We report on the shape resonance spectra of phenol-water clusters, as obtained from elastic electron scattering calculations. Our results, along with virtual orbital analysis, indicate that the well-known indirect mechanism for hydrogen elimination in the gas phase is significantly impacted on by microsolvation, due to the competition between vibronic couplings on the solute and solvent molecules. This fact suggests how relevant the solvation effects could be for the electron-driven damage of biomolecules and the biomass delignification [E. M. de Oliveira et al., Phys. Rev. A 86, 020701(R) (2012)]. We also discuss microsolvation signatures in the differential cross sections that could help to identify the solvated complexes and access the composition of gaseous admixtures of these species.
Much of the current knowledge on electron-driven damage to DNA and other biomolecules1,2 was gained from studies on subunits, such as bases and sugars.1,3–5 Despite the extensive work on gas-phase collisions, only a few theoretical studies addressed the solvation effects on transient anion states (resonances).6–9 We recently investigated the role of hydrogen bonding in microsolvated molecules,8,9 finding out that resonances can be either stabilized or destabilized, depending on the role played by the water molecules in the H-bonds. The π* resonances located on carbonyl8 and carboxyl9 groups were shifted to lower (higher) energies in case the water molecules act as proton donors (acceptors).
While our previous studies addressed target molecules having a single π* resonance (formaldehyde8 and formic acid9), we presently extend the investigations to phenol. This system is an interesting prototype, sharing a couple of features with the nucleobases: three π* resonances arising from the six-membered aromatic ring,10,11 and an indirect dissociative electron attachment (DEA) mechanism involving the second π* resonance and a
Structures of the hydrogen-bonded phenol...(H2O)n complexes with n = 1 (A and B) and n = 2 (C and D). The isolated phenol (Ph) and water molecules are also shown. The thick arrows are oriented from the proton donors to the acceptors along the H-bonds, while the thin arrows indicate the OH bond lengths. Oxygen, carbon, and hydrogen atoms are shown in red (gray), black, and white, respectively.
Structures of the hydrogen-bonded phenol...(H2O)n complexes with n = 1 (A and B) and n = 2 (C and D). The isolated phenol (Ph) and water molecules are also shown. The thick arrows are oriented from the proton donors to the acceptors along the H-bonds, while the thin arrows indicate the OH bond lengths. Oxygen, carbon, and hydrogen atoms are shown in red (gray), black, and white, respectively.
The elastic scattering calculations were performed with the parallel version13 of the Schwinger multichannel method with pseudopotentials (SMCPP).14 This variational approach to the scattering amplitude, which relies on a discrete trial set to expand the scattering state, was discussed in detail elsewhere.13–15 The cross sections were computed in the static-exchange (SE) approximation. While the neglect of correlation-polarization effects in the SE approximation overestimates the resonance energies, some essential aspects of the resonance spectra can be learned from this less computationally expensive approach. The previous studies on microsolvated molecules8,9 indicated that SE-level cross sections show the same (de)stabilization trends as those obtained in the SE plus polarization (SEP) approximation that accounts for the target dynamical response.
The structures of the complexes (Fig. 1) were obtained from liquid-phase classical Monte Carlo (MC) simulations in the isothermal-isobaric NPT ensemble (P = 1 atm and T = 298 K), as described by Barreto et al.16 The selected phenol...(H2O)n clusters are thus representative of H-bonding in the liquid. The intramolecular degrees of freedom were kept frozen during the MC simulations, though the geometries of the complexes and the isolated phenol molecule were subsequently optimized at the second-order Møller-Plesset perturbation (MP2) level employing the aug-cc-pVDZ basis set. The target ground states were then calculated at the Hartree-Fock (HF) level with the basis sets described elsewhere.10 The canonical HF orbitals were employed to build the SE-approximation scattering states.
Even though the target molecules are polar, the results were not corrected to account for long-ranged dipole potential contribution to the higher partial waves, since the latter only affects the background, having no significant influence on the shape resonance spectra. The impact of the dipole correction on the MTCS is also limited by the weighting factor (1 − cos θ), where θ is the scattering angle. The calculated dipole moment magnitudes are 4.072 D for complex A, 3.199 D for complex B, 2.399 D for complex C, 1.494 D for complex D, and 1.423 D for the isolated phenol molecule. The latter is overestimated with respected to the experimental value of 1.224 D,17 as usual in HF estimates.
The MTCS’s for isolated phenol and the four complexes are shown in Fig. 2, where the signatures of the three π* resonances are clear in all cases. The energies (Eres) and widths (Γ) of the lowest-lying anion states,
Momentum transfer cross sections (MTCS), obtained in the SE approximation, for elastic electron scattering by phenol (thin solid line) and the phenol-water complexes shown in Fig. 1.
Momentum transfer cross sections (MTCS), obtained in the SE approximation, for elastic electron scattering by phenol (thin solid line) and the phenol-water complexes shown in Fig. 1.
System . | $E_{\rm res}(\pi ^{\ast }_1)$
. | $\Gamma (\pi ^{\ast }_1)$
. | $E_{\rm res}(\pi ^{\ast }_2)$
. | $\Gamma (\pi ^{\ast }_2)$
. |
---|---|---|---|---|
Phenol | 2.81 | 0.47 | 3.43 | 0.60 |
Complex A | 3.20 | 0.51 | 3.83 | 0.64 |
Complex B | 2.74 | 0.41 | 3.30 | 0.60 |
Complex C | 2.81 | 0.38 | 3.35 | 0.58 |
Complex D | 2.82 | 0.47 | 3.41 | 0.64 |
System . | $E_{\rm res}(\pi ^{\ast }_1)$
. | $\Gamma (\pi ^{\ast }_1)$
. | $E_{\rm res}(\pi ^{\ast }_2)$
. | $\Gamma (\pi ^{\ast }_2)$
. |
---|---|---|---|---|
Phenol | 2.81 | 0.47 | 3.43 | 0.60 |
Complex A | 3.20 | 0.51 | 3.83 | 0.64 |
Complex B | 2.74 | 0.41 | 3.30 | 0.60 |
Complex C | 2.81 | 0.38 | 3.35 | 0.58 |
Complex D | 2.82 | 0.47 | 3.41 | 0.64 |
Figure 3 shows the lowest unoccupied virtual orbitals (LUMO’s) of phenol and the complexes A and B, as obtained from the compact 6-31G(d) basis set built into GAMESS,23 which are routinely employed to support electron transmission spectroscopy (ETS) assignments.22 In all cases, the LUMO and LUMO+1 have π* characters and can be associated with the
Lowest-lying virtual orbitals obtained with compact basis sets for phenol (left column), complex A (center), and complex B (left). The π* and σ* characters are indicated in the panels. The plots were generated with MacMolPlt.31
Lowest-lying virtual orbitals obtained with compact basis sets for phenol (left column), complex A (center), and complex B (left). The π* and σ* characters are indicated in the panels. The plots were generated with MacMolPlt.31
Electron interactions with microsolvated molecules would be expected to pose a number of challenges to experimentalists, such that pointing out signatures of clusterization is an essential point. To address the microsolvation signatures in the DCS, we consider electron scattering in three scenarios. (i) A gas of phenol-water complexes comprising the same number of water molecules (nw), i.e., with the same 1:nw (solute:solvent) ratio. (ii) A cluster-free gaseous admixture of phenol and water, with the phenol concentration also given by 1:nw, i.e., the admixture that would be obtained if all the intermolecular bonds in (i) were broken. (iii) A fictitious gas where the electrons would be incoherently scattered by “groups” with the same 1:nw ratio. The cross sections would be affected by the group size, although no interference between scattering from different molecules within the group would take place. In the case (ii), the DCS is given by the concentration average over the components, DCSadm = 1/(1 + nw)DCSphenol + nw/(1 + nw)DCSwater, while in the case (iii) it is given by the weighted sum DCSgrp = DCSphenol + nwDCSwater. Comparing DCSadm and DCSgrp thus points out size effects (arising from the groups of molecules), while the comparison between the latter and the complex cross section (DCScmp) would highlight interference effects. Since DCS measurements are intrinsically averaged over the target orientations, the differences between DCSgrp and DCScmp also arise from averaging over the complex orientations (the molecules rotate together, ideally as a rigid body) and over the orientations of individual molecules in the group (the molecules rotate independently).
The DCS results at 10 eV are shown in Fig. 4 for scattering angles (θ ⩾ 30o) not significantly affected by the dipole potential (see the supplementary material18 for other energies). We present calculations for the A–D complexes (DCScmp), along with the predictions for the phenol-water admixtures (DCSadm) and the fictitious gas (DCSgrp), as obtained from the DCS’s of the isolated species. In general, the DCSadm magnitudes are significantly smaller than those of DCScmp and DCSgrp, thus suggesting the size effects could be helpful to distinguish clusters with different numbers of water molecules. While interference and orientation averages give rise to discrepancies between DCScmp and DCSgrp at lower scattering angles (≲80o), both results are often very similar above θ ≳ 100o. At least for small clusters, this finding could provide valuable guidance to identify the microsolvated species: the DCS measurements are fairly precise, with errors below 20%, and the DCSgrp estimates can be easily obtained from the DCS’s of the isolated solute and solvent molecules. Although we dot not further explore the microsolvation signatures, DCS estimates for more realistic gases, i.e., admixtures of complexes comprising different numbers of water molecules, could be easily obtained from concentration averages of the appropriate DCSgrp predictions. We also mention that the optimized clusters addressed in this work would reflect those present in supersonic jet expansions, also pointing out the donor/acceptor role of the water molecules in H bonds and their influence in the experimental data.16,32
Differential cross sections (DCSs) at 10 eV for elastic electron scattering by the phenol-water complexes shown in Fig. 1 (solid lines), obtained in the SE approximation. The dashed lines are the DCSs for the molecular “groups” described in the text (DCSgrp). The dotted lines are the DCSs for gaseous admixtures (DCSadm).
Differential cross sections (DCSs) at 10 eV for elastic electron scattering by the phenol-water complexes shown in Fig. 1 (solid lines), obtained in the SE approximation. The dashed lines are the DCSs for the molecular “groups” described in the text (DCSgrp). The dotted lines are the DCSs for gaseous admixtures (DCSadm).
In summary, the SE-level calculated cross sections indicate the expected (de)stabilization trends for the H-bonded phenol-water clusters. These results, along with the VO analysis, also point out that microsolvation impacts the
The authors acknowledge support from the Brazilian agencies FAPESP and CNPq. This work was supported by the CNPq/NSF Cooperative Research Program and by the FAPESP Bioenergy Program (BIOEN 08/58034-0). The present calculations were performed at CENAPAD/SP, CTBE/CNPEM, IFGW/UNICAMP, LFTC-DFis-UFPR, and LCPAD-UFPR. M.H.F.B. acknowledges computational support from Professor C. de Carvalho.