Frustrated vacancy ordering creates novel quantum properties in Kutinaite, $\mathrm{Ag}_{6}\mathrm{Cu}_{14.4}\mathrm{As}_7$

Bo Brummerstedt Iversen; Emilie Skytte Vosegaard; Karl Omer Rimon Juul; Kasper Rasmussen Borup; Kim-Khuong Huynh; Rasmus Baden Stubkj{\ae}r; Ventrapati Pavankumar; Yong P. Chen

REVIEW 4 major objections 5 minor 1 references

Frustrated vacancy ordering creates novel quantum properties in Kutinaite, $\mathrm{Ag}_{6}\mathrm{Cu}_{14.4}\mathrm{As}_7$

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Cooling below about 300 K drives the copper vacancies in Kutinaite into a frustrated, degenerate superstructure with a 6.2 Å nearest-vacancy spacing, and this ordered defect state carries metallic Landau diamagnetism comparable to bismuth.

desk verdict A well-designed multi-method study: the frustrated 5th-NN vacancy ordering in 20%-vacant Kutinaite is plausible and mostly well supported, but the 'Dirac-like' electronic interpretation outruns the transport evidence. read the letter →

arxiv 2505.24447 v1 pith:3A444HZY submitted 2025-05-30 cond-mat.dis-nn cond-mat.mes-hallcond-mat.mtrl-scicond-mat.str-el

Kim-Khuong Huynh , Rasmus Baden Stubkj{ae}r , Ventrapati Pavankumar , Emilie Skytte Vosegaard , Karl Omer Rimon Juul , Kasper Rasmussen Borup , Yong P. Chen , Bo Brummerstedt Iversen This is my paper

classification cond-mat.dis-nncond-mat.mes-hallcond-mat.mtrl-scicond-mat.str-el

keywords frustratedvacancyorderingKutinaiteLandaudiamagnetism5thnearest-neighbourrule3D-ΔPDFresidualentropyDirac-likesemimetalcorrelateddisorder

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

The paper sets out to show that the 20% missing copper sites in the mineral Kutinaite (Ag6Cu14.4As7) are not random impurities but an ordered, frustrated subsystem that appears below about 300 K. Using single-crystal and powder X-ray diffraction, a 3D-difference-PDF-style analysis of the weak "forbidden" reflections, and Monte-Carlo simulations, the authors infer a single ordering rule: every vacancy's nearest neighbouring vacancies sit at the 5th nearest-neighbour distance of about 6.2 Å, which still leaves many energetically equivalent configurations and therefore residual entropy. The same defect-ordered crystal shows metallic resistivity yet a large negative magnetic susceptibility, a Landau-Peierls diamagnetism comparable to bismuth, which they attribute to a small pocket of very light, Dirac-like electron carriers. The broader claim is that vacancies can act as a quasi-chemical species that generates emergent electronic states, so correlated disorder should be treated as part of the crystal's design rather than as an imperfection.

What carries the argument

The central object is the 5th nearest-neighbour rule for vacancy order on the Cu2 sublattice: in the low-temperature state, every vacancy's closest other vacancies are at the 5th NN distance of about 6.2 Å, enforced in Monte-Carlo simulations by interaction energies $J_1 = 2J_2 = 4J_3 = 10J_4 = 10^3 J_5$ (where $J_i$ is the positive repulsion between a vacancy and its $i$-th neighbours). Because each Cu4 tetrahedron carries one vacancy and a vacancy can pair with either of two equivalent 5th-NN sites in each neighbouring tetrahedron, the rule leaves many degenerate configurations — the source of frustration and residual entropy. The real-space evidence comes from 3D-ΔPDF-like maps obtained by Fourier transforming only the F-forbidden reflections, which show a strong negative 1st-NN correlation, weaker negative 3rd/4th-NN features, and a positive 5th-NN correlation at 6.2 Å. These maps are the machinery that connects the weak forbidden peaks to the real-space ordering rule.

What would settle it

A total-scattering 3D-ΔPDF analysis that includes all diffuse scattering (not just the F-forbidden peaks) would settle the question: if the strong negative 1st-neighbour and positive 5th-neighbour (6.2 Å) correlations disappear or change sign once anharmonic As2 motion and other sources are included, the vacancy-ordering interpretation fails. Alternatively, measuring a fully ordered Pa-3 twin model, or testing a second composition with a different vacancy concentration, would show whether the 5th NN rule is generic.

Watch

Extended reading notes

Core claim

The paper claims that Kutinaite's Cu2 sublattice hosts a frustrated vacancy-ordered state below $T_{\mathrm{vo}} \approx 300$ K. In the high-temperature phase the vacancies are random except for the constraint of one vacancy per Cu4 tetrahedron; on cooling, weak reflections forbidden by the $Fm\bar{3}m$ average structure appear, and the authors show these arise from coherent scattering of the vacancy superstructure. The deduced ordering rule is that each vacancy has its nearest vacant neighbours only at the 5th nearest-neighbour distance (about 6.2 Å) within the Cu2 network. The rule does not select a unique configuration: a vacancy in one tetrahedron can choose between two equivalent 5th-neighbour positions in each neighbouring tetrahedron, so the ordered state is geometrically frustrated, with a residual entropy $S_{\mathrm{vo}} \approx 18.45$ J/(K·mol) and a broad, first-order-like transition with hysteresis. The same material is a metal with a large negative susceptibility ($\chi \approx -5.7 \times 10^{-5}$ SI), which after subtracting Pauli, core and ring-current terms leaves a Landau diamagnetism $\chi_L \approx -1.2 \times 10^{-4}$, on the scale of bismuth, attributed to a light electron pocket with Dirac-like character.

Load-bearing premise

The whole real-space argument rests on the assumption that the very weak extra diffraction peaks that appear below about 300 K come only from ordered copper vacancies; if anything else—such as thermal motion of arsenic atoms—also contributes to those peaks, the deduced 6.2 Å spacing rule would not be supported by the data.

Editorial extensions

If this is right

If the 5th NN rule is correct, the vacancy order in Kutinaite is inherently frustrated, so the low-temperature state is a frozen collection of degenerate domains rather than a unique superstructure; this explains the residual entropy of about 18.45 J/(K·mol), the broad specific-heat hump, and the first-order-like hysteresis in resistivity and susceptibility.
The coherence length of the vacancy order can be controlled by cooling rate: faster cooling fragments the ordered domains to roughly one-third of their relaxed size (about 40 unit cells when relaxed), and the resulting domain boundaries scatter electrons, so fast-cooled samples are more resistive and less diamagnetic.
The large Landau diamagnetism ($\chi_L \approx -1.2 \times 10^{-4}$ SI) requires a small, very light electron pocket; the authors connect it to a Dirac-like band with $E_F \approx 16$ meV and $v_F \approx 3.12 \times 10^5$ m/s, consistent with quasi-linear magnetoresistance following the Abrikosov model.
Because a fully occupied Ag6Cu16As7 would be charge-imbalanced, the ~20% Cu2 vacancies are chemically necessary; the paper's picture implies that the defect sublattice is an intrinsic part of the material's identity, and that the same structure type without vacancies (M6Ni16Si7) should show conventional metallic properties.

Reading between the lines

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

If the 5th NN rule truly governs the vacancy order, the same rule should predict ordering in other M6Ni16Si7-type or kutinaite-like compounds at different vacancy concentrations; a testable prediction is that compositions away from ~20% vacancies will either fail to satisfy the rule or select different superstructures, changing the residual entropy.
The vacancy-as-chemical-species framing suggests a materials-design strategy: choose a host lattice whose defect sites have a degenerate set of equivalent local configurations, then use cooling protocols to tune domain size and, through domain-wall scattering, the transport and magnetic response; this could be a general knob for engineering defect-ordered metals.
The identification of pocket 1 as Dirac-like rests on indirect transport signatures; a direct test would be angle-resolved photoemission or quantum oscillation measurements on single crystals to confirm the small Fermi pocket and its linear dispersion.
If the F-forbidden-only Fourier maps are the right real-space evidence, the same "difference Fourier" procedure could be applied to other partially occupied sublattices to reveal hidden ordering rules in materials where the superstructure peaks are too weak for conventional refinement.

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

Colleague,

Worth your time: this is the first real experimental characterization of Kutinaite, and it claims a frustrated 5th-NN vacancy-ordering rule (vacancies ~6.2 Å apart) that sets in below ~300 K, together with Landau diamagnetism on the bismuth scale from a metal that is 20% vacant. The vacancy-ordering result is genuinely new — absent from Karanovic 2002 and Holman 2008 — and the evidence base is serious. Temperature-dependent SCXRD and PXRD show weak F-forbidden reflections that appear reversibly below ~300 K. The MC study discriminates between models: 1NN–4NN interaction hierarchies fail to reproduce the low-T reflections; the k=5 model does. The 3D-ΔPDF maps give real-space signatures of the 5th-NN correlation, and the entropy accounting nearly closes (measured ΔS ≈ 4.6 vs counted 4.32 J/(K·mol)). The quenching experiments are the strongest part: cooling rate controls VO domain size, and FWHM broadening, resistivity, and diamagnetism all move together. That internal consistency is persuasive. The authors also handle alternatives honestly — the Pa-3 twinned solution is presented and rejected because it implies the wrong vacancy fraction.

Soft spots, in order of severity. First, the F-forbidden peaks are weak (mean I/σ = 0.4 at 25 K), and the 3D-ΔPDF maps are computed from those peaks alone, assuming anharmonic As2 diffuse scattering contributes nothing. Peak sharpness and the MC/Rietveld agreement partially compensate, but the direct real-space support for the rule rests on peaks below conventional detection significance. A referee should ask for a quantitative comparison of diffuse scattering computed from the MC supercells against the full observed pattern, with alternatives masked. Second, the MC hierarchy J1=2J2=4J3=10J4=10^3 J5 is built to maximize vacancy separation, so the 5th-NN rule is parameterized in; the failure of k=4 is a fair control, but the rule is not independently predicted. Third, the "Dirac-like" claim is the weakest link: no band structure, no quantum oscillations; the light pocket mass is inferred from μ1 = eτ/m1 with τ unmeasured, and it sits awkwardly beside the heavy m*/me ≈ 17 from specific heat. Linear MR plus an Abrikosov fit is suggestive, not probative, and "novel quantum properties" in the abstract outruns the transport evidence.

One stress-test point I would push back on: the entropy counting is not internally inconsistent. The stated count (4 × 2^3 × 4^4 configurations per unit cell) gives exactly S_vo^cal = 3.25R ln 2 ≈ 18.73 J/(K·mol), and the ~1.5% gap versus the measured residual entropy is what the paper calls "oversimplified." That is fine.

Who it is for: crystallographers and materials people working on correlated disorder and defect engineering. It deserves a serious referee: the evidence is rich, the claims are significant, and the gaps are addressable. Send it to review, and push on the diffuse-scattering control and the electronic-structure claims.

Referee Report

4 major / 5 minor

Summary. This manuscript reports a combined X-ray scattering, Monte Carlo (MC) simulation, thermodynamic, and transport study of synthetic Kutinaite, Ag6Cu14.4As7, a cubic (Fm-3m) compound with ~20% vacancies on the Cu2 sublattice. The authors observe weak, temperature-dependent reflections that are forbidden by the F-centering extinction rules, appearing below about 300 K, and interpret them as evidence of a vacancy-ordering (VO) transition. They propose a '5th nearest-neighbor (NN) rule' in which each Cu2 vacancy has its nearest vacancy at a distance of about 6.2 Å, leading to a frustrated, degenerate superstructure. Supporting evidence includes MC simulations with a hierarchy of interaction energies (J1=2J2=4J3=10J4=10^3 J5) that reproduce the low-T powder diffraction pattern, 3D-difference-PDF-like maps built by Fourier transforming only the F-forbidden peaks, and an entropy release of ~4.6 J/(K·mol) that is compared with a configurational entropy calculation. The same VO is linked to the material's large metallic diamagnetism (χ ≈ -5.7×10^-5 SI), attributed to Landau-Peierls diamagnetism from light carriers in a Dirac-like pocket. The paper claims that frustrated vacancy ordering creates novel quantum properties in this material.

Significance. If the central claim holds, this is an interesting demonstration that correlated vacancy ordering, rather than a conventional superstructure, can control electronic properties in a metallic compound. The experimental dataset is extensive: temperature-dependent single-crystal and powder diffraction, a transparent MC model, heat capacity, resistivity, Hall effect, magnetoresistance, and magnetization. The temperature dependence and reversibility of the F-forbidden peaks are well documented, and the interpretation in terms of Cu2 vacancy ordering is plausible. The MC model is simple and the paper honestly describes its limitations. However, the load-bearing evidence for the specific '5th NN rule' is weaker than the narrative suggests: the F-forbidden peaks used to construct the real-space maps have a mean I/σ of only 0.4, and the MC simulation enforces the 5th-NN separation by construction. The entropy counting is inconsistent with the refined vacancy concentration. These gaps prevent the paper from being accepted in its current form. The electronic-structure part (Dirac-like pocket, Landau diamagnetism) is suggestive but not conclusive.

major comments (4)

[Results - X-ray scattering and 3D Difference Pair Distribution Function] The F-forbidden reflections at 25 K have a mean SCXRD intensity of 0.8 and mean I/σ of 0.4 (main text), far below the conventional detection threshold of I/σ ≥ 2. These very weak peaks are the sole basis for the 3D-ΔPDF-like maps, which are constructed by Fourier transforming only the F-forbidden reflections under the assumption that the VO signal is fully contained in them. This assumption is not tested. The manuscript should provide a quantitative comparison of the measured diffuse scattering (including the regions between sharp peaks) with simulated scattering from the proposed VO models, rather than relying on the selected peaks alone. Without such a comparison, the 5th-NN rule deduced from these maps is not directly supported by the data.
[Results - Monte-Carlo simulation] The MC simulation imposes an interaction hierarchy J1=2J2=4J3=10J4=10^3 J5 designed to maximize vacancy separation. Including terms up to the 4th NN forces the minimum vacancy distance to be the 5th NN; hence the '5th NN rule' is embedded in the model by construction. The agreement of the k=4 model with the 240 K PXRD therefore does not independently validate the rule. The authors should test alternative ordering schemes (e.g., different J-ratio hierarchies, isotropic repulsion, or a random vacancy distribution) and compare their calculated F-forbidden intensities and full diffuse scattering to the experimental data, using a quantitative criterion such as Rwp or a goodness-of-fit parameter.
[Results - Signatures of frustrated vacancy order] The entropy counting is internally inconsistent with the refined vacancy concentration. The calculation S_vd = 4R ln2 = 23.05 J/(K·mol) assumes exactly one vacancy per Cu4 tetrahedron, i.e., 8 vacancies per unit cell, corresponding to 25% vacancies. The refined Cu2 occupancy of ~0.798 gives 6.4 vacancies per unit cell (20% vacancies). The configurational entropy should be recomputed for the actual vacancy count, and the sensitivity of ΔS_cal to the constraint 'one vacancy per tetrahedron' should be discussed. In addition, the experimental ΔS is obtained by integrating Cp/T after subtracting an unspecified linear background; the robustness of the integrated entropy to the background choice should be demonstrated.
[Results - Diamagnetic uncompensated semimetal] The deduction of the large Landau diamagnetism (χ_L ≈ -1.2×10^-4) relies on subtracting the Pauli paramagnetic contribution estimated from n_tot ≈ 2.38×10^21 cm^-3 and m*_av ≈ 17 m_e. The alternative estimate with m* = m_e yields χ_L ≈ -6.15×10^-5, i.e., a factor-of-two uncertainty. The claim of 'unprecedented metallic diamagnetism' is also not supported by Table S4, where several materials show comparable or larger negative susceptibilities. The linear magnetoresistance is interpreted via the Abrikosov model without ruling out classical mechanisms for linear MR in inhomogeneous conductors. While these issues do not directly invalidate the vacancy-ordering story, they affect the paper's broader claim of 'novel quantum properties' and should be toned down or quantified.

minor comments (5)

[Results - 3D Difference Pair Distribution Function] The method described is not the standard 3D-ΔPDF, which removes all Bragg peaks and Fourier transforms the remaining diffuse scattering. The authors should clarify that they are constructing a difference Fourier map from the F-forbidden peaks only, and discuss the limitations of this approach (e.g., loss of off-peak diffuse information).
[Results - Monte-Carlo simulation] The MC simulation details are not fully given: the number of MC steps, the cooling schedule, and how the 5×5×5 supercell with 20% vacancies is initialized should be described in the Methods or Supplementary Information.
[Supplementary - Transport properties] The Supplementary Information contains the placeholder text 'Error! Reference source not found.' in the transport section; this should be corrected.
[Results - Structure and vacancy ordering] The statement that the sharpness of the F-forbidden peaks 'implies a rather long-range correlation' is not quantitatively supported; the FWHM analysis in Fig. 4b is more convincing and should be referenced there.
[Discussion] The phrase 'unprecedented metallic diamagnetism' in the abstract and introduction is too strong in light of the values listed in Table S4 (e.g., LaV2Al20, bismuth); consider rephrasing to 'large metallic diamagnetism'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the 5th-NN rule is an emergent model output tested against independent diffraction and thermodynamic data, not a fitted input.

full rationale

I walked the derivation chain for the vacancy-ordering claim. The Monte Carlo simulation imposes a repulsion hierarchy (J1 = 2J2 = 4J3 = 10J4 = 10^3 J5) that encodes the physical assumption that vacancies repel and want to maximize their separation; the specific 5th-NN distance of 6.2 Å is an emergent consequence of that assumption combined with the 20% vacancy density on the Cu2 network, not a pre-specified input. The 3D-ΔPDF-like maps are built from F-forbidden peaks under the explicitly stated assumption that the VO signal is contained in those peaks; this is an assumption, and the subsequent MC/Rietveld refinement against the full PXRD pattern provides an independent test because the MC model parameters were not fitted to the F-forbidden peak intensities. The entropy comparison uses measured heat capacity and a combinatorial count based on the 5th-NN rule, yielding 4.6 vs 4.3 J/(K·mol), which is a consistency check rather than a definitional identity. The electronic analysis fits the Abrikosov linear-magnetoresistance model to extract Fermi energy and velocity; this is a standard model inference, and the paper does not claim the Dirac parameters were predicted before assuming the model. The cited refs. 7 and 34 are methodological self-citations but are not load-bearing uniqueness arguments. The main weakness is the low I/σ of the F-forbidden peaks (0.4 at 25 K) and the assumption that anharmonic As2 vibrations and other diffuse scattering sources do not contribute to them; this is a data-quality and alternative-model concern, not circularity. Therefore no step reduces by construction to its own input.

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

The central claim rests on fitted occupancies, imposed MC interaction ratios, transport-derived effective masses, and model assumptions about the origin of weak reflections and linear magnetoresistance. No new physical entity is introduced; the 'vacancy as chemical species' framing is conceptual.

free parameters (4)

Cu2 site occupancy = 0.798(3) across 25-350 K
Refined from SCXRD and used as the fixed vacancy concentration in MC simulations and entropy counts; the 20% vacancy level is the central input to the ordering model.
MC interaction ratios J1:J2:J3:J4:J5 = 1 : 1/2 : 1/4 : 1/10 : 1/1000
Chosen so vacancies repel into maximally separated positions; the 5th NN distance (6.2 Å) is effectively enforced by the 10^3 penalty rather than derived from a physical interaction potential.
Average effective mass ratio m*/m_e = about 17 (from gamma = 9 mJ mol-1 K-2 and n2 = 2.38e21 cm-3)
Used to compute Pauli and Landau susceptibilities; the paper notes microstrain in Cp may overestimate gamma and also gives the m* equals m_e alternative.
Abrikosov model parameters (E_F, v_F) = E_F approx 16 meV, v_F approx 3.12e5 m/s
Fitted to the temperature dependence of the linear-magnetoresistance crossover field H* using Eq. S5; used as evidence for Dirac-like pocket 1.

assumptions (4)

domain assumption The F-forbidden Bragg peaks are produced entirely by the vacancy-ordered Cu2 sublattice, with no contribution from anharmonic As2 motion or other diffuse sources.
The 3D-ΔPDF-like map Fourier transforms only F-forbidden peaks; this assumes the VO signal is isolated in those weak reflections. Methods, difference Fourier map paragraph.
domain assumption In the disordered high-temperature state, each Cu4 tetrahedron hosts at most one vacancy.
Used for the entropy count and MC constraints; this prevents nearest-neighbor vacancies but is not independently measured.
domain assumption The Abrikosov quantum linear-magnetoresistance model applies to pocket 1.
Linear magnetoresistance is fitted with Eq. S5 to extract E_F and v_F; alternative classical origins of linear MR are not excluded. SM III.iv.
standard math Landau-Peierls and Pauli free-electron formulae describe the carrier susceptibility with the measured effective mass and carrier density.
Used to derive chi_L as a residual; the paper also considers the m* equals m_e case. SM IV.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Frustrated vacancy ordering creates novel quantum properties in Kutinaite, $\mathrm{Ag}_{6}\mathrm{Cu}_{14.4}\mathrm{As}_7$." pith.science (2026). https://pith.science/paper/3A444HZY

@misc{pith2026250524447,
  author       = {Pith},
  title        = {Pith review of: Frustrated vacancy ordering creates novel quantum properties in Kutinaite, $\mathrmAg_6\mathrmCu_14.4\mathrmAs_7$},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3A444HZY}},
  note         = {Machine review of arXiv:2505.24447}
}

read the original abstract

Ideal crystals are fully ordered, but real-world crystals always contain defects breaking translational symmetry. Random defects in crystals have important implications and they e.g. provide the foundation for semiconductor-based electronic devices. Structurally correlated defects introduce an additional level of complexity, which may lead to novel materials properties, but rationalization of relations between correlated disorder and the emergent material properties are very rare. Here we report that the defect structure of the mineral Kutinaite, Ag6Cu14.4As7, exhibits unprecedented metallic diamagnetism, a hallmark of non-trivial electronic states that require delicate symmetrical protection. Using a combination of X-ray scattering methodologies, simulations, and physical property measurements, we deduced and verified subtle frustrated vacancy ordering of the Cu sublattice when cooling crystals below ~300 K. The vacancy frustration in Kutinaite leads to unique quantum properties, and our study calls for a reconsideration of the role of vacancies as quasi-chemical species in crystals.

Figures

Figures reproduced from arXiv: 2505.24447 by the authors.

**Figure 4.** Thermal cycling effects of disorder. (a) Normalized scale factor and (b) FWHM and coherence length resulting from [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗

**Figure 5.** Non-trivial electronic properties of Kutinaite. (a) Coexistence of metallicity and diamagnetism in Kutinaite. The measured magnetic susceptibility 𝜒𝜒 is negative, and the resistivity ρ monotonically increases with increasing 𝑇𝑇. 𝜒𝜒L is diamagnetic susceptibility after subtracting other magnetic contributions. (b) and (c) show the mobilites and numbers of electron-like (1) and hole-like (2) carrier types resulting fr… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

1 extracted references · 1 canonical work pages

[1]

2 Trueblood, K

1 XD2016 - A Computer Program Package for Multipole Reﬁnement, Topological Analysis of Charge Densities and Evaluation of Intermolecular Energies from Experimental or Theoretical Structure Factors (2016). 2 Trueblood, K. N. et al. Atomic displacement parameter nomenclature - Report of a subcommittee on atomic displacement parameter nomenclature. Acta Crys...

work page doi:10.1107/s0108767396005697 2016

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

2 Trueblood, K

1 XD2016 - A Computer Program Package for Multipole Reﬁnement, Topological Analysis of Charge Densities and Evaluation of Intermolecular Energies from Experimental or Theoretical Structure Factors (2016). 2 Trueblood, K. N. et al. Atomic displacement parameter nomenclature - Report of a subcommittee on atomic displacement parameter nomenclature. Acta Crys...

work page doi:10.1107/s0108767396005697 2016