Quantum Systems at The Brink. Existence and Decay Rates of Bound States at Thresholds; Atoms

Dirk Hundertmark; Markus Lange; Michal Jex

REVIEW 2 major objections 8 minor 12 references

Quantum Systems at The Brink. Existence and Decay Rates of Bound States at Thresholds; Atoms

T0 review · 2 major / 8 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read At the critical nuclear charge $Z_c$, every ground state of an $N$-electron atom decays at least as fast as $\exp(-C\sum\sqrt{|x_k|})$ over the $K$ outermost electrons, with no spectral gap or Born-Oppenheimer approximation.

desk verdict New threshold decay result for atoms that is probably right, but Theorem 2.3 is missing the hypotheses its own proof needs; fixable, worth refereeing. read the letter →

arxiv 1908.05016 v1 pith:SYKRE43K submitted 2019-08-14 math-ph math.MP

Dirk Hundertmark , Michal Jex , Markus Lange This is my paper · ORCID

classification math-phmath.MP MSC 81Q1081V45

keywords thresholdboundstatescriticalnuclearchargeexponentialdecayofeigenfunctionsCoulombmany-bodysystemsSchrödingeroperatorsnospectralgapBorn-Oppenheimerapproximationtightness

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

An atom with $N$ electrons stops binding when the nuclear charge is lowered to a critical value $Z_c$; at exactly that coupling the ground-state energy lies at the edge of the essential spectrum. This paper claims that the ground state still decays at least as fast as $\exp(-C\sum_{k=1}^{K}\sqrt{|x_k|})$, where the sum runs over the $K$ largest electron–nucleus distances and $K$ is the number of electrons whose removal leaves the ground-state energy unchanged, provided $Z_c

What carries the argument

The engine of the proof is a weighted $L^2$ estimate built from a piecewise-$C^1$ version of the IMS localization formula. Configuration space is split according to how many inner electrons lie within a $\delta$-fraction of the outermost distance, and an upper bound $F=\sum_{m=N+1-K}^{N}(C_m\sqrt{|x|_m}+K_m|x|_m\chi^{\perp}_{0,2\delta})$ is introduced for the weight $G$. After subtracting localization errors of the form $|\nabla F|^2$, the task reduces to proving positivity of three quadratic forms $A_1,A_2,A_3$. Positivity is driven by $Z_c<N-K$, which makes $W_2=(N-K)/(1+\delta)-Z$ positive, so the outer electrons feel net repulsion large enough to dominate $|\nabla F|^2$; the square-root decay comes from $|\nabla \sqrt{|x|}|^2\approx 1/(4|x|)$ while the repulsive term is also of order $1/|x|$.

What would settle it

For the two-electron atom at its critical charge $Z_c\approx 0.91$ (a case where a threshold ground state is known to exist), compute the ground state numerically and test whether $\int_{|x_1|,|x_2|>R}|e^{C(\sqrt{|x_1|}+\sqrt{|x_2|})}\psi(x)|^2\,dx$ stays bounded as $R\to\infty$; if it diverges for every $C>0$, Theorem 2.3 would be false for that system.

Watch

Extended reading notes

Core claim

At $Z=Z_c$, the paper's central result (Theorem 2.3) gives an $L^2$ decay bound: for the Hamiltonian $H_Z^{(N)}=\sum_{j=1}^{N}(-\Delta_j-Z/|x_j|)+\sum_{j\neq k}1/|x_j-x_k|$ on the antisymmetric subspace $L^2_a(\mathbb{R}^{3N})$, every normalized eigenfunction $\psi_Z$ with $H_Z^{(N)}\psi_Z=E_Z^{(N)}\psi_Z$ satisfies $e^G\psi_Z\in L^2_a(\mathbb{R}^{3N})$. The weight $G$ is a sum of square roots, $G=\sum_{m=N+1-K}^{N}C_m\sqrt{|x|_m}$, whenever the $K$ outermost electrons are well separated from the $N-K$ inner ones, and a linear sum of the same outer distances otherwise. Thus, although the eigenvalue sits exactly at the threshold of the essential spectrum and no gap is available, the wavefunction decays no slower than $\exp(-C\sqrt{r})$ in the outer coordinates. The proof localizes the eigenvalue equation with cutoff functions and uses the repulsive Coulomb terms to make the relevant quadratic forms positive; no Born-Oppenheimer approximation is used.

Load-bearing premise

The proof relies on $Z_c<N-K$, with $K$ the largest number of electrons whose removal leaves the ground-state energy unchanged, so that the repulsive Coulomb terms dominate the localization error; if that inequality fails, the final positivity step that produces the decay bound is not available.

Editorial extensions

If this is right

At the critical charge, the ground-state wavefunction is square-integrable against $\exp(2C\sum\sqrt{|x|_m})$, so the electron density in the $K$ outermost channels decays super-polynomially in the outer coordinates.
Because the decay estimate is uniform for every $Z>Z_c$, it combines with tightness criteria to give an alternative proof that a threshold ground state exists.
The argument does not use the fermionic sign, so the same decay bound applies to bosons and to distinguishable particles.
For a finite-mass nucleus whose mass is at least the total electron mass, the same estimate holds after straightforward modifications.
When the outer and inner electrons are not well separated, the theorem yields only linear exponential decay in the outer coordinates, reflecting stronger screening in that configuration.

Reading between the lines

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

One natural extension is to ask whether the same repulsion-driven sqrt-decay appears at thresholds set by cluster breakup rather than single-electron loss; the positivity trick may carry over to Coulomb systems with several nuclei.
The strict inequality $Z_c<N-K$ suggests a transition as $Z_c$ approaches $N-K$ from below: the constants in the decay exponent should deteriorate, possibly marking the change from a threshold bound state to no bound state. A quantitative version of that blow-up is not in the paper.
The $L^2$ decay statement could likely be upgraded to pointwise bounds via standard elliptic estimates once $e^G\psi\in L^2$ is known; the paper does not attempt this.
Since the method avoids the Born-Oppenheimer approximation, it may extend to molecular Hamiltonians with finite nuclear masses and give decay rates at autoionization thresholds, where the relevant outer coordinates are electron distances from the center of mass.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Short version: this is a genuinely new result — sqrt-exponential decay for N-electron atoms at critical charge — and the method is interesting. The main theorem as stated, however, is missing its own hypotheses. The proof outline is plausible and the flaw is fixable; this needs a serious revision, not a rejection.

What is actually new: Theorem 2.3 gives, for an atom at Z = Z_c, decay of the ground state like exp(-C Σ sqrt(|x|_k)) over the K outermost electron coordinates, with no spectral gap and without Agmon-type assumptions. Earlier threshold existence results ([3], [5]) did not provide decay rates. The trick of using the repulsive inter-electron potential to gain positivity without a gap is real and worth taking seriously. The generalized IMS localization formula for piecewise C^1 cutoffs in the appendix is also a useful piece of work.

Where it is soft: the stress-test note is correct. Theorem 2.3 says 'let H_Z^(N) be given and let ψ_Z be a normalized eigenfunction,' defines G with K, and never states Z = Z_c, never says K is the maximal integer from Theorem 2.1, never states Z < N-K. In the proof, positivity of A1 uses W2 = (N-K)/(1+δ) - Z > 0, and positivity of A2 uses E_Z^(N-K-1) - E_Z^(N) > 0. Both enter as silent assumptions. Take the theorem literally and it is false: for Z > N-K, W2 is negative, and the conclusion is not established. The fix is simple — add the standing hypotheses from Theorem 2.1 to the statement of Theorem 2.3 — and the proof outline becomes consistent.

Smaller issues: K is never defined in the theorem; key technical estimates are deferred to the companion paper [7], including the alternative existence proof; the abstract says the method needs no Born-Oppenheimer approximation, but Section 2 assumes an infinitely heavy nucleus and Remark 2.5 says finite mass requires extra work; there are typos (Lemma 3.1's U is missing a sum, and the proof of Theorem A.2 cites Theorem A.2 itself).

Net: the core idea is sound, the result is new, and once the theorem is restated correctly it is a solid contribution. It deserves a referee. I'd want the referee to insist on the hypothesis fix and to flag the companion-paper dependency before publication. I would not cite it in its current form.

Referee Report

2 major / 8 minor

Summary. The manuscript studies the decay of ground states of N-electron atoms at the critical nuclear charge Z_c, i.e., at the edge of the essential spectrum where the ground-state energy coincides with the ionization threshold. The main result (Theorem 2.3) asserts that if ψ_Z is a normalized ground state of the Hamiltonian H_Z^{(N)} at energy E_Z^{(N)}, then e^G ψ_Z is square-integrable, where G is a weighted sum of the square roots of the K largest electron distances in the region where the K outer electrons are well separated from the inner ones, and a weighted linear sum otherwise. The proof uses a weighted IMS localization method developed in the companion paper [7], exploiting the repulsive part of the interelectronic potential to obtain positivity of certain quadratic forms without assuming a spectral gap. Existence of the threshold ground state is imported from [3]; the paper also sketches an alternative existence proof via tightness. The theorem is stated for an arbitrary ground state of H_Z^{(N)} without specifying the critical value Z=Z_c or the definition of K.

Significance. If the stated result is correct, it gives a quantitative decay estimate for a bound state exactly at the threshold of a many-body Coulomb system, with no spectral gap and no dependence on the Born-Oppenheimer approximation (though the main theorem as written assumes an infinitely heavy nucleus; the finite-mass case is only addressed in Remark 2.5). This is a genuinely interesting and nontrivial question, and the method of using repulsive interactions to compensate for the absence of a gap is a useful complement to Agmon-type techniques. The estimates in Lemma 3.1 are plausible and supported by an iterative computation, and the argument is presented in a self-contained way apart from the deferred companion-paper details. A strength is that the constants in the theorem have an explicit form, and the argument is adaptable to bosons or distinguishable particles. However, the main theorem statement omits essential hypotheses used in the proof, which is a serious correctness issue that must be addressed.

major comments (2)

[Theorem 2.3 (Section 2), p. 5] Theorem 2.3 is not correctly stated. The integer K appears in the definition of G in Eq. (3) and throughout the proof, but it is never defined in the theorem. The proof silently takes K from Theorem 2.1—the largest integer with E_{Z_c}^{(N)}=E_{Z_c}^{(N-K)}—and also assumes Z=Z_c. The final positivity step on p. 12 additionally requires W2=(N-K)/(1+δ)-Z>0 and the strict gap E_Z^{(N-K-1)}-E_Z^{(N)}>0, neither of which appears among the theorem's hypotheses. As written, the theorem asserts a decay statement for an arbitrary ground state of H_Z^{(N)} for arbitrary Z and an arbitrary natural number K, for which the operators A1 and A2 in (8) need not be positive. The theorem should be restated with Z=Z_c, with K the specified maximal integer, and with the assumptions Z_c<N-K (or equivalently W2>0 for a suitably small δ) and E_{Z_c}^{(N-K-1)}>E_{Z_c}^{(N)} made explicit.
[Proof of Theorem 2.3, final step (p. 12)] The positivity argument for A2 is incomplete as written. The lower bound for |∇F|^2 in Lemma 3.2 on the relevant region contains a constant term d_k^2 that does not vanish as R→∞. The sentence 'the assumption that E_Z^{(N-K-1)}-E_Z^{(N)}>0 and the fact that we take sufficiently large R' does not account for this constant negative contribution; only the terms W1/|x|_k and e_k/sqrt(|x|_k) are small for large R. The proof needs to state explicitly that the constants C_m and K_m in the definition of F (Eq. (6)) are chosen small enough so that sum_k d_k^2 < E_Z^{(N-K-1)}-E_Z^{(N)}, which is possible but is not mentioned. Additionally, the text at this point says 'with N-K > Z', but the condition needed for A1>0 is (N-K)/(1+δ) > Z (i.e., W2>0); the weaker condition N-K>Z is insufficient because of the factor 1/(1+δ) in the definition of W2.

minor comments (8)

[Abstract] There are grammatical errors ('atoms undergoes', 'we derive upper bound for the bound state', 'Our method do not require') that should be corrected.
[Section 2, first paragraph] The statement 'We make the standard assumptions that the nucleus is infinitely heavy and at the origin, i.e. Born-Oppenheimer approximation' contradicts the abstract's claim that the method does not require the Born-Oppenheimer approximation. The abstract should be qualified, since the main theorem as stated assumes an infinitely heavy nucleus; the finite-mass extension appears only in Remark 2.5 and with an additional mass condition.
[Theorem 2.3] The constants C_m and K_m in Eq. (3) are left unspecified; the statement should mention that they are chosen as in the proof, or that they are sufficiently small positive constants.
[Lemma 3.1] The first line 'Let U = -Z/|x_j| + sum_{j≠k} 1/|x_j-x_k|' should read 'U = sum_{j=1}^N -Z/|x_j| + sum_{j≠k} 1/|x_j-x_k|'.
[Proof of Lemma 3.2] The computation '|∇|x|_m|^2 = |sum_j ∂_j |x|_m|^2 = |∂_k |x_k||^2 = 1' is written in nonstandard notation; it should say that on Ω, the gradient of |x|_m is the unit vector pointing in the direction of the corresponding coordinate, so its squared norm is 1.
[Throughout the proof] The letter K is used both for the number of outer electrons and as a generic constant in inequalities (e.g., '≤K' on pp. 10-12). This is confusing; a different symbol (e.g., C) for the constants would help.
[End of proof of Theorem 2.3] The sentence 'it remains to take the limit ε→0 in (8)' should also address the role of R; after (8), a sentence explaining that for a fixed large R the tail is controlled and the inner region is harmless because e^F is bounded on {|x|_outer≤R} would be useful.
[References] The reference to [7] as 'will appear on arXiv shortly' is not a stable reference; the authors should update it if available, or state the necessary facts from [7] explicitly when they are used.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the decay estimate follows from stated inequalities; the omitted hypotheses in Theorem 2.3 are a rigor gap, not circularity.

full rationale

I walked the claimed derivation chain and found no reduction of the conclusion to the inputs. The proof starts from the eigenfunction equation H_Z^(N) psi_Z = E_Z^(N) psi_Z (Section 3.2, Step 1), applies the IMS-type localization formula proved in Appendix A, and uses the potential estimate Lemma 3.1 and gradient estimate Lemma 3.2 to obtain the quadratic-form inequalities (8). The final positivity step of A1, A2, A3 uses 'N-K > Z' and 'the assumption that E_Z^(N-K-1)-E_Z^(N)>0' (p. 12). These are hypotheses that must be supplied (at Z=Z_c with K maximal and Z_c<N-K they follow), not the desired conclusion e^G psi in L^2; the constants C_m, d_m, e_m, K_m are chosen for positivity and do not encode the decay rate. Existence of the threshold ground state is imported from the external paper [3], and the method is credited to the authors' companion [7], but Theorem 2.3's proof is self-contained here; neither citation fixes the decay statement or assumes it. I flag as a non-circular correctness risk that Theorem 2.3 as stated omits Z=Z_c, the maximality of K, and Z_c<N-K, which the proof silently uses in the final positivity step. Remark 2.2 also sketches an alternative existence proof deferred to [7]; this is an omitted-proof limitation, not circularity. Thus the score is 0 because no equation or fitted parameter reduces to the claimed result by construction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The proof depends on no fitted data parameters. The listed free parameters are arbitrary positive constants and the geometric delta chosen in the proof; they do not encode measured physics. The main external inputs are the existence theorem at threshold from [3], the companion method in [7], and standard N-body spectral theory. The unpublished companion [7] is a reproducibility gap but not circular: Theorem 2.3 is proved here from stated estimates.

free parameters (3)

delta = small positive, not specified
Geometric separation parameter in the partition of Lemma 3.1 and the cutoff (5); it must be small enough that W2=(N-K)/(1+delta)-Z>0, which is possible only under Z_c<N-K.
C_m, K_m = not specified
Positive constants defining the decay function G in Theorem 2.3; the proof only needs their existence, and no numerical values are fitted.
c_m, d_m, e_m = not specified
Constants in the gradient bounds of Lemma 3.2; they are chosen small or large to make the expressions A1, A2, A3 positive.

assumptions (5)

standard math IMS localization formula, including the weak form in Theorem A.2.
Used in Step 2 of the proof of Theorem 2.3; cited to [4] and proved in the appendix for piecewise C^1 cutoffs.
domain assumption Existence of a normalized ground state at the critical coupling Z_c, imported from Bellazzini, Frank, Lieb and Seiringer [3].
Theorem 2.1 states this existence but gives no proof; Remark 2.2 sketches an alternative via tightness and defers details to the companion paper [7].
standard math Standard N-body spectral facts: concavity and monotonicity of E_Z^(N), the inequality E_Z^(N) <= E_Z^(M) for M<N, and Zhislin/HVZ results for subcritical binding.
Used to justify form lower bounds by comparison with M-particle ground state energies; cited to [9], [11], [12].
domain assumption Infinite nuclear mass with the nucleus fixed at the origin (Born-Oppenheimer approximation).
Section 2 makes this standard assumption; Remark 2.5 says finite-mass extension is possible under a mass condition, with modifications as in [7, Appendix B].
domain assumption Antisymmetric fermionic subspace with spin degrees of freedom omitted.
Introduced in Section 1; the proof does not use fermionic statistics and is stated to work for bosons or distinguishable particles.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Quantum Systems at The Brink. Existence and Decay Rates of Bound States at Thresholds; Atoms." pith.science (2026). https://pith.science/paper/SYKRE43K

@misc{pith2026190805016,
  author       = {Pith},
  title        = {Pith review of: Quantum Systems at The Brink. Existence and Decay Rates of Bound States at Thresholds; Atoms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SYKRE43K}},
  note         = {Machine review of arXiv:1908.05016}
}

abstract

It is well known that $N$-electron atoms undergoes unbinding for a critical charge of the nucleus $Z_c$, i.e. the atom has eigenstates for the case $Z> Z_c$ and it has no bound states for $Z<Z_c$. In the present paper we derive upper bound for the bound state for the case $Z=Z_c$ under the assumption $Z_c<N-K$ where $K$ is the number of electrons to be removed for atom to be stable for $Z=Z_c$ without any change in the ground state energy. We show that the eigenvector decays faster as $\exp\left(-C\sum\sqrt{|x|_{k}}\right)$ where we sum K largest values of $|x_j|$, $j\in\{1,\ldots,N\}$. Our method do not require Born-Oppenheimer approximation.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 12 canonical work pages

[3]

Frank, Elliott H

Jacopo Bellazzini, Rupert L. Frank, Elliott H. Lieb, and Robert Seiringer, Existence of ground states for negative ions at the binding threshold, Rev. Math. Phys. 26 (2014), no. 1, 1350021, 18. MR 3162796

work page 2014
[5]

Gridnev, Bound states at threshold resulting from Coulomb repulsion, J

Dmitry K. Gridnev, Bound states at threshold resulting from Coulomb repulsion, J. Math. Phys. 53 (2012), no. 10, 102108, 16. MR 3050564

work page 2012
[7]

existence and decay rates of bound states at thres holds; helium , will appear on arXiv shortly (2019), 25

Dirk Hundertmark, Michal Jex and Markus Lange, Quantum systems at the brink. existence and decay rates of bound states at thres holds; helium , will appear on arXiv shortly (2019), 25

work page 2019
[1]

29, Princeton University Press, Princeton, NJ; University of Tokyo Press, Tokyo, 1982

Shmuel Agmon, Lectures on exponential decay of solutions of second- order elliptic equations: bounds on eigenfunctions of N-body Schrödinger operators, Mathematical Notes, vol. 29, Princeton University Press, Princeton, NJ; University of Tokyo Press, Tokyo, 1982. MR 74 5286

work page 1982
[2]

Hans Wilhelm Alt, Linear functional analysis , 6th edition (2012)

work page 2012
[4]

H. L. Cycon, R. G. Froese, W. Kirsch, and B. Simon, Schrödinger oper- ators with application to quantum mechanics and global geom etry, study ed., Texts and Monographs in Physics, Springer-Verlag, Ber lin, 1987. MR 883643

work page 1987
[6]

Maria Hoﬀmann-Ostenhof and Thomas Hoﬀmann-Ostenhof, Absence of anL2-eigenfunction at the bottom of the spectrum of the Hamilton ian of the hydrogen negative ion in the triplet S-sector, J. Phys. A 17 (1984), no. 17, 3321–3325. MR 771622

work page 1984
[8]

Nonlinear Sci

Dirk Hundertmark and Young-Ran Lee, On non-local variational prob- lems with lack of compactness related to non-linear optics , J. Nonlinear Sci. 22 (2012), no. 1, 1–38. MR 2878650

work page 2012

Show all 12 references

[9]

Lieb, Bound on the maximum negative ionization of atoms and molecules , Phys

Elliott H. Lieb, Bound on the maximum negative ionization of atoms and molecules , Phys. Rev. A 29 (1984), 3018–3028

1984
[10]

Thomas Hoﬀmann-Ostenhof, Maria Hoﬀmann-Ostenhof and Barry Si- mon, A multiparticle Coulomb system with bound state at threshol d, J. Phys. A 16 (1983), no. 6, 1125–1131. MR 706689

1983
[11]

Phan Thành Nam, New bounds on the maximum ionization of atoms , Comm. Math. Phys. 312 (2012), no. 2, 427–445. MR 2917171 15

2012
[12]

Zhislin, Discussion of the spectrum of schrödinger operators for systems of many particles , Trudy Moskovskogo matematiceskogo ob- scestva 9 (1960), 81–120

Grigorii M. Zhislin, Discussion of the spectrum of schrödinger operators for systems of many particles , Trudy Moskovskogo matematiceskogo ob- scestva 9 (1960), 81–120. 16

1960

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [3]

Frank, Elliott H

Jacopo Bellazzini, Rupert L. Frank, Elliott H. Lieb, and Robert Seiringer, Existence of ground states for negative ions at the binding threshold, Rev. Math. Phys. 26 (2014), no. 1, 1350021, 18. MR 3162796

work page 2014

[2] [5]

Gridnev, Bound states at threshold resulting from Coulomb repulsion, J

Dmitry K. Gridnev, Bound states at threshold resulting from Coulomb repulsion, J. Math. Phys. 53 (2012), no. 10, 102108, 16. MR 3050564

work page 2012

[3] [7]

existence and decay rates of bound states at thres holds; helium , will appear on arXiv shortly (2019), 25

Dirk Hundertmark, Michal Jex and Markus Lange, Quantum systems at the brink. existence and decay rates of bound states at thres holds; helium , will appear on arXiv shortly (2019), 25

work page 2019

[4] [1]

29, Princeton University Press, Princeton, NJ; University of Tokyo Press, Tokyo, 1982

Shmuel Agmon, Lectures on exponential decay of solutions of second- order elliptic equations: bounds on eigenfunctions of N-body Schrödinger operators, Mathematical Notes, vol. 29, Princeton University Press, Princeton, NJ; University of Tokyo Press, Tokyo, 1982. MR 74 5286

work page 1982

[5] [2]

Hans Wilhelm Alt, Linear functional analysis , 6th edition (2012)

work page 2012

[6] [4]

H. L. Cycon, R. G. Froese, W. Kirsch, and B. Simon, Schrödinger oper- ators with application to quantum mechanics and global geom etry, study ed., Texts and Monographs in Physics, Springer-Verlag, Ber lin, 1987. MR 883643

work page 1987

[7] [6]

Maria Hoﬀmann-Ostenhof and Thomas Hoﬀmann-Ostenhof, Absence of anL2-eigenfunction at the bottom of the spectrum of the Hamilton ian of the hydrogen negative ion in the triplet S-sector, J. Phys. A 17 (1984), no. 17, 3321–3325. MR 771622

work page 1984

[8] [8]

Nonlinear Sci

Dirk Hundertmark and Young-Ran Lee, On non-local variational prob- lems with lack of compactness related to non-linear optics , J. Nonlinear Sci. 22 (2012), no. 1, 1–38. MR 2878650

work page 2012