Review of the foundations of time-dependent density-functional theory (TDDFT)

Jochen Schirmer

REVIEW 2 major objections 4 minor 46 references

Review of the foundations of time-dependent density-functional theory (TDDFT)

T0 review · 2 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Time-dependent density-functional theory, as an exact density-based theory, currently has no valid foundation, this review argues: the Runge-Gross action functional is trivial, the fixed-point route lacks a convergence guarantee, and…

desk verdict A clear synthesis of the author's longstanding attack on TDDFT foundations, with one airtight algebraic point and a concluding claim that outruns the evidence. read the letter →

arxiv 2411.16607 v1 pith:PVIBQZPG submitted 2024-11-25 physics.chem-ph

Jochen Schirmer This is my paper · ORCID

classification physics.chem-ph

keywords time-dependentdensity-functionaltheoryTDDFTfoundationsRunge-Grosstheoremaction-integralfunctionalKohn-Shamequationsfixed-pointiterationon-the-flypropagationrandom-phaseapproximation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Time-dependent density-functional theory (TDDFT) is usually presented as an in-principle exact way to follow many-electron time evolution and compute excitations from densities alone. This review argues that the foundations offered for that claim do not hold up. The original Runge-Gross action-integral functional collapses to a trivial expression that cannot support a stationarity principle and a time-dependent equation of motion. The replacement route, built on two Runge-Gross-type mappings, yields time-dependent Kohn-Sham equations only as an ad hoc fixed-point iteration with no guaranteed convergence. The paper concludes that TDDFT should be seen as a pragmatic, empirically tuned modification of the random-phase approximation rather than as a formally exact theory.

What carries the argument

The load-bearing object is the Runge-Gross (RG1) mapping, which assigns to a time-dependent density $n(\mathbf{r},t)$ an external potential $v_{\mathrm{ext}}[n](\mathbf{r},t)$ whose Schr\"odinger evolution reproduces that density. The analysis turns on two constructs built from that mapping: the action-integral functional $A[n]$ of Eq. (78), which reduces to Eq. (82) and thereby loses all dynamical content; and the radical Kohn-Sham (rKS) scheme, a single-orbital model in which the potential functional is constructed explicitly as $w[n] = \frac{\nabla^2 \phi}{2\phi} - \frac{1}{2}(\nabla k)^2 - \dot{k} - \dot{\alpha}(t)$, with $\phi = (n/N)^{1/2}$ and $k$ the orbital phase, exposing the dependence on $n$, $\dot n$, and $\ddot n$. This rKS construction is what makes the fixed-point iteration and the on-the-fly objections concrete rather than merely verbal.

What would settle it

Solve a small two-electron system exactly, read off its time-dependent density, and reconstruct a local potential that reproduces that density; if the reconstructed potential can be expressed using only the density and its first time derivative, the paper's second-order-propagation objection loses its force. A second check is to run the proposed fixed-point iteration on the same model and see whether it converges.

Watch

Extended reading notes

Core claim

The central claim, stated plainly, is that none of the existing justifications for exact TDDFT survives close inspection. Substituting the Runge-Gross mapping into the action integral removes the time derivative and leaves $A[n] = \int_{t_0}^{t_1} dt \int d\mathbf{r}\, (v_{\mathrm{ext}}[n](\mathbf{r},t) - u(\mathbf{r},t))\, n(\mathbf{r},t)$, so the proposed stationarity principle cannot yield a time-dependent equation of motion. The later mapping-based derivation of the time-dependent Kohn-Sham equations is shown to be a fixed-point iteration whose convergence is guaranteed by nothing, since it is not backed by any variational or stationarity principle. By constructing the single-orbital radical Kohn-Sham mapping explicitly, the paper shows the exact Kohn-Sham potential functional depends on $n(t)$, $\dot n(t)$, and $\ddot n(t)$, which turns exact on-the-fly propagation into an implicit second-order problem. From these three failures the author concludes that there is currently no valid justification for TDDFT as an exact theory.

Load-bearing premise

The load-bearing premise is that a valid density-based theory must supply a variational or stationarity principle or a convergent constructive scheme—and, as a supporting assumption, that the second-order dependence found in the single-orbital radical Kohn-Sham potential carries over to the many-orbital case.

Editorial extensions

If this is right

If the Runge-Gross action functional is trivial as claimed, the original 1984 foundation cannot be repaired by adjusting that functional; no stationarity principle for a density equation of motion exists along that route.
If exact time-dependent Kohn-Sham equations are only a fixed-point iteration without a convergence guarantee, then a purely mapping-based derivation does not qualify as a rigorous foundation for TDDFT.
If the exact Kohn-Sham potential depends on $n$, $\dot n$, and $\ddot n$, then exact TDDFT cannot be propagated by on-the-fly first-order stepping; only the adiabatic approximation makes that possible, and it is then an uncontrolled approximation.
Linear-response TDDFT would lose its status as a controlled approximation to an exact density-based theory and would instead be an empirical RPA-like scheme for singly excited states.
The burden of proof for any future exact density-based time-evolution theory lies with its proponents; the present arguments set a standard that a successor theory must meet.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: the argument's standard of proof—variational principle or convergent constructive scheme—is what does the work; a reader who accepts a non-constructive existence theorem for the density-to-potential mapping as foundational would not be forced to the paper's conclusion.
Editorial inference: the explicit $n$, $\dot n$, $\ddot n$ dependence is demonstrated for the single-orbital radical Kohn-Sham potential and only expected to carry over to the many-orbital Kohn-Sham potential; settling that generalization is a natural next step.
Editorial inference: the review's negative conclusion concerns formal exactness, not computational utility; even if correct, it would leave LR-TDDFT and adiabatic time propagation as useful approximations whose empirical success is not evidence for the exact theory.
Editorial inference: a different variational construction—for instance one with proper boundary conditions or a coupled density-potential action—might still produce a rigorous TDDFT-like theory, so the paper reads as a challenge to construct one rather than a proof of impossibility.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Jochen — quick read of Schirmer's TDDFT foundations review. My take: the paper is worth taking seriously, but the headline conclusion ('currently no valid justification for TDDFT') is not supported by the arguments as written. The load-bearing algebraic point is Eq. (82): given the RG1 wavefunction satisfies the td Schrödinger equation, the Runge-Gross action collapses to a potential-difference integral, so stationarity can't produce an equation of motion. That is clean, and it deserves acknowledgment. The rKS example in Sec. IV.C is also a useful pedagogical device, showing explicitly how a density-to-potential mapping can be constructed while remaining trivial in the sense that it just reproduces whatever density you feed in.

What's actually new is mainly organization: the paper collects Schirmer and Dreuw 2007, the 2012 two-pager, and the Reply into one coherent narrative, and it makes the FPI/OTF distinction explicit. That is useful for people who want the critique in one place. But it is a synthesis, not new mathematics.

Soft spots, in order of softness. First, the many-orbital extrapolation: the claim that the exact KS potential depends on n, n-dot, n-double-dot at time t is proven only for the single-orbital rKS scheme (Eq. 95). The transfer to the real KS system is stated as 'must be expected.' The OTF impossibility argument rests on that transfer, so as written it is an unsupported extrapolation. The stress-test note got this right. Second, the paper's standard for a valid foundation—either a stationarity principle or a guaranteed-convergent fixed-point scheme—is imposed, not argued for. Non-constructive existence proofs like van Leeuwen's mapping are dismissed by implication, but not refuted in detail. If one accepts non-constructive existence as a foundation, the conclusion doesn't follow. Third, the paper relies heavily on the author's own earlier papers for both the key equations and the standard of justification. Self-citation isn't a flaw when results are correct, but here it means the review is partly an extended restatement of a prior dispute. Fourth, the LR-TDDFT critique repeats the familiar RPA/adiabatic concerns; it's fair but not novel.

Verdict: the specific criticism of the RG AIF is strong and should be engaged by anyone defending TDDFT foundations. The global conclusion, though, overreaches. A referee should ask for either a much more careful statement of what counts as justification, or a real treatment of the many-orbital case. I'd send it to review—it's an important challenge from a knowledgeable source, and the algebraic core is checkable. I just wouldn't let the abstract's last sentence stand as is.

Referee Report

2 major / 4 minor

Summary. The paper is a critical review of the foundational claims of time-dependent density-functional theory (TDDFT). It argues that the original Runge-Gross action-integral functional is a trivial construct because the wavefunction used in it already satisfies the time-dependent Schrödinger equation with a potential that reproduces the density, reducing the action to a potential difference (Eq. 82). It further argues that the mapping-based route to the time-dependent Kohn-Sham equations, understood as a fixed-point iteration, lacks any convergence guarantee because it is not backed by a variational or stationarity principle (Sec. IV.D). Finally, it claims that on-the-fly propagation is impossible for the exact theory because the KS potential functional would need to depend on n(t), n-dot(t), and n-ddot(t), making the equations implicit second-order differential equations (Sec. IV.C). The paper concludes that there is currently no valid justification for TDDFT and that LR-TDDFT should be regarded as a pragmatic modification of RPA rather than an exact density-based method.

Significance. The paper addresses a foundational question that is central to a widely used electronic-structure method. The critique of the Runge-Gross action functional is mathematically clean and the explicit construction of the radical Kohn-Sham scheme in Sec. IV.C usefully exposes properties of one-orbital TDDFT. The paper also consolidates the author's prior arguments into a single self-contained account. However, the overarching conclusion that 'there is currently no valid justification for TDDFT' is stronger than what the presented arguments support: the many-orbital KS potential dependence on n, n-dot, and n-ddot is only asserted as an expectation, and the epistemic standard used to rule out non-constructive existence proofs is not itself justified. Thus the paper is a thought-provoking critical review rather than a definitive disproof.

major comments (2)

[Sec. IV.C (Eq. 95) and Sec. IV.D] The OTF-impossibility argument, which is a central pillar of the conclusion in Sec. V, relies on the statement that the derivative dependence found for the single-orbital rKS potential 'must be expected as well in the KS potential functional vKS[n(t)]' in the many-orbital case. No proof, model calculation, or known result is provided for the many-orbital system. Since the assertion is load-bearing for the claim that the td KS equations are implicit second-order differential equations that cannot be propagated in first-order fashion, the paper should either supply a derivation for a non-interacting many-particle system or clearly mark this step as a conjecture and weaken the conclusions in Sec. IV.D and Sec. V accordingly.
[Sec. I and Sec. IV.D] The paper's standard for a 'valid foundation' is that it must come with a variational or stationarity principle, or at least a constructive fixed-point scheme with a proven convergence guarantee. This criterion is asserted rather than argued, and it is used to dismiss mapping-based existence proofs such as the van Leeuwen type construction cited earlier. A reader who accepts non-constructive existence proofs as a legitimate foundation will not be convinced by the paper's negative conclusion. To make the conclusion rigorous, the authors need to defend this epistemic criterion, or else restate the conclusion in a more limited form such as 'TDDFT lacks a constructive variational foundation with a proven convergence guarantee.'

minor comments (4)

[Abstract and Sec. V] The phrase 'expectations of finding a remedy here are hardly justified' is a value judgment that goes beyond the technical analysis; consider replacing it with a more neutral statement of the scope of the conclusions.
[Sec. II.C] The statement 'There are no double excitations' in the context of LR-TDDFT is too categorical; the literature contains discussions of double excitations in TDDFT with frequency-dependent kernels, and the paper should either cite those exceptions or qualify the claim as applying to the adiabatic RPA-type formulation presented.
[Sec. IV.C] The notation in Eq. (95), specifically the term with the time derivative of the phase k[n], is not introduced with sufficient precision; please define the total time derivative and the domain of the functional dependence explicitly.
[General] The paper draws heavily on the author's earlier publications (Refs. 17, 19, 20) and the exposition would benefit from a clearer statement of what new insight the present review adds beyond those references, beyond being a synthesis.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the critique is self-contained algebra; self-citations are accompanied by derivations.

full rationale

The paper is a critical review, not a predictive scheme, so the usual circularity patterns (fitted inputs called predictions, self-definitional derivations) do not arise. The central claim that the Runge-Gross action functional is trivial follows from an explicit algebraic evaluation: because the RG1 construction defines Psi[n] as the solution of the Schroedinger equation with potential vext[n], substituting Eq. (74) into Eq. (78) reduces A[n] to an integral over the potential difference; this is a direct consequence of the defining property, not a circular use of the conclusion. The fixed-point iteration critique is a convergence concern, not a reduction to inputs: the paper proves that n0 is a fixed point by construction and then argues that no convergence guarantee exists, which is an epistemic claim rather than a circular derivation. The rKS analysis in Sec. IV.C explicitly constructs w[n] and shows n, ndot, nddot dependence for the single-orbital case; the extension to the many-orbital KS potential is explicitly flagged as an expectation ('This feature must be expected as well'), so it is an unsupported extrapolation and a correctness risk, not a circular step. The paper does cite the author's prior work (Refs. 17, 19, 20), but the load-bearing arguments are reproduced in the text (e.g., Eq. 82 and the rKS construction), so the self-citations are not load-bearing in the logical derivation. The author-imposed standard for a 'valid foundation' is a philosophical choice, not a fitted parameter or a hidden input. Overall, no circularity is exhibited.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The ledger contains no numerical fits and no new physical entities. The paper depends on background mapping theorems from earlier literature and, critically, on an author-imposed epistemic standard about what should count as a valid foundation. The fourth entry is an extrapolation that is load-bearing for the on-the-fly propagation criticism.

assumptions (4)

domain assumption RG1 mapping: for a given time-dependent density n(r,t) and initial state, there exists a local time-dependent external potential vext[n] whose Schroedinger evolution reproduces n(r,t), unique up to c(t) (Ref. 16).
Used to evaluate the action A[n] in Eq. (78) and to obtain the trivial form in Eq. (82). The paper critiques consequences of this mapping but does not prove it.
domain assumption Non-interacting RG1 mapping: for a given time-dependent density there is a Kohn-Sham potential vKS[n] such that non-interacting orbitals reproduce n(r,t) (Refs. 41, 45).
Underpins the fixed-point rewriting of the td KS equations in Eqs. (99)-(105).
ad hoc to paper A valid foundation must come with a variational or stationarity principle, or at least a constructive fixed-point scheme with a proven convergence guarantee.
This standard is imported by the author in Section I and Section IV.D. It is not demonstrated and is rejected by several TDDFT foundational works that accept non-constructive existence mappings.
ad hoc to paper The functional-dependence structure of the exact many-orbital KS potential mirrors the single-orbital rKS case, requiring n, ṅ, and n̈.
Derived explicitly for the one-orbital model in Eq. (95), then extrapolated to the many-orbital KS potential in Section IV.C with the phrase 'must be expected', not by proof.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Review of the foundations of time-dependent density-functional theory (TDDFT)." pith.science (2026). https://pith.science/paper/PVIBQZPG

@misc{pith2026241116607,
  author       = {Pith},
  title        = {Pith review of: Review of the foundations of time-dependent density-functional theory (TDDFT)},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PVIBQZPG}},
  note         = {Machine review of arXiv:2411.16607}
}

read the original abstract

Time-dependent density-functional theory (TDDFT) is deemed to be a formally rigorous way of dealing with the time-evolution of a many-electron system at the level of electron densities rather than the underlying wavefunctions, which in turn provides an, in principle, exact density-based approach to the treatment of electron excitations in atoms and molecules. This claim has not remained unchallenged, and a detailed account of the relevant criticism is given in this paper. In view of our analysis one has to face the conclusion that there is currently no valid justification for TDDFT, and expectations of finding a remedy here are hardly justified.

Figures

Figures reproduced from arXiv: 2411.16607 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 43 canonical work pages

[1]

Burke, J

K. Burke, J. Werschnik, and E. K. U. Gross, J. Chem. Phys. 123, 062206 (2005)

work page 2005
[2]

Elliott, K

P. Elliott, K. Burke, and F. Furche, Rev. Comp. Chem. 26, 91 (2008)

work page 2008
[3]

M. E. Casida, J. Mol. Struct.: THEOCHEM 914, 3 (2009)

work page 2009
[4]

E. K. U. Gross and N. T. Maitra, in Fundamentals of Time-dependent Density Functional Theory, edited by M. A. L. Marques, N. T. Maitra, F. M. S. Nogueira, and E. K. U. Gross (Springer, Heidelberg, 2012)

work page 2012
[5]

Ruggenthaler, M

M. Ruggenthaler, M. Penz, and R. van Leeuwen, J. Phys.: Condens. Matter 27, 203202 (2015)

work page 2015
[6]

N. T. Maitra, J. Chem. Phys. 144, 220901 (2016)

work page 2016
[7]

Huix-Rotllant, N

M. Huix-Rotllant, N. Ferr´ e, and M. Barbatti, inQuantum Chemistry and Dynamics of Excited States: Methods and Applications , edited by L. Gonz´ ales and R. Lindh (Wiley, Hoboken, New Jersey, 2021)

work page 2021
[8]

Engel and R

E. Engel and R. Dreizler, Density Functional Theory (Springer, Berlin Heidelberg, 2011)

work page 2011

Show all 46 references

[9]

C. A. Ullrich, Time-Dependent Density-Functional Theory (Oxford University Press, Oxford, 2012)

2012
[10]

M. R. Silva-Junior, M. Schreiber, S. P. A. Sauer, and W. Thiel, J. Chem. Phys. 129, 104103 (2008)

2008
[11]

Jacquemin, V

D. Jacquemin, V. Wathelet, E. A. Perp´ ete, and C. Adamo, J. Chem. Theory Comput.5, 2420 (2009)

2009
[12]

Caricato, G

M. Caricato, G. W. Trucks, M. J. Frisch, and K. B. Wiberg, J. Chem. Theory Comput. 6, 370 (2010)

2010
[13]

Hohenberg and W

P. Hohenberg and W. Kohn, Phys. Rev. 136, B 864 (1964)

1964
[14]

Kohn and L

W. Kohn and L. J. Sham, Phys. Rev. 140, A 1133 (1965)

1965
[15]

Zangwil and P

A. Zangwil and P. Soven, Phys. Rev. A 21, 1561 (1980)

1980
[16]

Runge and E

E. Runge and E. K. U. Gross, Phys. Rev. Lett. 52, 997 (1984)

1984
[17]

Schirmer and A

J. Schirmer and A. Dreuw, Phys. Rev. A 75, 022513 (2007)

2007
[18]

N. T. Maitra, R. van Leeuwen, and K. Burke, Phys. Rev. A 78, 056501 (2008)

2008
[19]

Schirmer and A

J. Schirmer and A. Dreuw, Phys. Rev. A 78, 056502 (2008)

2008
[20]

Schirmer, Phys

J. Schirmer, Phys. Rev. A 86, 012514 (2012)

2012
[21]

Schirmer, M

J. Schirmer, M. Braunstein, M.-T. Lee, and V. McKoy, in VUV and Soft X-Ray Photoioniza- 36 tion, edited by U. Becker and D. A. Shirley (Plenum Press, New York, 2001), p. 105

2001
[22]

Schirmer, Many-Body Methods for Atoms, Molecules and Clusters (Springer, Heidelberg, 2018)

J. Schirmer, Many-Body Methods for Atoms, Molecules and Clusters (Springer, Heidelberg, 2018)

2018
[23]

A. D. McLachlan and M. A. Ball, Rev. Mod. Phys. 36, 844 (1964)

1964
[24]

T. H. Dunning and V. McKoy, J. Chem. Phys. 47, 1735 (1967)

1967
[25]

D. J. Rowe, Rev. Mod. Phys. 40, 153 (1968)

1968
[26]

Ring and P

P. Ring and P. Schuck, The Nuclear Many-Body Problem (Springer, Heidelberg, 2000)

2000
[27]

M. E. Casida, in Recent Advances in Density Functional Theory, Part I , edited by D. P. Chong (World Scientific, Singapore, 1995)

1995
[28]

Bauernschmitt and R

R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett. 256, 454 (1996)

1996
[29]

M. E. Casida, C. Jamorski, K. C. Casida, and D. R. Salahub, J. Chem. Phys. 108, 4439 (1998)

1998
[30]

D. J. Tozer and N. C. Handy, J. Chem. Phys. 109, 10180 (1998)

1998
[31]

D. J. Tozer, R. D. Amos, N. C. Handy, B. Roos, and L. Serrano-Andr´ es, Mol. Phys. 97, 859 (1999)

1999
[32]

Dreuw, J

A. Dreuw, J. L. Weisman, and M. Head-Gordon, J. Chem. Phys. 119, 2943 (2003)

2003
[33]

Dreuw and M

A. Dreuw and M. Head-Gordon, J. Am. Chem. Soc. 126, 4007 (2004)

2004
[34]

A. L. Sobolewski and W. Domcke, Chem. Phys 294, 73 (2003)

2003
[35]

R. G. Parr and W. Yang, Density-Functional Theory of Atoms and Molecules (Oxford Uni- versity Press - Clarendon Press, New York, Oxford, 1989)

1989
[36]

Nagy, Phys.Rep

A. Nagy, Phys.Rep. 298, 1 (1998)

1998
[37]

M. Levy, J. P. Perdew, and V. Sahni, Phys. Rev. A 30, 2745 (1984)

1984
[38]

Holas and N

A. Holas and N. H. March, Phys. Rev. A 44, 5521 (1991)

1991
[39]

E. K. U. Gross, J. F. Dobson, and M. Petersilka, in Density Functional Theory , edited by R. F. Nalewajski (Springer, New York, 1996)

1996
[40]

A. K. Rajagopal, Phys. Rev. A 54, 3916 (1996)

1996
[41]

van Leeuwen, Phys

R. van Leeuwen, Phys. Rev. Lett. 80, 1280 (1998)

1998
[42]

M. K. Harbola and A. Banerjee, Phys. Rev. A 60, 5101 (1999)

1999
[43]

van Leeuwen, Int

R. van Leeuwen, Int. J. Mod. Phys. B 14, 1969 (2001)

2001
[44]

Vignale, Phys

G. Vignale, Phys. Rev. A 77, 062511 (2008)

2008
[45]

van Leeuwen, Phys

R. van Leeuwen, Phys. Rev. Lett. 82, 3863 (1999). 37

1999
[46]

M. A. L. Marques and E. K. U. Gross, Annu. Rev. Phys. Chem. 55, 427 (2004). 38

2004

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Burke, J

K. Burke, J. Werschnik, and E. K. U. Gross, J. Chem. Phys. 123, 062206 (2005)

work page 2005

[2] [2]

Elliott, K

P. Elliott, K. Burke, and F. Furche, Rev. Comp. Chem. 26, 91 (2008)

work page 2008

[3] [3]

M. E. Casida, J. Mol. Struct.: THEOCHEM 914, 3 (2009)

work page 2009

[4] [4]

E. K. U. Gross and N. T. Maitra, in Fundamentals of Time-dependent Density Functional Theory, edited by M. A. L. Marques, N. T. Maitra, F. M. S. Nogueira, and E. K. U. Gross (Springer, Heidelberg, 2012)

work page 2012

[5] [5]

Ruggenthaler, M

M. Ruggenthaler, M. Penz, and R. van Leeuwen, J. Phys.: Condens. Matter 27, 203202 (2015)

work page 2015

[6] [6]

N. T. Maitra, J. Chem. Phys. 144, 220901 (2016)

work page 2016

[7] [7]

Huix-Rotllant, N

M. Huix-Rotllant, N. Ferr´ e, and M. Barbatti, inQuantum Chemistry and Dynamics of Excited States: Methods and Applications , edited by L. Gonz´ ales and R. Lindh (Wiley, Hoboken, New Jersey, 2021)

work page 2021

[8] [8]

Engel and R

E. Engel and R. Dreizler, Density Functional Theory (Springer, Berlin Heidelberg, 2011)

work page 2011