line shows a large concentration of ions with a velocity distribution in the molecular beam (MB) part of the ion image indicative for direct ionization (DI) of radical 17 produced by FVP in the SiC reactor.Those molecules exhibit a ms-TPES of radical 17 at nearly the temperature of the reactor, thus showing low resolution (see below). 1 However some kinetic energy release is also present in the MB part of the ion image, which is a sign of dissociative ionization, especially when compared to the 298 K data (Figure S3, right).On the other hand, the background (BG, in blue) at m/z = 217 contains the information of the molecules that are rethermalized upon collisions with the wall and diffuse back into the ionization region of the PEPICO spectrometer. 1 This includes radical 17 if it is stable enough to survive the wall collisions but vibrationally cooled to (nearly) room temperature as well as fragments formed upon DPI of the unpyrolyzed precursor 15.Integration of the different areas (MB or BG) leads to ms-TPE spectra differentiating DI from DPI processes, that will be discussed in the next section.On the right part of Figure S3 we observe the ion image of m/z = 217 with pyrolysis off, which shows a broad velocity distribution of ions in the molecular beam, clearly indicating that those ions are purely formed upon DPI with large translational access energies.The same behavior was observed for precursor 16 in correspondence with its mass spectrum (Figure S1) and VMI images at 850 K and RT (Figure S4).In this case, the signal corresponding to DPI is even larger than that for isomer 15 (Figure S3).Since precursor 15 is not fully pyrolyzed at 850 K, remaining 15 undergoes DPI resulting in very broad and intense bands at photon energies higher than 8.85 eV (Figure S5, black trace).For this reason, we attempted to remove the contribution of DPI processes by recording a reference spectrum of precursor 15 at RT (Figure S5, red trace).The background VMI data, containing the rethermalized molecules, was selected as the region of interest (ROI) for both spectra so they have comparable temperatures.Moreover, the BG spectrum benefits from higher resolution after rethermalization of the molecules thanks to wall collisions. 1 The resulting vibrationally-cooled difference spectrum (Figure S5, blue trace) shows an improved triplet band system (8.7-9.0 eV) with better vibrational resolution, allowing to assign the vibrational progression (see Figure 2 of the manuscript).Moreover, above 9 eV, three sets of peaks with maxima at 9.10, 9.68 and 10.49 eV are now observable and are assigned to higher-energy excited states of 17 + .
The difference BG spectrum (Figure S6, blue trace) was also compared to the spectrum obtained upon selecting the whole ion image (WI, Figure S6, black trace) as well as to the vibrationally hot molecular beam (MB, Figure S6, red trace).The spectrum obtained from the whole ion image exhibits a poor resolution and is hampered by hot and sequence band transitions (red-shifted), which are not satisfactorily reproduced by FC simulations even when set to the pyrolysis temperature.In the highenergy region above 9 eV, only a broad feature centered at 9.63 eV could be identified, which reasonably fits to the band at 9.68 eV observed in the difference BG spectrum assigned to excited states of 17 + .However, the other two high-energy transitions are not clearly observable in the MB spectrum.

Vibrationally-cooled ms-TPE spectrum of 4-iodobenzyl radical 18
The same approach was also used for obtaining a vibrationally-cooled ms-TPE spectrum of 4iodobenzyl radical 18.In this case, the DPI process starts at a lower photon energy (8.70 eV, Figure S7, red trace).This fact, together with the blue shift of the triplet transition cause a more severe overlap of that band.The benefit of the background correction in this case is even more pronounced than for radical 17.However, the lower signal-to-noise ratio prevents the observations of high-energy transitions that might correspond to excited states.The comparison of the difference BG spectrum with that of the whole ion image also shows improvements in the resolution of the singlet and triplet bands and a lower influence of the hot bands (Figure S8).Due to the low intensity of the signal, the spectrum of the MB was not very revealing, and the signals were close to the noise level.Since no higher-energy bands could be observed in the difference BG spectrum, its comparison with the MB spectrum is less relevant.

Optimized Geometries
Geometry optimizations at the B3LYP-D3/def2-TZVP level of theory in the gas phase.

Figure S1 .
Figure S1.Mass spectrum of 4-iodobenzyl iodide 16 at RT recorded at a photon energy of 8.5 eV (a) and 9.0 eV (b).Mass spectrum of the FVP of 16 at 850 K (c) and 1260 K (d), recorded at a photon energy of 9.0 eV.

Figure S3 .
Figure S3.VMI image showing the direct ionization of radical 17 (left) and dissociative photoionization of precursor 15 (left) leading to cation 17 + at a photon energy of 9.0 eV.

Figure S4 .
Figure S4.VMI image showing the direct ionization of radical 18 (left) and dissociative photoionization of precursor 16 (left) leading to cation 18 + at a photon energy of 9.0 eV.

Figure S6 .
Figure S6.ms-TPE spectra comparing the signal (m/z = 217) obtained by integrating the whole ion image (WI, black trace), the hot molecular beam (MB, red trace) and the difference BG spectrum (blue trace) shown in Figure S4.The arrow indicates a band at 9.63 eV that is tentatively assigned to an excited state.

Figure S7 .
Figure S7.ms-TPE spectra comparing the signal (m/z = 217) obtained by integrating the rethermalized background obtained upon FVP of 16 at 850 K (black trace) and DPI of 16 at 298 K (red trace).Their difference BG spectrum is shown in blue.

Figure S8 .
Figure S8.ms-TPE spectra comparing the signal (m/z = 217) obtained by integrating the whole ion image (WI, black trace) with the difference BG spectrum (red trace).

Figure S9 .
Figure S9.ms-TPE spectrum of m/z = 217 obtained upon FVP of precursor 16 at 850 K (black trace).This spectrum was obtained with the same approach as Figure2, subtracting the background spectrum obtained from FVP at 850 K and DPI at 298 K of precursor 16.Simulations were performed at 0 K (blue sticks), and at 300 K (red trace) by convolution with 25 meV fwhm Gaussians.

Figure S11 .
Figure S11.IR spectra of the FVP of 2-iodobenzyl iodide 15 in Ar matrices at 4 K. a) Deposition of 15 at RT (pyrolysis off).b) FVP of 15 at 700 K c) Calculated IR spectrum of radical 17 at the B3LYP-D3/def2-TZVP level of theory.

Figure S13 .
Figure S13.IR spectra of the FVP of 4-iodobenzyl iodide 16 in Ar matrices at 4 K. a) Deposition of 16 at RT (pyrolysis off).b) FVP of 16 at 700 K c) Calculated IR spectrum of radical 18 at the B3LYP-D3/def2-TZVP level of theory.

Table S1 . Relative energies (in kcal/mol) and adiabatic ionization energies (in eV) of C7H6 isomers at different levels of theory.
Bold: Relative energies (in kcal/mol); and parenthesis: adiabatic ionization energies (in eV)

Table S2 . Experimental and calculated IR spectroscopic data of 2-iodobenzyl radical 17.
a FVP in Ar matrix at 700 K. b Calculated at the B3LYP-D3/def2-TZVP level of theory.c Absolute intensities.d Relative intensities based on the strongest absorption.

Table S3 . Experimental and calculated IR spectroscopic data of 4-iodobenzyl radical 18.
c Absolute intensities.d Relative intensities based on the strongest absorption.