1976 93Refinement of the Crystal and Molecular Structure of Potassium Tetra-nitritomercurate(i1) Nitrate by Neutron DiffractionBy Leslie F. Power and John A. King, Chemistry Department, James Cook University of North QueenslandFrank H. Moore, Australian Institute for Nuclear Science and Engineering, Lucas Heights, New SouthTownsville, Queensland 481 1, AustraliaWales, AustraliaA neutron diffraction study of the title compound has been carried out. Crystals are orthorhombic, space groupfnma,Z = 4,a = 12.290(6), b = 10.644(1), c = 9.377(1) A. Full matrix least-square refinements have producedR 0.034. The neutron-diffraction results agree well with results from an earlier X-ray diffraction study. Thestructure consists of K+, [Hg(N02),12-, and (NO,)- as discrete ions, with the four nitrite groups co-ordinated tothe mercury through both oxygens as bidentate chelates.The arrangement of the eight oxygens about the mercurycan be described in terms of a highly distorted square antiprism or alternatively as a highly distorted undecahedron.A NEUTRON diffraction refinement of potassium tetra-nitritomercurate(I1) nitrate has been carried out becausethe presence of heavy atoms gives rise to a high per-centage of scattering in the X-ray method, resulting inlow accuracy of positional co-ordinates of the lightatoms. Also, the presence of four bidentate nitritegroups, giving rise to a highly distorted square antiprismor undecahedron about the mercury ion, presents anarrangement of light oxygen atoms around a large metalion for which only a few structures have been previouslydetermined.A previous X-ray study of the structure of potassiumtetranitritomercurate(I1) nitrate has been carried out.laEXPERIMENTALCrystal D~~U.-K,[H~(NO~)~]NO,, M = 563.95, Ortho-rliombic, a = 12.290(6), b = 10.644(1), G = 9.377(1) A,c ~ P (ref.l a ) . Calculated neutron absorption coefficient,p = 0.75 f 0.1 cm-l.Large single crystals of potassium tetranitritomer-curate(I1) nitrate were grown by slow evaporation of asolution of mercury(I1) nitrate with an excess of potassiumnitrite. A well formed sample 1.28 mm3 in volume wasmounted on a coniputer-controlled four-circle neutrondiffractometer 2TanA a t the Lucas Heights High-IntensityFlux Atomic Reactor. Data were recorded automaticallyby use of a crystal-monochromated neutron beam of wave-length A = 0.981 A.The orthorhombic cell constantswere refined by least-square techniques from the diffracto-meter setting angles observed for 87 reflections well-distributed in reciprocal space. No significant change inthe intensity of one standard reflection, measured a t regularintervals, was observed.Four sets of equivalent reflections were recorded by use ofthe standard w-20 step-scan technique. Estimatedstandard deviations were assigned to individual reflectionson the basis of a least squares polynomial regression of theform: o 2 ~ ( I ) = 2 ,4,F1 to determine A,, where c2&) isthe variance of the mean obtained for each quartet ofequivalent reflections.The analysis was performed on theabsorption-corrected intensities before the application ofthe Lorentz correction. This approach was adopted inorder t o estimate non-counting errors in the data set.t See Notice to Authors No. 7 , in J.C.S. Dalton, 1975, Indexissue.U = 1226.7 A3, D, = 3.14, 2 = 4, D, = 3.0 & 0.1 gSpace group Pnma (No. 62, Di!).Nn = lEquivalent reflections were then averaged to yield a uniqueset of 1 045 reflections, all of which were used in the refine-ment.Structure Refineunerzt.-The atomic co-ordinates from theX-ray analysis la were refined isotropically by block-diagonalleast-squares t o a conventional R of 12% ; introduction ofanisotropic temperature factors for each atom then reducedthis t o 8%.The structure was further refined by use of the Brook-haven full-matrix least-squares programme LINUS withexperimental weights.Neutron scattering lengths usedwere: bg 0.370, bEg 0.266, bN 0.940, bo 0.577 (1O-la cm).The quantity minimized in the refinements was Zw(lFola -1F,12)2. Each reflection was assigned a weight w inverselyproportional t o the estimated variance of the observationIn the last cycle of refinement where 104 parameters werevaried including an isotropic extinction yarameter,lb noparameter varied by >0.001a. The final agreement fac-tors are Rp (= XllFo21 - IFc2~~/Z~Fo2~) 0.049, and R’p( =[Cw(r(lFo21 - F,21)2/2h[Fo4~]d-) 0.052, and the conventionalR factor is 0.034.Positional and thermal parameters are shown in Table 1,and observed and calculated structure factors are listed inSupplementary Publication No.SUP 21558 (10 pp., 1microfiche). f‘w-1 = c2(F2),DISCUSSIONThe neutron refinement has confirmed the gross resultsof the X-ray analysis, Le., the formula of the compoundis K,[Hg(NO,),]NO, with four nitrite groups acting asbidentate ligands. The arrangement of the nitrogen andoxygen atoms about mercury is shown in Figure 1, and aunit-cell projection down the x axis is shown in Figure 2.The mirror plane contains the mercury atom and theatoms 0(4), 0(5), N(2), and N(3) of the [Hg(N0,),I2- ion.Atoms N(l), 0(1), 0(2), and O(3) reflect across this plane,and this results in eight oxygen and four nitrogen atomsfor the complex ion. Atoms N(4) and O(7) of the nitrateion arein the plane, as is K(1). O(6) of the nitrate ionreflects across the plane as does K(2).This can be seenclearly in Figure 2.The dimensions of the complex are listed in Table 2together with the dimensions for the NO3- ion, andcorresponding results from the X-ray diffraction study of(a) D. Hall and R. V. Holland, 1not.g. Chirn. Acta, 1969,3,235;(b) W. H. Zachariasen, Acta Cryst., 1967, 23, 55894 J.C.S. Daltonref. la. There is reasonable agreement between the bond ever the nitrate angles are much closer to the ideal valuedistances. Standard deviation ranges given in ref. l a (120") expected for a discrete ion. Also, the threeare: Hg-0, Hg-N 0.027-0.03 9& and for 0,N-O,N 0-N-0 nitrate angles listed from the neutron study0.0184.058 A.The neutron results show the bond compare more favourably with each other and with theTABLE 1Final atomic co-ordinates and thermal parameters * ( x lo4), with estimated standard deviations in parenthesesX0.012 75(23)0.299 92(18)0.088 64(7)0.037 50(7)0.331 02(10)0.248 22(9)0.022 18(14)0.078 36(13)0.277 16(13)0.070 87(22)0.221 28(12)0.297 61(18)-0.030 19(12)-0.082 29(16)Y0.250 OO(0)0.006 42(16)0.250 OO(0)0.004 73(6)0.250 OO(0)0.260 O O ( 0 )0.250 OO(0)0.028 82(11)0.092 70(11)0.152 26(10)0.250 OO(0)0.250 OO(0)0.149 21(10)0.250 OO(0)z0.599 45(24)0.437 34(17)0.188 18(6)0.303 30(7)0.158 36(11)0.734 04(9)0.176 39(13)0.373 12(12)0.166 61(13)0.031 04(18)0.678 86(12)0.849 03(18)-0.081 15(9)-0.066 lO(17)* The temperature factor is of the form exp[-((pllh2 + pzzk2 + P3J2 + 2P12hk + 2P& + 2P&Z)J.lengths in the nitrate ion to be essentially equal andclose to the accepted value of (1.245 & 0.01) A for aregular environment.% The range of standard devi-ations given in ref.l for 0-"-0 angles is 2.0-3.9" andTABLE 2Molecular bond distances and angles of the[Hg(NO,)Jz- and the NO,- ions(a) Bond lengths (A)Neutron X-ray *1.23 1 (I) 1.221.248( 1) 1.211.232(2) 1.25N(1)-0(1)N(1)-0(2)N(2)-0(3)N(3)-0 (4)N(3)-0(5)Hg-O(l)Hg-0 (2)Hg-0(3)Hg-0(4)Hg-0 (5)Hg-NP)Hg-N(2)Hg-N(3)N(4)-0(7)N(4)-0 (6)1.2 3 6 ( 1) 1.281.2 50 (3) 1.302.494( 1) 2.492.414( 1) 2.462.648 (2) 2.522.5 6 6 (2) 2.582.394(2) 2.342.894(1) 2.862.9 92 (2) 3.042.91 7 (1) 2.961.237 (2) 1.301.236( 1) 1.261 14.34 (09) 118.51 14.72 (1 7) 107.5 0 (3)-N( 2)-0 (3 )N( l)-Hg-N(2) 1 04.64 (0 3) 104.9N(1)-Hg-N( 1) 1 28.86 (04) 131.3N(2)-Hg-N(3) 114.68(05) 113.20 (6)-N (4)-0 ( 7) 1 19.74(04) 121.60(6)-N(4)-0(6') 120.4 1 ( 16) 116.7* From ref.la.(b) Angles (")O(l)-N(1)-0(2)0(4)-N(3)-0(6) 114.84(15) 110.2N( l)-Hg-N(3) 102.3 7 (03) 101.2for N-Hg-N 0.5-1.4". The angles computed from theneutron study generally lie outside these ranges; how-C. C. Addison, N. Logan, S. C. Wallwork, and C . D. Garner,M. I. Kay and B. C. Frazier, Acta Cryst., 1961, 14, 56.M. G. B. Drew and D. Rogers, Chem. Comm., 1965, 476.Qztarl. Rev., 1971, 25, 289.value found for sodium nitrite [114.9(5)0].3 The O-N-0angle in the nitrite ion would be expected to be lessthan the ideal when co-ordinated in a bidentate arrange-ment,4*6 and this is indicated here.Evidence for thisexists also for the bidentate nitrate ion for which a reduc-tion in O-N-0 angle has been noted on co-ordinati~n.~?~The effect of co-ordination upon the structures of thenitrite ions appears slight. The mean of the N-0FIGURE 1 Numbering scheme for the [Hg(N02)J2- ion;thermal ellipsoids scaled a t 50% probabilitydistances involving the oxygen atoms which form theshorter bonds to mercury [1.249(3) A] is greater than themean of the five N-0 distances involving oxygen atomswhich form long bonds to mercury [1.233(1) A]. This5 I.M. Procter and F. S. Stephens, J . Chem. SOC. (A), 1969,1248.6 G. Bergman, jun., and F. ,4. Cotton, Inorg. Chem., 1966, 5,1208, and refs. therein.7 (a) F. A. Cotton and J. G. Bergman. J . Amer. Chem. SOL, 1964,86, 2941; (b) G. A. Jeffrey and D. W. Cruickshank, Quart. Rev.,1963, 4, 3351976 95appears to be significant (A/. 7.8).'~ Nitrate groupsshow this same effect to a lesser degree for uni- and bi-dentate co-ordination through the oxygen atom(s).gAn inverse relation also exists between the nitrito-bondlength to the mercury ion and the 0-N-0 bond angle.Hall and Holland have referred to the eight oxygenatoms of the nitrito-groups as forming a highly distortedsquare antiprism about mercury. The angles N(1)-Hg-N (2) [N (1')-Hg-N (Z)] , N (1)-Hg-N (3) [N (1')-Hg-N(3)], N(Z)-Hg-N(3), and N(1)-Hg-N(1') are 104.54,102.37, 114.68, and 128.86".Hence, if one regards eachnitrito-group as forming only one bond to mercury, withthe bond pointing towards the nitrogen atom, then thefour such bonds form a slightly distorted tetrahedron.The sum of these six angles is (657.36 -J= 0.21)' as com-pared to 656.82", the sum of six angles of size 109.47'.The tendency for constrained pairs of atoms to interactsuch that the mean positions of the pairs are roughly atthe vertices of one of the usual co-ordination polyhedrahas been noted previ~usly.~*~The idealized arrangements of the eight ligand oxygensFIGURE 2plane.the height of the atom from the projection planeUnit-cell projection of the crystal structure on the xyThe thickness of the boundary ellipse is a function ofare shown in Figure 3.In arrangement (a), the squareantiprismatic type (&, s2m) in which the mirror planeis perpendicular to the paper and contains the line(6)-(8), the chelates span the positions (6)-(8) [0(5)-N(3)-0(4)], (3)-(7) and the symmetry-related (2)-(5)[0( 1) -N (1) -0 (2)], and (1)-(4) [0 (3)-N (2)-0(3')].Figure 3(b) shows how the oxygen arrangement can bederived from an undecahedron (Cz,, mm2), containingone rectangular and ten triangular faces. This Figure islabelled similarly to the antiprism, with the mirror planeperpendicular to the paper through (6)-(8) so that thegeometric similarities between the two distorted arrange-ments can be seen.The undecahedron was accorded* G. E. Kimball, J. Chem. Phys., 1940,8, 188.s J. L. Hoard and J. V. Silverton, Inorg. Chem., 1963, 2, 236.lo D. G. Blight and D. L. Kepert, Inorg. Chem., 1972, 11, 1666.major theoretical status by Kimballs but no experi-mental evidence has been forthcoming. The discussionby Hoard and Silverton of the possible stereoisomers ofa tetrakis-bidentate molecule in the square antiprismaticconfiguration assumed that bridging as across (6)-(8)in Figure 3(a) could not occur. Similarly, studiesevaluating the relative merits of the possible stereoiso-mers of eight-co-ordinated metal complexes for identicalFIGURE 3 Relationship of the oxygen arrangement to (a) asquare antiprism, and (b) to an undecahedron. The similarityof the two final arrangements on the right of the Figure can beseen by rotation of the square antiprismatic arrangementclockwise through 90"and symmetric bidentate ligands by mapping the ligand-ligand repulsive energy for a range of normalized ligandbites assume spanning of the polyhedral edges.1°For a perfect bicapped trigonal prism, the triangularfaces [faces 148 and 567 in Figure 3(b)] are parallel,TABLE 3Close approaches (>3.2 A) to the potassium ionsK(l) * * * O(7VII) 2.688 K(2) * * * O(6) 2.894K(l) .* 0(2) 2.821 K(2) * - - 2.936K(l) * * * O(2VI) K(2) * - ;\%I) 2.942K ( l ) . * * 0(6& 2.876 K(2) * - O(61II) 2.947K(l) * . * O(6 ) 2.876 K(2) * * * O(4nI) 2.984K(l) * * * N(1W) 2.926 K(2) * * * O(3) 2.990K(l) .* - N(1II) 2.926 K(2) * * O(7nI) 3.094K(l) . . N(3) 3.041 K(2) * * - O(1nI) 3.164K(l) * . - N(4) 3.167 K(2) * - - 0(6=1) 3.168K(2) - * * O(3III)2.8213.893Roman numeral superscripts refer to the following equivalentpositions, relative to the reference molecule a t x , y, z :I & + x, + - y, Q - 2 v 9 - x , Q + J j , 4 + zVI x , Q - y, zVII & + x, y, 5 - z11 - x , 4 + y, -2111 8 - x , -y, Q + IV - x , -y, --zwhereas an angle of 25.7" is expected between the cor-responding faces in the most-favoured square anti-prism.ll The angle found between these two triangularfaces is 17.70'.Only one plane could be drawn through four of theoxygen atom positions, namely the plane 1457, and this isrequired by symmetry.l1 T. J. Pinnavaia, G. Podolsky, and P. W. Codding, J.C.S.Chem. Comnz.. 1973, 24296 J.C.S. DaltonIt would appear that the distorted square antipris-matic description of the observed stereochemistry is abetter approximation to the observed stereochemistrythan a bicapped trigonal prism.12The closer approaches of the potassium ions are listedin Table 3. K(l) has nine oxygen or nitrogen neigh-bours at (3.2 A while K(2) has ten. K(l) lies in themirror plane and its approaches to atoms not in the planeduplicate. There is no evidence for a particular poly-hedral arrangement about these ions, each occupying amore or less spherical cavity.We acknowledge financial assistance (to L. F. P. andJ. A. K.) and use of equipment fromThe Australian Institutefor Nuclear Science and Engineering, and a grant from TheAustralian Research Grants Committee.[4/2536 Received, 6th December, 19743E. L. Rluetterties, J . Amev. Cheuz. SOC., 1969, 91, 1636