UV-induced dissociation of CH₂BrI probed by intense femtosecond XUV pulses

Figure 2 shows ion time-of-flight mass spectra of CH₂BrI recorded with the PImMS2 camera at photon energies of 67 eV (black), 70 eV (blue), and 139 eV (red). Overall, the spectra look similar and contain the same species for all energies. Dominant fragmentation products are the singly-charged halogens, I⁺ and Br⁺, and ${\mathrm{C}\mathrm{H}}_{n}^{+}$ with n = 0, 1, 2 bound hydrogens. Moreover, the two radicals CH₂Br⁺ and CH₂I⁺, as well as higher charged atomic halogens up to I⁵⁺ and Br³⁺ are visible.

Zoom In Zoom Out Reset image size

At the higher photon energies, the mass spectra reveal a tendency for the formation of higher charge states and dissociation into smaller fragments than at lower energy. In particular, in the (partially overlapping) CH_n ⁺ peaks one can see a tendency to lose a higher number of hydrogens as the photon energy increases. These trends can be explained by a combination of two effects: (i) the change in photoabsorption cross section (see figure 3), and (ii) the accessibility of the Br(3d), and the I(4p) levels at 77 eV and 129 eV, respectively, (see table 1) at the highest photon energy.

Zoom In Zoom Out Reset image size

Figure 3 depicts literature values for the absolute absorption cross sections as a function of photon energy measured for CH₃I [43] (dotted line) and CH₃Br [44] (dashed line), as well as their sum (dash-dotted line). The latter provides a good estimate for the photoionization cross section of the CH₂BrI molecule. A high-resolution XUV absorption spectrum measured for gaseous CH₂BrI between 46 and 73 eV [16] is in good qualitative agreement with the data shown in figure 3. The broad, intense peak centered around 94 eV corresponds to photoionization of the I(4d) shell—the so-called giant resonance. The features around 53 and 71 eV (labelled A and B in figure 3) can be assigned to core excitations of I(4d) and Br(3d) electrons, respectively [16]. The averaged spectral profiles of the FEL pulse used in this work are also indicated in figure 3. We estimate cross sections of 12 Mb at 67 eV, 16 Mb at 70 eV, and 10 Mb at 139 eV for CH₂BrI. At 139 eV, the strongest fragmentation is observed, which can be rationalized when considering the average charge state produced by single photoabsorption at each of these photon energies.

To estimate the average charge created by a single photoabsorption at a given photon energy, we consider the relative absorption cross sections of iodine and bromine, as well as charge-state distributions of comparable rare gases at the same photon energy. Ionization of krypton (isoelectronic to Br⁻) just below the (3d) edge results in about 80% of Kr⁺ and 20% of Kr²⁺, while 50 eV above the edge, 60% of Kr²⁺ and about 30% of Kr³⁺ are created, see table 2 [45]. For xenon (isoelectronic to I⁻), the majority of photoionizations between the (4d) and the (4p) shells results in Xe²⁺ (70%) and Xe³⁺ (20–30%), with a small contribution of Xe⁺ from valence ionization. Above the (4p) edge, Xe⁴⁺ rises rapidly and Xe⁺ is virtually absent [45]. Considering that photoabsorption at 67 eV and 70 eV leads predominantly to inner-shell ionization at the iodine site in CH₂BrI, we estimate an average molecular charge of about 2.2 per inner-shell photoabsorption (see table 2). At 139 eV, the partial absorption cross sections at the bromine and iodine site are approximately equal (see figure 3). Absorption at iodine leads to an average charge of about 2.65, and absorption at bromine to about 2.2, thus yielding an average charge of approximately 2.4 per photoabsorption at 139 eV, contributing to the increased fragmentation observed at this photon energy. In addition to the single-photon absorption cross sections and average charge states, we also need to consider multi-photon absorption for the intense FEL pulses. In figure 2, the spectra at all photon energies contain Br³⁺,I⁴⁺, and I⁵⁺, indicating a significant contribution from multi-photon absorption.

The use of the PImMS2 time-stamping camera allows us to obtain the velocity-map images of all fragments simultaneously, in addition to the mass spectrum. Figure 4 shows the kinetic energy distributions (KEDs) of several species. To a large extent, they are independent of the photon energy. The KEDs of the singly-charged halogens, X⁺, and the corresponding radicals, CH₂X⁺, are dominated by a peak at approximately 1–3 eV. By comparing the KEDs to the results of a simple, classical Coulomb explosion simulation for partners of different charge states n (indicated by the arrows at the top of each panel), this peak can be assigned to a Coulomb explosion with a singly-charged co-fragment. We also confirmed the existence of these fragmentation channels by ion–ion covariance analysis, which is not addressed further in this manuscript. The photoabsorption of CH₂BrI at 67 and 70 eV happens predominantly at the iodine site, but the energies simulated by placing the charges at iodine and bromine, respectively, match the data well. This suggests that the second charge (initially created at the iodine site) is rapidly transferred to the bromine atom. We will elaborate further on this in the analysis of the time-resolved data presented in section 3.2. The additional peak at $<$1 eV visible for some of the fragments is attributed to a dissociation including a neutral fragmentation partner, resulting from valence ionization. The KEDs of halogens with higher charge states show an increasing contribution of co-fragments with two, three or four charges, consistent with an overall stronger charging of the entire molecule.

Zoom In Zoom Out Reset image size

When comparing the three photon energies, several differences are discernible in the KEDs in figure 4. The Br²⁺,I³⁺, and CH_n ⁺ distributions, although somewhat noisy, appear to show higher-energy shoulders for 139 eV compared to 67 and 70 eV. The corresponding partner of I³⁺ in a simple Coulomb explosion model would need to have three to five charges to explain an energy of 15–25 eV, implying a total molecular charge of at least +6 to +8. Such charge states can be reached by absorption of only two photons at 139 eV (see table 2), but at least three photons are required to reach charge states higher than +6 for the lower energies. The strongly enhanced yield of CH₂I⁺ with Br⁺ as a partner at 139 eV is slightly surprising, but can potentially be explained by a higher average molecular charge leading to this channel, in comparison to the other energies. The corresponding CH₂I⁺ peak in the mass spectrum in figure 2 appears slightly shifted to lower masses, suggesting that more hydrogens may be lost (not resolved). We also note that the valence ionization cross section is lower at 139 eV.

The influence of the bromine resonance around 70 eV on the molecular fragmentation has been investigated by Olney et al, who recorded photoion branching ratios for CH₃Br up to 80 eV [44]. Although of limited energy resolution, these indicate that the main differences at 70 eV (compared to 67 eV) are: (i) an increase of Br⁺, (ii) a trend towards overall enhanced fragmentation and hydrogen loss, and (iii) the appearance of a small contribution (≈2%) of Br²⁺. Our KEDs for 67 and 70 eV in figure 4 look very similar for most fragments. CH₂Br⁺ and CH₂I⁺ appear to occur slightly more often with a charged partner at 70 eV compared to 67 eV—consistent with the findings of Olney et al [44]; Br²⁺ occurs less often with a singly-charged cofragment, and more often with higher charged partners; and I³⁺ and Br³⁺ show an increase in combination with a partner of three or four charges. The peaks of I²⁺ and Br²⁺ are slightly enhanced for 70 eV in figure 2. We attribute those changes to an increased probability of multi-photon absorption at 70 eV, due to the higher cross section compared to 67 eV, in combination with a slightly higher pulse energy for the case of 70 eV.

In previous work, we have studied ultrafast charge transfer in different dissociating molecules [8, 9, 22, 23]. In those studies, the center of mass of the molecular fragment lay along the axis of the dissociating bond, and thus could be modeled well as two point-like particles undergoing prompt Coulomb explosion. A classical over-the-barrier model [46, 47] was found to describe the critical distance for electron transfer towards the iodine surprisingly well. In ring-like molecules, it was necessary to revisit the effective distance of the repelling charges [23], but the dissociation could still be described by two point-like charges.

In the case of CH₂BrI, the initial UV photoexcitation is very similar to the other studied systems containing a C–I bond. However, as described in the introduction, a large fraction of the total available energy is not converted into kinetic energy of the dissociation, but into rotation of the CH₂Br radical since the C–I dissociation is not symmetric with respect to the center of mass of the molecule. This influences the transfer of electrons to the iodine atom in different aspects. While it is intuitive that the dissociation effectively becomes slower, it is not a priori clear whether electron transfer to the iodine is best described as occurring from the carbon or the bromine site (or from the bromine via the carbon, or from a (delocalized) molecular orbital). One would expect two aspects to play an important role: (i) the preferential localization of the lowest bound electron in the CH₂Br radical, and (ii) the orientation of the radical with respect to the iodine. We will discuss the influence of both in the following.

Figure 5 shows a sketch of the molecular orbital energies of the CH₂Br radical. As in other halomethyl radicals, a C–X π-bond is formed upon interaction of the 4p_y orbital with the 2p_y orbital on the carbon atom, while the 4p_x orbital forms a non-bonding lone-pair on the bromine atom [48, 49]. The π* (C–Br) and 4p_x orbitals, calculated at the density functional level of theory using the B3LYP functional and an aug-cc-pVTZ basis set [50–52], are also shown in figure 5. The unpaired electron in the π* (C–Br) HOMO is localized on the carbon atom, which is in accordance with electronic structure calculations [49] and experimental studies employing hyperfine-resolved microwave spectroscopy [53]. The next occupied orbital is the 4p_x lone pair [49], which is localized at the bromine site. Both, the unpaired electron in the HOMO and the bromine lone pair, could potentially serve as a source of electrons that may be transferred to the iodine site, as long as the iodine is closer than the critical distance. In order to model the charge-transfer probability as a function of the pump-probe delay, it is therefore necessary to take the rotation of the radical into account.

Zoom In Zoom Out Reset image size

Figure 6 shows the time evolution of the CH₂BrI molecular geometry after UV-induced dissociation of the C–I bond, retrieved according to the procedure described in appendix A. Displayed are: (a) three molecular geometries at early delays up to 110 fs, when a linear molecular geometry is reached; (b) the C–I distance; and (c) the I–Br distance. With the exception of the first ≈100 fs, the I–Br distance increases approximately linearly, while the C–I distance exhibits an additional pronounced oscillation due to the rotation of the CH₂Br radical. This rotation proceeds with an average speed of 1.34°/fs. The center of mass of CH₂Br is located close to the bromine atom, therefore the CH₂ approximately revolves around the bromine and thus the I–Br distance barely oscillates. In addition to the average geometry (solid blue line in figure 6(b)), we also plot for comparison the time-dependent C–I distance for a slower and faster angular velocity (±0.5°/fs, orange and magenta dashed lines in figure 6(b)) to indicate the possible spread due to the zero-point vibration (see also figure 8 in appendix A). This illustrates that over time, the initially well-defined geometry exhibits a dephasing, leading to an ensemble of molecules with different geometries being probed at a given pump-probe delay time.

Zoom In Zoom Out Reset image size

To understand how this geometry evolution would be reflected in the electron transfer probability, we employ the classical over-the-barrier model [46, 47] for both the carbon and the bromine sites. This requires 'site-specific' ionization energies (IEs) as an input. The IE of the CH₂Br radical, corresponding to the IE of the HOMO located at the carbon (see figure 5), has been reported as E_C = 8.72 eV [54]. We determine the IE for an electron in the 4p_x bromine lone pair by calculating the zero-point corrected energy of the CH₂Br radical in its doublet ground electronic state, as well as the energy of the singly-charged CH₂Br⁺ ion in its triplet configuration (at the geometry of the neutral radical). The calculations have been performed using the Gaussian 09 software suite [55] at the CCSD(T)/aug-cc-pVTZ level of theory [50–52]. The difference between the two zero-point corrected energies is the vertical IE of the CH₂Br radical from the 4p_x − E_Br = 10.97 eV.

The critical distances for electron transfer from the bromine or the carbon atom of a neutral CH₂Br radical to the iodine can be calculated as [9]: ${R}_{\text{crit}}(q)=\frac{3\sqrt{q}}{{E}_{\mathrm{C}/\mathrm{B}\mathrm{r}}}$, with q being the iodine charge state. This distance corresponds to the case in which the potential barrier equals the electron binding energy. The results for different iodine charge states q are indicated with dashed lines in figures 6(b) and (c). The critical distances for electron transfer from carbon to I²⁺–I⁵⁺ (blue filled circles in figure 6(b)) are reached for pump-probe delays between 600 and 900 fs, and for delays between 450 and 750 fs for electron transfer from bromine (black filled circles in figure 6(c)). Also shown for comparison are the hypothetical R_crit for electron transfer from carbon when ignoring the rotation of the cofragment (red filled circles in figure 6(b)), which lie between 700 and 1100 fs. This illustrates that although a considerable amount of energy is spent on the rotational excitation, somewhat counter-intuitively, the rotation effectively shortens the delay at which the critical distance to the carbon atom is first reached by 100–200 fs, depending on the charge state. For almost all pump-probe delays, the delay-dependent interatomic distance is larger for the case of a rotating cofragment than a model ignoring the rotation would suggest (the blue curve in figure 6(b) lies above the red one for most delays). Rotation of a molecular co-fragment has previously been identified as a reason for unexpectedly fast dissociation of Br${({\mathrm{C}\mathrm{H}}_{2})}_{2}$Cl [56] and CHBr₃ molecules [57] after an initial halogen-carbon bond fission. Furthermore, the rotation leads to the fact that the critical distance for C–I can be crossed twice or even three times for certain charge states, which accordingly is expected to lead to a non-trivial oscillation of the charge-transfer probability.

The electron transfer towards the iodine ion is reflected in the delay-dependent yield of low-kinetic-energy iodine ions following UV-dissociation of the C–I bond. This channel only occurs for cases where the internuclear distance between iodine and the rest of the molecule is sufficiently large upon XUV-photoabsorption, such that charge transfer between them is inhibited. As observed in earlier experiments [8, 9, 22, 23], the step-wise increase in the yield of these low-energy ions therefore does not occur at time overlap between the pulses, but is shifted towards the UV-early side (i.e. longer delay times). The delay-dependent ion yields for different iodine charge states and photon energies are plotted in figure 7. A step-like increase of the ion yield is visible in the yields of I²⁺ and I³⁺ ions for all photon energies, and in addition for I⁴⁺ and I⁵⁺ ions at 139 eV. At the two lower photon energies, one XUV photoabsorption at iodine results in either two or three charges (see section 3.1). Charges of +4 or +5 would require two-photon absorption, which we consider unlikely, because the I(4d) binding energies of the I²⁺ and I³⁺ ions are most likely higher than 70 eV, thus the second photoabsorption can only take place in the valence shell. For the case of 139 eV, the situation is different. One photoabsorption at the I(4p) shell can result in four charges (see table 2), therefore I⁴⁺ can be created by a single photon. The I(4p) binding energy of the neutral atom is 137.7 eV [58], therefore a charged iodine can likely not be ionized at the I(4p) level. However, the I(4d) and the Br(3d) levels are both still accessible and have a significant absorption cross section such that a second photoabsorption can take place in an I²⁺ or I³⁺ ion, contributing to the I⁴⁺ and I⁵⁺ ion yield. Moreover, the data at 139 eV were recorded at a higher XUV pulse energy, but for a shorter total acquisition time, leading to lower statistics in this data set. We therefore doubt the significance of the seemingly slower rise time of the step functions at 139 eV, since our charge-transfer model gives no reason to expect a photon-energy dependence for a given charge state.

Zoom In Zoom Out Reset image size

The colored triangles in each panel of figure 7 indicate the delays at which critical internuclear distances for electron transfer are reached in different scenarios (see also figure 6). Blue triangles correspond to electron transfer from the carbon atom of a rotating CH₂Br, black triangles to electron transfer from the bromine atom (the rotation is practically irrelevant in this case, as the center of mass lies very close to Br). The red triangle shows the expectation for the hypothetical case ignoring the rotation. The latter case provides the worst fit to the data, as all steps clearly occur at shorter delays in the experiment. This is expected, as the rotation shortens the delay at which the critical distances are reached, as described above. The delay times for electron transfer from carbon or bromine (blue and black triangles) lie relatively close to each other, but the black triangles match the onset of the step better. The calculated delay at which the black triangles are located depends on the precise IE of the 4p_x bromine lone pair orbital. As shown in figure 5, the HOMO of the CH₂Br radical is located at the carbon site. However, the HOMO is only occupied by one electron, while the 4p_x lone pair, located at the bromine, contains two electrons, which might be reflected in an increased probability for electron transfer from the HOMO-1.

We note that the blue triangles correspond to the first time at which the critical internuclear distance is reached. However, charge transfer can again become possible up to 100–200 fs later due to the rotation (see figure 6). This would lead to a temporary decrease in the low-energy ion yield, and therefore an oscillation in the ion yields as a function of the pump-probe delay might be expected, reflecting the oscillating electron transfer probability due to the rotation of the radical with a period of about 250 fs. However, the overall temporal resolution of the experiment and the limited statistics for each delay point were not sufficient to clearly distinguish this in the present experiment. A further point to consider is the dephasing of the trajectories over time, leading to the fact that an ensemble of molecules with different geometries is probed at a given delay. As is visible in figure 6, significant deviations already start to occur for delays between 500 and 900 fs, which is when the critical distances for electron transfer are reached and the neutral channel starts to be populated.