On transition path times for condensed-phase non-adiabatic electron transfer reactions under a two-parabola model
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "On transition path times for condensed-phase non-adiabatic electron transfer reactions under a two-parabola model".
Mira: Here's the gist of the paper. Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science.
Kai: The study performs non-adiabatic simulations based on the Zhu-Nakamura theory, combining the two-parabola model with Kramers-like Langevin dynamics to analyze the dependence of macroscopic reaction time and microscopic transition process timescale on diabatic coupling and friction parameter.
Mira: Although microscopic transfer processes are generally regarded as instantaneous, the timescale of the microscopic transition process can still be a non-negligible fraction of the macroscopic reaction time.
Kai: Model System Description The condensed-phase electron transfer reaction is described by a two-parabola model where two diabatic parabolic potential energy surfaces represent the reactants and products, coupled by a diabatic coupling V.
Title and authors: Kai: So, we're talking about the title, "On transition path times for condensed-phase non-adiabatic electron transfer reactions under a two-parabola model." It sets up exactly what they are doing: looking at how the time it takes for an electron to jump between states in a coupled system behaves when you use this specific two-parabola setup.
Mira: What I find important about the authors is that they are explicitly combining the ZhuNakamura theory with Kramers-like Langevin dynamics to get this unified picture of how friction and coupling affect the timescale.
Lev: I'm thinking, if we were to try and run this on a real quantum system, we’d have to worry about accurately modeling that friction parameter gamma in our noise terms, which is always a challenge when moving from theory to experiment.
Kai: That's true. And the paper sets up the math clearly by defining how they diagonalize the Hamiltonian into those adiabatic energies E1 and E2, and then defining those basis states using coefficients like eta of x.
Mira: Exactly, because they show us that these coefficientseta are directly related to the coupling V, which is what drives this whole non-adiabatic process.
Lev: So when you look at equation two, that expression for the adiabatic energy E1 looks a bit messy with those terms involving r and p, which hints at how complex the underlying diabatic surfaces really are.
Kai: It’s complicated notation, but what matters is that it leads to these specific expressions for the coefficients used to define the adiabatic basis states phi one and phi two.
Mira: And those coefficients are key because they determine the hopping probability Phop, which is how often that transition actually happens in their model.
Lev: The hopping probability itself involves an exponential term with some terms like a squared s a squared and b squared, so it shows how sensitive the transition rate is to those effective coupling and collision energy parameters.
The paper's summary: Kai: So, the summary says they numerically analyzed how the macroscopic reaction time and that microscopic transition process timescale depend on two main things: diabatic coupling V and the friction parameter in Langevin dynamics.
Mira: And they found something interesting—although they generally assume these microscopic transfers are instantaneous—they show that the TPT itself can still be a non-negligible fraction of the macroscopic reaction time.
Lev: That's a big deal because it challenges the idea that you can just ignore the timescale difference between how fast things look macroscopically and how fast they actually happen at the molecular level.
Kai: Right, so they are quantifying this TPT distribution, and they show that as V increases, the distribution of these TPTs shifts toward shorter times because their hopping probability decreases with increasing momentum at the crossing point.
Mira: That dependence on V is consistent with what Fermi's golden rule predicts for coupling strength; it shows that stronger coupling leads to faster microscopic transitions, which makes sense.
Lev: But then they also found that the mean of the TPT, or mTPT, doesn't just depend on V; it's mainly determined by the friction parameter gamma in the diffusive regime.
Kai: That’s a contrast. So while V strongly affects how fast individual transitions happen, gamma governs how long you spend overall moving along that reaction coordinate.
Mira: And they pointed out that when V is large, this difference between the mean TPT and the macroscopic time constant tau can be about one order of magnitude, which suggests those two timescales are not completely separated.
The paper's improvements: Kai: The authors suggest a few things to improve or clarify their analysis. They focus on how they characterize these timescales by looking at the TPT distribution and how it responds to changes in V and gamma.
Mira: One improvement is that they analyze the dependence of the TPT distribution itself on both V and gamma, which gives a richer picture than just looking at one parameter in isolation.
Lev: From an engineering standpoint, I think analyzing the dependence on gamma helps us understand how noise affects the timing structure, which is exactly what we need when building robust systems that operate under thermal fluctuations.
Kai: Another point they make is that they show how increasing friction parameter gamma makes the TPT distribution less dependent on V. This suggests that for certain conditions, you can decouple those two effects somewhat in terms of timing structure.
Mira: That leads to their conclusion that the mean of the TPT, mTPT, is primarily governed by gamma when you are in the diffusive regime, while tau depends more on V.
Lev: So what this means for running simulations on actual hardware is that if we can control our noise level well enough, we might be able to get a clearer idea of how coupling V drives the microscopic events.
Conclusion: Kai: So to wrap up, the main implication of "On transition path times for condensed-phase non-adiabatic electron transfer reactions under a two-parabola model" is that we can’t treat the microscopic and macroscopic timescales as entirely separate things when coupling V is large.
Mira: They show that while the TPT distribution shifts toward shorter times with increasing coupling V, the mean of the TPT remains mainly controlled by gamma.
Lev: And they quantify this separation by showing that when V is big, tau and mTPT can differ by about a factor of ten. That tells us the timing structure is governed by different mechanisms depending on what you are measuring.
Kai: So, we have this detailed look at how non-adiabatic effects and thermal fluctuations structure the temporal behavior of electron transfer reactions using this two-level avoided crossing model.
Mira: It gives a clearer picture for anyone working in condensed matter to understand how coupling strength and friction parameters interplay in these kinds of dynamic processes.
Lev: And it points us toward needing better control over the friction parameter if we want our simulations to accurately reflect the true physical limits of the reaction timescale on real systems.
Department of Chemical Science and Technology, Graduate School of Engineering, Kyoto University · PRESTO, Japan Science and Technology Agency · Fukui Institute for Fundamental Chemistry, Kyoto University
physics.chem-ph
Submitted: 2026-07-28
Updated: 2026-11-12
DOI: 10.1063/5.0354372
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 85/100
The gist: Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science.
Key concepts
- Two-Parabola Model
- This model describes the electron transfer system using two distinct parabolic potential energy surfaces representing reactants and products. These surfaces are coupled by a diabatic coupling term ($V$), which represents the electronic interaction between them, allowing for non-adiabatic transitions between states.
- Transition Path Time (TPT)
- TPT quantifies the timescale of microscopic transition processes—the time it takes for a trajectory to cross an energy barrier. This metric reveals how long the actual physical movement or hopping event lasts, which can be different from the overall reaction rate constant.
- Diabatic Coupling ($V$)
- This parameter represents the strength of the electronic coupling between the two potential energy surfaces. Increasing $V$ enhances non-adiabatic effects, leading to faster microscopic transitions and influencing how quickly the macroscopic reaction time changes.
- Friction Parameter ($\gamma$)
- The friction parameter governs the dynamics through Langevin equation, representing environmental coupling or dissipation. It dictates whether the system moves ballistically (small $\gamma$) or diffusively (large $\gamma$), significantly affecting the mean TPT.
Terminology
Summary
Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science. p. 0 On transition path times for condensed-phase non-adiabatic electron
transfer reactions under a two-parabola model
Ryo Nihei,1 Hiroki Uratani,1, 2, a) and Hirofumi Sato1, 3
1)Department of Chemical Science and Technology, Graduate School of Engineering, Kyoto University, Kyoto, 615-0510,
Japan
2)PRESTO, Japan Science and Technology Agency, Kawaguchi, Saitama, 332-0012, Japan
3)Fukui Institute for Fundamental Chemistry, Kyoto University, Kyoto, 606-8103, Japan
Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science. The characteristics of the microscopic transition processes involved in these reactions and their dynamic behavior remain poorly understood. In this study, we performed non-adiabatic simulations based on the Zhu-Nakamura
theory, combining the two-parabola model with Kramers-like Langevin dynamics. We numerically analyzed the dependence of the macroscopic reaction time and the timescale of the microscopic transition process on the diabatic
coupling between the diabatic states and on the friction parameter in Langevin dynamics. We numerically analyzed the dependence of the macroscopic reaction time and the timescale of the microscopic transition process on the diabatic
coupling between the diabatic states and on the friction parameter in Langevin dynamics. Although the microscopic
transfer process is generally regarded as instantaneous in the non-adiabatic regime, we found that the timescale of
the microscopic transition process can still be a non-negligible fraction of the macroscopic reaction time.Where the paper says it
Where the paper says it
Where the paper says it
Model System Description
The condensed-phase electron transfer reaction is described by a two-parabola model where two diabatic parabolic potential energy surfaces represent the reactants and products, coupled by a diabatic coupling V. The electronic Hamiltonian is given as H el = 1/2ω2(x + xa)2 V/V 1/2ω2(x − xa)2! p. 0 21,33–38 Two diabatic parabolic potential energy
surfaces represent the reactants (Er) and products (Ep), with
the diabatic coupling V between these states(Fig. The
electronic Hamiltonian is
H
el =
Er V
V Ep
! , the reaction driving force is 0. The adiabatic energies, which are given by diagonalizing the
Hamiltonian (1), are
E1(x) =
1
2
Er(x) + Ep(x)
−
1
2
r
Er(x) − Ep(x)
2
+ 4V2
(2)
E2(x) =
1
2
Er(x) + Ep(x)
+
1
2
r
Er(x) − Ep(x)
2
+ 4V2
. (1) and can be expressed as a linear combination of the diabatic basis |Ξr⟩ and |Ξp⟩
|Φ1⟩ = cr,1(x) |Ξr⟩ + cp,1(x) |Ξp⟩ (4)
|Φ2⟩ = cr,2(x) |Ξr⟩ + cp,2(x) |Ξp⟩, (5)
where cr,1(x), cp,1(x), cr,2(x), and cp,2(x) are
cr,1(x) =
s
1
2
−
1
2
1
p
1 + η(x)
2
, cp,1(x) = −
s
1
2
+
1
2
1
p
1 + η(x)
2
cr,2(x) =
s
1
2
+
1
2
1
p
1 + η(x)
2
, cp,2(x) =
s
1
2
−
1
2
1
p
1 + η(x)
2
η(x) =
2V
Er(x) − Ep(x)
. (6)
We will use these coefficients in Sec.Where the paper says it
Where the paper says it
Where the paper says it
Dynamics Simulation Setup
The dynamics are simulated using the Langevin equation of motion, x¨(t) = -dEi(x)/dx - γx˙(t) + ξ(t), where γ is the friction parameter and ξ is a Gaussian random force satisfying the fluctuation–dissipation relation. The standard deviation of ξ is proportional to √γ, indicating that larger γ corresponds to the diffusive regime, while smaller values correspond to the ballistic regime. p. 1 (8), the standard deviation of ξ is proportional to √
γ. The larger γ corresponds to the diffusive
regime, while the smaller to the ballistic regime. To account for non-adiabatic transitions, we used the ZhuNakamura theory. 24–28 This theory provides an exact solution
for transitions in a one-dimensional two-level system. (9)
In this study, because diabatic gradients have opposite signs
at the avoided crossing point, the sign after b
4
is minus. The
parameters a and b represent the effective coupling and the
effective collision energy, respectively, where
a
2 =
ℏ
2
2
√
|F2F1||F2 − F1|
(2V)
3
, (10)
b
2 = (Et − EX)
|F2 − F1|
√
F2F1(2V)
, (11)
where F1 and F2 are the diabatic gradients.Where the paper says it
Where the paper says it
Where the paper says it
Characterization of Timescales
The transition path time (TPT) quantifies the timescale of microscopic transition processes, representing the time required for a trajectory to cross the energy barrier region along the reaction coordinate. The population of reactants R evolves according to R = 1/2(1 + exp(-2kt)), where τ (= 1/k) corresponds to the mean first-passage time (MFPT). The reaction rate constant k is proportional to V2. p. 2 The population of reactants [R] is given by averaging this probability over all trajectories at each time step. Since the potential energy surfaces
are symmetric with respect to reactant–product permutation,
the reaction scheme of this system is
Reactants
k
⇌
k
Products, (13)
where k is a rate constant.Where the paper says it
Key Findings on Time Scales
The dependence of the time constant τ on diabatic coupling V shows that as V increases, τ decreases monotonically, consistent with Fermi’s golden rule predictions. The distribution of TPT shifts toward shorter times as V increases because the non-adiabatic hopping probability decreases with increasing momentum at the crossing point. p. 4 6 shows the dependencies of the TPT distribution on
V and γ, indicating that the increase in V shifts the TPT distribution toward shorter times, and this trend is consistently
observed in any γ. To investigate the origin of these changes,
we compared the mean momentum at the potential crossing
point versus the TPT (Fig. The results
showed that while the TPT distribution depends on both V
and γ, the mean of TPT (mTPT) is mainly determined by γ. In contrast, the reaction time τ depends primarily on V, indicating that the microscopic and macroscopic time scales
are governed by different mechanisms. In contrast, the reaction time τ depends primarily on V, indicating that the microscopic and macroscopic time scales
are governed by different mechanisms. Furthermore, we
found that the difference between τ and mTPT is approximately one order of magnitude when V is large, suggesting
that the timescales of the microscopic reaction events and
the macroscopic reaction rate are not completely separated.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion
The study elucidates the influence of non-adiabatic effects and fluctuations on the temporal structure of condensed-phase ET reactions by analyzing TPT distributions under a two-level avoided crossing model. The results show that while the TPT distribution depends on both V and γ, the mean of TPT is mainly determined by γ, in contrast to the reaction time τ which depends primarily on V. p. 6 The results
showed that while the TPT distribution depends on both V
and γ, the mean of TPT (mTPT) is mainly determined by γ. In contrast, the reaction time τ depends primarily on V, indicating that the microscopic and macroscopic time scales
are governed by different mechanisms. The results
showed that while the TPT distribution depends on both V
and γ, the mean of TPT (mTPT) is mainly determined by γ. In contrast, the reaction time τ depends primarily on V, indicating that the microscopic and macroscopic time scales
are governed by different mechanisms. In contrast, the reaction time τ depends primarily on V, indicating that the microscopic and macroscopic time scales
are governed by different mechanisms. Furthermore, we
found that the difference between τ and mTPT is approximately one order of magnitude when V is large, suggesting
that the timescales of the microscopic reaction events and
the macroscopic reaction rate are not completely separated.Where the paper says it
Where the paper says it
Where the paper says it
Appendix Details
The simulation parameters were fixed for xa, T, ω, and ∆t according to previous studies. The choice of x0 was made where x0 is at its narrowest value of 2. p. 6 Comparion of various x0
As x0 increases, the TPT distribution shifts toward the
long-duration side when γ is large, however, the trend resulting from changes in V remains largely unchanged, and
it can be seen that the overall shape of the distribution itself
does not change depending on the choice of x0. Therefore,
we chose x0 to be the point where x0 is at its narrowest, i. This figure also shows that, with respect
to τ, the influence of V is much greater than that of γ. Comparion of various x0
As x0 increases, the TPT distribution shifts toward the
long-duration side when γ is large, however, the trend resulting from changes in V remains largely unchanged, and
it can be seen that the overall shape of the distribution itself
does not change depending on the choice of x0. 11 shows the time evolution of the reactant population
under all conditions. This figure also shows that, with respect
to τ, the influence of V is much greater than that of γ. Moreover, it can be noticed that the increase in γ makes
the TPT distribution less dependent on V (Fig. 6) and weakens the correlation between TPT and the mean momentum
at the potential crossing point (Fig. The mean momentum at the potential crossing point ultimately approaches that
obtained from the simple Maxwell–Boltzmann distribution. These behaviors can be understood that when the contribution of the random force ξ is significant the TPT is not substantially influenced by the momenta at the potential crossing
point, i. 9, V has a complicated influence on the mTPTs, while
the influence of γ is monotonic, and the latter is more significant than the former. Thus, the mTPT is primarily determined by γ in contrast to the TPT distribution. 9, V has a complicated influence on the mTPTs, while
the influence of γ is monotonic, and the latter is more significant than the former. Thus, the mTPT is primarily determined by γ in contrast to the TPT distribution. This figure also shows that, with respect
to τ, the influence of V is much greater than that of γ. Comparion of various x0
As x0 increases, the TPT distribution shifts toward the
long-duration side when γ is large, however, the trend resulting from changes in V remains largely unchanged, and
it can be seen that the overall shape of the distribution itself
does not change depending on the choice of x0. Time evolution of the reactant population for various values of diabatic coupling V and friction parameter γ (The horizontal axis
is time and the vertical axis is [R]. The graph are sorted in ascending order by the value of γ.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
References
The paper cites numerous works, including those by Kramers (21), Zhu and Nakamura (24–28), and others related to non-adiabatic dynamics. The references listed cover topics from chemical physics to molecular dynamics. p. 2 To obtain statistically converged results, we used 10,000
trajectories for the population dynamics and 100,000 trajectories for the TPT analysis. The parameters employed in
the simulations are listed in Table I. A microscopic understanding of these reactions is
important for new technologies such as photoredox catalysis,
solar energy conversion, and artificial photosynthesis. 1–9 Although the mechanisms and dynamics of condensed-phase
ET reactions have been widely investigated from both theoretical and experimental perspectives, the coupling among
nuclear motion, solvent fluctuations, and electronic transitions continues to pose fundamental problems. On transition path times for condensed-phase non-adiabatic electron
transfer reactions under a two-parabola model
Ryo Nihei,1 Hiroki Uratani,1, 2, a) and Hirofumi Sato1, 3
1)Department of Chemical Science and Technology, Graduate School of Engineering, Kyoto University, Kyoto, 615-0510,
Japan
2)PRESTO, Japan Science and Technology Agency, Kawaguchi, Saitama, 332-0012, Japan
3)Fukui Institute for Fundamental Chemistry, Kyoto University, Kyoto, 606-8103, Japan
Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science. The characteristics of the microscopic transition processes involved in these reactions and their dynamic behavior remain poorly understood. RESULTS AND DISCUSSION
A. Population dynamics and reaction rates
Fig.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI can simulate condensed-phase non-adiabatic electron transfer reactions by combining Kramers-like Langevin dynamics with Zhu-Nakamura theory to provide a unified microscopic description of coupled nuclearelectronic motion in condensed-phase ET.
-
The improved system can analyze the dependence of the macroscopic reaction time and the timescale of microscopic transition processes on diabatic coupling and friction parameter using numerical simulations.
-
The AI can quantify transition path times (TPT) to characterize the timescales of microscopic transition processes, providing information that differs from macroscopic reaction rates, even when a process is generally regarded as instantaneous in the non-adiabatic regime.
-
The system can determine the reaction rate and time constant tau by averaging trajectory probabilities and considering the Markov process assumption for reactant population evolution.
-
The AI can analyze TPT distributions to show how increasing diabatic coupling V shifts the distribution toward shorter times, and how increasing friction parameter gamma makes the distribution less dependent on V, suggesting that
the timescale of the overall reaction is governed by V.
-
The system can distinguish between the influence of diabatic coupling and friction parameter on mean TPT (mTPT), finding that while mTPT is mainly determined by gamma, it remains largely unchanged with respect to V.
-
The AI can assess when microscopic and macroscopic time scales are not completely separated by observing that the difference between tau and mTPT can be approximately one order of magnitude when V is large, contradicting the assumption that microscopic transitions are instantaneous.
Abstract
Condensed-phase non-adiabatic electron transfer (ET) reactions play an important role in various areas of science. The characteristics of the microscopic transition processes involved in these reactions and their dynamic behavior remain much room for further understanding. In this study, we performed non-adiabatic dynamics simulations based on Zhu-Nakamura theory, combining the two-parabola model with Kramers-like Langevin dynamics. We numerically analyzed the dependence of the macroscopic reaction time and the timescale of the microscopic transition process on the diabatic coupling between the diabatic states and on the friction parameter in Langevin dynamics. Focusing on resulting two quantities, that is, the reaction rate and the transition path times (TPTs), which characterize the macroscopic population transfer and the individual transition process, respectively, we discuss how non-adiabatic transitions influence the reaction dynamics.
Related papers
- A System-Independent Metadynamics Strategy for Reactive Training Data: Application to Gas-Phase Organic Reactions
- A Machine Learning Model for the Chemistry of a Solvated Electron
- Compression of virtual spaces in transcorrelated methods via singular value decomposition: application to the G2 set
- Charge Transfer with a Spin. II: A Framework for Diabatization which Localizes Charge and Spin
- Diabatic Seam Space Sampling for Hydrogen Tunneling Systems with Nuclear-Electronic Orbital Theory
- Augmented Roothaan-Hall Hessian Applied to Spin-Restricted Open-Shell Density-Functional Theory