Dielectric Spectroscopy: Yesterday, Today and Tomorrow

Abstract

The theory of orientational polarization and dielectric relaxation was developed by P. Debye more than 100 years ago. It approximates a molecule by a sphere having one or more dipole moments. While in the beginning the experimentally accessible spectral range was limited to roughly 6 decades in frequency, at the end of the last century, novel spectroscopic techniques were developed and dielectric spectroscopy became broadband, nowadays covering 18 decades with no gaps.This paved the avenue for a multitude of novel fields of research in soft matter and solid-state physics including fundamental questions like the scaling of relaxation processes or the dynamics of glasses. Yet the analysis of dielectric spectra is still based on the classical approach by Debye which does not consider the multitude of intra- and inter-molecular interactions within a molecular system. To experimentally overcome these principal limitations, it is suggested to take advantage of the molecular specificity of the infrared spectral range. This offers the unique possibility to realize a novel “Orientational Polarization Spectroscopy”, in which the orientational response of a molecular system can be analyzed on an atomistic scale. For that, the theory will be outlined and the first experimental results will be presented.

1. Introduction

The theory of orientational polarization [1,2,3] was developed 100 years ago by P. Debye referring to A. Einstein’s theory of Brownian motion [4]. At that time, dielectric measurements were only possible at a few frequencies in the kilohertz region. Today, dielectric spectroscopy covers the entire spectral range from ${10}^{-6}$ Hz to ${10}^{12}$ Hz without any gap, and it has become a highly versatile tool in modern molecular physics with a multitude of applications in academia and applied research [5,6,7,8,9,10,11]. The purpose of this review is to provide a brief summary of this development and to outline future perspectives.

2. Dielectric Spectroscopy-Yesterday

Following Debye’s fundamental work, the theory of orientational polarization was pushed forward, and refined models for dielectric relaxation were developed [12,13,14,15,16,17,18,19,20,21,22]. However, until the 1990s, due to technical limitations, dielectric measurements were carried out in narrow frequency ranges of about 3–4 decades, and data were often displayed only in dependence on temperature for fixed frequencies [13,16,17]. The analysis of charge transport and electrode polarization played only a minor role, if any. Furthermore, all theoretical attempts to describe relaxation processes remained mean-field descriptions.

3. Dielectric Spectroscopy-Today

Beginning in the 1990s, dielectric spectroscopy became broadband [23], and the term “Broadband Dielectric Spectroscopy” (BDS) was established as the title of a conference series. Research topics like the “scaling of relaxation processes”, “glassy dynamics in low molecular weight and polymeric systems”, “charge transport in ion-or electron-conducting materials”, “collective and molecular dynamics of liquid crystals”, and even “polarization at (inner and outer) interfaces” became emerging topics of worldwide research in both academic and industrial research. In addition, relaxations in biological systems entered the focus of dielectric research [24]. An important step forward was also the fact that alternative spectroscopic methods such as mechanical spectroscopy [25], photon correlation spectroscopy [26], and AC-calorimetry [27] became broadband, nowadays allowing a comparison of the dynamics of molecular systems. In addition, the combination of Fourier Transform Infrared Spectroscopy (FTIR) [28], Nuclear Magnetic Resonance (NMR) [29] and neutron scattering [30,31] with BDS turned out to be highly instructive.

4. The Current Understanding of Dielectric Relaxation Processes

A refined analysis of dielectric spectra requires, in addition to dielectric measurements in a wide frequency and temperature range, knowledge of vibrational dynamics as measured with IR or Raman spectroscopy, calorimetric data, and the temperature dependence of the density. Such combined information is often only available for a few molecular systems, e.g., the homologous series of polyalcohols glycerol, threitol, xylitol and sorbitol (

Table 1. Characteristics of several polyalcohols: chemical sum and structure, molecular mass $M_W$, mass density $\rho$, glass transition temperature $T_g$ and—calculated from the latter two—the average volumes per molecule $V_m$, as well as per COH unit $V_{COH}$.

Name	Sum	Structure	${\mathit{M}}_{\mathit{W}}$ [g/mol]	$\mathit{\rho }$ [g/cm3]	${\mathit{T}}_{\mathit{g}}$ [K]	${\mathit{V}}_{\mathit{m}}$ [Å3]	${\mathit{V}}_{\mathbf{COH}}$ [Å3]
glycerol	C3H8O3		92.09	1.259 * [36]	192 [37]	121	40.4
DL-threiol	C4H10O4		122.1	1.386 ** [38]	226 [37]	146	36.6
xylitol	C5H12O5		152.2	1.402 *** [39]	248 [37]	180	36.0
D-sorbitol	C6H14O6		182.2	1.449 * [40]	268 [37]	209	34.8

* At 292 K. ** At 293.15 K. *** At 300 K.

) [32,33,34,35].

Polyalcohols consist of several different polar moieties (Table 1) with different polarity (

Table 2. Characteristics of different covalent bond types in glycerol taken from [41]: bond length and dipole moment.

Bond Type	Bond Length	Dipole Moment
	[pm]	[ ${\mathbf{10}}^{-\mathbf{30}}$ cm]	[Debye]
O-H	96	15.1	4.5
C-H	110	2.4	0.72
C-O	143	17.5	5.3
C-C	150	0.65	0.19

, Figure 1) which are interlinked by intra- and inter-molecular bonds with widely varying atomic potentials (

Table 3. Force field constants of different types of bond vibrations employed in simulations of glycerol [41].

Bond Type	Vibrational Type	Force Field Constant
C-C/C-O	stretching	∼1300 kJ mol−1 Å−2
CH2	bending	∼150 kJ mol−1 rad−2
OCCO	torsion	∼0.6 kJ mol−1

), having quite different temperature dependencies.

IR spectroscopy enables one to trace in detail the evolution of the different vibrational bands with temperature far above and below the (calorimetrically determined) glass transition temperature $T_g$ of the specific molecular system (Figure 2).

For the $\nu$(O-H) stretching vibration for all four polyalcohols, a similar temperature dependence is observed, showing a pronounced blue-shift with increasing temperature (Figure 3i). It is caused by the weakening of the inter-molecular H-bonds with increasing temperature, and hence an increase in the strength of the $\nu$(O-H) [32]. At the calorimetric glass transition temperature for all polyalcohols, a kink in the spectral position and the oscillator strength is observed, indicating that at this temperature, the inter-molecular H-bonds change (Figure 3j). In contrast, for the intra-molecular C-O and C-C-O bonds, a red-shift is found without discontinuity at $T_g$ (Figure 3c,g). This can be explained semi-quantitatively by the increase in the potential width caused by a decrease in density with increasing temperature [28]. However, also the rescaled oscillator strengths of the $\nu$(C-O) and $\nu$(C-C-O) stretching vibrations exhibit, in part, a kink at $T_g$ (Figure 3d,h), which cannot be explained at the moment. In summary, it has been proven that the different polar groups in polyalcohols can be separated into two fractions, intra- and inter-molecular bonds, which vary quite differently with temperature.

It stands to reason that these different temperature dependencies are connected to different dynamics of the respective molecular moieties. Due to inherent (partial) charges, the timescale of certain molecular fluctuations can be deduced from their polarization in oscillating electrical fields. The classical theory of orientational polarization according to Debye [2,3] assumes a single dipole with dipole moment $\overrightarrow{\mu }$ interacting with an external electric field $\overrightarrow{E}\left(\omega \right)$; using Boltzmann statistics, the averaged mean orientation counterbalanced with the thermal fluctuation of the dipolar system is determined. This yields the dielectric relaxation strength

$$\Delta \epsilon ={\displaystyle \frac{{\mu }^2}{3{\epsilon }_o{k}_{B}T}}{\displaystyle \frac{N}{V}}$$

(1)

where $\mu =|\overrightarrow{\mu }|$ is the absolute value of the dipole moment, ${\epsilon }_o=8.85\dots \times {10}^{-12}{\displaystyle \frac{\mathrm{As}}{\mathrm{Vm}}}$ is the electrical field constant, $k_B$ is the Boltzmann constant, T is the absolute temperature and ${\displaystyle \frac{N}{V}}$ is the dipole concentration. Clearly, this approach can only hold for dilute systems; if the dipoles are more concentrated, their impact on each other and the surrounding field (inducing a so-called reaction field) changes the response of the system. Typically, $\mu$ is replaced by an effective average dipole moment $\langle {\mu }_{eff}\rangle$ that includes all these effects [23]. Consequently, $\langle {\mu }_{eff}\rangle$ cannot be deduced from the structure of a single molecule.

Onsager suggested to decompose the field which acts upon a molecule in a polarized dielectric into a cavity field G given by the shape of the molecule and a reaction field R, which is proportional to the total electric moment [43]; its impact is described by the Onsager factor $F={\displaystyle \frac{{\epsilon }_s{\left({\epsilon }_{\infty }+2\right)}^2}{3\left(2{\epsilon }_s+{\epsilon }_{\infty }\right)}}$, where ${\epsilon }_{\infty }$ and ${\epsilon }_s$ are the limiting values of the permittivity at high and low frequencies, respectively. Later, Kirkwood/Fröhlich [12] expanded this approach by introducing a correlation factor $g_K$ that indicates the degree of correlation between the permanent dipole moments of the molecules, resulting in the relation

$$\Delta \epsilon ={\displaystyle \frac{{\mu }^2}{3{\epsilon }_o{k}_{B}T}}F{g}_K{\displaystyle \frac{N}{V}}$$

(2)

where technically, the correlation factor can be calculated as a sum of the pairwise angular correlations of the test dipole with its neighbors

$$g_K={\displaystyle \frac{\langle {\sum }_i{\mu }_i{\sum }_j{\mu }_j\rangle }{\tilde{N}{\mu }^2}}=1+{\displaystyle \frac{\langle {\sum }_{i=1}^{\tilde{N}}{\sum }_{j>i}{\mu }_i{\mu }_j\rangle }{\tilde{N}{\mu }^2}}=1+z\langle \cos \left(\psi \right)\rangle$$

(3)

which is often simplified using the number of nearest neighbor molecules z and their angle to the test dipole $\psi$. This correlation factor is supposed to describe the ratio of the macroscopically observable or effective average dipole moment $\langle {\mu }_{eff}\rangle$ and the average molecular dipole $\langle \mu \rangle$. Theoretically, one could calculate $g_K$ if the coordinates of the molecular dipole moments were known. Practically, the reverse is attempted; $g_K$ is calculated from the dielectric relaxation strength as a measure of $\langle {\mu }_{eff}\rangle$, an assumed molecular dipole moment, to gain insight into molecular correlations [33,44,45,46].

These modifications of Debye’s original relation appear necessary to overcome its limitation to dilute systems. Nevertheless, Onsager’s approach still considers only a single species of dipoles—an assumption which is certainly questionable for molecules with internal degrees of freedom that permit librations of distinct subunits with different dipole moments. This is already demonstrated by the presence of a $\beta$-relaxation, commonly understood as localized motion of a small submolecular unit, in addition to the structural or $\alpha$-relaxation in the polyalcohols (Figure 3a,b). Consequently, this correction might not be specific enough to accurately estimate the impact on a single relaxation in a system of various types of fluctuating species.

In contrast, the theoretical foundation of the Kirkwood correlation factor seems to capture interactions with neighboring dipoles regardless of their species or even the type of interaction (besides electrostatic interactions, one might also expect steric interactions). However, the concept covers angular (e.g., structural) correlations only, while dynamic correlations are not included, even though these must be highly relevant in a fluctuating system.

These problems seem to become relevant in more recent studies, which have taken into account intra-and inter-molecular interactions, as well as the conformations of the molecules [47], and have already obtained pronounced differences between theory and experiment for the low molecular weight liquid methanol, which cannot be explained on the basis of current theories. Further inconsistencies are found in the published correlation factors of glycerol: early studies find values of $g_K$∼5–5.5 (based on a molecular dipole moment of 2.67 D), which surprisingly dropped to a value of ∼1.5 at a temperature of 263 K [45]. Later studies found (using the same value for the molecular dipole moment) $g_K$∼2.5–3 [44].

Considering these results, we analyze the polyalcohol series glycerol, xylitol and sorbitol, and specifically calculate an effective dipole moment using their molar mass M and their mass density $\rho$ according to

$${\langle {\mu }_{eff}\rangle }_{mol}=\sqrt{{\displaystyle \frac{3{\epsilon }_o{k}_{B}T\Delta \epsilon M}{\rho N_A}}}=\langle \mu \rangle \sqrt{F{g}_K}$$

(4)

which exhibits a large variation among these similarly composed molecules (Figure 4). The resulting values are on the order of 6–10 D which exceed the interatomic dipole moments obtained from simulation studies [41] for the O-H and the C-O group of 4.5 D ($15.1\times {10}^{-30}$ cm) and 5.3 D ($17.5\times {10}^{-30}$ cm), respectively (Table 2).

In all these classical approaches, the detailed chemical structure of the molecule under study, its conformations [48] and the various chemical potentials of the bonds interconnecting the different polar moieties are not taken into consideration. Additionally, inter-molecular (e.g., dipole–dipole) interactions are neglected. It is evident that an atomistic description of the molecular dynamics based on quantum mechanics is required. On the experimental side, methods which enable to measure the effect of orientational polarization on the level of the specific (polar) moieties within a molecular system are called for.

In fact, it has been argued that the intra-molecular degrees of freedom in these polyalcohols enable conformational changes of the molecules that correspond to structural relaxation rather than the orientational relaxation of the whole molecule [33]. Consequently, the dipole moment of a submolecular structure, in particular the OH-group with 1.69 D, has been used to obtain Kirkwood correlation factors of ∼2–2.5 [33].

Following this approach, Equation (4) must be modified to refer to the concentration of the respective dipole species (instead of the concentration of molecules), i.e.,

$${\langle {\mu }_{eff}\rangle }_{COH}=\sqrt{{\displaystyle \frac{3{\epsilon }_o{k}_{B}T\Delta \epsilon M}{\rho N_A{N}_{COH}}}}$$

(5)

where $N_{COH}$ is the number of COH units per molecule.

Concerning the molecular origin and composition of dielectric relaxation peaks, it was recently suggested that generally two types of contributions are present: self-correlations (i.e., the fluctuation/libration dynamics of individual dipoles with respect to their earlier orientation), and cross-correlations (i.e., the fluctuation/libration dynamics with respect to the orientation of neighboring dipoles) [49]. This was concluded from comparative measurements of BDS and photon correlation spectroscopy which respond to molecular dynamics via different physical principles [46]. Such investigations of the polyalcohol series discussed here revealed a complex shape of the dielectric peak shape in sorbitol that indicates pronounced intra-molecular dynamics, while for glycerol a rather simple peak shape led to the conclusion of a rather rigid molecule [34,35]. This contrasts with an attempt to simulate the dielectric response of several types of glass-forming molecules which succeeded for most of them but failed particularly in glycerol [50]; the reason for this was found in the flexibility of the glycerol molecule, which was not taken into account in that model.

Further evidence for the impact of minute details of the molecular structure (and thus of intra-molecular fluctuations) on the inter-molecular dynamics is found by comparing the isomers xylitol, adonitol, D-arabitol and L-arabitol: although they have an identical molecular formula and nearly the same structure, their glass transition temperatures are 250 K, 252 K, 261 K and 261 K, respectively [51]. Consequently, also the temperature dependence of their structural relaxation as probed by BDS (often quantified in the fragility parameter) varies considerably [52]. We note that the described complexity of dynamics on the intra- and inter-molecular levels is very likely not restricted to H-bonding materials but may be expected to be inherent to any molecular system that contains several interactions of highly varying strengths and pronounced directionality, e.g., combining van der Waals forces with Coulombic interactions or steric dipoles.

This demonstrates the theoretical and experimental demand to access dynamics on a submolecular level. A potential hint may be recent studies employing terahertz spectroscopy: although earlier work on glycerol, xylitol and sorbitol does not provide particular insight into the molecular origin of the signal (only characteristic changes at their respective $T_g$ were found) [53], recent experiments on methanol used an optical pumping mechanism to identify the molecular and dynamical origin of different dielectric relaxation processes (in the THz region) [54].

5. Dielectric Spectroscopy-Tomorrow

The (linear) interaction of electro-magnetic waves is described by convention in the optical spectral range (IR, visible and UV) by the complex index of refraction $n^{*}$ as

$$n^{*}=n^{\prime }+in″$$

(6)

where $n^{\prime }$ describes the real and $n″$ the imaginary part of the index of refraction ($i=\sqrt{-1}$ represents the imaginary unit), whereas for Tera-Hz frequencies and the spectral range of Broadband Dielectric Spectroscopy, the complex dielectric function ${\epsilon }^{*}$

$${\epsilon }^{*}={\epsilon }^{\prime }-i\epsilon ″$$

(7)

is used with ${\epsilon }^{\prime }$ and $\epsilon ″$ being the real and imaginary parts. The two quantities are related by Maxwell’s relation

$$n^{*}=\sqrt{{\epsilon }^{*}}$$

(8)

The various absorption processes in the IR spectral range can be described within the Lorentz-oscillator model by

$${\epsilon }^{*}\left(\omega \right)={\epsilon }^{\prime }\left(\omega \right)-i\epsilon ″\left(\omega \right)=\left[{\epsilon }_{\infty }+\sum _j{\displaystyle \frac{s_j\left({\omega }_{o,j}^2-{\omega }^2\right)}{{\left({\omega }_{o,j}^2-{\omega }^2\right)}^2+{\left({\Gamma }_j\omega \right)}^2}}\right]-i\left[\sum _j{\displaystyle \frac{s_j\left({\Gamma }_j\omega \right)}{{\left({\omega }_{o,j}^2-{\omega }^2\right)}^2+{\left({\Gamma }_j\omega \right)}^2}}\right]$$

(9)

where $\omega$ is the angular frequency, ${\epsilon }_{\infty }$ is the value of the dielectric function at a very high frequency, ${\omega }_{o,j}$ is the resonance frequency of the jth oscillator, $s_j={\omega }_p^2{f}_j$ is related to the strength of the jth absorption mechanism with parameter $f_j$ and ${\Gamma }_j$ describes the damping in the jth oscillator.

Hereby the real parts of the index of refraction and of the complex dielectric function have the character of a memory function, in which all absorption processes add up with decreasing frequency (Figure 5), respectively, increasing wavelength. Hence, the different contributions of the polar moieties to the net orientational polarization can be determined by measuring their response separately in the IR.

In the fingerprint range of the IR spectrum, glycerol and sorbitol show several characteristic absorptions (Figure 2) representing the stretching vibrations of the (CO), (OH), (CH) and (CCO) moieties and the bending of the CH group. Applying an external, frequency-dependent electric field to a sample at widely varying temperatures and determining the field-induced IR-dichroism of specific absorption bands enables one to measure the orientational polarization of the respective molecular units separately.

The net dipole moment of a molecule ${\langle \overrightarrow{\mu }\rangle }_{mol}$, which is classically considered to be determined by dielectric spectroscopy, is the result of the superposition of all atomistic polar moieties $\overrightarrow{p_j}$:

$${\langle \overrightarrow{\mu }\rangle }_{mol}=\sum _j{\overrightarrow{p}}_j$$

(10)

In the classical view, an external electric field acts on a dipole $\overrightarrow{\mu }$, trying to orient it against the thermal fluctuations of the molecular system as a whole. The counterbalance between the two results in an ensemble-averaged orientational polarization $\langle \overrightarrow{\mu }\rangle$ given by the Langevin equation. However, as established above, in molecules with intra-molecular degrees of freedom, a more refined view on the submolecular scale is required. For an atomistic moiety having a transition moment $\langle {\psi }_i\left|{\overrightarrow{p}}_j\right|{\psi }_e\rangle$, the absorption A in the quantum mechanical dipole approximation [55] is proportional to

$$A\sim \langle {\psi }_i|{\overrightarrow{p}}_j{|{\psi }_e\rangle }^2$$

(11)

where ${\psi }_i$ and ${\psi }_e$ are the wave functions of the initial and excited states of the IR absorption event. If an additional external electrical field oscillating at a frequency well below the IR range is applied, the dipole of an atomistic moiety can be divided into a stationary component ${\overrightarrow{p}}_j^o$ (corresponding to that without a field) and a varying one $\Delta {\overrightarrow{p}}_j^{or}$

$${\overrightarrow{p}}_j={\overrightarrow{p}}_j^o+\Delta {\overrightarrow{p}}_j^{or}$$

(12)

Combining Equations (11) and (12) yields

$$A\sim \langle {\psi }_i|{\overrightarrow{p}}_j^o+\Delta {\overrightarrow{p}}_j^{or}{|{\psi }_e\rangle }^2={\left(\langle {\psi }_i|{\overrightarrow{p}}_j^o|{\psi }_e\rangle +\langle {\psi }_i|\Delta {\overrightarrow{p}}_j^{or}|{\psi }_e\rangle \right)}^2$$

(13)

where for small perturbations the quadratic term $\langle {\psi }_i|\Delta {\overrightarrow{p}}_j^{or}{|{\psi }_e\rangle }^2$ can be neglected, and thus

$$A\sim \langle {\psi }_i|{\overrightarrow{p}}_j^o|{\psi }_e{\rangle }^2+2\langle {\psi }_i|{\overrightarrow{p}}_j^o|{\psi }_e\rangle \langle {\psi }_i|\Delta {\overrightarrow{p}}_j^{or}|{\psi }_e\rangle =A_o+\Delta A_{or}$$

(14)

Due to the orientational polarization, this absorption is modulated if the polar unit is involved in a relaxation process. This depends on thermodynamic parameters such as temperature and pressure, and, of course, on the frequency and strength of the external electric field. Hereby it is assumed that the wave functions ${\psi }_i$ and ${\psi }_e$ are not seriously changed, which has to be tested experimentally by checking, for instance, the temperature dependence of the effect. For $\Delta A_{or}$ it holds that

$$\Delta A_{or}(\omega ,T)\sim \langle {\psi }_i|\Delta {\overrightarrow{p}}_j({\overrightarrow{E}}_{ext}\left(\omega \right),T)|{\psi }_e\rangle$$

(15)

with the external electrical field ${\overrightarrow{E}}_{ext}$. To measure this effect, the technical difficulty lies in the fact that conventional metal electrodes (as used in dielectric spectroscopy) are not transparent in the IR-spectral range. However, employing evaporated mesh structures [56] or nanometric (∼6 nm) sputtered platinum layers, it is possible to find a compromise between the requirement of sufficient IR transparency and the need to apply high electric fields of typically ${10}^5$ to ${10}^6$ V/cm. As an IR source, quantum cascade laser diodes or tunable IR lasers can be used (Figure 6). The sample material, having a thickness of a few μm, is contained between two mesh electrodes in a crossed arrangement.

At a wavelength of $\lambda$ = 2966.8 nm (i.e., $\overline{\nu }$ = 3370.6 cm⁻¹), the effect of orientational polarization on the O-H moiety can be measured in dependence on the frequency and strength of the external electrical field ${\overrightarrow{E}}_{ext}\left(\omega \right)$ and on temperature (Figure 7). The response normalized by the noise level of the detector signal for ${\overrightarrow{E}}_{ext}\left(\omega \right)=0$ shows a strong decrease in ${\displaystyle \frac{\Delta A_{or}}{\Delta A_0}}$ with temperature. It is caused by the formation of H-bonds which reduce and finely suppress the fluctuation of the polar O-H unit [32]. This relaxation is much slower than the dynamic glass transition ($\alpha$-relaxation) that represents the libration of the C-O groups in the core of the glycerol molecules (in agreement with other studies [46]), which is well decoupled from the fluctuation of the H-bonds in the periphery.

As a rough estimate, we compare (Figure 8) the interaction energy of a dipole with a point charge $W\left(r\right)$ and the thermal energy $E_{th}$. For each degree of freedom, the latter is ${\displaystyle \frac{1}{2}}{k}_{B}T\approx 1.7\times {10}^{-21}\mathrm{J}$ per entity or ∼$1{\displaystyle \frac{\mathrm{kJ}}{\mathrm{mol}}}$ at a temperature of 250 K. The former can be calculated as [57]

$$W\left(r\right)={\displaystyle \frac{q\mu cos\theta }{4\pi {\epsilon }_o\epsilon r^2}}$$

(16)

where q is the point charge, $\mu$ is the absolute value of the dipole moment, $\theta$ is the angle of the dipole axis with the connection line between the charge and the center of the dipole, $\epsilon$ is the (real part of the) permittivity of the medium between charge and dipole, and r is the distance along this connection line. For an OH moiety (treated as a dipole) interacting with another OH as bond acceptor (treated as a point charge), we use $q={\displaystyle \frac{1}{2}}e$ (where e is the elemental charge), $\mu =4.5\mathrm{D}$, $\theta =0°$, $\epsilon =3$, and $r=230\mathrm{pm}$ to obtain an interaction energy of $6.8\times {10}^{-21}\mathrm{J}$ per OH unit or ∼$41{\displaystyle \frac{\mathrm{kJ}}{\mathrm{mol}}}$. It is evident that this approximation includes many uncertainties of the involved parameters and thus can only provide a qualitative picture.

Orientational Polarization Spectroscopy (OPS) offers completely novel perspectives to selectively study molecular dynamics: (i) By the use of tunable IR sources it will be possible to determine the response of atomistic transition moments separately in dependence of temperate and pressure. By that, detailed insights into the intra- and inter-molecular interactions of the system under study can be achieved. (ii) It is expected that also non-polar groups linked to a polar moiety will respond to the external electrical field. (iii) It is furthermore anticipated that the OPS-response of different polar moieties is complex and shifted in time for the different moieties of a molecular system depending on the potentials of the chemical bonds. (iv) The Debye theory of orientational polarization describes the response of a non-interacting single dipole connected to a heat bath. This delivers an exponential decay in the time domain and the Debye relaxation formula in the frequency domain. It is not expected that this idealized picture holds for molecular systems at the atomistic scale.

6. Conclusions and Outlook

Dielectric spectroscopy is nowadays one of the leading experimental tools for exploring molecular dynamics. Unfortunately, it still lacks a theoretical foundation in quantum mechanics, and its central measurement quantities, the complex dielectric function or the complex conductivity, have a mean-field character. However, on the basis of modern laser technology, it is possible to combine vibrational (IR and Raman) and dielectric spectroscopies to overcome these limitations and to pave the avenue for novel and much more refined atomistic insights into the orientational polarization of soft and solid materials.