Ab initio spectroscopy of water under electric fields

Giuseppe Cassone*a, Jiri Sponera, Sebastiano Trussob and Franz Saija*b
aInstitute of Biophysics of the Czech Academy of Sciences, Královopolská 135, 61265 Brno, Czech Republic. E-mail: cassone@ibp.cz; sponer@ncbr.muni.cz
bCNR-IPCF, Viale Ferdinando Stagno d'Alcontres 37, 98158 Messina, Italy. E-mail: trusso@ipcf.cnr.it; saija@ipcf.cnr.it

Received 31st May 2019 , Accepted 23rd July 2019

First published on 23rd July 2019

Whereas a broad range of literature exists on the spectroscopy of water in disparate conditions, infrared (IR) and Raman spectra of water subjected to electric fields have never extensively been investigated so far. Based on ab initio molecular dynamics simulations, here we present IR and Raman spectra of bulk liquid water under the effect of static electric fields. A contraction of the entire frequency range is recorded upon increasing the field intensity both in the IR and in the Raman spectra. Whilst the OH stretching band is progressively shifted toward lower frequencies – indicating a field-induced strengthening of the H-bond network – all the other bands are up-shifted by the field. Furthermore, an evident modification of the librational mode band appears in all the spectra. Finally, the order-maker action of the field emerges also from the increase of the water orientational tetrahedral order. Upon field exposure, the water structure becomes more “ice like”.

I Introduction

Water holds a central place in very diverse branches of science such as inter alia biology, geology, chemistry, physics, and materials science and engineering.1–3 Despite the huge amount of fundamental and applied spectroscopic investigations of water in disparate conditions of temperature, pH, and chemical environment,4 infrared (IR) spectra of bulk liquid water under static electric fields have only partially been characterized in ref. 5, whereas the field-induced changes to the Raman features, to the best of our knowledge, have never been reported so far. Since local electric fields, additionally to quantum mechanics, ultimately rule the behavior of matter on a microscopic scale,6–9 understanding the reactivity and the spectroscopy of water subjected to the effect of electrostatic potential gradients is crucial in several contexts. By employing state-of-the-art ab initio molecular dynamics (AIMD) simulations, here we explore the effects due to the application of static electric fields on the IR and Raman spectra of water over the entire spectral range.

Mainly because of the difficult-to-grab connections between the emerging water properties and its molecular interactions – which are based on a complex dynamical H-bonded network – a complete understanding of water's unique behavior is still lacking.10 Vibrational spectroscopy is widely used as a powerful probe of the structure and dynamics of water and aqueous solutions.11 Albeit deriving an unambiguous molecular-level interpretation of the experimental spectral features still represents a challenge due to the subtle interplay between molecular dipole moments, local electric fields, long-range electrostatics, van der Waals interactions and so on and so forth, nowadays it has become possible to deploy reliable AIMD simulations to predict IR and Raman water spectra under diverse conditions.12,13 Since in first-principles techniques molecular interactions are evaluated on the fly from electronic structure calculations, it is possible to ascribe a given mode to the respective spectral feature.14,15

Electric fields bring about a range of effects by interacting with atoms, molecules, and complex matter, modifying the activation barriers of chemical reactions,6,8 and shaping their free-energy landscape,16 and hence hold a crucial place in catalysis, electrochemistry and biological processes.7,17 Most – if not all – of these phenomena occur in water; however, the effects of the application of external electric fields on the properties of the water molecules, although of extraordinary relevance in all research fields, have been preliminarily explored by means of first-principles computational approaches only in the last decade. Owing to developments of the modern theory of polarization,18,19 nowadays AIMD20–23 and density functional theory-based17,24,25 studies have succeeded in describing complex behavior of molecular systems under strong electric fields, quantitatively confirmed by experiments,6,26–29 consolidating those first-principles numerical techniques as robust predictors of laboratory experimental results. Moreover, although of extraordinary fundamental as well as applicative relevance in industrial processes, electrochemistry and electrofreezing-related research,30,31 the reactivity-enhancing properties carried by electric fields16,32,33 and their effects on liquid5,21,34,35 and solid36,37 water have only recently been studied by means of AIMD methods. It is well-known that field intensities equal to 0.01 V Å−1 are necessary to induce detectable molecular alignments in water.38 Moreover, extreme field strengths of ∼2–3 V Å−1 are naturally present in proximity to the atomic sites both in confined39 and in bulk liquid water,40,41 and also accompanying perpetual local dipole (and quadrupole) fluctuations.42 Here we report on IR and Raman spectra of water under the action of electric field intensities up to 0.30 V Å−1, the latter representing the computational21 and experimental26–28 molecular dissociation threshold.

II Methods

We used the software package CP2K,43,44 based on the Born–Oppenheimer approach, to perform AIMD simulations of a sample of liquid water under the action of static and homogeneous electric fields applied along a given direction (corresponding to the z-axis). The implementation of an external field in numerical codes based on Density Functional Theory (DFT) can be achieved by exploiting the modern theory of polarization and Berry's phases18,19,45 (see, e.g., ref. 20). Thanks to those seminal studies, nowadays AIMD simulations under the effect of static electric fields with periodic boundary conditions are almost routinely carried out (see, e.g., ref. 23). The reader who is interested in the implementation of static electric fields in atomistic simulations can refer to the following literature: ref. 18–20 and 46–50. The liquid water sample contained 128 H2O molecules (i.e., 384 atoms) arranged in a cubic cell with side parameter a = 15.82 Å, so as to reproduce a density of 0.97 g cm−3. As usual, in order to minimize spurious and non-physical surface effects, the structures were replicated in space by employing periodic boundary conditions. The intensity of the electric field was gradually increased with a step increment of 0.05 V Å−1 from zero up to a maximum of 0.30 V Å−1, the latter being the computational21 and experimental26–28 water dissociation threshold. Furthermore, also during our simulations, the first significant molecular dissociation events have been recorded for such a field strength. In the zero-field case we performed dynamics of 50 ps, whereas, for each other value of the field intensity, we ran dynamics of 20 ps, thus accumulating a total simulation time equal to 150 ps, where a time-step of 0.5 fs has been chosen.

The wavefunctions of the atomic species have been expanded using the TZVP basis set with Goedecker–Teter–Hutter pseudopotentials using the GPW method.51 A plane-wave cutoff of 400 Ry has been imposed. Exchange and correlation (XC) effects were treated with the gradient-corrected Becke–Lee–Yang–Parr (BLYP)52 density functional. Moreover, in order to take into account dispersion interactions, we employed the dispersion-corrected version of BLYP (i.e., BLYP-D3(BJ)).53,54 The adoption of the BLYP-D3 functional has been dictated by the widespread evidence that such a functional, when dispersion corrections are taken into account, offers one of the best agreements with the experimental results among the standard GGA functionals.55,56 It is well-known indeed that neglecting dispersion corrections leads to a severely over-structured liquid (see, e.g., ref. 57 and references therein). Moreover, a nominal temperature slightly higher than the standard one has been simulated in order to better reproduce the molecular structure (i.e., T = 350 K). Furthermore, an additional simulation at T = 300 K has been executed under a zero-field regime in order to evaluate the effects induced by a temperature decrease and compare them with those stemming from the application of an electric field.

Albeit the BLYP-D3 functional represents a reasonably good choice, computationally more expensive hybrid functionals, such as revPBE0, when used along with a quantum treatment of the nuclei, perform excellently well for water, as very recently demonstrated by Marsalek and Markland.58 However, the inclusion of nuclear quantum effects is beyond the scope of the present work. The IR absorption line shapes of liquid water (and ice) are overall reproduced remarkably well by standard AIMD simulations, which include by their nature the explicit quantum adiabatic response of the electrons.14 Moreover, the agreement of the IR and, to a lesser extent, of the Raman spectra evaluated under zero-field conditions with recent experimental results59,60 justifies a posteriori the classical treatment of the nuclei, rendering nuclear quantum effect inclusion in such calculations the subject of future work. As a consequence, the dynamics of ions was simulated classically within a constant number, volume, and temperature (NVT) ensemble, using the Verlet algorithm, whereas the canonical sampling has been executed by employing a canonical-sampling-through-velocity-rescaling thermostat61 set with a time constant equal to 10 fs. IR and Raman spectra have been determined by means of the software TRAVIS,62,63 as detailed in the ESI.

III Results and discussion

As shown in Fig. 1 (black curve), not only are the shape and the position (i.e., located at 3220 cm−1) of the OH stretching band evaluated by means of the employed DFT framework in fairly good agreement with the experimental data,59 but also the low-frequency libration mode band position (i.e., placed at 560 cm−1) is well-reproduced. On the other hand, as displayed in Fig. 1, a slight – but measurable – blue-shift of the bending (located at 1630 cm−1) and of the combined bending and libration (at 2170 cm−1) modes is observed.64,65
image file: c9cp03101d-f1.tif
Fig. 1 Infrared (IR) absorption spectra of liquid water determined in the absence of a field (black line) and under different field intensities (colored lines) as detailed in the legend. The experimental result59 is shown as gray shading for reference. Arrows are guides for the eye qualitatively following the field-induced shifts of the bands.

Similar considerations hold also for the zero-field regime anisotropic and isotropic Raman spectra shown in Fig. 2a and b, respectively (black curves).

image file: c9cp03101d-f2.tif
Fig. 2 Anisotropic (a) and isotropic (b) Raman scattering spectra of liquid water in the absence of a field (black lines) and under different field intensities (colored lines) as detailed in the legends compared to recent experimental data60 shown as gray shading. All spectra have been multiplied by the Bose-Einstein correction image file: c9cp03101d-t1.tif.

However, notwithstanding that the shape of the bands is generally well-reproduced, the position of some peaks of the anisotropic Raman spectrum (Fig. 2a) is sizably shifted. On the one hand, the computed OH stretching band is shifted toward lower frequencies with respect to one of the most recent experimental Raman spectra60 (i.e., 3330 vs. 3448 cm−1, respectively) and, on the other, the low-frequency libration mode band appears to be shifted toward higher frequencies (i.e., 800 vs. 433 cm−1, respectively). However, all the spectral features are remarkably well-caught, including the subtle bending and libration mode combination band at 2130 cm−1, accordingly to recent AIMD simulations.58 Besides the band shape, also the line positions of the isotropic Raman spectrum in the zero-field regime are in excellent agreement with the experimental data60 over the entire spectral range, as displayed in Fig. 2b.

Upon exposure of the sample to progressively higher field intensities, several features of the IR and Raman spectra appreciably change, as reported for a portion of the IR spectrum (and up to relatively moderate field strengths) by Futera and English.5 Among them, the most striking field-induced frequency shift is exhibited by the prominent OH stretching band, as highlighted by the high-frequency arrow in Fig. 1 for the IR spectrum and as visibly caught in Fig. 2a and b for the anisotropic and isotropic Raman spectra, respectively. By isolating the OH stretching band contribution and by evaluating the mean values of each distribution, it is possible to trace the trend of the mean frequency as a function of the field strength, as quantitatively shown in Fig. 3. In particular, both the mean frequencies of the IR OH stretching band (black curve) and those stemming from the anisotropic and isotropic Raman spectra (red and blue curves, respectively) exhibit a monotonic decrease as the field intensity is increased. Because of a more direct coupling of the field with the water dipole moments than that established with the polarizabilities, the slope of the IR curve in Fig. 3 is more negative than that characterizing the curves corresponding to the Raman OH stretching band (i.e., ∼−780 vs. ∼−560, respectively). Besides, both the IR and Raman spectra clearly indicate that the application of external static and homogeneous electric fields to bulk liquid water leads to a sizable decrease of the frequencies at which the OH covalent bonds stretch. As a direct consequence, the water H-bond network is significantly strengthened upon the field exposure of the sample, an effect that can be considered as due to a temperature decrease. In water, indeed, the red-shift of the OH stretching frequencies is generally associated with stronger H-bonding66 and with more “ice-like” molecular structures.67 Not surprisingly hence a monotonic decrease of the location of the average OH stretching frequency is recorded upon cooling the water sample in the zero-field regime, as shown in the inset of Fig. 3. However, the electric field effects appear to be more prominent than those induced by the temperature. In fact, as displayed in Fig. S2 of the ESI, the shift of the IR and Raman OH stretching band caused by a temperature decrease of 50 K – within the time-scales affordable by means of the present AIMD simulations – is slightly lower than that produced by a field strength of 0.10 V Å−1. It is worth pointing out that, similarly to previous investigations on the water H-bond dynamics in aqueous solutions of amphiphiles68 and according to recent second-generation Car–Parrinello molecular dynamics simulations exploiting an opportune decomposition of the water binding energy,69 the field-induced OH stretching frequency decrease of ∼200 cm−1 recorded in the IR spectrum at the maximum field intensity can be ascribed to an increase equal to ∼10–12 kJ mol−1 of the H-bond strength.

image file: c9cp03101d-f3.tif
Fig. 3 Mean frequency of the IR (black line and dots), anisotropic (blue line and squares) and isotropic (red line and squares) Raman OH stretching band as a function of the electric field strength. The mean frequency value has been determined by isolating each OH stretching band contribution from the rest of the respective spectrum. Open symbols refer to the zero-field case. In the inset, the average OH stretching frequency as a function of the temperature is reported. For the sake of clarity, the temperature axis is reversed.

Albeit less visible than in the IR and Raman spectra, the field-induced structuring of the intermolecular interactions is visible from the atomistic radial distribution functions (RDFs). In fact, as shown in Fig. 4, the height of the first and the second peak of the oxygen–oxygen (O–O) RDF is increased whilst that of the first dip is sizably lowered upon the field action. Similarly, an electric-field-driven structuring of the first solvation shell is also observed from the oxygen–hydrogen (O–H) and hydrogen–hydrogen (H–H) RDFs. As far as the latter is concerned, the evident field-induced modification of the second peak is discussed later.

image file: c9cp03101d-f4.tif
Fig. 4 From the top-left to the bottom: oxygen–oxygen (O–O), oxygen–hydrogen (O–H), and hydrogen–hydrogen (H–H) radial distribution functions of bulk liquid water determined from the zero-field regime (black curves) up to a field strength of 0.25 V Å−1 (red curves).

Following a descending frequency order, another interesting field-induced feature is the enhancement of the bending and libration mode combination band, clearly visible in the IR spectrum (Fig. 1) and in the anisotropic Raman spectrum (Fig. 2a). In addition, the application of the field shifts the location of the respective peak toward higher frequencies. As an example, whereas in the zero-field regime a peak at ∼2130 cm−1 is recorded, e.g., in the anisotropic intensity of the water Raman spectrum, at 0.25 V Å−1 its location is shifted to ∼2320 cm−1. Interestingly, beneath the frequencies of the OH stretching band in the IR and Raman spectra, the frequency shift due to the field follows the opposite trend, in that all the modes exhibit increased collective frequencies with respect to their own zero-field regime counterparts (i.e., they are blue-shifted). This is true, of course, also for the bending mode band where, however, the frequency shift is reduced to a few tenths of cm−1; a modest change if compared with those of the already discussed bands, as is visible from Fig. 1 and 2. The only relevant effect ascribable to the field in this frequency range (i.e., at ∼1650 cm−1) is the intensity enhancement of the bending mode band of the IR and Raman spectra.

In conjunction with the field-induced strengthening of the intermolecular interactions highlighted by the frequency shifts of the IR and Raman OH stretching bands, another spectral feature is drastically influenced by the external electrostatic potential gradient. In particular, the low-frequency band – which is unambiguously attributed to librational modes15,70 – exhibits the growth of an additional peak besides being slightly blue-shifted, as it is manifestly visible from the spectra of Fig. 1 and 2. More specifically, as shown in Fig. 5, the progressive increment of the field intensity leads to a more and more pronounced high-frequency side mode of the librational mode band, both in the IR (Fig. 5a) and in the Raman (Fig. 5b) spectra. By evaluating the point-by-point difference between the spectral intensity under the field action and the zero-field intensity spectrum, it is possible to notice that even the weakest field strength is able to induce a manifest change of such a peak, as displayed in Fig. 5. From the analysis of the respective areas, it turns out that seemingly a librational mode is converted into another one upon the field action. Thus, such a sub-band – caught both by the IR (Fig. 5a) and by the Raman (Fig. 5b) spectra – represents the onset of a field-induced mode in the water system. In order to attribute such a feature to a molecular mode, it is sufficient to evoke the most fundamental definition of libration. In fact, the application of the field induces not only a strengthening of the topological constraints on the molecular motions through reinforced H-bonds, but also imposes a preferential direction to the molecular orientations. Upon exposure of the sample to progressively higher field intensities, increasingly larger fractions of the water dipole moment vectors tend to align toward the field direction, as shown in Fig. S1 of the ESI. Because of such a molecular alignment, the intermolecular H–H RDF shown in Fig. 4 exhibits inter alia a prominent shift of the second peak toward larger distances. Thus, the mode is mainly due to the growth of a libration caused by the electrostatic coupling between the field and the molecular dipole moments, whilst the slight blue-shift of the standard libration mode band is a direct consequence of the enhanced restrictions imposed by the H-bonds. The fact that such a peak is smaller (higher) than the typical one in the IR (Raman) spectrum remains unexplained since relative intensities can be compared only within the same kind of experiment (i.e., IR vs. IR and Raman vs. Raman). However, one can speculate that it can be ascribed to the fact that only the water molecule rocking and wagging librations are IR – and Raman – active modes, whereas the twisting libration, consisting of a rotation of the hydrogen atoms around the dipole axis, does not affect the dipole moment of the molecule but contributes to the polarizability and hence to the Raman spectrum in that frequency range.

image file: c9cp03101d-f5.tif
Fig. 5 IR (a) and isotropic Raman (b) point-by-point differences between the librational mode band (see the insets) determined under the action of different field intensities (see the legends) and the same band in the zero-field regime. Arrows are guides for the eye highlighting the growth of the high-frequency side of the band.

The enhancement of the intermolecular connectivity is, of course, due to several side-effects of the application of the external field. In fact, concurrently with the molecular alignments shown in Fig. S1 of the ESI, the molecular dipole moments’ magnitude increases as well, as displayed in Fig. 6a, leading to an overall increment of about 13% of the mean dipole moment magnitude when passing from the zero-field regime to the pre-dissociation field strength 0.25 V Å−1. Such an increment leads to a better electrostatic coupling with the external field and thus to a larger water dipole alignment along the field direction; a circumstance that, in turn, contributes to the statistical increase of the dipole moments’ magnitude.

image file: c9cp03101d-f6.tif
Fig. 6 (a) Average molecular dipole moment magnitude as a function of the field strength. Calculated points are plotted as full red circles whereas the linear fit is presented as a solid red line. (b) Probability distribution P(q) of the tetrahedral order parameter q72 for different field strengths (see the legend). Arrows indicate the most significant field-induced changes.

The structuring induced by the field through enhanced intermolecular interactions (i.e., H-bonds) and collective molecular alignments can be partially probed through well-established tetrahedral translational (Sk)71 and orientational (q)71,72 order parameters which are sensitive to the topology of the oxygen atoms' sub-system. If, on the one hand, Sk typically changes in the fourth decimal place,73 as also displayed in Table S1 of the ESI, q more promptly reacts to modifications of the external conditions.74 As shown in Fig. 6b, indeed, a progressive narrowing of the distribution P(q) is recorded as the field strength is increased. Moreover, the increment of the peak height and of its location – achieving a mean value of 0.82 at 0.25 V Å−1 – indicates a clear tendency of the system toward more orientationally ordered configurations. This way, the electrostatic gradient not only re-orients the hydrogen sub-lattice in such a way to align the dipole vectors, but also it is able to arrange the oxygen sub-lattice with increased orientational tetrahedral order. Incidentally, such a trend – similar to what we have observed for the field-induced H-bond strengthening – is exactly opposite to that induced by a temperature increment.74,75 In other words, notwithstanding that the tetrahedral ordering of the oxygen atoms is less prominent than the intermolecular interaction reinforcement, the electric field acts as an order-maker. Roughly speaking, the application of the field somehow renders the aqueous environment more “ice-like”.

IV Conclusions

In summary, although the coupling of the external field and the water molecular dipoles takes place anisotropically along the field direction, field-induced effects on the three-dimensional H-bond network clearly emerge in the IR and Raman spectra under the field influence and are progressively enhanced as its strength is increased. In particular, whereas all the other mode bands are blue-shifted by the field action, the OH stretching band is sizably red-shifted. Moreover, the growth of a high-frequency side mode of the librational mode band appears both in the IR and in the Raman spectra because of the additional topological constraints imposed on the molecular dipoles not only with the H-bond reinforcement but mainly because of the induced molecular dipoles’ alignments. Finally, the order-maker activity of the external field emerges, concurrently with the H-bond strengthening and the ordering of the molecular orientations along the field direction, also from the increase of the water orientational tetrahedral order. As here explicitly demonstrated – by first-principles – through a direct comparison with lower temperature simulations performed in the zero-field regime, all these findings indicate that the effects caused by the field conform to those induced by a water temperature decrease.

Conflicts of interest

There are no conflicts to declare.


G. C. thanks Martin Brehm and Martin Thomas for useful discussions on the evaluation of the Raman spectra.


  1. G. E. Walrafen, Water: A Comprehensive Treatise, ed. F. Franks, Plenum, New York, vol. 1, 1972 Search PubMed.
  2. M. Gerstein and M. Levitt, Sci. Am., 1998, 279, 100–105 CrossRef CAS PubMed.
  3. K. Kaila and B. R. Ransom, pH and brain function, ed. K. Kaila and B. R. Ransom, Wiley, New York, 1998 Search PubMed.
  4. H. J. Bakker and J. L. Skinner, Chem. Rev., 2010, 110, 1498–1517 CrossRef CAS PubMed.
  5. Z. Futera and N. J. English, J. Chem. Phys., 2017, 147, 031102 CrossRef PubMed.
  6. A. C. Aragones, N. L. Haworth, N. Darwish, S. Ciampi, N. J. Bloomfield, G. G. Wallace, I. Diez-Perez and M. L. Coote, Nature, 2016, 531, 88–91 CrossRef CAS PubMed.
  7. S. Shaik, D. Mandal and R. Ramanan, Nat. Chem., 2016, 8, 1091–1098 CrossRef CAS PubMed.
  8. F. Che, J. T. Gray, S. Ha, N. Kruse, S. L. Scott and J.-S. McEwen, ACS Catal., 2018, 8, 5153–5174 CrossRef CAS.
  9. C. J. Fecko, J. D. Eaves, J. J. Loparo, A. Tokmakoff and P. L. Geissler, Science, 2003, 301, 1698–1701 CrossRef CAS PubMed.
  10. P. Gallo, et al., Chem. Rev., 2016, 116, 7463–7500 CrossRef CAS PubMed.
  11. B. M. Auer and J. L. Skinner, J. Chem. Phys., 2008, 128, 224511 CrossRef CAS PubMed.
  12. T. Morawietz, O. Marsalek, S. R. Pattenaude, L. M. Streacker, D. Ben-Amotz and T. E. Markland, J. Phys. Chem. Lett., 2018, 9, 851–857 CrossRef CAS PubMed.
  13. V. Rosza, D. Pan, F. Giberti and G. Galli, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 6952–6957 CrossRef PubMed.
  14. M. Sharma, R. Resta and R. Car, Phys. Rev. Lett., 2005, 95, 187401 CrossRef PubMed.
  15. P. L. Silvestrelli, M. Bernasconi and M. Parrinello, Chem. Phys. Lett., 1997, 277, 478–482 CrossRef CAS.
  16. G. Cassone, F. Pietrucci, F. Saija, F. Guyot and A. M. Saitta, Chem. Sci., 2017, 8, 2329–2336 RSC.
  17. Z. Wang, D. Danovich, R. Ramanan and S. Shaik, J. Am. Chem. Soc., 2018, 140, 13350–13359 CrossRef CAS PubMed.
  18. R. D. King-Smith and D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 47, 1651–1654 CrossRef CAS PubMed.
  19. R. Resta, Rev. Mod. Phys., 1994, 66, 899–915 CrossRef CAS.
  20. P. Umari and A. Pasquarello, Phys. Rev. Lett., 2002, 89, 157602 CrossRef CAS PubMed.
  21. A. M. Saitta, F. Saija and P. V. Giaquinta, Phys. Rev. Lett., 2012, 108, 207801 CrossRef PubMed.
  22. A. M. Saitta and F. Saija, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 13768–13773 CrossRef CAS PubMed.
  23. N. J. English and C. J. Waldron, Phys. Chem. Chem. Phys., 2015, 17, 12407–12440 RSC.
  24. S. Shaik, S. P. de Visser and D. Kumar, J. Am. Chem. Soc., 2004, 126, 11746–11749 CrossRef CAS PubMed.
  25. R. Meir, H. Chen, W. Lai and S. Shaik, ChemPhysChem, 2010, 11, 301–310 CrossRef CAS PubMed.
  26. E. M. Stuve, Chem. Phys. Lett., 2012, 520, 1–17 CrossRef.
  27. Z. Hammadi, M. Descoins, E. Salançon and R. Morin, Appl. Phys. Lett., 2012, 101, 243110 CrossRef.
  28. W. K. Lee, et al., Nano Res., 2013, 6, 767–774 CrossRef CAS.
  29. S. L. Miller, Science, 1953, 117, 528–529 CrossRef CAS PubMed.
  30. I. M. Svischchev and P. G. Kusalik, J. Am. Chem. Soc., 1996, 118, 649–654 CrossRef.
  31. P. K. Jha, E. Xanthakis, V. Jury, M. Havet and A. Le-Bail, Curr. Opin. Food Chem., 2018, 23, 85–89 Search PubMed.
  32. G. Cassone, F. Pietrucci, F. Saija, F. Guyot, J. Sponer, J. E. Sponer and A. M. Saitta, Sci. Rep., 2017, 7, 6901 CrossRef PubMed.
  33. G. Cassone, J. Sponer, J. E. Sponer, F. Pietrucci, A. M. Saitta and F. Saija, Chem. Commun., 2018, 54, 3211–3214 RSC.
  34. G. Cassone, G. Calogero, J. Sponer and F. Saija, Phys. Chem. Chem. Phys., 2018, 20, 13038–13046 RSC.
  35. G. Cassone, F. Creazzo, P. V. Giaquinta, J. Sponer and F. Saija, Phys. Chem. Chem. Phys., 2017, 19, 20420–20429 RSC.
  36. G. Cassone, P. V. Giaquinta, F. Saija and A. M. Saitta, J. Phys. Chem. B, 2014, 118, 4419–4424 CrossRef CAS PubMed.
  37. G. Cassone, P. V. Giaquinta, F. Saija and A. M. Saitta, J. Phys. Chem. B, 2014, 118, 12717–12724 CrossRef CAS PubMed.
  38. M. Shaflei, M. von Domaros, D. Bratko and A. Luzar, J. Chem. Phys., 2019, 150, 074505 CrossRef PubMed.
  39. S. Laporte, F. Finocchi, L. Paulatto, M. Blanchard, E. Balan, F. Guyot and A. M. Saitta, Phys. Chem. Chem. Phys., 2015, 17, 20382–20390 RSC.
  40. J. D. Smith, J. D. Cappa, K. R. Wilson, R. C. Cohen, P. L. Geissler and R. J. Saykally, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 14171–14174 CrossRef CAS PubMed.
  41. B. Sellner, M. Valiev and S. M. Kathmann, J. Phys. Chem. B, 2013, 117, 10869–10882 CrossRef CAS PubMed.
  42. P. L. Geissler, C. Dellago, D. Chandler, J. Hutter and M. Parrinello, Science, 2001, 291, 2121–2124 CrossRef CAS PubMed.
  43. J. Hutter, M. Iannuzzi, F. Schiffmann and J. VandeVondele, WIREs Comput. Mol. Sci., 2014, 4, 15–25 CrossRef CAS.
  44. J. Vandevondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing and J. Hutter, Comput. Phys. Commun., 2005, 167, 103–128 CrossRef CAS.
  45. M. V. Berry, Proc. R. Soc. London, Ser. A, 1984, 392, 45 CrossRef.
  46. R. W. Nunes and D. Vanderbilt, Phys. Rev. Lett., 1994, 73, 712 CrossRef CAS PubMed.
  47. R. W. Nunes and X. Gonze, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 155107 CrossRef.
  48. R. Resta, Phys. Rev. Lett., 1998, 80, 1800 CrossRef CAS.
  49. X. Gonze, Ph. Ghosez and R. W. Godby, Phys. Rev. Lett., 1995, 74, 4035 CrossRef CAS PubMed.
  50. X. Gonze, Ph. Ghosez and R. W. Godby, Phys. Rev. Lett., 1997, 78, 294 CrossRef CAS.
  51. M. Krack, Theor. Chem. Acc., 2005, 114, 145–152 Search PubMed.
  52. A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098 CrossRef CAS PubMed; C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785 CrossRef PubMed.
  53. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104 CrossRef PubMed.
  54. S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465 CrossRef CAS PubMed.
  55. I.-C. Lin, A. P. Seitsonen, I. Tavernelli and U. Rothlisberger, J. Chem. Theory Comput., 2012, 8, 3902–3910 CrossRef CAS PubMed.
  56. A. Bankura, A. Karmakar, V. Carnevale, A. Chandra and M. L. Klein, J. Phys. Chem. C, 2014, 118, 29401–29411 CrossRef CAS.
  57. M. J. Gillan, D. Alfé and A. Michaelides, J. Chem. Phys., 2016, 144, 130901 CrossRef PubMed.
  58. O. Marsalek and T. E. Markland, J. Phys. Chem. Lett., 2018, 8, 1545–1551 CrossRef PubMed.
  59. J. E. Bertie and Z. Lan, Appl. Spectrosc., 1996, 60, 1047–1057 CrossRef.
  60. S. R. Pattenaude, L. M. Streacker and D. Ben-Amotz, J. Raman Spectrosc., 2018, 49, 1860–1866 CrossRef CAS.
  61. G. Bussi, D. Donadio and M. Parrinello, J. Chem. Phys., 2007, 126, 014101 CrossRef PubMed.
  62. M. Brehm and B. Kirchner, J. Chem. Inf. Model., 2011, 51, 2007–2023 CrossRef CAS PubMed.
  63. M. Thomas, M. Brehm, R. Fligg, P. Voehringer and B. Kirchner, Phys. Chem. Chem. Phys., 2013, 15, 6608–6622 RSC.
  64. L. R. Pestana, O. Marsalek, T. E. Markland and T. Head-Gordon, J. Phys. Chem. Lett., 2018, 9, 5009–5016 CrossRef PubMed.
  65. P. K. Verma, A. Kundu, M. S. Puretz, C. Dhoonmoon, O. S. Chegwidden, C. H. Londergan and M. Cho, J. Phys. Chem. B, 2018, 122, 2587–2599 CrossRef CAS PubMed.
  66. S. A. Rice, Topics in Current Chemistry, ed. P. Shuster, et al., Springer, New York, 1975, vol. 60, p. 109 Search PubMed.
  67. Z. Wang, A. Pakoulev, Y. Pang and D. D. Dlott, J. Phys. Chem. A, 2004, 108, 9054–9063 CrossRef CAS.
  68. G. Stirnemann, J. T. Hynes and D. Laage, J. Phys. Chem. B, 2010, 114, 3052–3059 CrossRef CAS PubMed.
  69. D. Ojha, K. Karhan and T. D. Kühne, Sci. Rep., 2018, 8, 16888 CrossRef PubMed.
  70. Q. Wan, L. Spanu, G. A. Galli and F. Gygi, J. Chem. Theory Comput., 2013, 9, 4124 CrossRef CAS PubMed.
  71. P.-L. Chau and A. J. Hardwick, Mol. Phys., 1998, 93, 511–518 CrossRef CAS.
  72. J. R. Errington and P. G. Debenedetti, Nature, 2001, 409, 318–321 CrossRef CAS.
  73. E. Duboué-Dijon and D. Laage, J. Phys. Chem. B, 2015, 119, 8406–8418 CrossRef PubMed.
  74. X. Wu, W. Lu, L. M. Streaker, H. S. Ashbaugh and D. Ben-Amotz, J. Phys. Chem. Lett., 2018, 9, 1012–1017 CrossRef CAS PubMed.
  75. S. K. Reddy, D. R. Moberg, S. C. Straight and F. Paesani, J. Chem. Phys., 2017, 147, 244504 CrossRef PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cp03101d

This journal is © the Owner Societies 2019