Low energy (1–19 eV) electron scattering from condensed thymidine (dT) I: absolute vibrational excitation cross sections

V. Lemelin*, A. D. Bass, P. Cloutier and L. Sanche
Groupe en Sciences des Radiations, Département de Médecine Nucléaire et Radiobiologie, Faculté de Médecine et Sciences des radiations, Université de Sherbrooke, Québec J1H 5N4, Canada. E-mail: Vincent.Lemelin@usherbrooke.ca

Received 18th June 2019 , Accepted 3rd September 2019

First published on 3rd September 2019


Absolute cross sections (CSs) for vibrational excitation by electrons of energy between 1–19 eV scattering from condensed thymidine (dT) were measured by means of high-resolution electron energy loss spectroscopy (HREELS). The CSs were extracted from electron energy loss spectra of dT condensed on multilayers film of Ar held at about 20 K under ultra-high vacuum (∼1 × 10−11 Torr). dT is one of the most complex molecules to be studied in condensed phase by HREELS. The magnitudes of the vibrational CSs lie within the 10−17 cm2 range. Structures observed in the energy dependence of the vibrational CSs under 3 eV and around 4 eV were compared with previous results of gas- and solid-phase studies on dT and related molecules (e.g., thymine and tetrahydrofuran). These structures were attributed to the formation of shape resonances.


I. Introduction

Low-energy electron (LEE) scattering by molecules is a topic of importance in many areas, such as astrochemistry,1–4 plasma science,5 nanolithography,6 fast-laser photochemistry7–11 and the radiobiological sciences.12–14 In the latter field, LEEs, which are produced abundantly15 by the passage of high-energy ionizing radiation (HEIR) in biological matter, were found to play a major role in the transfer of energy from primary radiation tracks into a medium.16–19 LEEs are highly reactive and interact within biological media via inelastic collisions that ionize and excite vibrationally and electronically biomolecules. They can additionally attach temporarily to biomolecules to form transient anion (TA) states that can modulate and enhance excitation and dissociation cross sections (CSs). As first demonstrated by Boudaïffa et al.,12 single and double strand breaks can be induced in DNA by the formation of TAs and their decay into bond-breaking channels. Since nuclear DNA is considered the primary target for radiotherapy,20,21 it is crucial in radiobiological applications to fundamentally understand the interactions of LEEs with the biomolecules that comprise the basic units of DNA, such as the bases (adenine, cytosine, guanine and thymine) and the phosphate and deoxyribose moieties of the backbone. Furthermore, to properly quantitate the contribution of LEEs to the damage imparted to irradiated biomolecules, the energy they transfer to the biological medium must be known, as well as the number and nature of the ensuing species.

Monte-Carlo (MC)-based simulations can perform nanodosimetric calculations, since they describe event-by-event modifications of the biological medium, predict the number and nature of secondary species created and calculate the deposited energy after the passage of HEIR.22–25 These radiobiological simulations must incorporate all processes, interactions and reactive secondary species within biological matter, including LEEs and the chemical species they produce. To achieve this description, they should incorporate parameters such as CSs (i.e., interaction probabilities) for vibrational and electronic excitations, dissociative electron attachment (DEA), ionization and elastic scattering, to accurately deduce the dose and possible damages to vital biomolecules such as DNA.26,27 Therefore, MC simulations partially rely on absolute CSs or interaction probabilities for the interactions of LEEs with biomolecules as input parameters (e.g., CSs for LEE scattering from the basic units of DNA).28,29 Since biological media exists in a condensed phase, such CSs should be preferably measured in that phase.30

In targeted radionuclide therapy (TRT) and metal nanoparticle (MNP)-aided radiotherapy, absolute LEE CSs are particularly needed to estimate the local dose administered and the resulting biological effectiveness. In MNP-aided radiotherapy with incident photons, the most efficient primary energy ranges from 10 to 80 keV, where the absorption coefficient is much larger, by roughly two orders of magnitude, than that of biological tissue.31 When exposed to such radiation, MNPs emit copious amounts of photo- and Auger electrons, which are produced mainly through the photoelectric effect.32,33 The distribution of electron energies extends from a few eV up to energies close to that of the primary photon energy.34 Electrons with energies greater than the ionization potential of the medium generate further LEEs. In the case of primary charged-particle radiation, the secondary electron energy distribution lies essentially in the low-energy (0–30 eV) range due to their production via the excitation of MNP plasmons;35 this distribution is not limited to a specific range of primary particle energy.35 Considering the penetration depth of electrons in water36 and the spherical geometrical factor, a huge LEE density is created within a submicrometer distance from the surface of the MNP irradiated with either type of radiation.37 These LEEs are the most numerous reactive species created around the MNP and carry a large portion of the extra energy absorbed by the metal. In biological media, this energy is transferred to neighboring biomolecules.37 Therefore, when radionuclides and MNPs are targeted to cancer cells and their components, a large radiation dose can be administered within a nanoscopic volume.33,38–44 Calculation of such a localized dose and the related biological effectiveness requires nanodosimetric models, which heavily rely on LEE CSs.38,39,45,46

High-resolution electron energy loss spectroscopy (HREELS) is a powerful technique to study the interactions of LEEs with molecules and generate absolute CSs. Using HREELS, our laboratory has provided vibrational and electronic excitation CSs for LEEs scattering from several DNA constituents and/or their analogs in the condensed phase. These include the DNA bases, cytosine, adenine, thymine and the analog pyrimidine,47–52 as well as tetrahydrofuran (THF) and dimethyl phosphate, which represent respectively the sugar ring and phosphate group within the backbone of DNA.53–56 These earlier results provided LEEs scattering CSs from isolated fundamental units of DNA. However, since DNA is an assembly of different constituents, the interactions of LEEs with a base or the backbone are likely affected by the presence of the differing neighboring, constituent molecules. To create more realistic nanodosimetric calculations with MC simulation, information is required on how the absolute CSs for more complex assemblies of multiple DNA sub-units are modulated by this increased complexity. Hence, the measurement of absolute CSs for more complex molecular assemblies such as a nucleoside (i.e., a DNA base linked to the deoxyribose) is of considerable interest in radiotherapy. Here, we report absolute CSs for vibrational excitation of condensed thymidine (dT) by electrons of 1–19 eV. As seen in Fig. 1, the nucleoside dT consists of the base thymine linked to the sugar moiety of the DNA backbone (i.e., deoxyribose). There exist some studies on the interactions of LEEs with dT,57,58 but absolute CSs for nanodosimetry are unavailable in any phase. While preliminary results on dT were presented at a conference,59 this is the first comprehensive HREELS study on the measurements of absolute vibrational CSs for LEE scattering from a nucleoside.


image file: c9cp03447a-f1.tif
Fig. 1 Chemical structure of a short DNA strand. The thymine base and the sugar moiety (deoxyribose) linked together corresponds to thymidine (dT). Chemical structures of the two constituents of thymidine: (A) tetrahydrofuran (THF – model of deoxyribose) and (B) thymine.

II. Experiment

Measurements were performed using a HREELS apparatus, which has been described in detail previously.60 Briefly, the spectrometer is contained in a cryopumped ultra-high vacuum (UHV) chamber that reaches a base pressure of ∼1 × 10−11 Torr. It consists of two hemispherical electrostatic deflectors, which comprise the monochromator and the analyzer. These latter have a combined resolution of about 23 meV full width at half maximum (FWHM). The monochromator is fixed at an incidence angle θ0 = 15° from the normal of the sample surface, while the analyzer is set at analyzing angle of θd = 45° in the opposite azimuth for all measurements.

dT in powder form (Sigma-Aldrich, 99%) was sublimated under UHV conditions with a double-stage oven system that was described previously.49,61 This double-stage oven is kept at a base pressure of 3 × 10−10 Torr in a load-lock chamber attached to the principal chamber that houses the HREELS. The first step requires raising the temperature of a crucible housed in the load-lock. The crucible contains the dT, whose temperature is elevated to the sublimation point of the powder (127 °C in our experiments). This temperature is as low or lower than others previously reported to volatilize dT without its degradation.57,58,62,63 Afterwards, the oven is moved 2 mm in front of a ceramic tip for approximately 90 seconds. Sublimated dT escapes from the crucible via an aperture and condenses upon the ceramic tip at ambient temperature. The tip is then translated into the principal chamber and placed at about 5 mm from a clean Pt substrate, where it is heated to evaporate the dT molecules, which condense onto the substrate cooled at 20 K.49 Before use, the dT sample is degassed by heating the crucible at ∼50 °C for several hours to remove residual solvent or water impurities. In the present experiments, 1.7 monolayer (ML) of dT is deposited onto a 3 ML spacer film of Ar already condensed on the Pt substrate from a gas handling manifold.60 The Ar film reduces image-charge polarization and dT–Pt interactions.64 Also, the large band gap of Ar in the solid state ensures that electron scattering events within the spacer are mainly elastic, creating a background in the electron energy loss (EEL) spectra.48,49,54,55,65 Increasing the number of Ar layers will only induce a higher background that can easily be removed. Furthermore, it is to be noted that the intensity of Ar exciton formation at 12.06 eV64 is many times lower than those of the vibrational excitation of a molecule. Competition between exciton formation and molecular vibrational excitation is expected to be negligible and hence does not affect significantly the vibrational cross section measurements. Generally, vibrational modes are 15–20 times more intense than electronic modes in the EEL measurements.47,51,53–56

To determine the number of layers of dT deposited by sublimation, we measure the attenuation of the specular reflectivity of the Ar substrate.49 The 1.7 ML thickness used in these experiments corresponds to an attenuation of ∼82% of the specular reflectivity of the Ar substrate. To generate absolute vibrational CSs from HREEL spectra, this thickness must be converted to a surface number density (ns). Previously, for gases and liquids, the molecular coverage was obtained by the pressure drop of the calibrated volume of the gas handling manifold.48,49,54,55 Here, since the dT is deposited by sublimation, we need to rely on the literature and simple calculations to approximate the surface number density of our films.49 Using the unit cell parameters of self-assembly of dT on Au(111) in the work of Yang et al.,66 we calculate a ns of 1.3 × 1014 molecules cm−2 for a monolayer of dT. A second calculation can be made using the conformation data for dT on Au(100) of Bogdan and Morari,67 from which we obtain a value of about 1.5 × 1014 molecules cm−2. As proposed by Hervé du Penhoat et al.68 for thymine layers, the number density of dT can also be estimated from an average thickness of ∼3 Å for a single disordered ML, the density and the molar mass of dT and the Avogadro's number NA. This calculation gives ns = 1.12 × 1014 molecules cm−2. This is consistent with the value of 2.1 × 1014 molecules cm−2 for thymine, considering that the geometry of dT is roughly twice as large as that of thymine, which does not have the deoxyribose unit. Considering these estimates, we choose an average value of 1.3 × 1014 molecules cm−2 for the surface number density of one ML of dT which is close to the ns of 9.36 × 1013 molecules cm−2 for a nucleoside in DNA strands.69

III. Scattering model

Absolute vibrational CSs of dT are generated by the single collision model previously developed from a multiple scattering model.48,70 In the single collision model, for a near normal (θ0 = 15°) incident electron current I0 of energy E0, it has been shown that the analyzer measures at a given angle (e.g., θd = 45°) a current of electrons I that transfers part of their energy (i.e., EE0) into the film such that48,51
 
image file: c9cp03447a-t1.tif(1)
where σr(E0,EE0) corresponds to the CS for an incident electron with E0 to deposit an energy of EE0 in the film and be backscattered into vacuum. The factor 1/cos[thin space (1/6-em)]θ0 accounts for the projection of the incident electron current onto the surface. The quantity ns is the surface number density already discussed in the previous section. The term I0(θd, E0) represents the effective incident electron current. It can be seen as that fraction of the total incident electron current I0 backscattered in the direction of the analyzer (θd) (i.e., in the direction of the measured EEL spectrum I(θd, E0, EE0)) by a material with a diffuse elastic reflectivity of one.48 With the method previously described,48 the value for the effective current I0(θd, E0) is 16[thin space (1/6-em)]500 count s−1 eV (±3%).

To obtain the absolute CS values as functions of the incident electron energy, EEL spectra (i.e., I(θd, E0, EE0)) are recorded at multiple incident energies E0. The vibrational excitation-energy losses in each spectrum are identified and fitted with a Gaussian function. From the area under each Gaussian, the absolute CS (σr(E0,EE0)) value for the corresponding vibrational modes are obtained with eqn (1). As explained in previous investigations,48–50,55 dT molecules in the films are expected to be randomly orientated due to rapid cooling of their vapor onto the cold Ar spacer.49,55 Accordingly, the backscattered intensity is considered to be isotropic.70–72 This condition permits EEL measurements to be made at a single scattering geometry, such that by normalizing to the effective incident current I0(θd, E0), the single scattering model (eqn (1))48 leads to the CS values corresponding to the electrons backscattered over the whole half-angular space.48 The CSs generated in the present work thus correspond to integral and absolute values. Since the instrument is equipped with double-zoom lenses on both the monochromator and the analyzer, the focus and the analyzing solid angle are constant over a large energy range.60 These lenses are also adjusted at each incident energy for a maximum signal, which gives a fairly uniform performance over the energy range of the present work. The performance of the apparatus is reduced at low energies (<1.5 eV) and at energies greater than 16 eV. However, at these energies, few data are recorded and the vibrational CSs are expected to be small compared to CSs in the range 4–10 eV.

IV. Results and discussion

A. Vibrational spectra

Vibrational EEL spectra for 1.7 ML of dT deposited on 3 MLs of Ar were recorded for electron incident energies E0 between 1 and 19 eV. For incident energies from 1 to 12 eV, spectra were recorded at every 0.5 eV and for energies between 12 and 19 eV at every 1 eV. Fig. 2 shows the EEL spectrum of condensed dT. This representative spectrum shows multiple structures related to excitation of different vibrational modes, which are consistent with the Raman study of Tsuboi et al.73 on dT and thymine. We should note that while HREELS is an excellent technique for measuring LEE scattering cross sections, the spectra it generates are unlike from those obtained with photons for two reasons: selection rules for excitation by LEEs are different, so that optically forbidden energy loss processes may be present, and the energy loss resolution is far below that of optical techniques. These considerations were taken into account in our assignments of each vibrational mode. In the present work, our spectra were assigned by reference to the work of Tsuboi et al.73 The strong peak at zero-eV energy loss corresponds to electrons that experienced elastic collisions with dT, as well as those that have undergone small energy transfers to phonons.74 This elastic peak also contains vibrational excitations of dT at very small energy losses (<80 meV) that cannot be resolved. The thin solid line passing through the data points represents the fit obtained from the sum of the Gaussian functions assigned to the various vibrational modes of dT. The background arising from electrons backscattered from the Ar spacer and the Pt substrate74 is represented by the dashed line. The energies, assignments and FWHM of each vibrational mode observed in EEL spectra are shown in Table 1 and compared to their excitation energies measured with Raman spectroscopy by Tsuboi et al.73 We have previously studied two constituents of dT using HREELS; thymine49 and tetrahydrofuran54 separately (see Fig. 1), most vibrational modes observed of dT were previously observed in spectra for one or the other of the earlier molecules. As seen from Fig. 1, thymine is the DNA base present in dT and THF approximates the ring of the deoxyribose to which the thymine is joined. Small discrepancies between the measured energy loss of certain vibrational excitations and the values reported in the Raman study are observed. These differences are likely related to the fitting process and uncertainty in localization of the structures, particularly within energy loss ranges where several vibrational modes are near each other. Furthermore, differences are also observed between the energies of vibrational modes in dT and those of the isolated constituents, thymine and THF. These differences may also be partially due to the fitting and localization of vibrational structures, but it is possible that they result from the chemical bonding between the two constituents, which can shift the energy of certain vibrational modes. Such changes are the subject of the following publication [V. Lemelin, A. D. Bass, P. Cloutier and L. Sanche, DOI: 10.1039/c9cp03448j]. Also, it must be noted that dT contains the entire sugar moiety (deoxyribose – C5H10O4) rather than THF (C4H8O) which lacks the CH2OH and OH groups; this also can be expected to alter the energy of vibrational modes. Nevertheless, the energies of the vibrational modes found in dT generally agree well with previous studies.
image file: c9cp03447a-f2.tif
Fig. 2 A representative electron energy loss spectrum recorded with electrons of 3.5 eV incident on 1.7 monolayer (ML) of thymidine deposited on 3 MLs of Ar. The fit represented by the thin solid line passing through the data points is produced by the sum of the Gaussian functions associated with each vibrational energy lost. The energies and widths of the vibrational modes are listed in Table 1. The dashed line accounts for the background signal produced by the Ar spacer and the Pt substrate.
Table 1 Energies (meV) of vibrational modes of thymidine and thymine from Raman data73 compared to those obtained from the present measurements of condensed dT and those of the two constituents (thymine and tetrahydrofuran (THF)) obtained with HREELS.49,54 All listed results were recorded in the condensed phase. Abbreviations: umbr. – umbrella motion, sciss. – scissoring motion
Assignment dT Raman73 Thymine Raman73 Thymine HREELS49 THF HREELS54 dT HREELS (this work) Mode FWHM (this work)
THF ring bend 83 76 81 80 23
Thymine ring breath 95 91 95 98 23
98 99
THF C–C stretch 115 114 114 23
THF C–O–C stretch 132 129 131 131 23
Thymine ring bend or THF CH2 twist and wag 148 150 150 146 150 23
152 154, 155
THF CH2 twist and wag 169 169 163 168 23
THF CH3 degenerate def and CH bend, umbr. and sciss 180 187 180 181 182 24
Thymine C[double bond, length as m-dash]O and C[double bond, length as m-dash]C stretch 206 207 213 210 25
209 210
THF CH2 stretch 357 361 23
Thymine and THF CH stretch 375 371 375 30
Thymine NH stretch 417 417 30
Deoxyribose OH stretch (this study) 439 25


Complete details of each vibrational mode are found in the Raman study of Tsuboi et al.73 Briefly, starting at the lowest resolved energy loss, we find the ring bending mode of THF at 80 meV. The next structure at 98 meV is ascribed to the breathing mode of the thymine ring. The vibrational modes around 114 and 131 meV are due to the stretching of C–C and C–O–C respectively in the THF molecule. The peak observed at 150 meV can be attributed to one mode of each molecule: the bending and deformation of the thymine ring and the twisting and wagging of CH groups of THF. The mode found around 168 meV is also related to the twisting and wagging of CH groups of THF. The structure around 182 meV correlates with the degenerate deformations and motions of CH groups in both molecules. The next structure found at 210 meV is ascribed to the stretching of C[double bond, length as m-dash]O and C[double bond, length as m-dash]C bonds of thymine. The peak at 357 meV corresponds to the stretching mode of CH in THF. The structure around 375 meV is related to the CH stretching modes in both molecules. The mode found at 417 meV is also observed in thymine and it is due to the stretching mode of NH. Finally, the structure at 439 meV correlates with the stretching mode of OH groups.

In the ESI, Fig. S1 shows EEL spectra measured for various incident electron energies between 1 and 19 eV. Each of these spectra were fitted with Gaussian functions as shown in Fig. 2. The energy and FWHM of the Gaussian functions associated with each vibrational mode were kept constant between spectra and only the amplitude was allowed to vary. Generally, the FWHM of a given mode corresponds to the resolution of the spectrometer (i.e., 23 meV). However, since the resolution of HREELS is not as good as that of IR or Raman spectroscopies, closely spaced vibrational modes can appear as single energy loss feature. Thus, in some cases, to produce the best fit possible, certain Gaussian functions will have a FWHM that is larger than the experimental resolution of the apparatus.

B. Vibrational scattering cross sections

Fig. 3 presents the CSs values as a function of electron incident energy for each vibrational mode identified in the EEL spectra of dT. The numerical values of these CSs are tabulated in Table 2. The ±5% error bars on data points of Fig. 3 indicate the uncertainty related to the Gaussian curve fitting as well as the reproducibility of the experimental measurements. These error bars neither include uncertainties associated with the stability of the incident current over time (±5%) nor systematic error in the effective incident current of ±3% or those in the determination of the molecular coverage ns (±15%). When summed, the total absolute error on the present CSs amounts to ±28%. In Fig. 3, similarities are observed between the incident energy dependences of the CSs of several vibrational modes. Effectively, maxima at specific energies in multiple CSs are observed and these structures are usually related to the formation of resonant compound states of the molecule and incident electron, i.e., TAs.30,75 Considering the broadness of the features in Fig. 3, they are likely to be produced from the decay of pure shape resonances. Core-excited resonances are rarely seen as broad maxima in vibrational CSs, since their lifetimes are usually too long.76 Furthermore, shape resonances are coupled only to the vibrational excited states of the target in the ground state, since the molecule and the TA are not excited electronically. In a shape resonance, temporary capture of an electron in an unfilled molecular orbital (at a particular energy) immediately produces a displacement of the nuclear distances between the constituent atoms of the molecule, when the orbital retaining the extra electron is bonding or anti-bonding. Thus, during the TA lifetime, these atoms are set in motion and, when the extra electron autodetaches, they move back toward their initial equilibrium position. This process imparts considerable nuclear-motion energy into various vibrational modes, causing an enhancement of the vibrational CSs at the energy of the TA. The transferred energy depends on the lifetime of the TA, up to about half of a vibrational period of the mode being considered.
image file: c9cp03447a-f3.tif
Fig. 3 Incident electron energy dependence of the scattering cross sections for various vibrational excitations of thymidine. Numerical values of these cross sections are presented in Table 2. The vertical error bars account only for the reproducibility of the measurement and uncertainty in the Gaussian fit of about ±5%. When accounting for all experimental uncertainties, the total absolute error amounts to ±28%. The interpolation line serves only for better visualisation.
Table 2 Absolute cross sections σ(10−17 cm2) for vibrational excitation by 1–19 eV (E0) electrons impinging onto a 1.7 monolayer of thymidine deposited on an inert Ar spacer
E0 Vibrational modes
υring bend (THF) (80 meV) υbreath (T) (98 meV) υC–C (THF) (114 meV) υC–O–C (THF) (131 meV) υC–O–C (T) (150 meV) υCH2 (THF) (168 meV) υbend[thin space (1/6-em)]CH+CH2 (T + THF) (182 meV) υC=O (T) (210 meV) υCH2 (THF) (361 meV) υCH (T + THF) (375 meV) υNH (T) (418 meV) υOH (dT) (439 meV)
  σ
1 1.4 0.98 0.48 0.79 0.67 0.67 0.78 0.93 0.28 0.59 0.42 0.23
1.5 1.4 1.1 0.68 1.1 1.1 1.3 1.2 1.3 0.90 1.5 0.79 0.34
2 1.3 1.1 0.68 1.1 1.1 1.2 1.4 1.4 0.79 1.7 0.70 0.34
2.5 1.1 1.3 0.70 1.4 1.3 1.4 1.7 1.5 1.1 2.1 0.74 0.40
3 1.0 1.5 0.73 1.7 1.5 1.7 1.9 1.7 1.1 2.4 0.74 0.38
3.5 1.2 1.5 0.85 1.9 1.6 2.0 1.9 1.9 1.6 2.1 0.80 0.31
4 1.0 1.2 0.77 1.8 1.5 1.8 1.5 1.5 1.3 1.6 0.59 0.23
4.5 0.93 1.1 0.76 1.9 1.6 1.8 1.8 1.5 1.1 1.8 0.59 0.25
5 0.98 0.99 0.82 1.8 1.6 1.8 1.5 1.4 1.1 1.4 0.55 0.20
5.5 0.95 0.88 0.75 1.6 1.4 1.6 1.3 1.3 1.1 0.99 0.54 0.19
6 0.79 0.74 0.61 1.3 1.2 1.3 1.2 1.0 0.73 0.86 0.38 0.16
6.5 0.75 0.71 0.60 1.3 1.2 1.2 1.1 0.93 0.70 0.77 0.38 0.16
7 0.63 0.63 0.51 1.1 1.1 1.1 1.0 0.82 0.61 0.70 0.32 0.15
7.5 0.62 0.61 0.52 1.1 1.0 1.1 0.91 0.73 0.64 0.59 0.31 0.13
8 0.52 0.53 0.44 0.90 0.85 0.91 0.84 0.68 0.52 0.56 0.24 0.12
8.5 0.50 0.51 0.44 0.84 0.78 0.87 0.78 0.56 0.52 0.51 0.23 0.11
9 0.42 0.50 0.37 0.79 0.71 0.73 0.80 0.53 0.37 0.54 0.18 0.11
9.5 0.48 0.44 0.40 0.70 0.63 0.73 0.63 0.47 0.44 0.41 0.18 0.08
10 0.46 0.43 0.37 0.67 0.59 0.68 0.58 0.44 0.42 0.39 0.16 0.06
10.5 0.45 0.42 0.36 0.60 0.54 0.62 0.55 0.41 0.42 0.35 0.14 0.05
11 0.44 0.41 0.34 0.56 0.54 0.55 0.62 0.45 0.33 0.46 0.14 0.06
11.5 0.45 0.41 0.34 0.53 0.49 0.53 0.57 0.43 0.31 0.40 0.13 0.04
12 0.42 0.37 0.30 0.46 0.43 0.47 0.51 0.39 0.27 0.34 0.10 0.03
13 0.38 0.33 0.27 0.36 0.36 0.41 0.41 0.33 0.22 0.23 0.08 0.02
14 0.38 0.34 0.26 0.34 0.34 0.37 0.41 0.34 0.17 0.20 0.07 0.01
15 0.35 0.32 0.23 0.29 0.29 0.32 0.38 0.29 0.12 0.17 0.06 0.01
16 0.31 0.28 0.20 0.25 0.25 0.27 0.32 0.26 0.09 0.13 0.05 0.005
17 0.27 0.25 0.17 0.21 0.21 0.24 0.26 0.22 0.08 0.10 0.04 0.005
18 0.23 0.22 0.15 0.17 0.17 0.20 0.21 0.18 0.06 0.07 0.03 0.005
19 0.23 0.21 0.13 0.16 0.15 0.18 0.18 0.15 0.04 0.06 0.02 0.005


Two TAs appear in the curves of Fig. 3: a common weak structure at low energies below E0 = 3 eV and a strong maximum around 4 eV. These two resonances do not decay into all vibrational modes. For example, the highest vibrational mode at an energy loss of 439 meV (υOH) only exhibits the low-energy resonance. Since resonances can be observed in many decay channels (DEA, vibrational and electronic excitations, etc.), the ones observed here can be compared to those found in other earlier experimental or theoretical studies on electron interactions with dT. DEA to gas-phase dT has been studied by Ptasinska et al.;57 a broad resonance at an incident electron energy around 5.5 eV, was detected. This resonance could be related to our broad feature at 4 eV, if the downward shift in energy of about 1.5 eV can be explained by the induced polarization potential of the TA in the solid.54 However, such a shift requires a TA lifetime of the order of the vibrational period of the phonon modes, which is unlikely true for shape resonances that are usually responsible for vibrational excitation.77 In the same study, Ptasinska et al.40 observed two low-lying resonances at 1.2 and 1.8 eV, which may be related to most of the structures observed under 3 eV in the present study. These authors57 concluded that the low-lying resonances were shape resonances with the 1.8 eV electron localized on thymine and the 1.2 eV electron localized on the sugar moiety. A similar resonance at low energy was found by Abdoul-Carime et al.,58 who investigated the fragmentation of gas-phase dT via DEA. They found a TA under 3 eV that was also attributed to a shape resonance; it also can be correlated with some of the structures observed under 3 eV in Fig. 3.

Since dT is composed of the sugar and thymine moiety, we can also compare the resonances observed in the present work to those seen in earlier HREEL studies of both isolated molecules. For instance, Lévesque et al.49 measured the absolute vibrational CSs of thymine and they observed a resonance at the same energy as our second one around 4 eV. They correlated this TA to the shape resonance found in electron transmission spectroscopy78 and concluded that it corresponds to an extra electron quasi-bound to the third empty anti-bonding molecular orbital (π*) of thymine at 4.05 eV.78 Lévesque et al.49 did not observe any resonance below 3 eV. Our low-lying anion state may therefore be related to the one seen in measurements with THF (deoxyribose analog) at around 2.5 eV in our previous condensed-phase investigations on the vibrational CSs of the molecule.54 This TA was also detected by Allan et al.76 in their excitation functions of THF and attributed to a shape resonance. Another resonance observed by HREELS at 4.5 eV in THF54 could contribute to the maxima found at 4 eV in the present CSs of dT. Two other resonances were also found in condensed THF near 9.5 and 12.5 eV,54 but no maximum is observed in the present CSs of dT at these energies. Since these resonances are already weak in THF, they may have become imperceptible in the CSs of dT. Furthermore, chemical bonding of the two moieties to form dT may delocalize the extra electron orbital and thus considerably modify the electron capture CSs and resonance energies.

On the theoretical level, Winstead et al.79 calculated the elastic integral CSs (ICSs) for LEEs scattering from thymine and dT. They observed multiple resonances at low energy around 0.3, 1.9 and 5.7 eV for thymine. The first two resonances correlate well with those found by Aflatooni et al.78 at 0.29, 1.71 eV, but the third resonance is far from the one measured experimentally. Their ICS calculations for dT79 showed a close resemblance to those of thymine. Hence, the same resonances resulted from the computation with a slight upward shift of 0.1–0.3 eV for the two low-lying TAs. These two anion states correlate well with those seen under 3 eV in our vibrational CSs of dT. However, the third shape resonances observed around 5.7 eV does not agree with our measurement of a maximum around 4 eV. Since multiple experimental studies, including those of the present work, have shown a resonance around 4 eV for thymine and dT,49,78 further theoretical studies of LEEs scattering from dT are needed to elucidate these discrepancies.

Table 3 summarizes the resonances found in the literature for dT, thymine and THF as well as the assignments made by the authors. Our results correlate well with most resonances previously found for dT as well as those of the two constituents. We agree with all the authors on the assignments of the resonances; we attribute the structures found in the vibrational CSs of dT to the decay of shape resonances into the autodetachment channel, leaving the molecule in a multitude of well-defined vibrational modes of the ground state.

Table 3 Resonances energies of thymidine, thymine and tetrahydrofuran observed in gas and solid phase studies
Source Molecule studied Method Er (eV) Character
Gas phase
Ptasinska et al.57 Thymidine DEA 1.2 Shape
1.8
5.5–12
Abdoul-Carime et al.58 Thymidine DEA <3 Shape
 
Solid phase
Lévesque et al.49 Thymine Absolute vibrational CSs 4
Aflatooni et al.78 Thymine Electron transmission spectroscopy 4.05 Shape
Lemelin et al.54 Tetrahydrofuran Absolute vibrational CSs 2.5 Shape
4.5 Shape
9.5
12.5
This present study Thymidine Absolute vibrational CSs <3 Shape
4 Shape
 
Calculations
Winstead and McKoy79 Thymine Integral elastic CSs 0.3 Shape
1.9
5.7
Thymidine   0.4
2.2
5.7


V. Summary

We reported absolute vibrational CSs for electron scattering from condensed dT molecules. dT is a nucleoside composed of thymine (a DNA base) and the deoxyribose (sugar in the DNA backbone). EEL spectra were acquired with various incident electron energies between 1 and 19 eV for 1.7 ML of dT deposited on an inert Ar spacer previously condensed on a Pt substrate. dT is one of the most complex molecules studied with HREELS since it is composed of two DNA constituents. The vibrational excitations identified in the EEL spectra compared favorably to those reported by Raman spectroscopy study. They also correlate well with previous HREELS studies on vibrational excitations of the two isolated constituents (thymine and THF). A deconvolution of the vibrational peaks in each spectrum has been made and the absolute CSs of dT were calculated using the area under the Gaussian associated with each excitation mode. Furthermore, the resonances observed in the incident electron-energy dependence of the CSs correlated well with those observed in other gas and solid phase studies of dT, thymine and THF.

In the energy dependence of these CSs, a common strong maximum is observed around 4 eV and a weaker resonance under 3 eV. The structures were interpreted to result from the decay of shape resonances into the autoionization channel, as did the authors that also observed resonances for dT, thymine and THF at similar energies. In the following paper, we undertake a comparative, quantitative study between the absolute CSs of dT and those of its two constituents to better elucidate the effects on CS values when two molecules are combined to form another. In other words, to what extent are the vibrational CSs of dT related to those of its constituents?

Finally, we should note that when considering the authentic biological environment of DNA, the interactions of LEEs will be modified due to the presence of other biomolecules. For instance, studies on larger biomolecules such as nucleosides, nucleotides and DNA strands are of particular relevance for radiobiology not least because the complex structure of DNA may, via diffraction and other cooperative effects, further modulate electron interactions.75 Additionally, surrounding targets with water molecules (e.g., to simulate the effects of structural water in DNA) to observe the effect on absolute CS values, represents another useful refinement.80,81 It would therefore be interesting to incorporate water molecules into our target in future HREELS absolute CSs measurements, to move closer to biological reality, as many theoretical calculations and experimental studies have already attempted to do.82–89 Since the biological conditions are not accounted in the present work, the CSs values presented here represent a first approximation. However, they are the most relevant values available, since they are measured in condensed phase and for a molecule with a level of complexity higher than ever before.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

V. Lemelin gratefully acknowledges the financial support from the Fonds de Recherche du Québec – Santé (FRQS) in the form of a doctoral research award. Thanks are extended to Marc Michaud for his help in the experimental part of this project. This research is supported by CIHR grant PJT-162325.

References

  1. S. Esmaili, A. D. Bass, P. Cloutier, L. Sanche and M. A. Huels, J. Chem. Phys., 2017, 147, 224704 CrossRef PubMed.
  2. S. Esmaili, A. D. Bass, P. Cloutier, L. Sanche, M. A. Huels and M. A. Huels, J. Chem. Phys., 2018, 164702 CrossRef PubMed.
  3. P. Auburger, I. Kemeny, C. Bertram, M. Ligges, M. Bockstedte, U. Bovensiepen and K. Morgenstern, Phys. Rev. Lett., 2018, 121, 206001 CrossRef CAS PubMed.
  4. C. J. Bennett and R. I. Kaiser, Astrophys. J., 2007, 661, 899 CrossRef CAS.
  5. A. I. Lozano, K. Krupa, F. Ferreira da Silva, P. Limão-Vieira, F. Blanco, A. Muñoz, D. B. Jones, M. J. Brunger and G. García, Eur. Phys. J. D, 2017, 71, 226 CrossRef.
  6. C. Wu, Y. Makiuchi and C. Chen, Lithography, 2010 Search PubMed.
  7. M. V. Pavliuk, S. Gutiérrez Álvarez, Y. Hattori, M. E. Messing, J. Czapla-Masztafiak, J. Szlachetko, J. L. Silva, C. M. Araujo, D. L. A. Fernandes, L. Lu, C. J. Kiely, M. Abdellah, P. Nordlander and J. Sá, J. Phys. Chem. Lett., 2019, 1743 CrossRef CAS PubMed.
  8. D. Zhu, L. Zhang, R. E. Ruther and R. J. Hamers, Nat. Mater., 2013, 12, 836 CrossRef CAS PubMed.
  9. J. Szczerbiński, L. Gyr, J. Kaeslin and R. Zenobi, Nano Lett., 2018, 18, 6740 CrossRef PubMed.
  10. Y. Dong, D. Parobek, D. Rossi and D. H. Son, Nano Lett., 2016, 16, 7270 CrossRef CAS PubMed.
  11. S. Simoncelli, E. L. Pensa, T. Brick, J. Gargiulo, A. Lauri, J. Cambiasso, Y. Li, S. A. Maier and E. Cortés, Faraday Discuss., 2019, 214, 73 RSC.
  12. B. Boudaïffa, P. Cloutier, D. Hunting, M. A. Huels and L. Sanche, Science, 2000, 287, 1658 CrossRef PubMed.
  13. M. A. Huels, B. Boudaiffa, P. Cloutier, D. Hunting and L. Sanche, J. Am. Chem. Soc., 2003, 125, 4467 CrossRef CAS PubMed.
  14. W. Chen, S. Chen, Y. Dong, P. Cloutier, Y. Zheng and L. Sanche, Phys. Chem. Chem. Phys., 2016, 18, 32762 RSC.
  15. V. Cobut, Y. Frongillo, J. P. Patau, T. Goulet, M. J. Fraser and J. P. Jay-Gerin, Radiat. Phys. Chem., 1998, 51, 229 CrossRef CAS.
  16. D. Bouchiha, J. D. Gorfinkiel, L. G. Caron and L. Sanche, J. Phys. B: At., Mol. Opt. Phys., 2006, 39, 975 CrossRef CAS.
  17. H. V. Duque, T. P. T. Do, M. C. A. Lopes, D. A. Konovalov, R. D. White, M. J. Brunger and D. B. Jones, J. Chem. Phys., 2015, 142, 124307 CrossRef CAS PubMed.
  18. M. Inokuti, Appl. At. Collis. Phys., Academic Press, New York, 1983, pp. 179–236 Search PubMed.
  19. S. M. Pimblott, J. A. Laverne and A. Mozumder, J. Phys. Chem., 1996, 100, 8595 CrossRef CAS.
  20. A. McNamara, C. Geng, R. Turner, J. R. Mendez, J. Perl, K. Held, B. Faddegon, H. Paganetti and J. Schuemann, Phys. Med., 2017, 33, 207 CrossRef PubMed.
  21. M. Rezaee, R. P. Hill and D. A. Jaffray, Radiat. Res., 2017, 188, 123 CrossRef CAS PubMed.
  22. A. Muñoz, F. Blanco, G. Garcia, P. A. Thorn, M. J. Brunger, J. P. Sullivan and S. J. Buckman, Int. J. Mass Spectrom., 2008, 277, 175 CrossRef.
  23. D. Alloni, A. Campa, W. Friedland, L. Mariotti and A. Ottolenghi, Int. J. Radiat. Biol., 2012, 88, 77 CrossRef CAS PubMed.
  24. D. Liljequist, T. Liamsuwan and H. Nikjoo, Int. J. Radiat. Biol., 2012, 88, 29 CrossRef CAS PubMed.
  25. A. G. Sanz, M. C. Fuss, A. Muñoz, F. Blanco, P. Limão-Vieira, M. J. Brunger, S. J. Buckman and G. Garcia, Int. J. Radiat. Biol., 2012, 88, 71 CrossRef CAS PubMed.
  26. W. Liu, Z. Tan, L. Zhang and C. Champion, Radiat. Environ. Biophys., 2017, 56, 99 CrossRef CAS PubMed.
  27. M. Dingfelder, Radiat. Prot. Dosimetry, 2006, 122, 16 CrossRef CAS PubMed.
  28. M. Michaud, M. Bazin and L. Sanche, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2013, 87, 032701 CrossRef.
  29. H. Nikjoo, S. Uehara, D. Emfietzoglou and F. A. Cucinotta, Radiat. Meas., 2006, 41, 1052 CrossRef CAS.
  30. L. Sanche, in Radic. Radic. Ion React. Nucleic Acid Chem., ed. M. M. Greenberg, John Wiley & Sons, Inc., 2010, pp. 239–294 Search PubMed.
  31. J. H. Hubbell and S. M. Seltzer, Natl. Inst. Stand. Technol., 2004 DOI:10.18434/T4D01F.
  32. S. Her, D. A. Jaffray and C. Allen, Adv. Drug Delivery Rev., 2017, 109, 84 CrossRef CAS PubMed.
  33. W. Ngwa, R. Kumar, S. Sridhar, H. Korideck, P. Zygmanski, R. A. Cormack, R. Berbeco and G. M. Makrigiorgos, Nanomedicine, 2014, 9, 1063 CrossRef CAS PubMed.
  34. H. N. McQuaid, M. F. Muir, L. E. Taggart, S. J. McMahon, J. A. Coulter, W. B. Hyland, S. Jain, K. T. Butterworth, G. Schettino, K. M. Prise, D. G. Hirst, S. W. Botchway and F. J. Currell, Sci. Rep., 2016, 6, 1 CrossRef PubMed.
  35. A. V. Verkhovtsev, A. V. Korol and A. V. Solov’yov, J. Phys. Chem. C, 2015, 119, 11000 CrossRef CAS.
  36. J. Meesungnoen, J.-P. Jay-Gerin, A. Filali-Mouhim and S. Mankhetkorn, Radiat. Res., 2002, 158, 657 CrossRef CAS PubMed.
  37. S. J. McMahon, W. B. Hyland, M. F. Muir, J. A. Coulter, S. Jain, K. T. Butterworth, G. Schettino, G. R. Dickson, A. R. Hounsell, J. M. O’Sullivan, K. M. Prise, D. G. Hirst and F. J. Currell, Sci. Rep., 2011, 1, 18 CrossRef PubMed.
  38. M. N. Gaze, Phys. Med. Biol., 1996, 41, 1895 CrossRef CAS PubMed.
  39. M. R. Zalustky, in Handb. Nucl. Chem., ed. A. Vertes, S. Nagy, Z. Klencsar, R. G. Lovas, and F. Rosch, Springer, Heidelberg, 2013, pp. 2179–2203 Search PubMed.
  40. C. Brazzale, F. Mastrotto, P. Moody, P. D. Watson, A. Balasso, A. Malfanti, G. Mantovani, P. Caliceti, C. Alexander, A. T. Jones and S. Salmaso, Nanoscale, 2017, 9, 11137 RSC.
  41. S. Yook, Z. Cai, Y. Lu, M. A. Winnik, J.-P. Pignol and R. M. Reilly, J. Nucl. Med., 2016, 57, 936 CrossRef CAS PubMed.
  42. M. Shi, B. Paquette, T. Thippayamontri, L. Gendron, B. Guerin and L. Sanche, Int. J. Nanomed., 2016, 11, 5323 CrossRef CAS PubMed.
  43. Z. Cai, N. Chattopadhyay, K. Yang, Y. L. Kwon, S. Yook, J.-P. Pignol and R. M. Reilly, Nucl. Med. Biol., 2016, 43, 818 CrossRef CAS PubMed.
  44. C. J. Yang and D. B. Chithrani, Curr. Top. Med. Chem., 2015, 16, 271 CrossRef PubMed.
  45. P. J. Eifel, Nat. Clin. Pract. Oncol., 2006, 3, 248 CrossRef CAS PubMed.
  46. Y. Zheng, D. J. Hunting, P. Ayotte and L. Sanche, Phys. Rev. Lett., 2008, 100, 198101 CrossRef PubMed.
  47. M. Bazin, M. Michaud and L. Sanche, J. Chem. Phys., 2010, 133, 155104 CrossRef CAS PubMed.
  48. P. L. Levesque, M. Michaud and L. Sanche, J. Chem. Phys., 2005, 122, 094701 CrossRef CAS PubMed.
  49. P. L. Levesque, M. Michaud and L. Sanche, Nucl. Instrum. Methods Phys. Res., Sect. B, 2003, 208, 225 CrossRef CAS.
  50. P. L. Levesque, M. Michaud, W. Cho and L. Sanche, J. Chem. Phys., 2005, 122, 224704 CrossRef CAS PubMed.
  51. M. Michaud, M. Bazin and L. Sanche, Int. J. Radiat. Biol., 2012, 88, 15 CrossRef CAS PubMed.
  52. R. Panajotović, M. Michaud and L. Sanche, Phys. Chem. Chem. Phys., 2007, 9, 138 RSC.
  53. V. Lemelin, A. D. Bass, P. Cloutier and L. Sanche, J. Chem. Phys., 2016, 145, 174703 CrossRef CAS PubMed.
  54. V. Lemelin, A. D. Bass, P. Cloutier and L. Sanche, J. Chem. Phys., 2016, 144, 074701 CrossRef CAS PubMed.
  55. V. Lemelin, A. D. Bass, J. R. Wagner and L. Sanche, J. Chem. Phys., 2017, 147, 234305 CrossRef CAS PubMed.
  56. V. Lemelin, A. D. Bass and L. Sanche, J. Chem. Phys., 2018, 149, 246101 CrossRef CAS PubMed.
  57. S. Ptasińska, S. Denifl, S. Gohlke, P. Scheier, E. Illenberger and T. D. Märk, Angew. Chem., Int. Ed., 2006, 45, 1893 CrossRef PubMed.
  58. H. Abdoul-Carime, S. Gohlke, E. Fischbach, J. Scheike and E. Illenberger, Chem. Phys. Lett., 2004, 387, 267 CrossRef CAS.
  59. R. Panajotovic and L. Sanche, J. Phys.: Conf. Ser., 2007, 88, 012074 CrossRef.
  60. M. Michaud and L. Sanche, Phys. Rev. B: Condens. Matter Mater. Phys., 1984, 30, 6078 CrossRef.
  61. P. L. Levesque, M. Michaud and L. Sanche, Rev. Sci. Instrum., 2005, 76, 103901 CrossRef.
  62. Z. Deng, I. Bald, E. Illenberger and M. A. Huels, Phys. Rev. Lett., 2005, 95, 1 CrossRef PubMed.
  63. S. A. Krasnokutski, A. Y. Ivanov, V. Izvekov, G. G. Sheina and Y. P. Blagoi, J. Mol. Struct., 1999, 482–483, 249 CrossRef CAS.
  64. M. Michaud and L. Sanche, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 4725 CrossRef CAS PubMed.
  65. M. Michaud, M. Bazin and L. Sanche, J. Chem. Phys., 2012, 137, 115103 CrossRef CAS PubMed.
  66. B. Yang, Y. Wang, G. Li, H. Cun, M. Ying, S. Du, M. Xu, N. Y. Song and H. J. Gao, J. Phys. Chem. C, 2009, 113, 17590 CrossRef CAS.
  67. D. Bogdan and C. Morari, J. Phys. Chem. C, 2012, 116, 7351 CrossRef CAS.
  68. M. A. Hervé Du Penhoat, M. A. Huels, P. Cloutier, J. P. Jay-Gerin and L. Sanche, J. Chem. Phys., 2001, 114, 5755 CrossRef.
  69. M. Rezaee, D. J. Hunting and L. Sanche, Med. Phys., 2014, 41, 072502 CrossRef PubMed.
  70. M. Michaud and L. Sanche, Phys. Rev. B: Condens. Matter Mater. Phys., 1984, 30, 6067 CrossRef CAS.
  71. M. Lepage, M. Michaud and L. Sanche, J. Chem. Phys., 2000, 112, 6707 CrossRef CAS.
  72. M. Michaud and L. Sanche, Phys. Rev. A: At., Mol., Opt. Phys., 1987, 36, 4672 CrossRef CAS PubMed.
  73. M. Tsuboi, M. Komatsu, J. Hoshi, E. Kawashima, T. Sekine, Y. Ishido, M. P. Russell, J. M. Benevides and G. J. Thomas, J. Am. Chem. Soc., 1997, 119, 2025 CrossRef CAS.
  74. T. Goulet, J.-M. Jung, M. Michaud, J.-P. Jay-Gerin and L. Sanche, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 5101 CrossRef CAS PubMed.
  75. L. G. Caron and L. Sanche, in Low-Energy Electron Scatt. from Mol. Biomol. Surfaces, ed. P. Carsky and R. Curik, CRC Press, 2011, pp. 161–230 Search PubMed.
  76. M. Allan, J. Phys. B: At., Mol. Opt. Phys., 2007, 40, 3531 CrossRef CAS.
  77. L. Sanche, J. Phys. B: At., Mol. Opt. Phys., 1990, 23, 1597 CrossRef CAS.
  78. K. Aflatooni, G. A. Gallup and P. D. Burrow, J. Phys. Chem. A, 1998, 102, 6205 CrossRef CAS.
  79. C. Winstead, V. McKoy and S. D’Almeida Sanchez, J. Chem. Phys., 2007, 127, 2 CrossRef PubMed.
  80. E. Alizadeh, A. G. Sanz, G. García and L. Sanche, J. Phys. Chem. Lett., 2013, 4, 820 CrossRef CAS PubMed.
  81. S. Ptasińska and L. Sanche, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2007, 75, 031915 CrossRef PubMed.
  82. M. McAllister, M. Smyth, B. Gu, G. A. Tribello and J. Kohanoff, J. Phys. Chem. Lett., 2015, 6, 3091 CrossRef CAS PubMed.
  83. P. M. Dinh, L. B. Bourg, C. Gao, B. Gu, L. Lacombe, M. Mcallister, M. Smyth, G. Tribello, M. Vincendon, J. Kohanoff, P. Reinhard, L. Sanche and E. Suraud, Nanoscale Insights into Ion-Beam Cancer Ther, Springer, 2017, pp. 277–309 Search PubMed.
  84. J. Kohanoff, M. McAllister, G. A. Tribello and B. Gu, J. Phys.: Condens. Matter, 2017, 29, 383001 CrossRef PubMed.
  85. M. McAllister, N. Kazemigazestane, L. T. Henry, B. Gu, I. Fabrikant, G. A. Tribello and J. Kohanoff, J. Phys. Chem. B, 2019, 123, 1537 CrossRef CAS PubMed.
  86. J. Kočišek, B. Sedmidubská, S. Indrajith, M. Fárník and J. Fedor, J. Phys. Chem. B, 2018, 122, 5212 CrossRef PubMed.
  87. J. Ma, S. A. Denisov, J.-L. Marignier, P. Pernot, A. Adhikary, S. Seki and M. Mostafavi, J. Phys. Chem. Lett., 2018, 9, 5105 CrossRef CAS PubMed.
  88. J. Ma, A. Kumar, Y. Muroya, S. Yamashita, T. Sakurai, S. A. Denisov, M. D. Sevilla, A. Adhikary, S. Seki and M. Mostafavi, Nat. Commun., 2019, 10, 102 CrossRef PubMed.
  89. M. B. Hahn, S. Meyer, M.-A. Schröter, H. Seitz, H.-J. Kunte, T. Solomun and H. Sturm, Phys. Chem. Chem. Phys., 2017, 19, 1798 RSC.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cp03447a

This journal is © the Owner Societies 2019