Structures of K2V3O8 and Rb2V3O8 mixed-valence vanadate fresnoites are studied with synchrotron single-crystal diffraction at the Swiss Norwegian Beamlines at the ESRF at low temperatures and ambient pressure. K2V3O8 exhibits a phase transition to an incommensurately modulated structure at about 115 K. At 100 K, the satellite reflections can be indexed with two q vectors: q1 = (α,α,0.5) and q2 (-α,α,0.5), where α ≈ 0.3. Although no mixed satellite reflection are observed, the modulated structure is better described in (3+2) than in (3+1) dimensional space -superspace groups P4bm(αα½)(-αα½)0gg and Cmm2(0β½)s00, respectively. The latter includes two twin domains originating from the loss of the 4-fold axis. The geometries of the VO4 and VO5 building units in both models are rigid and it is mainly slight rotations of these polyhedra and small variation of the inter-mediate K-O distances that are modulated [2] . As a large number of single crystal measurements was performed in the temperature range 88-298 K, the newly developed option of cyclic refinements for single crystal data, which was recently incorporated into Jana2020 [3], was used. The prolonged exposure to the high-brilliance synchrotron beam suppresses the incommensurate phase. The previously postulated phase transition to the incommensurate phase in Rb2V3O8 at 270 K [1] is not observed in our data. One of the reasons could be that the intense radiation also affects the modulation in this material. Our results imply that the detection of weak satellites in incommensurate phases is difficult as the prolonged exposure to a high-brilliance synchrotron beam could lead to the disappearance of subtle modulations. Strategies to collect and analyze single-crystal diffraction data measured with very intense synchrotron radiation using modern lownoise pixel area detectors are discussed.

Prussian blue analogues (PBAs) are versatile metal-cyanide framework materials with a broad range of potential applications. Besides mass transport and mass-storage applications enabled by their microporosity, many PBAs also experience electronic and magnetic bistability [2] . This bistability allows an optical and magnetic switching of PBAs between two distinct states by external stimuli, such as temperature, light, pressure, or electric field. The process responsible for switching between these two electronic configurations involves the transfer of an electron from one metal to another and is known as charge transfer (CT) [2] . Upon cooling from room temperature down to 200K, KCo[Fe]-PBA switches from a cubic paramagnetic Fe III Co II to a cubic diamagnetic Fe II Co III state, with metal-ligand bond shortening by 0.2 Å. CT transitions are known to follow one of two routes: first order or crossover type transition [3] . The first order transition is characterized by the high degree of cooperativity and an abrupt change of the unit cell parameters, while crossover type transition happens through a gradual change of unit cell parameters. Typically, these transitions are considered mutually exclusive. In our work, we report a mixed-type scenario in KCo[Fe]-PBA. The transition begins as the crossover type and continues as the first order transition. We also observe a self-healing effect, a recovery of crystallinity, during the subsequent warming-up cycle.

The Cambridge Structural Database (CSD) and the Crystallography Open Database (COD) contain thousands of structures. Large portions of these databases overlap, often because their entries originate from the same publication. Constructing a list of CSD-COD cross entries is difficult because data can be reformatted in the curation process, many data points do not reliably distinguish or identify crystals, and most ways of finding matches are slow, requiring on the order of 10 10 comparisons. Using geometry-based structural invariants (PDD [1]: Pointwise Distance Distribution and its simplified version AMD [2]: Average Minimum Distance), we compared all-vs-all entries in the CSD and COD, discovering the extent of their overlap for the first time. PDD is invariant under any choices of unit cells; if two crystals are rigidly equivalent, their PDDs are identical. Unlike the density of a crystal, which can coincide for some crystals by chance, PDD is stronger than the Pair Distribution Function and distinguishes all different periodic crystals in the CSD in under one hour. If all atoms are sightly displaced, PDD continuously changes and hence can find near duplicates in noisy data. Using these invariants, we compared 1,214,848 entries from the CSD against 508,392 from COD, taking only 17 minutes on a typical desktop computer. Over 400,000 crystals were matched, an overlap of 33% of the CSD and 80% of COD. We also found a significant overlap of COD with the ICSD (over 50,000 entries), as well as at least minor overlaps with the Materials Project database. We additionally searched for duplicate entries, finding several thousand in both the CSD and COD, many of which are not listed anywhere as known duplicates. The recent 'GNoME' dataset [3] of AI-predicted crystals has over 1,000 duplicates. Figure 1. Left: Proposed workflow for incorporating structural invariants into curation. Right: Venn diagram of the discovered extent of overlap between the CSD, ICSD and COD.