Pith. sign in

REVIEW 3 major objections 4 minor 27 references

A Fermi system relaxing toward equilibrium passes through two distinct relaxation regimes, so its short-time effective relaxation time (≈1.0×10⁻²³ s) and its asymptotic exponential time (≈3.2×10⁻²³ s) measure different stages and do not dis

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 16:51 UTC pith:FSSIGUAD

load-bearing objection The two-time-scale explanation is likely correct, but the paper's key asymptotic formula has a missing (1−f_eq) factor that undercuts the quantitative claims until fixed. the 3 major comments →

arxiv 2607.17837 v1 pith:FSSIGUAD submitted 2026-07-20 nucl-th

Dissipative properties of a Fermi system within the diffusion approximation of kinetic theory

Sergiy V. Lukyanov This is my paper
classification nucl-th
keywords Fermi systemdiffusion approximationkinetic theoryrelaxation timeFokker–Planck equationnonexponential relaxationBurgers equationnuclear dissipation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper claims that the factor-of-three difference between two previously reported relaxation times for a Fermi system is not an inconsistency but evidence that relaxation is nonexponential and has two temporal regimes. Working in the diffusion approximation of kinetic theory with constant kinetic coefficients, it solves the nonlinear Fokker–Planck equation in energy space exactly for an initially sharp Fermi sphere. The solution approaches the Fermi–Dirac distribution with an asymptotic time τeq ≈ 3.2×10⁻²³ s, while the area-based effective time τeff ≈ 1.0×10⁻²³ s is dominated by the early, faster, nonexponential phase. This matters because it tells modelers which time scale to use when extracting nuclear dissipation rates, and why two legitimate extractions can differ.

Core claim

The paper’s central claim is that relaxation in the diffusion approximation is governed by an exact solution of the nonlinear diffusion equation in energy space. For an initial step-function distribution, the solution is expressed in terms of error functions. In the long-time limit, the deviation from equilibrium decays as t^{-1/2} exp(-t/τeq), so the asymptotic exponential relaxation time is τeq = 4D/v². The normalized root-mean-square deviation Δ(t)/Δ0, when integrated over time to define τeff, yields a shorter time because the early decay is faster than exponential. An instantaneous relaxation time τ(t) = -t/ln(Δ/Δ0) grows monotonically toward τeq from below, while the asymptotic approxim

What carries the argument

The central object is the nonlinear Fokker–Planck equation in energy space with constant kinetic coefficients: ∂f/∂t = -v ∂[f(1-f)]/∂ε + D ∂²f/∂ε². Substituting w = v - 2vf turns it into the Burgers equation, and the Cole–Hopf transformation linearizes it to the heat equation. The exact solution (14) for a step initial condition is the workhorse: its long-time expansion yields the equilibrium Fermi distribution and the asymptotic decay with τeq = 4D/v². The integral definition of τeff and the logarithmic instantaneous time τ(t) are then used to expose the two regimes.

Load-bearing premise

The whole derivation rests on treating the diffusion and drift coefficients as constant and the single-particle density of states as constant, g(ε) ≈ g(εF), in the low-temperature limit; if the real kinetic coefficients vary appreciably near the Fermi surface, the exact solution and the two relaxation times may not describe an actual nucleus.

What would settle it

Run the same relaxation with the same parameters in a numerical Landau–Vlasov or Boltzmann–Uehling–Uhlenbeck solver with energy-dependent kinetic coefficients and a Fermi-gas density of states, then compare the normalized root-mean-square deviation Δ(t)/Δ0 with Eq. (40): if the instantaneous relaxation time τ(t) stays flat near 1.0×10⁻²³ s instead of climbing toward about 3.2×10⁻²³ s, the two-regime explanation fails.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Relaxation of a Fermi system in this model cannot be characterized by a single relaxation time: early-time decay is faster than exponential, while late-time decay is exponential with a longer time.
  • τeq ≈ 3.2×10⁻²³ s is determined purely by the kinetic coefficients (τeq = 4D/v²), whereas τeff ≈ 1.0×10⁻²³ s depends on the initial distribution, so different initial excitations can share the same asymptotic rate but different effective rates.
  • The previously reported factor-of-three discrepancy between momentum-space and energy-space relaxation times is reconciled: both extractions are correct but measure different parts of the relaxation curve.
  • For the model parameters (εF = 37 MeV, Teq = 4 MeV), the deviation from equilibrium becomes practically negligible after about 10×10⁻²³ s, and the asymptotic formula (40) is only valid after roughly 5.6×10⁻²³ s.
  • A single-exponential fit applied to early-time data will yield a relaxation time near 1.0×10⁻²³ s and will underestimate the asymptotic relaxation time.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A testable extension: if D and v are allowed to be energy-dependent, the ratio τeq/τeff should depart from the constant-model value of about 3.2 in a way that tracks the variation of the kinetic coefficients near the Fermi surface.
  • The two-time-scale structure is likely generic for nonlinear diffusion equations with a drift term that generate a t^{-1/2} algebraic prefactor in the asymptotic deviation; the distinction between an integral relaxation time and an asymptotic relaxation time should appear in any such system.
  • The paper implies that in nuclear-reaction or heavy-ion simulations, the fitting window used to quote a 'relaxation time' should always be reported, because the fitted value shifts continuously from about 1×10⁻²³ s at early times toward 3.2×10⁻²³ s at late times.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper studies the relaxation of a Fermi system described by a nonlinear diffusion equation in energy space with constant diffusion and drift coefficients, using a step-function initial distribution. It reproduces the exact analytical solution, derives the long-time asymptotic form of the deviation from equilibrium, and introduces two relaxation times: τeq from the exponential tail and τeff from the time integral of the normalized root-mean-square deviation. The numerical and analytical results are used to argue that τeff ≈ 1.0×10^-23 s and τeq ≈ 3.2×10^-23 s characterize different temporal regimes, thereby explaining the factor-of-three discrepancy reported in earlier work.

Significance. If the central claim holds, the paper provides a useful and physically plausible resolution of an apparent inconsistency between two definitions of relaxation time in a nonlinear kinetic model. The work has clear strengths: it uses the exact analytical solution, supports the qualitative conclusion by numerical integration of the exact Δ(t), and does not fit τeff or τeq to the values they are supposed to explain. The conclusion that nonexponential relaxation implies different effective and asymptotic time scales is falsifiable and independent of the algebraic error discussed below. However, the quantitative asymptotic analysis, especially Eqs. (23) and (36)–(40), contains a load-bearing mathematical error that must be corrected before the paper can be accepted.

major comments (3)
  1. [§III, Eq. (23)] The expansion leading to Eq. (23) is incorrect. From Eq. (15) with R≈1−r and x=(ε−εF)/Teq, one obtains f=(1−r)/(1−r+e^x) ≈ f_eq − r e^x/(1+e^x)^2 = f_eq − r f_eq(1−f_eq). Thus δf = −r f_eq(1−f_eq), not −r f_eq. The missing factor (1−f_eq) is essential: it suppresses the deviation deep inside the Fermi sea, whereas Eq. (23) predicts a spurious linear-in-energy deviation for ε≪εF. This error propagates into all subsequent asymptotic results.
  2. [§IV, Eqs. (36)–(40), Figs. 3–4] Because of the missing (1−f_eq) factor, the squared deviation integral in Eq. (36) has the wrong integrand. The correct integrand is x² e^{2x}/(1+e^x)^4, not x²/(1+e^x)^2. The former is integrable over (−∞,∞) and equals 1/6, giving an asymptotic Δ(t)/Δ0 that is independent of εF; the latter diverges as x→−∞ and only becomes finite through the artificial lower cutoff, producing the (εF/Teq)^3/3 term in Eq. (38). Consequently Eq. (40), the claimed threshold t<5.6×10^-23 s, and the asymptotic curves in Figs. 3–4 are quantitatively unsupported. The central qualitative distinction between τeff and τeq does not depend on this prefactor, but the derivation and the quantitative comparisons must be redone.
  3. [§IV, Eq. (38)] Independently of Eq. (23), the 'low-temperature limit' evaluation in Eq. (38) is not a valid asymptotic approximation: the integral ∫ x² dx/(1+e^x)^2 diverges at the lower limit, so replacing the lower limit by −εF/Teq makes the result cutoff-dominated rather than a genuine low-T expansion. This is a second indication that the asymptotic calculation needs to be reworked.
minor comments (4)
  1. [§I, Refs. [7,11]] References [7] and [11] appear to be self-citations to arXiv-only or very recent items; please update with published versions or clearly indicate status.
  2. [§IV, Eq. (24)–(26)] Δ(t) is called a root-mean-square deviation, but Eq. (25) is an unnormalized L2 norm. Since only ratios enter, this does not affect the results, but the terminology should be adjusted.
  3. [Figs. 3–4] After correcting Eq. (40), the asymptotic curves and the time ranges over which they exceed unity must be recomputed; the current captions and the quoted thresholds refer to the erroneous expression.
  4. [§V, Conclusions] The conclusion that 'the asymptotic expression (40) ... falls outside its range of validity by taking values greater than 1' is based on the incorrect prefactor; it should be revised once the correct asymptotic ratio is used.

Circularity Check

0 steps flagged

No significant circularity: τ_eff and τ_eq are computed from the exact solution rather than fitted, and the central explanation is derived, not assumed.

full rationale

The derivation chain from Eq. (2) onward is self-contained. The exact solution (14) is obtained from the Burgers/Cole–Hopf reduction, τ_eq emerges from the asymptotic expansion as 4D/v² (Eq. 20), and τ_eff is evaluated by direct integration of the normalized mean-square deviation (Eq. 24). Neither relaxation time is fitted to the value it is supposed to explain; both follow from the same kinetic coefficients and initial step distribution. The paper does adopt D, v, and the constant-density-of-states approximation from prior work, including several self-citations (Refs. [6,7,11]), but these are modeling inputs and background derivations, not the target conclusion. The factor-of-three discrepancy is a consequence of the exact solution, not an input: the paper shows that early-time relaxation is faster than the asymptotic exponential decay, making τ_eff < τ_eq. No equation is found to reduce to its own input by construction, and no fitted parameter is renamed as a prediction. One correctness concern exists—Eq. (23) appears to omit a (1−f_eq) factor in the asymptotic deviation, which would affect the normalization of Eqs. (36)–(40)—but this is an algebraic issue, not a circularity. Overall, the central claim is independently derived; the score reflects only minor self-citation for model setup, not load-bearing circularity.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 0 invented entities

The paper introduces no new particles, forces, or entities. Its central result rests on model choices (constant coefficients, constant density of states, step initial condition) and on standard mathematical facts (Burgers equation, Cole–Hopf transformation).

free parameters (3)
  • Diffusion coefficient D = 20 × 10^23 MeV² s^-1
    Adopted from Refs. [8,9]; not fitted here, but τeq = 4D/v² and τeff scale with D, so the numerical results depend on this input.
  • Drift coefficient v = -5 × 10^23 MeV s^-1
    Adopted from Refs. [8,9]; sets the equilibrium temperature Teq = -D/v and the asymptotic relaxation time τeq.
  • Fermi energy εF = 37 MeV
    Model parameter for a spherical nucleus; determines the initial step distribution and enters the asymptotic prefactors.
axioms (5)
  • domain assumption Landau–Vlasov kinetic equation with the diffusion approximation and constant kinetic coefficients describes dissipation in a Fermi system
    Introduction and Sec. II: the calculation starts from Eq. (1), assuming the diffusion approximation is valid for nuclear Fermi systems.
  • domain assumption Constant single-particle density of states, g(ε) ≈ g(εF)
    Sec. IV: used to replace momentum integrals by energy integrals and to preserve the structure of the diffusion equation; cancels in normalized ratios but is assumed throughout.
  • domain assumption Low-temperature limit Teq ≪ εF (4 MeV ≪ 37 MeV)
    Secs. III–IV: used to extend integration limits to −∞ and to approximate integrals; the ratio is about 9, which is plausible but not quantified.
  • domain assumption Infinite nuclear matter, spherical symmetry, no surface effects
    Introduction: the distribution function is taken independent of spatial coordinates and spherically symmetric in momentum space.
  • ad hoc to paper Initial state is a Heaviside step function f0(ε) = θ(εF − ε)
    Sec. II, Eq. (13): chosen for comparison with earlier work; τeff is known to be sensitive to the initial state, so this choice is not generic.

pith-pipeline@v1.3.0-alltime-deepseek · 9291 in / 23971 out tokens · 244839 ms · 2026-08-01T16:51:57.652032+00:00 · methodology

0 comments
read the original abstract

The dissipative properties of a Fermi system are studied within the diffusion approximation of kinetic theory for a model of a spherical atomic nucleus. An analytical solution of the nonlinear diffusion equation in energy space with constant kinetic coefficients is used to show that the distribution function asymptotically approaches the equilibrium Fermi distribution. It is found that the deviation from equilibrium at finite times decays with an effective relaxation time of $\tau_\mathrm{eff}\approx 1.0\times10^{-23}$ s, whereas the asymptotic regime is characterized by an exponential decay with a relaxation time of $\tau_\mathrm{eq}\approx 3.2\times10^{-23}$ s. These results explain the difference between the relaxation times extracted from integral characteristics of the relaxation process and from the asymptotic long-time evolution.

Figures

Figures reproduced from arXiv: 2607.17837 by Sergiy V. Lukyanov.

Figure 1
Figure 1. Figure 1: shows the results of numerical calculations for the difference δf(ǫ, t) between the exact distribution function (14) and the equilibrium Fermi distribution (21) as a function of the dimensionless energy ǫ/ǫF at different times. -0.5 -0.3 -0.1 0.1 0.3 0.5 0.4 0.6 0.8 1 1.2 1.4 1.6 δf (ϵ,t) ϵ /ϵF 0.00 0.01 0.10 1.00 10.0 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Time dependence of the ratio ∆(t)/∆0. The solid curve represents the numerical result, while the dashed curve corresponds to an exponential decay with the relaxation time τ = τeff ≈ 1.0 × 10−23 s. yields the effective relaxation time characterizing the dis￾sipation process. As seen in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: , the asymptotic expression (40) lies outside its range of validity, at least over the time interval t < 5.6 × 10−23 s, since it predicts values exceeding 1. As time increases, the asymptotic expression decays signif￾icantly more slowly than the exact result. As pointed out in Ref. [11], this difference reflects the distinct physi￾cal meaning of the two time scales: τeff characterizes the relaxation of the… view at source ↗
Figure 4
Figure 4. Figure 4: Time dependence of the quantity τ . The calculation parameters and curve styles are the same as those used in the caption of [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

27 extracted references · 6 linked inside Pith

  1. [1]

    Under certain simplifying assumptions, however, it admits an exact analytical solution [ 8, 9]

    can only be solved numerically [5]. Under certain simplifying assumptions, however, it admits an exact analytical solution [ 8, 9]. It is assumed that the kinetic coefficients are energy independent and remain constant, D = const and v = const. Under this assumption, they can be taken outside the differential operators, and Eq. (

  2. [2]

    (2) The following analysis is restricted to the limit of con- stant kinetic coefficients

    reduces to ∂f ∂t = −v ∂ ∂ǫ [f (1 − f )] + D ∂2f ∂ǫ2 . (2) The following analysis is restricted to the limit of con- stant kinetic coefficients. The analytical solution of this nonlinear diffusion equa- tion is well known [ 5, 8, 9, 12]. For completeness, one pos- sible derivation is briefly outlined below. After expanding the nonlinear term and introducing th...

  3. [3]

    (4) The Cole–Hopf transformation [ 13, 14] is then intro- duced to linearize this equation w(ǫ, t) = − 2D φ(ǫ, t) ∂φ(ǫ, t) ∂ǫ

    reduces to the classical Burgers equation [ 12], ∂w ∂t + w ∂w ∂ǫ = D ∂2w ∂ǫ2 . (4) The Cole–Hopf transformation [ 13, 14] is then intro- duced to linearize this equation w(ǫ, t) = − 2D φ(ǫ, t) ∂φ(ǫ, t) ∂ǫ . (5) Substituting Eq. (5) into Eq. ( 4) yields the heat equation, ∂φ ∂t = D ∂2φ ∂ǫ2 . (6) Let the initial condition be f (ǫ, 0) = f0(ǫ). (7) From Eqs. ...

  4. [4]

    V. M. Kolomietz and S. V. Lukyanov, Diffuse approxima- tion to the kinetic theory in a Fermi system, Int. Journ. Mod. Phys. E 24, 1550023 (2015) , arXiv:1504.00216 [nu- cl-th]

    Pith/arXiv arXiv 2015

  5. [5]

    Risken, The Fokker–Planck Equation: Methods of So- lution and Applications , 2nd ed., Springer Series in Syn- ergetics, Vol

    H. Risken, The Fokker–Planck Equation: Methods of So- lution and Applications , 2nd ed., Springer Series in Syn- ergetics, Vol. 18 (Springer, Berlin, 1989)

    1989

  6. [6]

    S. Lukyanov, Relaxation of particle–hole-type excitation in a Fermi system within the diffusion approximation of kinetic theory for the case of constant diffusion and drift coefficients, Acta Phys. Pol. B. 56, 1 −A2 (2025) , arXiv:2410.10492 [nucl-th]

    Pith/arXiv arXiv 2025

  7. [7]

    S. V. Lukyanov, Equilibrium state of a Fermi sys- tem in the diffusion approximation of kinetic theory, arXiv:2606.06321 [nucl-th] (2026)

    Pith/arXiv arXiv 2026

  8. [8]

    (36) Introducing the dimensionless variable x as before and changing the integration variable yields ∆2(t) = g(ǫF) T 3 eq πD e−2t/τeq t ∫ ∞ −ǫF/Teq [ x 1 + ex ] 2 dx

    into the definition ( 25) yields the following expression for the squared root-mean-square deviation: ∆2(t) = g(ǫF) πD e−2t/τeq t ∫ ∞ 0 [ ǫ − ǫF 1 + e(ǫ−ǫF)/Teq ] 2 dǫ. (36) Introducing the dimensionless variable x as before and changing the integration variable yields ∆2(t) = g(ǫF) T 3 eq πD e−2t/τeq t ∫ ∞ −ǫF/Teq [ x 1 + ex ] 2 dx. (37) The integral in E...

  9. [9]

    Bartsch and G

    T. Bartsch and G. Wolschin, Equilibration in fermionic systems, Ann. Phys. 400, 21 (2019) , arXiv:1806.04044 [cond-mat.stat-mech]

    Pith/arXiv arXiv 2019

  10. [10]

    ( 41) yields τ (t) = τeq 1 − τeq t ( ln(B) − 1 2 ln(t) )

    and absorbing all re- maining factors into the constant B, substitution into Eq. ( 41) yields τ (t) = τeq 1 − τeq t ( ln(B) − 1 2 ln(t) ) . Hence, in the limit t → ∞, τ (t) ≈ τeq + τ 2 eq ln(B) − 1 2 ln(t) t + O ( ln2(t) t2 ) . Thus, τ (t) indeed approaches τeq, but only slowly, with a correction of order ln( t)/t. Thus, in the limit t → ∞ , both the exac...

  11. [11]

    f (ǫ, t) = { 1 + erf ( ǫF − ǫ − vt√ 4Dt )} × { 1 + erf ( ǫF − ǫ − vt√ 4Dt ) + exp [ v(ǫF − ǫ) D ] erfc ( ǫF − ǫ + vt√ 4Dt ) } −1

    can be expressed in terms of error functions [ 8, 9]. f (ǫ, t) = { 1 + erf ( ǫF − ǫ − vt√ 4Dt )} × { 1 + erf ( ǫF − ǫ − vt√ 4Dt ) + exp [ v(ǫF − ǫ) D ] erfc ( ǫF − ǫ + vt√ 4Dt ) } −1 . (14) 3 III. ASYMPTOTIC BEHA VIOR OF THE SOLUTION Using the odd symmetry of the error function, Eq. ( 14) can be rewritten as f (ǫ, t) = R(t) R(t) + exp ( ǫ−ǫF Teq ) , (15) ...

  12. [12]

    In addition, the short-dashed horizontal line corresponds to the exponential dependence with the re- laxation time τ = τeq ≈ 3.2 × 10−23 s. solution ( 40) evaluated with the relaxation time set to τeq = τeff ≈ 1.0 × 10−23 s (dash-dotted curve) gradually approaches the exponential dependence with the same relaxation time (dashed curve) as time increases. Th...

  13. [13]

    As seen in the figure, the initial devi- ation is localized primarily in the vicinity of the Fermi surface and has opposite signs below and above ǫF

    and the equilibrium Fermi distribution ( 21). As seen in the figure, the initial devi- ation is localized primarily in the vicinity of the Fermi surface and has opposite signs below and above ǫF. With increasing time, the deviation decreases monotonically and becomes practically negligible for t > 10 × 10−23 s. IV. EFFECTIVE RELAXATION TIME A direct analys...

  14. [14]

    V. M. Kolomietz and S. Shlomo, Mean Field Theory (World Scientific, Singapore, 2020)

    2020

  15. [15]

    E. M. Lifshitz and L. P. Pitaevskii, Physical kinetics (Pergamon Press, Oxford, 1981) Chap. 2

    1981

  16. [16]

    V. M. Kolomietz and S. V. Lukyanov, Diffusion on the distorted Fermi surface, Ukr. J. Phys. 59, 764 (2014) , arXiv:1409.1391 [nucl-th]

    Pith/arXiv arXiv 2014

  17. [19]

    Wolschin, Equilibration in finite fermion systems, Phys

    G. Wolschin, Equilibration in finite fermion systems, Phys. Rev. Lett. 48, 1004 (1982)

    1982

  18. [21]

    S. V. Lukyanov, Properties of the diffusion and drift ki- netic coefficients in momentum space for a cold Fermi system, Nucl. Phys. and At. Energy 24, 5 (2023) , arXiv:2210.15299 [nucl-th]

    Pith/arXiv arXiv 2023

  19. [22]

    S. V. Lukyanov, Relaxation of a single-particle excita- tion in a fermi system within the diffusion approxima- tion of kinetic theory, Phys. Rev. C 113, 034312 (2026) , arXiv:2511.19689 [nucl-th]

    arXiv 2026

  20. [23]

    A more transparent characterization is obtained by introducing the effective relaxation time defined in Ref

    is neither convenient nor particularly informative. A more transparent characterization is obtained by introducing the effective relaxation time defined in Ref. [ 6]: τeff = ∫ ∞ 0 ∆(t) ∆0 dt, (24) where ∆(t) = √ ∫ [δf (p, t)]2 dp (25) is the root-mean-square deviation, and ∆(t = 0) ≡ ∆0 = √ ∫ [δfin(p)]2 dp, (26) is its initial value. The definition ( 24) invo...

  21. [24]

    J. M. Burgers, The Nonlinear Diffusion Equation (Rei- del, D., Dordrecht, 1974)

    1974

  22. [25]

    Hopf, The partial differential equation ut + uux = µuxx, Comm

    E. Hopf, The partial differential equation ut + uux = µuxx, Comm. Pure Appl. Math. 3, 201 (1950)

    1950

  23. [26]

    J. D. Cole, On a quasi-linear parabolic equation occurring in aerodynamics, Quart. Appl. Math. 9, 225 (1951)

    1951

  24. [27]

    L. C. Evans, Partial differential equations , 2nd ed., Grad- uate Studies in Mathematics, Vol. 19 (American Mathe- matical Society, Providence, RI, 2010)

    2010

  25. [28]

    Abramowitz and I

    M. Abramowitz and I. A. Stegun, Handbook of Mathe- matical Functions (Dover, New York, 1972)

    1972

  26. [34]

    ( 30) gives the final expression for the squared initial root-mean- square deviation ∆2 0 = (ln(4) − 1) g(ǫF)Teq

    into Eq. ( 30) gives the final expression for the squared initial root-mean- square deviation ∆2 0 = (ln(4) − 1) g(ǫF)Teq. (35) The quantity ∆( t) is evaluated numerically. To this end, the definition of the deviation from equilibrium, Eq. ( 22), together with the expressions for the equilib- rium Fermi distribution, Eq. ( 21), and the exact distribu- tion ...

  27. [40]

    The corresponding dependence is shown by the dash-dotted curve

    was performed, in which the relax- ation time was set to τeq = τeff ≈ 1.0 × 10−23 s instead of 3.2 × 10−23 s. The corresponding dependence is shown by the dash-dotted curve. It lies considerably closer to the exact result shown in Fig. 2, but still exceeds the physically admissible range for t < 2.3 × 10−23 s. At asymptotically large times, however, this c...