Structure of [Pd(NH 3 ) 4 ]Cr 2 O 7

Transcription

1 Powder Diffraction 10 (1995) Structure of [Pd(NH 3 ) 4 ]Cr 2 O 7 Y. Laligant and A. Le Bail Laboratoire des Fluorures, CNRS URA 449, Université du Maine, Avenue O. Messiaen, Le Mans Cedex, France 1

2 Abstract The structure of [Pd(NH 3 ) 4 ]Cr 2 O 7 has been determined ab-initio from conventional X-Ray powder diffraction data by the Patterson method. The cell is monoclinic (space group P2 1 /c, Z = 4) with a = 7.771(3) Å, b = (1) Å, c = (4) Å and β = (4). Refinements of 57 parameters by the Rietveld method, using 852 reflections lead to R B = 0.032, R P = and R WP = The structure is built up from PdN 4 square planes linked to Cr 2 O 7 groups by hydrogen bonds. Hydrogen atoms could not be located. Introduction Investigations of the Pd-Cr-O ammine system by Borgna, Moraweck and Fessler (1989) led to the synthesis of the palladium tetrammine diphosphate [Pd(NH 3 ) 4 ]Cr 2 O 7 by precipitation from [Pd(NH 3 ) 4 ]Cl. 2 H2 O and (NH 4 ) 2 Cr 2 O 7. The compound was characterized by EXAFS at both K edges of Pd and Cr, it was concluded that the surrounding of these atoms was retained during formation. A triclinic cell that should be considered as preliminary was proposed from tentative indexing of powder diffraction data. More recently, another compound containing simultaneously palladium and chromium, Pd(NH 3 ) 4 (CrO 4 ), was reported (Laligant, 1993a) showing isolated CrO 4 tetrahedra. The structure determination of the latter compound was performed from single crystal diffraction data. To solve the structure of the palladium tetrammine dichromate, we had no other choice than to make use of the recent developments in ab initio structure determination from powder diffraction data (Cheetham & Wilkinson, 1991, 1992). The study was successful from conventional X-rays, and the results are presented here. Experimental [Pd(NH 3 ) 4 ]Cr 2 O 7 was prepared following the general procedure given by Borgna et al. (1989). Palladium tetrammine chloride Pd(NH 3 ) 4 Cl 2 (Alfa - Johnson Matthey - Specpure) and ammonium dichromate (NH 4 ) 2 Cr 2 O 7 (Prolabo - R.P.) were dissolved in distilled water (respectively g in 10 ml and g in 20 ml). The latter solution was added to the 2

3 former very slowly with stirring (room temperature). A solid microcrystalline yellow-orange product precipitated. It was washed with distilled water and air dried at 80 C for two days. The powder diffraction pattern was collected on a Siemens D501 diffractometer (CuK α ; graphite diffracted beam monochromator ; 36KV, 28mA ; receiving slit : 0.05 ; angular range : (2θ) ; counting time : 42 sec by steps of 0.02 (2θ) ; T = 25 ± 1 C ; no sample rotation). Preferred orientation effects were considerable and almost unavoidable, even by using the McMurdie et al. (1986) sample holder with vertical loading. We used a method which consists in dusting the holder surface through a 63 µm sieve (the microcrystals size is much more smaller but they tend to form brittle agregates that we preferred not to mill). In this way, the preferred orientation was at its minimum, but the granulated surface had bad effects on the reflection line-widths. A compromise was obtained after one final very slight pressure on the sample by using the small celluloid card kindly offered by ICDD as "aide-mémoire" for X-Ray analysis. The X-ray powder diffractogram was confirmed as corresponding to [Pd(NH 3 ) 4 ]Cr 2 O 7 (JCPDS (quasi-) identical files n and ). Data Analysis The reflection positions were estimated by means of the program EVA (available in the Socabim PC software package DIFFRAC-AT supplied by Siemens ; derivative method) after stripping α 2. Auto-indexing was performed by using the program TREOR (Werner, Eriksson and Westdhal, 1985) applied to the first 20 detected lines. The most probable solution was monoclinic and characterized by the conventional figures of merit M 20 = 31 and F 20 = 51 (0.0077, 48) (De Wolff, 1968 ; Smith and Snyder, 1979). Extinction conditions were unambiguously compatible with the P2 1 /c space group. This result was in complete disagreement with the preliminary triclinic cell proposed by Borgna et al. (1989). No relation between this triclinic cell and the present one could be found. Moreover, excluding their first weak reflection at Å, we were able to retrieve our monoclinic cell from the application of TREOR on the 20 next lines, however the c parameter was halved and the figures of merit were poor (M 20 = 11 and F 20 = 19 (0.024, 46)). 3

4 In the (2θ) angular range, a total of 852 structure factors (' F obs ') were extracted by the method of iterating the Rietveld (1967, 1969) decomposition formula (Le Bail, 1988a, 1992) with the ARITB program (in the range, the pattern was almost flat and very noisy in spite of the long measuring time, this range was excluded). Used as well as for synchrotron (Morris et al., 1992) or conventional X-ray data, this ' Fobs ' extraction procedure has led to more than 50 ab initio structure determinations. It was not included as option in other Rietveld-derived softwares than ARITB before 1989 (GSAS, MPROF, FULLPROF). Then, the Patterson method (SHELXS-86 program, Sheldrick, 1986) allowed the location of two palladium and two chromium atoms from a reduced data set (271 reflections selected because they had no neighbouring reflection at less than 0.04 (2θ)). Atomic scattering factors (for neutral atoms) and anomalous dispersion terms were taken from the International Tables for X-ray Crystallography (1974). Refinements of the atomic coordinates (SHELX-76 program, Sheldrick, 1976) followed by Fourier difference synthesis led to the location of all N and O atoms from the same reduced data set, but the final R = 0.30 value was unacceptable. Application of the Rietveld method with the ARIT4 program (Le Bail, 1988a) led to the final result. The preferred orientation along the [100] direction was treated by the March (1932) -Dollase (1986) function. Previous data from Borgna et al. (1989) were also obviously affected by a very large preferred orientation, their most intense reflection would have been the 300 reflection according to our monoclinic cell for which we observed I/I o 25%. On the powder pattern, the full width at half maximum (FWHM) is minimal at 14 (2θ) with and reaches at 90 (2θ). Such values indicate considerable size and/or microstrain broadening effects because the FWHM measured with the same experimental conditions on a well crystallized standard (quartz) is minimal at 45 (2θ) with and takes a value of at 90 (2θ). The quality of the fit with a standard Rietveld treatment (R P < 10%) is an indication that line-broadening intrinsic to the sample is nearly isotropic. X-ray data for the palladium tetrammine dichromate are given in Table 1. The failure of the previous study to identify the correct cell could be due to the fact that, up to 40 (2θ), reflections with l odd do not exceed 2.7% of the most intense and none was located by Borgna et al. (1989). 4

5 Moreover, reflections with k odd are also very weak (less than 6% of the most intense), few are listed in the and JCPDS files. A total of 96 reflections should have been observed in this range, the EVA Socabim automatic peak search option detected 41 of them. On these data, the figure of merit after indexing is F(41) = 39 (0.0094,96). No PDF card was given in the form presented in Table 1. Either from the structure factor extracting stage (structure unknown but cell and eventually space group tentatively proposed) or the final Rietveld refinement stage (structure determined), we can list 'I obs ' values for which we have no corresponding d obs in the usual sense (that we could reasonably estimate/refine from the raw data by various methods not requiring the cell knowledge), we have only d calc. When the structure is unknown, I obs are extractable by whole pattern fitting techniques with cell constraint like in the Pawley (1981) or Le Bail (1992) methods (with the Rietveld 'I obs ' sense in the latter). Of course, several exactly overlapping reflections are generally given the same structure factor by such extracting methods, there is no miracle. 'Observed' intensities are also biased in the Rietveld method because they are obtained by a partition of the raw data according to the calculated intensities (i.e. two superposed reflections are given 'I obs ' values in the same ratio as the I calc ones, whereas the sums of these observed or calculated intensities may be different). In spite of these disadvantages, 'I obs ' are of great value. The R B factor is calculated from them compared to the I calc and they allow to perform more or less efficient Fourier difference syntheses. It is suggested here for the first time that best 'I obs ' for Fourier difference map would even be obtained after a few iterations of the Rietveld decomposition formula instead of one application at the last refinement cycle. Finally, this paper gives at least the 50th demonstration that 'I obs ' extracted by iterating the Rietveld decomposition formula are sufficiently well estimated to attempt a successful structure determination by the direct or Patterson methods. Discussion The diffraction pattern plot is shown on Fig.1(a,b,c). Residual errors appearing at low angles in the profiles are due to difficulties in reproducing their asymetry. On another hand, having not located hydrogen atoms is certainly the cause of residual errors on intensities. The 5

6 atomic coordinates and thermal parameters are reported in Table II. Selected interatomic distances and angles are listed in Table III. As usual for powder structure calculations, one cannot expect a very high precision on the light atoms parameters because of the presence of heavy scatterers. However, the mean Pd 2+ -N and Cr 6+ -O distances are in agreement with the sum of ionic radii (respectively 2.11 and 1.64 Å, Shannon, 1976). In fact, the Pd 2+ -N distances given in recent litterature are significantly shorter than the Shannon's value (2.040 Å in Pd(NH 3 ) 4 (CrO 4 ) and Å in Pd(NH 3 ) 4 (MoO 4 ), Laligant, 1993a,b ; and Å in Pd(NH 3 ) 4 I 8, Tebbe & Freckmann, 1982). Because the hydrogen atoms were not located and the NH 3 groups were not included as rigid body with constraints in the refinement, an hypothesis to explain such long Pd - N distances would be that we could have located the NH 3 electronic gravity center. Argument favouring this hypothesis is the fact that negative isotropic B factors were obtained for the N atoms when refined separately from those of the O atoms. Such an observation was made for N atoms belonging to NH 4 groups in several studies from powder diffraction data in which the H atoms were not located. For instance, this effect was observed for α- and β-(nh 4 ) 2 FeF 5 (Fourquet et al., 1989) and for (NH 4 )VO 2 (HPO 4 ) (Amorós et al., 1988). As in these studies, the missing electron density was tentatively reproduced by allowing the N occupation factor to be refined, whereas the temperature factor was constrained to be that of the oxygen atoms. In the previous EXAFS study (Borgna et al., 1989), the Pd atom was found coordinated to 4 ± 0.2 N atoms located at 2.05 ± 0.01 Å. The EXAFS curve at the Cr K-edge was modeled with 4 ± 0.2 oxygen atoms at 1.52, 1.56, 1.81 and 1.90, all ± 0.01 Å from the Cr atom. In our study, there are two sites for both Pd and Cr atoms, and we cannot claim such a precision on the distances. However, we access to the tridimensional arrangement. Also, a difference exceeding 10% is now evident between our calculated density (Table 2) and the previous measured value (D m = 2.29 ± 0.01). The crystal structure viewed along the a-axis, is shown in Fig.2. The palladium and chromium atoms build up, with the NH 3 groups and O atoms, classical PdN 4 square planes and CrO 4 tetrahedra. The weakness of the reflections with l or k odd is explained from the two special positions occupied by the Pd atoms leading to their no-participation to these reflections (existence of a subcell with half c and b parameters). Moreover, in the c and b 6

7 directions, the Cr 2 O 7 groups are translated at almost 1/2+z and 1/2+y coordinates, showing small displacements that lead to the c and b doubling. Two of the four independent N atoms (those at x 0) are responsible for different orientations of the square planes which are alternately nearly parallel to (ac) and (ab) planes along the b and c axes. Therefore, these two N atoms are the main contributors to the intensities of the reflections with l or k odd. Ideally, the square-planes orientation is such that Pd(NH 3 ) 6 octahedra could be considered to be built by adding two supplementary N atoms with extremely long Pd-N distances. From such a view, one can recognize disconnected perovskite pseudo-octahedra planes parallel to (bc). In spite of the lack of hydrogen atom positions, an examination of the N - O distances and O - N - O and Pd - N - O angles should have allowed the suggestion of the hydrogen bonding scheme. This examination proved to be difficult and ambiguous, so that bifurcated bonds or some disorder cannot be excluded on the basis of our data. The solution to this problem might be obtained from powder neutron diffraction data on a deuterated sample. With the [Pd(NH 3 ) 4 ] 2+ [Cr 2 O 7 ] 2- formulation, the structure may have been related to those of the ACr 2 O 7 compounds (A 2+ =Ba, Sr) (Blum, Averbuch-Pouchot & Guitel, 1979 ; Wilhelmi, 1967). The only evident relation is that in all three compounds, the two CrO 4 tetrahedra of a Cr 2 O 7 group point in almost the same direction, in a semieclipsed configuration (about 30 in the palladium compound). Conclusion A new demonstration of the powerful powder diffraction methodology for material characterization is made by the ab initio structure determination of [Pd(NH 3 ) 4 ]Cr 2 O 7. Due to line-broadening intrinsic to the sample, the high resolution provided by a synchrotron source would not have been more decisive. This result shows that conventional powder diffraction has still some future either with or without the presence of the α 2 doublet component. 7

8 References Amorós, P., Beltrán-Porter, D., Le Bail, A., Férey, G. and Villeneuve, G. (1988). " Crystal Structure of A(VO 2 )(HPO 4 ) (A=NH 4, K, Rb) Solved from X-ray Powder Diffraction," Eur. J. Solid State Inorg. Chem. 25, Blum, D., Averbuch-Pouchot, M.T. and Guitel, J.C. (1979). "Structure du Dichromate de Baryum, Forme α," Acta Crystallogr. B35, Borgna, A., Moraweck, B. and Fessler, P. (1989). "X-ray Diffraction and EXAFS Study of Palladium Tetrammine Dichromate, [Pd(NH 3 ) 4 ]Cr 2 O 7," Powder Diffract. 4, Cheetham, A.K. and Wilkinson, A.P. (1991). "Structure Determination and Refinement with Synchrotron X-ray Powder Diffraction Data," J. Phys. Chem. Solids 52, Cheetham, A.K. and Wilkinson, A.P. (1992). "Synchrotron X-ray and Neutron Diffraction Studies in Solid-State Chemistry," Angew. Chem., Int. Ed. Engl. 31, Dollase, W. A. (1986). "Correction of Intensities for Preferred Orientation in Powder Diffractometry : Application of the March Model," J. Appl. Crystallogr. 19, Fourquet, J.L., Le Bail, A., Duroy, H. and Moron, M.C. (1989). "(NH 4 ) 2 FeF 5 : Crystal Structures of its α and β forms," Eur. J. Solid State Inorg. Chem. 26, "International Tables for X-ray Crystallography," (1974). Vol.IV, Kynoch Press, Birmingham. Laligant, Y. (1993a). "On the First Palladium Chromate : Crystal Structure of Pd(NH 3 ) 4 (CrO 4 )," Eur. J. Solid State Inorg. Chem. 30, Laligant, Y. (1993b). "Crystal Structure of the First Palladium Ammine Molybdate : Pd(NH 3 ) 4 (MoO 4 )," Eur. J. Solid State Inorg. Chem. 30, Le Bail, A. (1988a). ARIT4/ARITB User Guide, Univ. of Maine (France). Le Bail, A., Duroy, H. and Fourquet, J.L. (1988b). "Ab Initio Structure Determination of LiSbWO 6 by X-ray Powder Diffraction," Mat. Res. Bull. 23, Le Bail, A. (1992). "Extracting Structure Factors from Powder Diffraction Data by Iterating Full Pattern Profile Fitting," NIST Special Publication 846, 213. March, A. (1932). "Mathematische Theorie der Regelung nach der Korngestalt bei Affiner Deformation," Z. Kristallogr. 81,

9 McMurdie, H.F., Morris, M.C., Evans, E.H., Paretzkin, B. and Wong-Ng, W. (1986). "Methods of Producing Standard X-ray Diffraction Powder Patterns," Powder Diffract. 1, Morris, R. E., Harrison, W. T. A., Nicol, J. M., Wilkinson, A. P. and Cheetham, A.K. (1992). "Determination of Complex Structures by Combined Neutron and Synchrotron X-ray Powder Diffraction," Nature 359, Pawley, G.S. (1981). "Unit-cell Refinements from Powder Diffraction Scans," J. Appl. Crystallogr. 14, Rietveld, H.M. (1967). "Line Profiles of Neutron Powder-diffraction Peaks for Structure Refinement," Acta Crystallogr. 22, Rietveld, H.M. (1969). " A Profile Refinement for Nuclear and Magnetic Structures," J. Appl. Crystallogr. 2, Shannon, R.D. (1976). "Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides,"Acta Crystallogr. A32, Sheldrick, G.M. (1976). "SHELX-76, A Program for Crystal Structure Determination," Univ. of Cambridge. Sheldrick, G.M. (1986). "SHELXS-86 User Guide," Univ. of Gottingen. Smith, G.S. and Snyder, R.L. (1979). "F N : A Criterion for Rating Powder Diffraction Patterns and Evaluating the Reliability of Powder-Pattern Indexing," J. Appl. Crystallogr. 12, Tebbe, K.F. and Freckmann, B. (1982). "Untersuchungen an Polyhalogeniden, V (1) Darstellung und Kristallstruktur des Tetrammin-palladium(II)-oktaiodids, Pd(NH 3 ) 4 I 8," Z. Naturforsch. B37, Werner, P.E., Eriksson, L. and Westdahl, J. (1985). "TREOR, a Semi-exhaustive Trial-anderror Powder Indexing Program for all Symmetries," J. Appl. Crystallogr. 18, Wilhelmi, K.A. (1967). "The Crystal Structure of Strontium Dichromate, SrCr 2 O 7," Arkiv Kemi 26, Wolf, P.M. de (1968). "A Simplified Criterion for the Reliability of a Powder Pattern Indexing," J. Appl. Crystallogr. 1,

10 Figure Captions Figure 1a,b,c. Oberved (dots), calculated (line) and difference profiles for [Pd(NH 3 ) 4 ]Cr 2 O 7. Figure 2. Projection of the crystal structure of [Pd(NH 3 ) 4 ]Cr 2 O 7 along the a axis. The x coordinates of N and Cr (underlined) atoms are given. The [Pd(NH 3 ) 4 ] square planes are almost parallel either to the (ac) or (ab) planes with the Pd atoms at x = 0, encapsulating the Cr 2 O 7 bitetrahedral groups. 10

11 Table 1. Powder diffraction data of [Pd(NH 3 ) 4 ]Cr 2 O 7 (Cu Kα 1 ; λ = Å). d obs I calc 'I obs ' h k l 2θ obs 2θ calc d obs I calc 'I obs ' h k l 2θ obs 2θ calc < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 < <1 <

12 Table 2. Atomic Parameters and R factors for [Pd(NH 3 ) 4 ]Cr 2 O 7 in space group P2 1 /c (No. 14). B is the Debye-Waller Factor (Å 2 ). Atom Site x y z B # Pd(1) 2a (3) Pd(2) 2c 0 0 1/2 1.48(3) Cr(1) 4e (4) (6) (4) 2.21(7) Cr(2) 4e (4) (5) (4) 2.21(7) N(1)* 4e (17) (10) (17) 3.12(11) N(2)* 4e (26) (19) (19) 3.12(11) N(3)* 4e (17) (15) (10) 3.12(11) N(4)* 4e (26) (18) (19) 3.12(11) O(1) 4e (15) (16) (13) 3.12(11) O(2) 4e (16) (13) (15) 3.12(11) O(3) 4e (17) (18) (13) 3.12(11) O(4) 4e (18) (15) (15) 3.12(11) O(5) 4e (13) (22) (11) 3.12(11) O(6) 4e (18) (17) (14) 3.12(11) O(7) 4e (14) (22) (12) 3.12(11) Cell parameters : a = 7.771(3) Å, b = (1) Å, c = (4) Å, β = (4) Volume : Å 3 D x (g.cm -3 ) Atomic coordinates refined : 39 Zeropoint ( 2θ) : 0.049(2) Preferred orientation : 0.875(2) Profile parameters : U 1 = 0.60(3), V 1 = -0.24(2), W 1 = 0.113(3) (for a=59, l=120 ; see equation U 2 = 0.0, V 2 = 0.08(5), W 2 = 1.26(1) n 1 in Le Bail et al. (1988b)) C = (1) D = (5) Discrepancy factors extracting ' F obs ' : R p = 0.057, R wp = Discrepancy factors from refinement : R p = 0.075, R wp = 0.092, R B = * the N occupancy factor was refined to 1.19(1) to take account empirically of the H atoms. # three B factors were refined : one for Pd atoms, one for Cr atoms and one for N and O atoms the discrepancy factors are the conventional Rietveld ones (calculated after background subtraction and from "peak-only" ranges for R p and R wp ). 12

13 Table 3. Selected Interatomic Distances (Å) and Angles ( ) for [Pd(NH 3 ) 4 ]Cr 2 O 7. Pd(1) square plane : Pd(1) - N(1) = 2.08(1) [x2] N(1) - Pd(1) - N(2) = 92(1) Pd(1) - N(2) = 2.15(2) [x2] N(1) - N(2) = 2.93(2) N(1) - N(2) = 3.05(2) Pd(2) square plane : Pd(2) - N(3) = 2.12(2) [x2] N(3) - Pd(2) - N(4) = 92(1) Pd(2) - N(4) = 2.08(2) [x2] N(3) - N(4) = 2.93(2) N(3) - N(4) = 3.01(2) Cr(1) tetrahedron : Cr(1) - O(1) = 1.52(2) O(1) - Cr(1) - O(2) = 118(1) Cr(1) - O(2) = 1.65(3) O(1) - Cr(1) - O(3) = 106(2) Cr(1) - O(3) = 1.67(2) O(1) - Cr(1) - O(7) = 108(1) Cr(1) - O(7) = 1.82(1) O(2) - Cr(1) - O(3) = 109(2) O(1) - O(2) = 2.71(2) O(2) - Cr(1) - O(7) = 106(1) O(1) - O(3) = 2.55(2) O(3) - Cr(1) - O(7) = 109(2) O(1) - O(7) = 2.71(2) O(2) - O(3) = 2.70(2) O(2) - O(7) = 2.77(1) O(3) - O(7) = 2.84(3) Cr(2) tetrahedron : Cr(2) - O(4) = 1.61(1) O(4) - Cr(2) - O(5) = 109(1) Cr(2) - O(5) = 1.63(1) O(4) - Cr(2) - O(6) = 112(1) Cr(2) - O(6) = 1.68(2) O(4) - Cr(2) - O(7) = 108(1) Cr(2) - O(7) = 1.85(2) O(5) - Cr(2) - O(6) = 113(1) O(4) - O(5) = 2.63(1) O(5) - Cr(2) - O(7) = 102(1) O(4) - O(6) = 2.73(2) O(6) - Cr(2) - O(7) = 112(1) O(4) - O(7) = 2.81(2) O(5) - O(6) = 2.77(2) O(5) - O(7) = 2.71(1) O(6) - O(7) = 2.93(2) Cr - Cr nearest neighbours : Cr(1) - Cr(2) = 3.112(6) N - O nearest neighbours : N(1) - O(5) = 2.95(2) N(2) - O(2) = 2.77(1) N(1) - O(1) = 2.96(2) N(2) - O(6) = 2.93(3) N(3) - O(3) = 3.03(2) N(4) - O(1) = 2.97(3) N(3) - O(5) = 3.08(3) N(4) - O(3) = 3.03(3) 13

14 1 c -28 b

15 Counts x (a) Theta (Degrees) 15

16 Counts x (b) Theta (Degrees) 16

17 Counts x (c) Theta (Degrees) 17