Prediction of Novel Li-AgII-F Compounds using Evolutionary Algorithms

Katarzyna Kuder; Wojciech Grochala

REVIEW 2 major objections 5 minor 1 cited by

Prediction of Novel Li-AgII-F Compounds using Evolutionary Algorithms

T0 review · 2 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read The paper predicts that a triclinic LiAgF3 phase hosts a record magnetic superexchange of −359 meV.

desk verdict A well-documented computational search for Li–Ag(II)–F phases with a striking predicted superexchange constant, but the record claim rests on a single DFT+U setting and needs a sensitivity check before it can be taken at face value. read the letter →

arxiv 2512.05048 v1 pith:B4PNHE2O submitted 2025-12-04 cond-mat.mtrl-sci cond-mat.str-el

Katarzyna Kuder , Wojciech Grochala This is my paper · ORCID

classification cond-mat.mtrl-scicond-mat.str-el

keywords Li–Ag(II)–FcompoundsevolutionaryalgorithmstructurepredictionDFT+Umagneticsuperexchangeantiferromagnetismsilver(II)fluoridesmetastablephasesconvexhull

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

This paper searches for previously unknown ternary lithium–silver(II)–fluoride compounds and predicts their structures, stability, and magnetism. It claims that all five lowest-energy LiAgF3 and Li2AgF4 polymorphs lie slightly above the LiF + AgF2 decomposition line, so they are metastable, but that formation from AgF, LiF, and fluorine gas or fluorine radicals is exothermic, making alternative syntheses plausible. All predicted phases prefer antiferromagnetic order. The central result is that the triclinic LiAgF3 type-2 polymorph contains [Ag2F7] dimers with short, nearly linear Ag–F–Ag bridges, giving a superexchange constant J ≈ −359 meV, which the authors place above the strongest magnetic exchange interactions known so far. If correct, this moves Ag(II) fluoride chemistry into the superexchange regime where magnetic-fluctuation-driven high-temperature superconductivity has been predicted for two-dimensional systems.

What carries the argument

The load-bearing machinery is density-functional theory with a Hubbard U correction (DFT+U, U = 5.0 eV, J_H = 1.0 eV on Ag 4d) used to relax candidate structures and to map total energies of ferromagnetic and antiferromagnetic spin arrangements onto a Heisenberg Hamiltonian H = −(1/2)ΣJ_ij S_i·S_j. The structural motif that carries the record claim is the [Ag2F7] dimer: two Jahn–Teller-distorted Ag(II) octahedra sharing a fluoride bridge with short Ag–F bonds of 2.014 Å and a 180° bridge angle. The same angle-short-bond combination is what earlier work identified in related Ag(II) fluorides, but here it is predicted to be even stronger.

What would settle it

Compute J for LiAgF3 type 2 at several Hubbard U values (say 3, 4, 5, 6, 7 eV) or with a hybrid functional / embedded-cluster quantum chemistry calculation. If |J| falls below about 240 meV (the Sr2CuO3 value), the record claim is refuted. Experimentally, synthesize the phase and measure magnetic susceptibility or neutron scattering: a singlet-dimer gap near 359 meV (or a Weiss temperature of that order) would confirm it.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a specific metastable phase, triclinic LiAgF3 type 2, should exhibit the largest antiferromagnetic superexchange constant reported for any material: J1 = −359 meV within its [Ag2F7] dimers, with a much weaker J2 = −11 meV between dimers. The magnitude exceeds the measured −240 meV of Sr2CuO3 and the calculated −250/−313 meV of related Ag(II) fluoride motifs. The strong coupling is attributed to very short Ag–F bonds (2.014 Å) at a 180° Ag–F–Ag angle. The same DFT+U treatment gives smaller J values for the other four predicted phases (−4 to −95 meV). Because all predicted phases sit 8–18 kJ/mol above the LiF + AgF2 hull, the paper does not claim

Load-bearing premise

The record coupling rests on one fixed Hubbard correction (U=5.0 eV, J_H=1.0 eV on silver's d-orbitals); if the real effective correlation strength is different, the numerical value of J — and the comparison with earlier records — could change substantially.

Editorial extensions

If this is right

If synthesized, LiAgF3 type 2 would be the strongest known magnetic superexchange material, with |J| ≈ 359 meV, above Sr2CuO3's 240 meV.
The predicted metastability means conventional sintering of LiF + AgF2 will not work; fluorination from AgF, LiF and F2 or F• is the energetically plausible route.
All five predicted Li–Ag(II)–F phases have antiferromagnetic ground states, so any successful synthesis should yield an antiferromagnet or a singlet-dimer system rather than a ferromagnet.
The result suggests that quasi-0D [Ag2F7] dimers can reach the coupling range (≈400–700 meV) where 2D magnetic-fluctuation-driven superconductivity is predicted, motivating searches for 1D and 2D Ag(II) fluoride lattices with linear short bridges.

Reading between the lines

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

A targeted computational scan varying the Hubbard U (e.g., 3–7 eV) or using hybrid functionals on the [Ag2F7] dimer would show how far the −359 meV value can move; the paper itself gives no such sensitivity test.
The polar and non-centrosymmetric space groups found for LiAgF3 type 1 (Cc) and Li2AgF4 type 3 (P21) suggest these metastable fluorides could combine antiferromagnetism with ferroelectricity or nonlinear optical response — properties the paper mentions but does not develop.
If a high-pressure or epitaxial route can stabilize the linear short-bridge motif in a 2D lattice, the same mechanism that gives J ≈ 359 meV in a dimer might push J into the 400–700 meV window proposed for magnetic-fluctuation-driven superconductivity; searching for such lattices is a natural extension.

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

The paper is a solid, carefully executed computational search of the Li–Ag(II)–F ternary phase space, and it deserves a serious referee. What it actually adds: a systematic XtalOpt exploration with convex-hull analysis, a set of new polymorphs (LiAgF3 1/2, Li2AgF4 1–3), and a particularly interesting result—LiAgF3 2, a triclinic structure with [Ag2F7] dimers, predicted to have a superexchange constant of about −359 meV. That would place it above the strongest exchange constants known in these systems. The SI is a real strength: CIFs, cutoff and k-point convergence, magnetic models, and the full set of extracted J values are all there. The mapping from spin-configuration energies to Heisenberg couplings is standard and not circular—the J's are fit to energy differences, not to experimental targets, so the internal methodology is fine.

The soft spots, in order of importance. First, the headline J is computed with PBEsol+U (U=5.0 eV, J=1.0 eV on Ag d) and there is no U-sensitivity scan. Superexchange scales roughly as t^2/U_eff, so a plausible range in U could easily shift J by tens of percent. The margin over the measured −240 meV of Sr2CuO3 is about 119 meV, so the record claim is not robust without knowing how J varies with U. This is not an internal contradiction, but it is load-bearing: if the effective U for Ag(II) is larger than 5 eV, the record may evaporate. The other comparisons to CsAgF3, AgF2-HPII, and Ag2ZnZr2F14 all used the same U, so they share any systematic bias and cannot calibrate the absolute scale. A calibration against an experimental J (e.g., KAgF3 or AgF2) or even a simple U scan would fix this. Second, the paper compares its computed J to measured values without uncertainty or a caveat that computed and measured values are not directly equivalent. Third, minor: the abstract says −358 meV, the text and Table S2 say −359 meV; trivial but should be caught. Fourth, the 'previously unexplored' framing is overstated—Li2AgF4 is already in refs. 21 and 31, so the novelty is in the new polymorphs and the systematic map, not in opening an untouched system.

Who is this for? People working on strongly correlated Ag(II) fluorides, magnetic superexchange, and predictive materials discovery. It is a strong candidate for a rigorous peer-reviewed venue, provided the authors can address the U sensitivity and moderate the record claim. I would send it to referees rather than desk-reject it.

Referee Report

2 major / 5 minor

Summary. The paper reports a computational structure search of ternary Li–Ag(II)–F compounds using the XtalOpt evolutionary algorithm, followed by PBEsol+U (Dudarev U=5.0 eV, JH=1.0 eV on Ag d) relaxations and total-energy calculations. Five low-energy polymorphs of LiAgF3 and Li2AgF4 are identified, all lying 8–18 kJ/mol above the LiF + AgF2 convex hull. The authors compute magnetic superexchange constants from FM/AFM energy differences and find an antiferromagnetic ground state in every case. The central result is the triclinic LiAgF3 2 phase, which is reported to contain [Ag2F7] dimers with a very large intradimer coupling J1 = −358/−359 meV, claimed to exceed the strongest known magnetic exchange interactions to date.

Significance. If the quantitative prediction holds, the paper is significant for Ag(II) fluoride chemistry and for the broader effort to realize very strong superexchange, a relevant ingredient in proposals for magnetic-fluctuation-driven superconductivity. The study has concrete strengths: a systematic evolutionary search, well-defined convex-hull and entropy analyses, and a supplementary information file with CIFs, convergence data, spin models, and DOS/band-structure information. The J extraction is a standard energy-mapping approach and is not circular. However, the headline record claim is not yet supported because it rests on a single DFT+U parameter choice and on a comparison set that mixes same-group DFT+U values with one experimental value.

major comments (2)

[§2, §3.2, Table S2] The headline J1 = −358.75 meV for LiAgF3 2 is computed at a single DFT+U setting (PBEsol+U, Dudarev U=5.0 eV, JH=1.0 eV). No U-sensitivity scan is reported, and no calibration against experimental J values for Ag(II) fluorides (e.g., AgF2 or KAgF3) is provided. For superexchange the relevant scale is roughly t^2/U_eff, so plausible variations of the effective U in a strongly correlated fluoride can shift J by a significant fraction of the margin between −359 meV and the −240 meV experimental value of Sr2CuO3 used as the record benchmark. Without a quantitative robustness check, the 'record-breaking' conclusion is not established. I request a U-scan (e.g., U=3–7 eV) or a calibration to measured superexchange in Ag(II) fluorides, and the text should report J as a function of the Hubbard parameter rather than as a single number.
[Abstract; §3.2] The claim that LiAgF3 2 lies 'above the strongest known magnetic exchange interactions reported to date' is stronger than the presented evidence. The only experimental value directly compared is Sr2CuO3 (−240 meV). The other comparison values (CsAgF3, AgF2-HPII, Ag2ZnZr2F14) are DFT+U calculations from the same group using the identical Hubbard parameters (ref 54). These calculations can establish an internally consistent ranking within this computational scheme, but they cannot validate the absolute scale or support a general record claim. The authors should either provide a broader survey of experimental and independent computational J values, or reframe the claim as 'largest calculated Ag(II)–F superexchange at this level of theory'.

minor comments (5)

[Abstract vs §3.2/Conclusions] The abstract reports J=−358 meV, while §3.2, the Conclusions, and Table S2 report −359 meV or −358.75 meV. Please unify the rounded value.
[Figure 3 caption] The caption of Figure 3a refers to 'LiAgF3 1', but the text and the magnetic structure under discussion concern LiAgF3 2. Please correct the caption.
[§3.3 and Figure 4b] The abstract states that the phases have negative formation energies relative to AgF, LiF, and F2/F, but the main text gives only the schematic Figure 4b without numeric values. A table of the computed formation energies would make the synthetic-feasibility argument quantitative and reproducible.
[SI Table S2] The magnetic models are presented compactly as spin strings with energy expressions. Please add an explicit diagram or a full Hamiltonian showing the numbering of the four Ag sites and the assignment of J1 and J2 edges, so the mapping from total energies to J values is unambiguous.
[Conclusions] The statement that LiAgF3 2 is a 'meaningful step towards' the J≈400–700 meV regime is not directly supported because the predicted J1 (−359 meV) lies below that window. Please soften the wording or add a clarifying argument about how the dimer coupling relates to the 2D models discussed.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the record J is obtained by a direct DFT energy-mapping onto a Heisenberg model, not by fitting or by self-referential definition; self-citations are methodological or comparative and not load-bearing.

full rationale

The central derivation chain is non-circular. The superexchange constants are extracted from explicit spin-configuration energy expressions in the SI (Eq. 1, Tables S1–S5). For LiAgF3 2, Table S2 lists three spin models (FM, AFM, AFM2) whose energies combine with the Heisenberg Hamiltonian to determine J1 and J2 directly from DFT+U total-energy differences; no experimental J is used as an input, and no parameter is fitted to reproduce a target J. The Hubbard U=5.0 eV and J=1.0 eV are stated methodological parameters, inherited partly from prior self-cited work (refs. 10, 11), but this is a standard parameter choice, not a quantity defined in terms of the predicted J. The comparison values for other known Ag(II) fluorides are indeed taken from ref. 54, the authors' own earlier work, and are described as 'calculated using identical methodology as the one used here'; this reduces the independence of the 'strongest known' benchmarking, but it does not make the new J prediction circular, because the LiAgF3 2 J value is independently computed in this paper and is also compared against the external Sr2CuO3 value of −240 meV (ref. 55). The thermodynamic stability analysis is separate and self-contained. The lack of a U-sensitivity scan is a robustness/correctness concern, not a circularity: a different U could change the magnitude of J, but the mapping from DFT energies to J remains a genuine calculation. Overall, the derivation is independent of its own conclusions, with only minor non-load-bearing self-citation.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

All physics input is GGA+U with user-set U and J; the search seeds come from self-cited prior structures; no new particles, forces, or conserved quantities are introduced.

free parameters (2)

Hubbard U (Dudarev) on Ag 4d = 5.0 eV
Chosen from prior Ag(II) literature; directly controls the magnitude of J and relative phase energies; no sensitivity analysis in this paper.
Hubbard exchange J_H on Ag 4d = 1.0 eV
Used with U in the Dudarev formulation; affects the DFT+U total energies from which J values are extracted.

assumptions (4)

domain assumption PBEsol+U accurately describes strongly correlated 4d electrons and magnetic superexchange in Ag(II) fluorides.
Invoked in §2 as the reason for applying DFT+U with U=5 eV, J=1 eV; the central J claim depends on this functional choice.
domain assumption The Heisenberg spin Hamiltonian with the small set of FM/AFM configurations used in SI S3 is sufficient to extract J1 and J2.
The J values are obtained from 2–4 spin configurations per structure; any additional magnetic couplings would change the extracted numbers.
domain assumption The Jenkins–Glasser volume-based entropy approximation correctly estimates stabilization temperatures for these phases.
Used in §3.3 to argue whether entropic stabilization could make metastable phases accessible; this is a rough ionic-crystal estimate, not a phonon calculation.
domain assumption The XtalOpt search over 2, 4, and 6 formula-unit cells and the chosen seed structures covers all relevant low-energy Li–Ag(II)–F polymorphs.
The conclusions about the lowest-energy and magnetic properties rest on the search having found the ground-state configurations.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Prediction of Novel Li-AgII-F Compounds using Evolutionary Algorithms." pith.science (2026). https://pith.science/paper/B4PNHE2O

@misc{pith2026251205048,
  author       = {Pith},
  title        = {Pith review of: Prediction of Novel Li-AgII-F Compounds using Evolutionary Algorithms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B4PNHE2O}},
  note         = {Machine review of arXiv:2512.05048}
}

read the original abstract

This work provides a theoretical exploration of the thermodynamic stability and magnetic behaviour of previously unknown ternary Li AgII F compounds. Convex-hull analysis shows that all predicted structures lie slightly above the LiF plus AgF2 decomposition line, indicating a natural tendency toward phase separation; nevertheless, their negative formation energies relative to AgF, LiF, and F2 or F suggest that alternative synthetic pathways may be feasible for these compounds. All studied structures show preference for antiferromagnetic ground state. Notably, the triclinic LiAgF3 type2 is predicted to exhibit an exceptionally large superexchange constant, J equal to minus 358 meV, within Ag2F7 dimers, placing it above the strongest known magnetic exchange interactions reported to date.

Figures

Figures reproduced from arXiv: 2512.05048 by the authors.

**Figure 2.** Crystal structures of the investigated Li [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** a - Local magnetic structure of LiAgF3 1 showing the orientations of the spin moments on the Ag(II) cations. The solid lines indicate the nearest-neighbor superexchange pathways (J1), and the dashed blue lines the next-nearest-neighbor ones (J2); bond lengths and angles are shown; b - Comparison of different structural motifs containing the [Ag2F7] unit, including chain- and dimer-type arrangements, together with th… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: a - Convex hull of the investigated systems as a function of the AgF [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Theoretical exploration of Be Ag(II) F phases and their magnetic properties using learning algorithms
cond-mat.supr-con 2026-07 conditional novelty 6.5 of 10

Global structure prediction plus DFT+U identifies five AgBeF4 polymorphs, two with unprecedented AFM superexchange J ≈ −460 and −359 meV from [Ag2F7] dimers or straight [AgF] chains.

Reference graph

Works this paper leans on

6 extracted references · 1 linked inside Pith · cited by 1 Pith paper

[1]

McMillan,Chemical Reviews, 1962,62, 65–80

(1) J. McMillan,Chemical Reviews, 1962,62, 65–80. (2) P. Fischer, D. Schwarzenbach and H. Rietveld,Journal of Physics and Chemistry of Solids, 1971,32, 543–550. (3) W. Grochala, R. G. Egdell, P. P. Edwards, Z. Mazej and B. ˇZemva,ChemPhysChem, 2003, 4, 997–1001. (4) C. Miller and A. S. Botana,Physical Review B, 2020,101, 195116. (5) P. Fischer, G. Roult a...

1962
[140]

Kurzyd lowski, T

(44) D. Kurzyd lowski, T. Jaro´ n, A. Ozarowski, S. Hill, Z. Jagliˇ ci´ c, Y. Filinchuk, Z. Mazej and W. Grochala,Inorganic Chemistry, 2016,55, 11479–11489. (45) K. Dupre, K. Recker and F. Wallrafen,Materials research bulletin, 1992,27, 311–318. (46) G. Calestani and F. Mezzadri, inPhotonic and Electronic Properties of Fluoride Materials, ed. A. Tressaud ...

2016
[458]

Grzelak, J

(51) A. Grzelak, J. Gawraczy´ nski, T. Jaro´ n, D. Kurzyd lowski, A. Budzianowski, Z. Mazej, P. J. Leszczy´ nski, V. B. Prakapenka, M. Derzsi, V. V. Struzhkin et al.,Inorganic Chemistry, 2017, 56, 14651–14661. 9 (52) R.-H. Odenthal and R. Hoppe,Zeitschrift f¨ ur anorganische und allgemeine Chemie, 1971, 385, 92–101. (53) D. Kurzyd lowski, Z. Mazej, Z. Jag...

2017
[2024]

Belak Vivod, Z

(15) M. Belak Vivod, Z. Jagliˇ ci´ c, G. King, T. C. Hansen, M. Lozinˇ sek and M. Dragomir,Journal of the American Chemical Society, 2024,146, 30510–30517. (16) R. Fischer and B. M¨ uller,Zeitschrift f¨ ur anorganische und allgemeine Chemie, 2002,628, 2592–2596. (17) B. Zemva, K. Lutar, A. Jesih, W. J. Casteel Jr, A. P. Wilkinson, D. E. Cox, R. B. Von Dre...

2024
[2781]

Koteras, J

(10) K. Koteras, J. Gawraczy´ nski, G. Tavˇ car, Z. Mazej and W. Grochala,CrystEngComm, 2022, 24, 1068–1077. (11) D. Kurzyd lowski and W. Grochala,Angewandte Chemie, 2017,129, 10248–10251. (12) T. Jaro´ n and W. Grochala,physica status solidi (RRL)–Rapid Research Letters, 2008,2, 71–73. (13) W. J. Casteel Jr, G. Lucier, R. Hagiwara, H. Borrmann and N. Bar...

arXiv 2022
[4198]

(18) C. Shen, B. Zemva, G. M. Lucier, O. Graudejus, J. A. Allman and N. Bartlett,Inorganic Chemistry, 1999,38. (19) M. Kraus, M. M¨ uller, R. Fischer, R. Schmidt, D. Koller and B. M¨ uller,Journal of Fluorine Chemistry, 2000,101, 165–171. (20) O. Graudejus, S. H. Elder, G. M. Lucier, C. Shen and N. Bartlett,Inorganic Chemistry, 1999,38, 2503–2509. (21) R....

1999

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

McMillan,Chemical Reviews, 1962,62, 65–80

(1) J. McMillan,Chemical Reviews, 1962,62, 65–80. (2) P. Fischer, D. Schwarzenbach and H. Rietveld,Journal of Physics and Chemistry of Solids, 1971,32, 543–550. (3) W. Grochala, R. G. Egdell, P. P. Edwards, Z. Mazej and B. ˇZemva,ChemPhysChem, 2003, 4, 997–1001. (4) C. Miller and A. S. Botana,Physical Review B, 2020,101, 195116. (5) P. Fischer, G. Roult a...

1962

[2] [140]

Kurzyd lowski, T

(44) D. Kurzyd lowski, T. Jaro´ n, A. Ozarowski, S. Hill, Z. Jagliˇ ci´ c, Y. Filinchuk, Z. Mazej and W. Grochala,Inorganic Chemistry, 2016,55, 11479–11489. (45) K. Dupre, K. Recker and F. Wallrafen,Materials research bulletin, 1992,27, 311–318. (46) G. Calestani and F. Mezzadri, inPhotonic and Electronic Properties of Fluoride Materials, ed. A. Tressaud ...

2016

[3] [458]

Grzelak, J

(51) A. Grzelak, J. Gawraczy´ nski, T. Jaro´ n, D. Kurzyd lowski, A. Budzianowski, Z. Mazej, P. J. Leszczy´ nski, V. B. Prakapenka, M. Derzsi, V. V. Struzhkin et al.,Inorganic Chemistry, 2017, 56, 14651–14661. 9 (52) R.-H. Odenthal and R. Hoppe,Zeitschrift f¨ ur anorganische und allgemeine Chemie, 1971, 385, 92–101. (53) D. Kurzyd lowski, Z. Mazej, Z. Jag...

2017

[4] [2024]

Belak Vivod, Z

(15) M. Belak Vivod, Z. Jagliˇ ci´ c, G. King, T. C. Hansen, M. Lozinˇ sek and M. Dragomir,Journal of the American Chemical Society, 2024,146, 30510–30517. (16) R. Fischer and B. M¨ uller,Zeitschrift f¨ ur anorganische und allgemeine Chemie, 2002,628, 2592–2596. (17) B. Zemva, K. Lutar, A. Jesih, W. J. Casteel Jr, A. P. Wilkinson, D. E. Cox, R. B. Von Dre...

2024

[5] [2781]

Koteras, J

(10) K. Koteras, J. Gawraczy´ nski, G. Tavˇ car, Z. Mazej and W. Grochala,CrystEngComm, 2022, 24, 1068–1077. (11) D. Kurzyd lowski and W. Grochala,Angewandte Chemie, 2017,129, 10248–10251. (12) T. Jaro´ n and W. Grochala,physica status solidi (RRL)–Rapid Research Letters, 2008,2, 71–73. (13) W. J. Casteel Jr, G. Lucier, R. Hagiwara, H. Borrmann and N. Bar...

arXiv 2022

[6] [4198]

(18) C. Shen, B. Zemva, G. M. Lucier, O. Graudejus, J. A. Allman and N. Bartlett,Inorganic Chemistry, 1999,38. (19) M. Kraus, M. M¨ uller, R. Fischer, R. Schmidt, D. Koller and B. M¨ uller,Journal of Fluorine Chemistry, 2000,101, 165–171. (20) O. Graudejus, S. H. Elder, G. M. Lucier, C. Shen and N. Bartlett,Inorganic Chemistry, 1999,38, 2503–2509. (21) R....

1999