On the connection between Gaussian statistics and excited-state linear response for time-dependent fluorescence

Solvation dynamics on the diffusion timescale elucidated using energy-represented dynamics theory

Photoexcitation of a solute alters the solute-solvent interaction, resulting in the nonequilibrium relaxation of the solvation structure, often called a dynamic Stokes shift or solvation dynamics. Thanks to the local nature of the solute-solvent interaction, the characteristics of the local solvent environment dissolving the solute can be captured by the observation of this process. Recently, we derived the energy-represented Smoluchowski-Vlasov (ERSV) equation, a diffusion equation for molecular liquids, which can be used to analyze the solvation dynamics on the diffusion timescale. This equation expresses the time development for the solvent distribution on the solute-solvent pair interaction energy (energy coordinate). Since the energy coordinate can effectively treat the solvent flexibility in addition to the position and orientation, the ERSV equation can be utilized in various solvent systems. Here, we apply the ERSV equation to the solvation dynamics of 6-propionyl-2-dimethylamino naphthalene (Prodan) in water and different alcohol solvents (methanol, ethanol, and 1-propanol) for clarifying the differences of the relaxation processes among these solvents. Prodan is a solvent-sensitive fluorescent probe and is thus widely utilized for investigating heterogeneous environments. On the long timescale, the ERSV equation satisfactorily reproduces the relaxation time correlation functions obtained from the molecular dynamics (MD) simulations for these solvents. We reveal that the relaxation time coefficient on the diffusion timescale linearly correlates with the inverse of the translational diffusion coefficients for the alcohol solvents because of the Prodan-solvent energy distributions among the alcohols. In the case of water, the time coefficient deviates from the linear relationship for the alcohols due to the difference in the extent of importance of the collective motion between the water and alcohol solvents.

Simulating the Fluorescence of the Locally Excited State of DMABN in Solvents of Different Polarities

4-(N,N-Dimethylamino)benzonitrile (DMABN) is a luminescent probe that can be used for tracking changes in the surrounding solvent due to the large change in polarity between its ground and excited states. An important characteristic of DMABN is that it exhibits dual fluorescence with two different emission energies that can be monitored, allowing for better characterization of the surrounding system. The first excited state is called the locally excited (LE) state and is characterized by the movement of charge over the conjugated ring structure. In nonpolar solvents and in the gas phase, the fluorescence of DMABN is entirely attributed to the transition from the near-planar LE state. In more polar environments, emission occurs from both the LE and a second excited state, corresponding to a twisted intramolecular charge-transfer (ICT) structure. For the sake of simplicity, this work considers transitions between only the ground and LE state. Molecular mechanical force field models of DMABN in its ground and LE states have been developed to investigate the sensitivity of the LE state to the polarity of the solvent. Both nonpolarizable and polarizable force fields were developed to simulate the molecule in a series of 10 different solvents of different polarities. The calculated Stokes shift of DMABN increases with increasing orientation polarizability of the surrounding solvent, which is the expected trend, as seen in experimental studies.

Diffusion theory of molecular liquids in the energy representation and application to solvation dynamics

The generalized Langevin equation (GLE) formalism is a useful theoretical fundament for analyzing dynamical phenomena rigorously. Despite the systematic formulation of dynamics theories with practical approximations, however, the applicability of GLE-based methods is still limited to simple polyatomic liquids due to the approximate treatment of molecular orientations involved in the static molecular liquid theory. Here, we propose an exact framework of dynamics based on the GLE formalism incorporating the energy representation theory of solution, an alternative static molecular liquid theory. A fundamental idea is the projection of the relative positions and orientations of solvents around a solute onto the solute-solvent interaction, namely the energy coordinate, enabling us to describe the dynamics on a one-dimensional coordinate. Introducing systematic approximations, such as the overdamped limit, leads to the molecular diffusion equation in the energy representation that is described in terms of the distribution function of solvents on the energy coordinate and the diffusion coefficients. The present theory is applied to the solvation dynamics triggered by the photoexcitation of benzonitrile. The long-time behavior of the solvation time correlation function is in good agreement with that obtained by the molecular dynamics simulation.

Explicit Projection of Stokes Shifts onto Solvent Motion in an Aqueous Liquid and Linear Response Theory

We investigate the molecular origin of the fluorescence Stokes shift in an aqueous liquid. By examining the speed of energy change, the solvation response function is explicitly projected onto the translational and rotational motions of water molecules for both nonequilibrium relaxation and equilibrium fluctuations. Molecular dynamics simulations of a tryptophan solution show that these two processes have highly consistent dynamics, not only for the total response function but also for the decomposed components in terms of specific molecular movements. We found that the rotational mode governs the relaxation of the Stokes shift, whereas the translational mode contributes non-negligibly with slower dynamics. This consistency implies the similarity of the underlying translational and rotational movements of water molecules as the system is far away from and at equilibrium, supporting the validity of the linear response theory at the molecular level. The decomposition methodology is also applicable to a rigid solvent.

Investigation of the Failure of Marcus Theory for Hydrated Electron Reactions

Reactions of the hydrated electron with a wide variety of substrates have been found to exhibit unusually similar activation energies in a manner incompatible with Marcus electron transfer theory. Given the fundamental linear response assumption of Marcus theory, one possible explanation for this apparent failure is that the underlying free energy surfaces governing the reactions are not harmonic; i.e., hydrated electron structural fluctuations exhibit non-Gaussian behavior. In this work, we test this hypothesis by using simulations to calculate the hydrated electron vertical detachment energy distribution. We consider both cavity and noncavity models for the hydrated electron, between which the actual hydrated electron behavior is expected to lie. Our results identify a possible origin for non-Gaussian behavior of the hydrated electron but show that it is not of sufficient magnitude to explain the failure of Marcus theory to describe its reactions. Thus, other explanations must be sought.

Bond-Breaking Reactions Encounter Distinct Solvent Environments Causing Breakdown of Linear Response

Solvent effects are important for understanding solution-phase chemical reactions. Surprisingly, very few studies have explored how solvent dynamics change during the course of a reaction with solutes that encounter a wide range of configurations. Here, we use quantum simulation methods to explore the solvent dynamics during a solution-phase bond-breaking reaction: the photodissociation of Na₂⁺ in liquid Ar. We find that the solute experiences a small number of distinct solvent environments that change in a discrete fashion as the bond lengthens. In characterizing the solvent environments, we show also that linear response fails by all measures, even when nonstationarity of solvent dynamics is considered. This observation of distinct solvent response environments with a solvent that can undergo only translational motions highlights the complexity of solute-solvent interactions, but that there are only a few environments gives hope to the idea that solvation dynamics can be understood for solution-phase reactions that explore a wide configuration space.

Ultrafast Rotational and Translational Energy Relaxation in Neat Liquids

The excess energy flow pathways during rotational and translational relaxation induced by rotational or translational excitation of a single molecule of and within each of four different neat liquids (H₂O, MeOH, CCl₄, and CH₄) are studied using classical molecular dynamics simulations and energy flux analysis. For all four liquids, the relaxation processes for both types of excitation are ultrafast, but the energy flow is significantly faster for the polar, hydrogen-bonded (H-bonded) liquids H₂O and MeOH. Whereas the majority of the initial excess energy is transferred into hindered rotations (librations) for rotational excitation in the H-bonded liquids, an almost equal efficiency for transfer to translational and rotational motions is observed in the nonpolar, non-H-bonded liquids CCl₄ and CH₄. For translational excitation, transfer to translational motions dominates for all liquids. In general, the energy flows are quite local; i.e., more than 70% of the energy flows directly to the first solvent shell molecules, reaching almost 100% for CCl₄ and CH₄. Finally, the determined validity of linear response theory for these nonequilibrium relaxation processes is quite solvent-dependent, with the deviation from linear response most marked for rotational excitation and for the nonpolar liquids.

Nonlinear measurements of kinetics and generalized dynamical modes. I. Extracting the one-dimensional Green's function from a time series

Often, a single correlation function is used to measure the kinetics of a complex system. In contrast, a large set of k-vector modes and their correlation functions are commonly defined for motion in free space. This set can be transformed to the van Hove correlation function, which is the Green's function for molecular diffusion. Here, these ideas are generalized to other observables. A set of correlation functions of nonlinear functions of an observable is used to extract the corresponding Green's function. Although this paper focuses on nonlinear correlation functions of an equilibrium time series, the results are directly connected to other types of nonlinear kinetics, including perturbation-response experiments with strong fields. Generalized modes are defined as the orthogonal polynomials associated with the equilibrium distribution. A matrix of mode-correlation functions can be transformed to the complete, single-time-interval (1D) Green's function. Diagonalizing this matrix finds the eigendecays. To understand the advantages and limitation of this approach, Green's functions are calculated for a number of models of complex dynamics within a Gaussian probability distribution. Examples of non-diffusive motion, rate heterogeneity, and range heterogeneity are examined. General arguments are made that a full set of nonlinear 1D measurements is necessary to extract all the information available in a time series. However, when a process is neither dynamically Gaussian nor Markovian, they are not sufficient. In those cases, additional multidimensional measurements are needed.

Linear-Response and Nonlinear-Response Formulations of the Instantaneous Marcus Theory for Nonequilibrium Photoinduced Charge Transfer

Instantaneous Marcus theory (IMT) offers a way for capturing the time-dependent charge transfer (CT) rate coefficient in nonequilibrium photoinduced CT processes, where the system was photoexcited from its equilibrated ground state vertically to the excitonic state, followed by an electronic transition to the CT state. As derived from the linearized semiclassical nonequilibrium Fermi's golden rule (LSC NE-FGR), the original IMT requires expensive all-atom nonequilibrium molecular dynamics (NEMD) simulations. In this work, we propose computationally efficient linear-response and nonlinear-response formulations for IMT rate calculations, which only require equilibrium molecular dynamics simulations. The linear- and nonlinear-response IMT methods were tested to predict the transient behavior in the photoinduced CT dynamics of the carotenoid-porphyrin-C₆₀ molecular triad solvated in explicit tetrahydrofuran. Our result demonstrated that the nonlinear-response IMT is in excellent agreement with the benchmark NEMD for all cases investigated here, whereas the linear-response IMT predicts the correct trend for all cases but overestimates the transient CT rate in one case involving a significant nonequilibrium relaxation. This mild breakdown of linear-response IMT is due to neglecting the higher-order terms in the exact nonlinear-response IMT. Taking advantage of time translational symmetry, the linear- and nonlinear-response approaches were demonstrated to be able to reduce the computational cost by 80% and 60% compared with NEMD simulations, respectively. Thus, we highly recommend the readily applicable and accurate nonlinear-response IMT approach for simulating nonequilibrium CT processes in complex molecular systems in the condensed phase.

Solvation Dynamics in Water. 4. On the Initial Regime of Solvation Relaxation

It is shown, by means of numerical and analytic work, that initial molecular momenta play little significant role in the initial fast solvation relaxation that follows electronic excitation of, and charge creation for, a standard model system of a solute in water. Instead, the nonequilibrium dynamics are predominantly described by noninertial "steering" by the torques directly generated by the newly created charge distribution. It is this process that largely overcomes inertia and drives the relaxation dynamics on a time scale of a few tens of femtoseconds in the key initial regime of the dynamics. These results are discussed in the context of commonly employed descriptions such as inertial, Gaussian, and underdamped dynamical behavior.

To unravel the connection between the non-equilibrium and equilibrium solvation dynamics of tryptophan: success and failure of the linear response theory of fluorescence Stokes shift

The connections between the non-equilibrium solvation dynamics upon optical transitions and the system's equilibrium fluctuations are explored in aqueous liquid. Linear response theory correlates time-dependent fluorescence with the equilibrium time correlation functions. In the previous work [T. Li, J. Chem. Theory Comput., 2017, 13, 1867], Stokes shift was explicitly decomposed into the contributions of various order time correlation functions on the excited state surface. Gaussian fluctuations of the solute-solvent interactions validate linear response theory. Correspondingly, the deviation of the Gaussian statistics causes the inefficiency of linear response evaluation. The above mechanism is thoroughly tested in this study. By employing molecular simulations, multiple non-equilibrium processes, not necessarily initiated from the ground state equilibrium minimum, were examined for tryptophan. Both the success and failure of linear response theory are found for this simple system and the mechanism is analyzed. These observations, assisted by the width dynamics, the initial state linear response approach, and the variation of the solvation structures, integrally verify the virtue of the excited state Gaussian statistics on the dynamics of Stokes shift.

Linear Response Theory for Decomposition Energies of Stokes Shift in Proteins

In aqueous solution, fluorescence Stokes shift experiments monitor the relaxation of the solute-solvent interactions upon photon excitation of the solute chromophore. Linear response (LR) theory expects the identical dynamics between the Stokes shift and the system's spontaneous fluctuations. Whether this identity guarantees similar dynamics between the nonequilibrated and equilibrium processes for the decomposition energy of the Stokes shift is the main focus of this study. In our previous work [Li, T. J. Chem. Theory Comput. 2017, 13, 1867-1873], Stokes shift is properly correlated with various order time-correlation functions. As a continuation, its decomposition energy from the subsystem is further represented as the full summation of all of the cross-time correlation functions between the decomposition energy and the total solute-solvent interactions. Gaussian statistics of the total solute-solvent interactions ensure the same decay rates among the odd orders not only for the time-correlation functions but also for the cross-time correlation functions, validating the LR of the Stokes shift and the decompositions, respectively. The above mechanism is verified by molecular dynamics simulations in the protein Staphylococcus nuclease and is robust even as the decomposed energy associated with an individual residue exhibits typical non-Gaussian properties. Further examinations reveal the consistent molecular motions for a specific residue over the nonequilibrium and equilibrium processes, which are responsible for the nonequilibrium dynamics of the associated decomposed energy. Our results show the appropriateness of LR on finer molecular scales.

Origin of Slow Solvation Dynamics in DNA: DAPI in Minor Groove of Dickerson-Drew DNA

The measurement and understanding of collective solvation dynamics in DNA have vital biological implications, as protein and ligand binding to DNA can be directly controlled by complex electrostatic interactions of anionic DNA and surrounding dipolar water, and ions. Time-resolved fluorescence Stokes shift (TRFSS) experiments revealed anomalously slow solvation dynamics in DNA much beyond 100 ps that follow either power-law or slow multiexponential decay over several nanoseconds. The origin of such dispersed dynamics remains difficult to understand. Here we compare results of TRFSS experiments to molecular dynamics (MD) simulations of well-known 4',6-diamidino-2-phenylindole (DAPI)/Dickerson-Drew DNA complex over five decades of time from 100 fs to 10 ns to understand the origin of such dispersed dynamics. We show that the solvation time-correlation function (TCF) calculated from 200 ns simulation trajectory (total 800 ns) captures most features of slow dynamics as measured in TRFSS experiments. Decomposition of TCF into individual components unravels that slow dynamics originating from dynamically coupled DNA-water motion, although contribution from coupled water-Na⁺ motion is non-negligible. The analysis of residence time of water molecules around the probe (DAPI) reveals broad distribution from ∼6 ps to ∼3.5 ns: Several (49 nos.) water molecules show residences time greater than 500 ps, of which at least 14 water molecules show residence times of more than 1 ns in the first solvation shell of DAPI. Most of these slow water molecules are found to occupy two hydration sites in the minor groove near DAPI binding site. The residence time of Na⁺, however, is found to vary within ∼17-120 ps. Remarkably, we find that freezing the DNA fluctuations in simulation eliminates slower dynamics beyond ∼100 ps, where water and Na⁺ dynamics become faster, although strong anticorrelation exists between them. These results indicate that primary origin of slow dynamics lies within the slow fluctuations of DNA parts that couple with nearby slow water and ions to control the dispersed collective solvation dynamics in DNA minor groove.

Theoretical Insights on Nonlinear Response Theory of Fluorescence Spectroscopy in Liquids

Correlations between the nonequilibrium solvation dynamics upon the photon excitation of the chromophore and a system's equilibrium fluctuations are deeply studied. As the linear response of the solvent has been linked with Gaussian statistics of the energy fluctuations in the literature, we specifically explore the cases beyond the regime of the linear response theory due to deviation from Gaussian fluctuations. As a continuation of our previous work, an analytical formalism is presented to project the energy shift with various order moments, where the non-Gaussian statistics arise from the overlap of the energy basins on the perturbed potential energy surface. It is shown that the nonequilibrium dynamics still correlate with the spontaneous regressions at equilibrium and are controlled by the decay rates of those higher order components with the prevailing contributions to the energy shift. Molecular dynamics simulations were performed in the protein Staphylococcus nuclease, in which even the dynamics of the high order moments are available. The results further verify the above relationship. Our scheme is used to evaluate Stokes shift using the information on non-Gaussian statistics at equilibrium, thus presenting a broad picture on the correlation between the nonequilibrium process and equilibrium properties in liquids.

Polarizability in ionic liquid simulations causes hidden breakdown of linear response theory

The validity of linear response theory (LRT) in computer simulations of solvation dynamics, i.e. the time-dependent Stokes shift, has been debated widely during the last decades. Since the use of LRT is computationally less expensive than the calculation of the true nonequilibrium response, it is often invoked for large systems exhibiting a particularly slow solvation response, e.g. ionic liquids. In the case of ionic liquids, LRT does not only need to capture the correct overall dynamics of the system, but also the contributions and timescales of the respective cation and anion movement. We show by large scale computer simulations that the contribution of the permanent dipoles to the solvation response obeys LRT to some extent, whereas the induced contributions in polarizable simulations lead to a failure of LRT for the respective ion contributions.

Changes in protein hydration dynamics by encapsulation or crowding of ubiquitin: strong correlation between time-dependent Stokes shift and intermolecular nuclear Overhauser effect

The local changes in protein hydration dynamics upon encapsulation of the protein or macromolecular crowding are essential to understand protein function in cellular environments. We were able to obtain a spatially-resolved picture of the influence of confinement and crowding on the hydration dynamics of the protein ubiquitin by analyzing the time-dependent Stokes shift (TDSS), as well as the intermolecular Nuclear Overhauser Effect (NOE) at different sites of the protein by large-scale computer simulation of single and multiple proteins in water and confined in reverse micelles. Besides high advanced space resolved information on hydration dynamics we found a strong correlation of the change in NOE upon crowding or encapsulation and the change in the integral TDSS relaxation times in all investigated systems relative to the signals in a diluted protein solution.

Changes in local protein hydration dynamics caused by encapsulation or crowding are reflected in the TDSS and the intermolecular NOE alike.

Origin of diverse time scales in the protein hydration layer solvation dynamics: A simulation study

In order to inquire the microscopic origin of observed multiple time scales in solvation dynamics, we carry out several computer experiments. We perform atomistic molecular dynamics simulations on three protein-water systems, namely, lysozyme, myoglobin, and sweet protein monellin. In these experiments, we mutate the charges of the neighbouring amino acid side chains of certain natural probes (tryptophan) and also freeze the side chain motions. In order to distinguish between different contributions, we decompose the total solvation energy response in terms of various components present in the system. This allows us to capture the interplay among different self- and cross-energy correlation terms. Freezing the protein motions removes the slowest component that results from side chain fluctuations, but a part of slowness remains. This leads to the conclusion that the slow component approximately in the 20-80 ps range arises from slow water molecules present in the hydration layer. While the more than 100 ps component has multiple origins, namely, adjacent charges in amino acid side chains, hydrogen bonded water molecules and a dynamically coupled motion between side chain and water. In addition, the charges enforce a structural ordering of nearby water molecules and helps to form a local long-lived hydrogen bonded network. Further separation of the spatial and temporal responses in solvation dynamics reveals different roles of hydration and bulk water. We find that the hydration layer water molecules are largely responsible for the slow component, whereas the initial ultrafast decay arises predominantly (approximately 80%) due to the bulk. This agrees with earlier theoretical observations. We also attempt to rationalise our results with the help of a molecular hydrodynamic theory that was developed using classical time dependent density functional theory in a semi-quantitative manner.

Solvation dynamics in polar solvents and imidazolium ionic liquids: failure of linear response approximations

This study presents the large scale computer simulations of two common fluorophores, N-methyl-6-oxyquinolinium betaine and coumarin 153, in five polar or ionic solvents. The validity of linear response approximations to calculate the time-dependent Stokes shift is evaluated in each system. In most studied systems linear response theory fails. In ionic liquids the magnitude of the overall response is largely overestimated, and linear response theory is not able to capture the individual contributions of cations and anions. In polar liquids, the timescales of solvation dynamics are often not correctly reproduced. These observations are complemented by a detailed analysis of Gaussian statistics including higher order correlation functions, variance of the energy gap distribution and its time evolution. The analysis of higher order correlation functions was found to be not suitable to predict a failure of linear response theory. Further analysis of radial distribution functions and hydrogen bonds in the ground and excited state, as well as the time evolution of the number of hydrogen bonds after solute excitation reveal an influence of solvent structure in some of the studied systems.

Efficient Criterion To Evaluate Linear Response Theory in Optical Transitions

The role of the Gaussian statistics on the solvation dynamics upon the photon excitation of the chromophore is deeply explored. The linear response theory for the fluorescence Stokes shift is investigated. An analytical formulism is presented to recast Stokes shift into the contributions of the equilibrium time correlation functions of the solute-solvent interactions on the excited-state surface, and the latter is further reformed and depicted by the time relaxation of the moment. As the first application of the formulism in the molecular dynamics simulations, it is verified that the efficiency of the linear response theory relies on the Gaussian characteristics of the dominant moments in terms of the Stokes shift, which is identified by the same relaxation dynamics between those moments and the linear order one. The comparisons between the above observations on the linearity of Stokes shift and the explanations in the literature are discussed. The key finding is the development of explicit criterion to measure the appropriateness of applying linear response theory.

On the validity of linear response approximations regarding the solvation dynamics of polyatomic solutes

Gaussian statistics and linear response predictions of the nonequilibrium solvation dynamics are tested for numerous solute/solvent combinations.

Vibrational energy relaxation of large-amplitude vibrations in liquids

Given the limited intermolecular spaces available in dense liquids, the large amplitudes of highly excited, low frequency vibrational modes pose an interesting dilemma for large molecules in solution. We carry out molecular dynamics calculations of the lowest frequency ("warping") mode of perylene dissolved in liquid argon, and demonstrate that vibrational excitation of this mode should cause identifiable changes in local solvation shell structure. But while the same kinds of solvent structural rearrangements can cause the non-equilibrium relaxation dynamics of highly excited diatomic rotors in liquids to differ substantially from equilibrium dynamics, our simulations also indicate that the non-equilibrium vibrational energy relaxation of large-amplitude vibrational overtones in liquids should show no such deviations from linear response. This observation seems to be a generic feature of large-moment-arm vibrational degrees of freedom and is therefore probably not specific to our choice of model system: The lowest frequency (largest amplitude) cases probably dissipate energy too quickly and the higher frequency (more slowly relaxing) cases most likely have solvent displacements too small to generate significant nonlinearities in simple nonpolar solvents. Vibrational kinetic energy relaxation, in particular, seems to be especially and surprisingly linear.

Solvation dynamics and proton transfer in nanoconfined liquids

Nanoconfined liquids are of interest because of both their fundamental properties and their potential utility in an array of applications. The structure and dynamics of the liquid can be dramatically impacted by the geometrical constraints and the interactions with the interface. Understanding the molecular-level origins of these changes and how they are determined by the characteristics of the confining framework is the subject of ongoing experimental and theoretical studies. The progress and remaining challenges in these efforts are reviewed in the context of solvation dynamics and proton transfer reactions, processes that are strongly affected by nanoscale confinement.

Polar solvation dynamics of coumarin 153 by ultrafast time-resolved fluorescence

Polar solvation dynamics of coumarin 153 in acetonitrile, methanol, and butanol are investigated by dynamic Stokes shift function, S(t). In small protic solvents, it is known that an initial ultrafast component below 50 fs constitutes more than half of the total solvation process. We use fluorescence up-conversion technique via two-photon absorption process, which can provide 40 fs time resolution for the whole emission wavelength range. Moreover, time-resolved fluorescence spectra are recorded directly without the spectral reconstruction. We observe a temporal oscillation in frequency of whole emission spectrum in the solvation curve. In the obtained S(t), initial solvation time scale is 120 fs, invariant to solvents used in this experiment, although its amplitude varies in different solvents.