User:Praseodymium-141/Gadolinium sesquichloride

Gadolinium sesquichloride is a lower chloride of gadolinium. It can be obtained by reduction of another lower chloride, Gd5Cl11. It is constructed with Gd6 octahedra sharing trans-edges and with chlorine atoms which bridge adjacent chains. It is a semiconductor, and was the first compound of a lanthanide with oxidation state lower than 2. It can be further reduced to obtain GdClHn.

<!--Introduction Among the rare earth elements the formation of the di-positive cation is relatively common among the halides (except fluorides) existing either in a stoichiometric dihalide or as the (probable) reduced component in various phases with average oxidation states between 2 and 3? The sole known exception to the formation of apparently simple M** ions occurs in the lower chloride of gadolinium for which a chloride to gadolinium ratio of 1.58 ‡ 0.06 has been deduced from analytical and powder pattern data on the purest samples obtainable.? Subsequent magnetic sus- ceptibility studies have shown that the compound exhibits only the 4f7 core magnetism of gadolinium(IlI).* The present paper reports the results of a single crystal X-ray study which establishes that the compound has the simple stoichiometry Gd,Cla and contains a remarkable metal atom chain. A short communication has appeared pre-viously.' Experimental Section Adequate single crystals could not be obtained from the reac- Imployed in the original solation Raaction of plces of the was Reaction of pieces of the metal with the trichloride via the vapor phase in a sealed tantalum tube at 610° for 9 days gave long chunks or bundles of the dark bronze to black product on the metal surface which exhibited exactly the same powder pattern as that obtained previously for Although these bundles were generally longer than optimal, efforts to cut them always led to fraying into "hairs." These were strong and flexible but attempts to cut them led only to still finer "hairs." Finally, several of the untrimmed chunks were mounted in 0.3-mm Lindemann glass capillaries in a drybox with the needle (b) axis parallel to the spindle axis. The single crystal finally selected from among these for data collection was prismatic with approximately pentagonal cross section and dimensions of 70 × 495 p. Other chunks from the same reaction were used for density measurements and for further work to ensure the absence of significant impurities. Emission spectroscopy measurements eliminated the presence of other heavy metals while an election microprobe analysis eliminated extraneous elements heavier than chlorine at the level of a few hundred parts per million and also eliminated oxygen or nitrogen at a higher level. The space group and approximate unit cell parameters were obtained from five levels (hkl, k = 0-4) of cquiinclination Weissenberg photographs taken with Cu Ka radiation. The final unit cell pa- rameters and their standard deviations were obtained by a least-squares fit to the 20 angles of 12 independent reflections whose centers were determined at room temperature by left-right, top-bottom beam splitting on a Hilger-Watts four-circle diffractometer using Mo Ka radiation (X 0.71069 A). Values of each reflection were measured at +20 and -20 and averaged in order to eliminate any error in the instrumental zero. The final values, with estimated standard deviations, are a = 15.237 (4) A, b = 3.896 (1) A, c= 10.179 (3) A, a = y = 90.009, and $ = 117.66 (3)°. The general extinctions for reflections with # + k = 2n indicated the monoclinic space groups C2/m, C2, or Cm. Although a broad distribution of intensities suggested the centric group C2/m, the correct space group was subsequently shown to be Cm (No. 8). A density of 5.14 = 0.3 g cm"" was determined on 36 mg of the crystalline material by micropyenometric methods using chloroform. This value compared quite well with the calculated density of 5.23 g cm"" for the composition GdCI,. so and Z = 8 that were subsequently established. Integrated intensities were measured for the same crystal at 1oom temperature using Z-filtered Mo radiation and the 6-20 scan technique on the automated diffractometer cited earlier. Data were collected for each reflection with thirty 0.4-sec counts separated by 0.01° in 0, with one additional counting added for each degree in 8 to allow for the increase in peak width with 0. Periodic counts were made on three standard peaks to check for instrument and crystal stability. Data were collected for the asymmetric unit HKL + HKL over the range 0 < 0 < 35°. The crystal was a good diffractor and 1331 of the possible 1378 reflections were observed to be greater than 3c g, The standard deviations were computed as previously described.* The data were corrected for background and then for Lorentz-polarization and for absorption (u = 266.7 cm*') using a general polyhedral crystal description. Transmission factors varied from 0.146 to 0.243. Structure Determination and Refinement A three-dimensional Patterson function generated from the intensity data contained peaks only in the sections v= 0 and 0.5, thus indicating a layering in the structure normal to the Short b axis. A structural solution was first sought in the centric space group C2/m. However, the unit cell could not accommodate all ap- parent gadolinium atoms with unit occupancy of the indicated fourfold general positions, and the chloride atoms could not be located by Fourier synthesis. On the other hand a solution was found readily in the acentric space group Cm (the alternate C2 is the same as C2/m with the indicated layering). Diagnosis of gadolinium positions from the Patterson function followed by successive Fourier syntheses yielded four gadolinium and six chlorine atoms, all in twofold general positions. Three cycles of full-matrix least-squares refinement using isotropic thermal parameters gave a conventional rellability index of 0.090 for R = 2 Fo'- fe/2 fol The difference Fourier computed at this point was quite flat indicating that the correct stoichiometry was indeed GdCl,.s Further least-squares refinements were carried out using absorption correc>