Studies of the Torbernite Minerals (Ii): the Crystal Structure of Meta-Torbernitei

THE AMERICAN MINERALOGIST, VOL. 49, NOVEMBER_DECEMBER, 1964

STUDIES OF THE TORBERNITE MINERALS (II): THE CRYSTAL STRUCTURE OF META-TORBERNITEI

M.q.rcorrrRoss,2 H. T. EveNs,JR. AND D. E. Arereuaw, U. S. GeologicalSurwy, Washi.ngton,D. C.

ABSTRAcT

The crystal structure of meta-torbernite, Cu(UOrPODz.8HgO, has been determined and refined by three-dimensional least-squares analysis. The structure consists of infinite (UOrPOt""- sheets isostructural rn'ith those of meta-autunite(I) and abernathyite. Be- tween the sheets lie squares of four water molecules. Two of the four squares per unit ceII coordinate the two copper atoms. The other two squares do not coordinate a cation but are hydrogen bonded together in the same manner as those in the abernathyite structure. Each water molecule of the Cu(H:O)a squares is hydrogen bonded to a water molecule of an adjacent (HzO)a square and to a phosphate oxygen atom. Each water molecule of the (H:O)a seuares is hydrogen bonded to two other water molecules within the square, to an adjacent Cu(H:O)r water molecule, and to a phosphate oxygen atom. The copper atoms are also bonded to two uranyl-oxygen atoms; thus the coordination polyhedra about the copper cations is in the form of an asymmetrical tetragonal dipyramid. X-ray powder data for meta-torbemite are given. Unit-cell data for meta-zeunerite, Cu(UOzAsOr)z.SHzO, are also presented.

IN'rnonucrroN In the first paper of this three-part series (Ross and Evans, 1964r hereafter referred to as Part I) the crystal structures of three members of the torbernite mineral group were given, namely those of abernathyite, NHa(UO2AsOa)' 3HzO, and K(H3O)(UOzAsOa)z'6HzO. The presentpaper (Part II) presentsthe crystal structure of a fourth member of this group, meta-torbernite, Cu(UOrPOn)r.8H2O.A general discussionof the torbernite minerals was given in Part I. In Part III of this series(Ross and Evans, 1965) the crystal chemistry of the torbernite minerals will be discussedin detail.

Cnysrer-rocRApHy oF META-roRBERNrrE A sample of meta-torbernitefrom Schneeberg,Germany, (U. S. Na- tional Museum No. 84318) was used in the crystal-structure study. The crystals are clear, transparent, and bright emerald green in color. The optical propertiesare: uniaxial negative,u:I.626*0.002. The value of € was not determined becauseof the small amount of material available. No birefrigenceis evident in the plane of the plate. Spectrographicanaly- sis of this specimen (K. V. Hazel, analyst) showed >107o ll , t-\To P,

1 Publication authorized by the Director, U. S. Geological Survey. 2 Based on a dissertation submitted by M. Ross to the Department of Geological Sci- ences, Harvard University (19621 in partial fulfillment of the requirements of the degree of Doctor of Philosophy.

1603 1604 -41.ROSS, rI. T. EVANS, lR. AND D. E. APPLEMAN

Cu, and O.14.57o Mn. No other elementswere detected. X-ray powder patterns (Daphne R. Ross, anaiyst) confirmed that the material is meta- torbernite. X-ray single-crystalstudies were made with the Buerger precession camera using zirconium-filtered molybdenum radiation. The spacegroup was determinedfrom the inspectionof the hkl, hkl, hkz, hk3, hk4, hks, 0fr1,and 1ftl photographs. The condition limiting the possiblereflections is: hk0:h*h:2n

The hkl, hk3, and 2ft5 photographsshow distinct 4f m Laue s) mmetry. Thus the spacegroup is P4/n (No. 85). The 4/m Laue symmetry is not apparent inthe hkJ, hh2,hk+,0k1, and 1ft1photographs, and if only these were used for the spacegroup determination it would appear that meta- torbernite possesses4f mmm Laue symmetry and belongs in spacegroup P4/nmm (No. 129). As will be discussedlater, the strong pseudo-s1'm- metry is due to the fact that the uranium, phosphorus,copper, and uranyl oxygen atoms occupy special positions compatible with the higher 4f mmm symmetry. LII *-ray reflectionsare sharp, giving no indication of possible atomic disorder or mechanical distortion in the crystals. The unit-cell and optical data found for meta-torbernite in the present studl', the unit-cell data found for this mineral by Donnay and Donnay (1955), and the data of Makarov and Tobelko (1960) are compared in Table 1. The unit-cell data for the present study were obtained from a least- squaresrefi.nement of the *-ray powder film data listed in Table 2 using a program written by Evans et al. (1963). Donnay and Donnay observed the 4/m Laue symmetry but did not observe the weak 003 and 005 reflections and thus assumedthat meta-torbernite possesseda screw axis parallel to c. As a result they assignedthe space group P42f n. Makarov and Tobelko did not observe the 4/m Laue symmetry and thus assigned the space group PLf nmm to this mineral. The indexed c-ray powder data for meta-torbernite(U.S.N.M. specimenNo. 8431&)are given in'I'able 2.

Cnvsrar Srnucrunn DerpnurNarrou A small tabular crystal measuring0.02 X 0.30X 0.42 mm, mounted with the thin direction parallel to the precessionaxis, was used to collect the hk}, hht, hkz, hk3, hk4, and 2ft5 intensity data. The thinness of the crystal in the c-direction minimizes the absorption error for these net planes (Donnay and Donnay, 1955). To collect the 0kl, lkl, and 2kl data, a pyramidal-shapedcrystal measuring0.05X0.06X0.10 mm was used. The hkl, hk3, and hkS photographs were made with unfiltered molybdenum radiation; all others with zirconium-filtered molybdenum radi- M ETA-TORB ERN I T I] ST RU CT U RE

Tenrr 1. X-nev nxo Oprrcer, Pnopnnrms ol MnrA-rcxeERNrrE. Cu(UOzPOr)g.8H:O

Donnay and Makarov and Present studyl Donnay,1955 | Tobelko,1960

System Tetragonal Tetragonal Tetragonal

o (A) 6.969+ 0.001 6.98 6.95 d (A) 17.306 + 0.005 17.4r tl.zo V (49 840.5 848.2 833.7 Z 2 2 2 LaueGroup 4/m 4/m 4/mmtn Space Group P4/n P$r/n P4/nmm Density, g,/cm3 (Obs) 3.40 3.79 Density, g/cm3 (calc) 3.70 J.O/ 3.73 c

@ 1.626 Forms [100],[001] Locality Schneeberg, Cornwall, Germany England (u.s.N.M.84318)

1 The unit-cell edges were obtained from a least-squares refinement of the X-ray powder film data listed in Table 2. These values are in excellent asreement with the values o:6.963+0.00? A and c: 17.2g7+o.or8 A obtained from single-crystal patterns taken with a quartz-calibrated Buerger precessioncamera.

ation. Lorentz and polarization corrections were applied to the observed intensities by means of a computer program based on the relations given by Waser (1951).No absorptioncorrections were made. The assumptionwas made at first that the structureof the (UOzpOn).^- sheetsof meta-torberniteis identical to that of the (UOzAsOs)^"-sheets of abernathyite(Part I). The four uranium atomswere placed tentatively in two 2c positions;two U at z:0.051, and two U at a:0.551 cyclesof space group P4/n. The four phosphorus atoms were placed in positions 2b and 2a. The 16 water molecules and the 16 phosphate oxygen atoms were placed in four eight-fold positions (8g) at coordinates equivalent to those given the water molecules and arsenate oxygens in abernathyite. The two copperatoms were placedin position 2c at z:0.815. This positioning permits each copper atom to be coordinated by four water mole- culesso as to form squareplanar Cu(H:O)a groups.Thus, with the excep- tion of copper,the tentative structure was given atomic coordinatesiden- tical to those of abernathyite. Instead of first preparing Fourier projections, as was done with the previous compounds (Part I), it was decided to subject the proposed 1606 11. ROSS,H. T. EVANS, IR. AND D. E. APPLEMAN

T.mln 2. X-nev Powoen Dera lon Met.q-toneenNrlrl

I2 d(meas.) d(calc.)a hkl 12 d(meas.) d(calc.)s hkl

17.3r 001 1.940 r.937 2t7 20 8 71* 8.65 002 1.929 305 8 6.48 6.46 101 1.923 009 5 J./J J.ll 003 1.921 32r 5.44* 5.43 102 1 888* 1.886 322 15 4.93+ 4.93 110 |.873 226 4 1.70 4.74 111 1 859 315 4.M 103 1.854 109

IJ 4 .3t* 4.33 004 10 1.838t 1.839 208 4.28 tt2 1.833 323 3.747 113 10 1.809* 1.809 306 20 3.678+ 3.676 104 1.791 119 l6 3.480* 3.48+ 200 9 1.778 r.777 218 3.461 005 12 1.766* 1.765 \)L 3.416 201 t.751 3t6 3 25r lt4 10 1.743* 1.742 400 t6 3.232* 3.232 202 8 r.732 3.100 105 10 1.710+ 1.708 402 8 3.063 3.067 2lr 5 1.658+ 1.659 4t2 3 2.979 2.982 203 .5 1.642 t4 2.93t* 2.932 212 10 1.634* 1.633 11,1,0 2.884 006 6 1.616 2 2.84A 2.832 I l.) 9 1.605* 1.606 326 2.742 2r3 7 1.585 1.583 308 8 2.7t4* 2.7t4 204 10 1.575+ |.574 414 t4 2.667* 2.665 106 10 1.560 1.558 420 t4 2.529+ 2.529 214 12 1.544" 1.544 318 o 2.488* 2.489 1t6 o 1.535 2.472 007 6 1.459* 1.458 416 t2 2 462+ 2.464 220 6 l.M3 2.456 205 11 t.414 2.439 22r I 1 386 13 2.368* 2.370 222 6 l.J/o 2.330 to7 11 1.361 2.316 215 3 1.351* I .350 512 2.304 2.302 301 5 I.JJJ

l CuKo raciiation, Ni filter (\:1.5413 A) Ca*e.a diameter: 114.59 mm. Lower limit 29 measurable:approximatelv 8'(11.0 Ar 2 Intensities were measured with a calibrated intensity strip. 3 d-spacings u'ere calculated from the following unit-cell data: tetragonal, P4fn, o:6.969 A, c:17.306 A. These unit cell data were obtained by least-squaresrefinement of the observed d-spacings marked with an asterisk in the table above. Allcalculated spacings > 1.750 A permitted by the space group are listed" calculated spacings lessthan 1.7.50A are listed only if uniquely indexed to an observed reflection' M ETA-TORBDRN ITN STRUCTURE 1607

T ant-r,2 - (co ntinu eil)

12 d(meas.) c(calc.)S hkl d(meas.) hkl

2.266 223 1.327 2.241t 2.2M 302 1.309 2.222 206 1 280 2.210 tt7 t.264 L4 2.2014 2.204 310 1.255 2.186 311 r.241 2.161+ 2.163 008 1.227 2.155 303 1.218 2.141 a)A r.207 t2 2.137* 2 136 312 1.191 14 2.115* 2.111 216 1.182 11 2.065* 2.066 108 1.172 2.O59 313 1.158 74 2 .045* 2.M7 304 1.150 2.016 207 1.138 2 007 225 14 1.981* 1.981 118 1.964 314 i

structure directly to full-matrix least-squaresanalysis. The \kl d.ata, consisting oI 126 non-zero terms, were subjected to four cycles of refinement using an overall temperature factor. The reliabilitv factor, R : Il tFot-i F"rl/Ir r.t, droppedfrom 30.0to 8.3/6 during this refinement.There weresome large shifts in the positions of the oxygen atoms and small shifts in the heavy atom positions.rnteratomic distancescalculated from thesenew atomic positions seemedplausible and with the low R-factor it appeared that most of the atoms had been correctly positioned. A heavy-atom subtraction map, projected on (100), was then pre- pared by subtracting the calculated contributions of the copper, uranium, and phosphorus atoms to the structure factors from the observed structure factors. This subtraction map showed clearly alr the oxygen atoms in the structure. rt appearedat this point that meta-torbernitehas essentiallythe same structure as abernathyitewith the following exceptions:(1) the centerof two of the four squaresof water moleculesare occupiedby copper; (2) no atomssubstitute for H2O; (3) the 16 water moleculesand 16phosphate oxygens lie in four eight-fold positions instead of two sixteen-fold positions, which permits a different configuration of water moreculesin the two Cu(H2O)4squares than is found in the two (HrO)usquares. Next, the least-squaresrefinement was continued using three-dimen- 1608 rl1..ROSS, IL T. EVANS,JR. AND D. E. APPLEMAN

sional data. The classesof data used in this ref,nement are given in Tabie 3. For this least-squaresanal)-sis the following atomic scatteringfactors wereused: U0 from Thomas and Umeda (1957)reduced unif ormly b1'5.6 electronsfor anomalousdispersion, O0 from Berghuisand others (1955), and P0 and Cuo from Internationale Tabellen (1935). After four cycles of three-dimensional least-squares refinement were completed using an overall temperature factor it was found that the hkl data with I odd would not refine below an R factor of 43/6. Since onll' the copper atoms, the phosphate oxygen atoms, and the water molecules contribute ap- preciably to these structure factors, it was apparent that there was a

Taer-n3. INraxsr:rvDere Usnnron FrNal Lnasr-SQuenns ANe,r,vsrs or Mpta-torgrnNttE

hkl F>0 Ir:0

hho 35 I hhl 9 hk2 79 J hk3 35 11 hk4 59 27 IIRJ 29 9 oktr| 100 OJ lkll 158 179 2kt\ 120 193

672 495

I Redundant reflections not included.

significant error in the positioning of one or more of these at.oms.It was observedthat the water molecule Oz was positioned with an r coordinate of 0.160 cycle. With this parameter,hydrogen-bonding of the Oz water molecules to the 05 phosphate oxygen atoms is not possible. Since in tetragonal structures there is often an ambiguous solution to the r or y parametersif two-dimensionaldata are used, it was thought that the r parameterof Oz shouldbe 0.340,not 0.160cycle. To test this new model, three-dimensional least-squaresanalysis was carried out using only the hkl, hk3, and hk1 data. With this new parameter the refinement of these data proceededsmoothly, the R-factor dropping from 43/6 to 19/6 in nine cycles. Least-squaresanalysis was then continued with the full three-dimensional data for seven more cycles starting with the coordinates and in- dividual temperature factors given by the 9th cycle of the previous refinement. The refinement then proceeded satisfactorily to completion. M ETA-TORBERNI TE STRUCTU RE 1609

The final reliability factor is 9.7/s for the complete set of non-zero three- dimensional data consisting of 672 terms. The final R-factor for the hkl, hk3, and hkS datais 16.0/o. The higher R-factor for these data is to be expectedbecause the structure factors of these reflectionsare very small. Table 4 gives the final atomic parameters for the meta-torbernite structure as given by the last cycle of refinement. Table 5 gives the bond dis- tancesand bond anglesand their standard errors. As mentioned previously, and also as discussedin connection with the abernathyite structure problem, many tetragonal structures give ambiguous solutions to the r or 1 parameters of certain equivalent positions if only two-dimensional data are used. The meta-torbernite structure offers a similar problem. Here Fourier projections on (100) cannot dis- tinguish between the positions r and (j) - x for atoms 05 and 06, and the positions r and (f) - * for atoms 07 and O_e.The low R-factor of t6.O/6lor lhe hkl, hh3, and hh\ data seemedto indicate that the final model, with hydrogenbonds between O3 and Oz,Os and 06, Oz and 07, and 07 and 05, is the correct one. Although the procedure describedabove was expected to give the correct structure model it was thought that an independent test of all the possiblemodels should be made, in view of the possibility of a false minimum in the refinement procedure. The sixteen possible models were tested by taking the parameters given by the last cycle of three-dimensional refinement and calculating the R-factor oI the hkl, hk3, and 2ft5 structure factors. The various models were formed by taking the various combinations of

Q5at r:O.7834or 0.7166(aA), Ooat rc:0.7038or 0.7962(bB), Ozat r:0.3476 or 0.1524(cC), and Qsat s:0.2225or O.2775cycles (dD).

Table 7 gives the various models, the R-factors obtained, and the H- bonds formed by the particular model. As can be seen,the model (abcd) gives by far the lowest reliability factor. A shift of the * parameter of Og by 0.055 cyclesraises the R-factor to 23/6-a significant increasefor such a small displacement of one atom. The other models give even higher R- factors. Four models permit hydrogen bonds between all water molecules and phosphateoxygen atoms but only one, (abcd), gives a low R-factor of 16.070.This test clearly shows that the model (abcd) given by the three- dimensional refinement is the correct one. As an additional test to see if the three-dimensional least-squaresrefinement trulv proceeded to convergence, the correlation matrix was examined for large coeffi.cients-indicativeof interactions between the positional parameters (Geller, 1961). The standard errors of the inter- 1610 M. ROSS, II. T, EVANS, JR, AND D. E. APPLEMAN

Tasln 4. FrNar, Arourc Peneurrrns lNo StAxoenn Ennons lon Mtre.-rorBERNlrE, Cu(UOzPOr):. 8HzO1

Parameters2 Standard Error3

--L

r-4",-L z:O.1564 0.0045 a:1.58 0.95 --L

)-1",-.1 z: 0. 6563 0.0013 B:1.22 084 --!

.,-l z:O.9488 0.0037 B --0.67 0.75 --r

",-1 z:0.M03 0.0041 B: | .54 1.O7 r:0.7834 0.0036 y:0.0802 0.0035 z:O.4466 0.0019 B:1.59 0.42 r:0 7038 0.0029 Y:0.0818 0.0026 z:O.9486 0.0014 B:0.40 028 r:0.3476 0.0039 Y:0.9814 0 0039 z:0.3105 0.0024 B:2.74 0.52 *:O.2225 0.0038 y:0.9768 0.0039 z: 0. 8095 0 0024 B:2.62 0.52 lC:+

",-!. z:o B:O.136 y:+

r Space group P4/n (No.85b), origin at 1. A complete list of the observed and calculated structure factors for meta-torbernite may be found in Ross (1962) or may be obtained by writing the authors. 2 Atomic coordinates in cycles, the temperature factor B in 4'9. 3 The standard errors of the atomic parameters which are refined by the method of least squares are evaluated with the relationships given by Clark et al. (1962, p.213). M ETA-TORBERN I TE STRUCT URE 1611

T mtn 4-(continued.)

Parameters2 Standard Errors --r B:0. 133 0.195 --l

",-l z:0.8099 0.0007 B:1.058 0.151 --l

s:0.0510 0.0002 B:O.7 571 0.0476 --L

z:0.5524 0.0002 B:0.6269 0.0443

atomic distances were computed both with and without the terms in- volving correlations between positional parameters. In no casewas there a significant difference between the two values thus obtained for each standard error. This shows that there is no significant interaction between positional parameters. The possibility of large correlations between temperature and positional parameters was not investigated directly, but the apparent convergenceof all parameters makes such correlations unlikelv.

Dnscnrprrolt or rHE MBre-roneBRNrrE STRUCTURE The structural schemeof meta-torbernite, projected on (100) and (001), is shown in Fig. I and 2, respectively. If we compare these figures to thoseof abernathyite (Figs. 4 and 5 of Part I) we seethat the two structures are quite similar. The most interesting feature of this structure is the copper coordination and the associatedhydrogen-bonding. The two copperatoms in the unit cell lie at the center of squaresof four water molecules,one at +, +, 0.810, and one at ?, t,0.190 cycles.The distance between the copper atom and each water molecule of the square is 1.91 A. l|tti. unusually short distance indicates a rather strong covalent copper-oxygen bond. The water molecules here must be highly polarized with a rather large positive charge directed away from the copper atom. Short copper-water bonds have been discoveredin the compoundsCuFz.2HzO (Geller and Bond, 1958),CuCl2.2H2O (Peterson and Levy, 1957)and CuSeOe.2H:O 1612 M. ROSS,H. T. DVANS, lR. AND D. lL. APPLEMAN

T,lnr.r 5. Borvp DrsraNcos eNo Bolrl ANcres tN Mnre-TonennrTtte

Multiplicity Length (angle)2 Standard Error3

I. Uranyl ions Ur-r-Or-r 1.82A 0.08A Ur-r-Oa-r r.77A 0.07A Or-rOa-r 3.seA 0.1sA Oa-r-Ur-rOr-r 1900 0 Uu-r-Oz-r 1.80A 0.07A Us-r-Oa-r 1.94A 0.07A Oz-r-Osr 3.74A 0.14A OrrU:-r-O:-r 180" 0

II. POn ions Pr-z-Opr 1.51A 0.02A Oe-r-Pr-z-Oo-e 107.6" r.7" Oe-r-Pr-z-Oo-e 110. 4' noo Oe-r-Ope 2.47A 0.04A Oo-r-Oo-a 2.$ A 0.04A Pz-r-Os-r r.s2A O.03A Os-r-Pz-s-Os-s 105.1" 2.20 Os-r-Pz-s-Os o lll.70 1.2" Os-rOs-r 2.sl4 0.06A O:-r-Or-s 2.4rA 0.04A

III. UOr-PO4 environment Ur-r-Opr 233;3A 0.02A Oe-e-Ur-r-Or-r 89.80 0.60 Uz-r-Os-o 23r L 0.03A Os-e-Ur-r-O:-r 89.6" 0.8" Oe-o-Or-r 2.96A, 0.0.sA Oe-e-Oa-t 2.%A 0.0sA Oo-t-O:-r s37A 0.02A Oe-o-Oe-n 3.304 O.ffiA Oo o-Our 3.s44 0.04A Ou-o-Ou-r 2.e2A 0.0sA Os roa-r 3.ffi A 0.0sA Os-rOl-t s.46A o.ffi A O;-rOs-n s.27A 0.03A

I Atomic positions are given in'Iable 6. 'zAll interatomic distancesless than 3.70 A are listecl. 3 The standard errors of the interatomic distances were calculated $/ith a computer program written by D. S. Handwerker which is based on relationships given by Rusing and Levy (1959, p. 4, 5). a Hydrogen bonds. M ETA-TORBERNI TE STRUCTU RE r613

Trsrn 5-(continued)

Bondl Multiplicity Length (angle)' Standard Error3

Os-rP.-r 3.seA 0.014 Ol-r-Pr-r 3.63A 0.01A

IV. Cu environment CurOs-e 8 1.e14 0.03A Os rCurOs-r 179.6" 2.60 OprCurOs-s 8 90 00 0.0" CurOa-r z 2.40A 0.07A Cur-Oz-r 2.66L 0.07A Cuz-Oz-s 2 3.84A O.ffiA

V. HzO environment Oz-o-Oz-n 2.81L4 0.04A Oz-e-Og-e 2.67A4 0.0sA Oz-u-Ou-n 2.77A4 0.0sA Oz:-Oz-e-Oz : 90" 0 Oz-n-Oz-o-Os-a 109.80 1.10 Oz-z-Oz-o-Os-a 122.0" 1.10 Oz-n-Ozo-Os-q 112.0" 1.00 Oz :-Oz-o-Os-r 712.0" I .0" Os-s-Oz-o-O;-a 109.50 1.30 Oe-a-Or-t 2.70L 0.o14 Ops-Oe-: 2.89h4 0 0sA Oa-+-Or:-Os-z 900 0 Os-s-Os-:-Oz-o 85.70 1.10 Oa-z-Oa r-Oz-o t29.7" 0. go Os-r-Os-r-Os-s 97.r" 1.00 Os-z-Os-s-Oo-s 122.6" 0. 80 Or-o-Op:-Ops 108.70 1.20 O's-rOs-a 3.18A 0.04A O'pr_pr_: 3.66A 0.03A Oz-r-Oz-a 3.s0A 0.04A Oz r-Os-z 3.6sA 0.0sA Oz-rOa-s 3.s6A 0.04A Oz-rOr-r 3.33A 0.07A Oz-s-Oa-r 3.00A 0.06A Oz-a-Oz-r 3.28A 0.04A Oz-e-O;-, s.s2L 0.04A Oa-a-Or-r 3.08A 0.06A Os-:-Or-r 3.27A 0.07A Os-rOr-z 3.70A 0.03A Os-'rOs-z 3.36A 0.04A Taer,a 6. Arourc PosrrroNsl ron Arous Lrsrco rN T-qnrc 5

Atom Atom Position

-_! Ir", t_" Urr Oe-l ^ 2t 2 | )t \ Urr r,ltZ Ou-u L_", - ._1 Prz ilr, v-i, z Ouu Ity, 1-r, l-z a_- Prr rc,y, I+z Out ", . Pz-z ilr, y-i, z Oz-r Cur x'i'z Oz-z 7-s,2-y,7-z !r- E_", . Orr s,y,z Oz-s 2 t ^t 2 ), c Or-t r'!'z Oz-..t ilr, y-i, l-z

Ort rc,y, z-l Oz-o ) 2t L On-r x'!,2 Or-r r, y-1, z Orz i'lx,|ly, r-z Ott 1-r,2-y, 1-z -! - q_", Ot-t frt!,z Or-a Z ^t 2 " Os-r 1_* 1_^, 1_. Os-r ?_., * - Os-a t-r, Os-z z i-y, " 1t-1,l-x, O;-r r-t,|i_y,1-z Oi-a ' )t - Osc \fy, r-r,l-z O'r-t Ou-t xr!tz Oo-s t-*, i-y, z

r The equivalent positions are taken ftom Internati,onal Tables, 19.12, p 175, Space grorupP4/n, No.85b.

Tasre 7. TBsr ol rrrE PossrBLE Srnucrunr Molnr,s loR META-ToRBERNTTE UsrNc /241.izA3. eNl hhS o*t

NIodel R-factor H-bonds formedl

abcd 16.07o Oz-Oz,OrOs, OrOa, OrOe ABCD 50.9 OrOz, Or-O3,OrOr, OrOu aBcd 34.0 OrOz, OrOa, OrOu AbCD 43.3 OrOz, OrOa, OrOs Abcd 29.9 OrOz, Oz-Or,OrOu aBCD 44.5 OrOz, Oz-Os,OrOs '.' ABcd ?7 OrOr, OrOi abCD 37.4 OrOr, OrOs abcD 23.1 Oz-Oz,OrO's, OrO; ABCd 423 Oz Oz, Oz-O'a,OrOs abCd 29.3 Oz-Oz,Oz-O's, Os-Oo ABcD 44.5 OrOz, Oz-O'a,OrOe aBcD J/.U Oz-Oz,Oz-O's, OrOs, OrOo AbCd 35.2 OrOz, OrO's, OrO;, OrOr aBCd 362 Oz-Oz,Oz-O's F; AbcD 33.8 OrOz, OrO'a

I The bond lengths for the various possible hydrogen bonds are as follows: Or0: (2.81 A), o7-o8 (2.67 A1, o-o', (2.89 A), o7-o5 (2.77 A), and 08-06 (2.89 A). The values of the interatomic distances for Oz-Os, Oz-O'r, Oz-Os, and OrOo in the unbonded settings are 3.38, 3.56, 3.30 and 3.22 A respectively. M ETA-TORBERN I TE STRUCTU RE

\rr \_-z ;

\ttr-. QnF@\'ttl.\ t\/t t I t

Frc. 1. Projection of the meta-torbernite structure on (100). The dotted lines indicate the uranium-phosphate oxygen bonds, the dashed lines indicate hydrogen bonds, and the solid lines indicate copper-water bonds. Large open circles, uranyl oxygensl smaller open circles, water molecules; small stippled circles, uranium atoms; small solid circles, phosphorus atomsl and small dashed circles, copper atoms. 1616 M. ROSS,H. T. EVANS,JR. AND D. D. APPLE]TAN

(Gattow, 1958).In the very preciselydetermined CuFz. 2HzO and CuCIz '2HrO structures,copper-water bonds were found to be 1.93and 1.925A respectively.A Cu-HzO bond of 1.94 hwas found in CuSeOa.2HrO. The copper atom in meta-torbernite is also bonded to two uranyl oxygensOz and Osat 2.66 and 2.40A respectivelv.The next nearestatom to Cu is Ozat 3.84A. fne copperatom is within0.007A of beingexactly in the plane of the four 03 water molecules. The Oa-Cu-O8angles are 179.6'and 90.0'. The planar coordinationof copper with four bonds of 1.9 to 2.0 A and two additional bondsof greaterlength to form an asymmetrical tetragonal dipyramid has been found in a large number of crystal structures: CuClz.2HzO, CuFz.2HzO, and CuSeOa.2HrOare ex- amples. Eight additional water moleculesform squaresat |, +, 0.311 and f;, f;, 0.689 cycles.The HzO moleculesof thesesquares do not coordinate a cation. They are, however, hydrogen-bonded together as in abernathyite. The Oz-Ozbond distancewithin these souaresis 2.81 A. l|tte Oz w&t€r

Frc. 2. Projection of the meta-torbernite structure on (001). The designationof the various bonds and atoms is the sameas that given in Figure 1. only ths uranyl ions at +0.0510cycles are shown. META-TORBERNITDSTRUCTURE T617 moleculesare also hydrogen-bondedto a phosphateoxygen Ob (2.77 ]) and to an adjacentOs water molecule(2.67 A). The Os water moleculesof the Cu(HzO)aseuares are eachhydrogen-bonded to an Oz water molecule and to an 06 phosphate oxygen atom (2.89 A;. Wittrin the Cu(HzO)a squarethe Os-Oadistance is2.70 A. This short distanceis not causedby hydrogen bonding but is the result of the close attraction of the HzO moleculesto the copper atom. The hydrogen bonds betweenOz and Oz, Oz and Or, O, and Os,and Os and Ooaccount for all of the32 protons in the unit cell. The next nearestatom to the Oz water moleculeis O+ at 3.00 A. The next nearestatom to the Oewater moleculeis Oaat 3.08 A. Thus, no hydrogen bonding between water moleculesand uranl'l oxygen atoms is indicated.About the Oz water molecuieslie two Oz water molecules, one Os water molecule and one 05 ox)gen, in a distorted tetrahedral arrangement. The tetrahedrai angles are 90o, 109.8o,t22.0", It2.0", 112.0", and 109.5o.The average tetrahedral angle is 109.2'. About the Oe water moleculeslie two Oa rvater molecules,an Oz water molecule, an 06 oxygen, and a copper atom. The water molecules and oxygen atoms are grouped about Oa in a highly distorted tetrahedral arrangement.This distortion is due to the fact that Osis bondedonly to the copper atom, the 07 water molecule,and the 06 ox|gen. The Oo-Os-Oz, Cu-Os-Oz,and Cu-Os-O6angles are 108.7o, 113.0o, and 117.7", respectively. The distribution of the positive charge on the copper atoms to the negatively charged (UOzPOn)^"- sheets is accomplished through the copper-uranyl oxygen bonds (Cu-Or, Cu-Oe)and through the various hydrogen bonds. The (UOzPO4)"'- sheet at z:0 contains the Oe uranyl oxygen atoms and the Oophosphate oxygen atoms, whereasthe sheet at s:I containsthe Oz uranyl oxygen atoms and the Osphosphate oxygen atoms. Sincethe Cu-Os bond is shorter (2.40 A) than the Cu-Oz bond (2.66 A) it is to be expected that a greater charge will be distributed through the uranyl oxygensto the sheet which lies at z:O. Each sheet must receive an equal charge; thus a greater positive charge must be transmitted through the hydrogen bonds_tothe sheet at 3:|. This ap- pearsto be so,for the Oz-Or,H-bond (2.77A) is shorterthan the Os-OoH- bond (2.89 A). The hydrogen bonding between the water moleculesand the phosphate oxygen atoms causesa clockwise15o rotation of the tetrahedra at z:0 and a counter-clockwise11o rotation of the tetrahedraat s:l as viewed down the c-axis.The system of hydrogen bonding is shown as dashedlines in Figs. t and2. A similar rotation of tetrahedra was found in the abernathyite structure. The two types of tetrahedra in meta-torbernite are ro- tated through slightly different angles because of the requirements of 1618 ],1.ROSS, H. T. EVANS,]R. AND D. E. APPLEMAN

hydrogen-bonding to the two different types of water molecules,Oz and Os.The Pr-Ooand the Pz-Orbond lengthsof 1.51 and 1.52A, respectively, compare very favorably with other determinations of this interatomic distance' The Poa tetrahedra are distorted. with the horizontar edges somewhatshortened (2.41,2.43 A) and the other edgessomewhat length- ened(2.51, 2.17 A). The averageOr-pz-Os angle is 10g.4oand the average Oo-Pt-Ooangle is 109.0'.The phosphateoxygen-uranium interatomic dis- tancesare2.33 and 2.31 A, in closeagreement with the value of 2.3SA found for the arsenateoxygen-uranium bond in abernathyite.The ur and Uz atoms are not significantly displaced from the plane of the four phosphate oxy-genatoms. The averageuranium-uranyl oxygen bond distance is 1.83A. sioce thesebonds have a large standard error associated.with them (0.07-0.08A), the variation of thlse interatomic distancescannot be consideredsignificant. The value of 194 A obtained for Uz-Or is anomalously large. The as1'mmetricalenvironment of the UOz2+ions found in abernathr.ite is retainedin meta-torbernite.01 is adjacent to four 06 atoms at 2.96 A, and eight rvater molecules(four at 3.33 A, and four at 3.70 A). O, is associatedrvith the copperatom at 2.66A, with four phosphateoxygens at2.92 A, and with eight HzO molecules(four at 3.27 and four at 3.itA). 03 is associatedwith the copper atom at 2.40 h, with four water mole- culesat 3.08 A, and with eight phosphateoxygen atoms (four at 2.93 A and four at 3.37 A). rne o+ atoms are adiacent to four water molecules (3.004), and to eight phosphateoxygen atoms (four at 3.03A and four at 3.46A). Makarov and Tobelko (1960)examined the crystal structure of meta- torbernite with 0fr1intensity data gathered by Weissenbergtechniques. 'lvorkers These confirm the sheetstructure proposedb). Beintema (1938) for these compounds.They also proposethat the copper atoms are coordinated in a square-planararrangement bv four water molecules.The placement of the uranium and phosphorus atoms is essentialll' the same as that found in the presentstudy. r{owever,the detailsof their structure cannot be consideredcorrect for, by not examiningthe hkl reciprocallat- tice plane, thev mistakenly assumedthat 4f mmm Laue symmetry exists and they thus assignedthe mineral to space group P4f nmm instead oI P4/n. This error forced them to propose a false statistical distribution of the two copper atoms over four sitesand also made it impossibleto ob- tain correct positions for the phosphate oxygen atoms and the water molecules.Ross (1963) discussesthe problems encounteredin deter- mining the true symmetries of the torbernite minerals and gives ex- amples of horv seriouserrors can be made in deducing the positions of the light atoms r,vhenthe correctsymmetrv is not found. In the caseof meta- M ETA -TORBIJRN I TE STRU CTU RD l619 torbernite the uranium, phosphorus, copper, and uranyl oxygen atoms occupy positions compatible with the higher 4f mmm Laue svmmetry. The contributions of these atoms to the structure factors so dominate the hkl,reflecti.onswhen l:2nthat the true 4f mLaue symmetr)-iscompletely obscured. Only by carefully examining the reflections where I is odd, to which only the copper atoms, the phosphate oxygen atoms and the water moleculesappreciably contribute, can the true symmetry be observed.

Mera-zeuNpnrrr The structure of the arsenate analogue of meta-torbernite, meta- zeunerite,Cu(UOzAsOn)r'8HrO, has been reportedby Hanic (1960).He

Tarr,r 8. UNrr-cplr- Dar.a. aNn PRopERTTEslon MBra.-zruNERrrE, Cu(UO,AsO.)r.SH:O

Presentstudy Hanic (1960)

System Tetragonal Tetragonal d (A) 7.r2 7.105 , (A) _t/ .4J 17.70t Laue Group 4/m (probable) 4/mmm SpaceGroup P4/n (probable) P4t/nmc V (A) 884.6 893.7 Z 2 z Densitl' (calc.)g/cm3 3 .85 381 1.635 + 0.005

@ 1.650 + 0.005 Forms Iroo],{oor} Locality Majuba Hill, Nevada Rotava, Czechoslovakia (u.s.N.M.R8738) (Nat. Iluseum No 30365) proposes a structure which differs significantly from that of meta-torbernite,reported here. The unit-cell data measuredby Hanic are given in Table 8. The choice of space group P4zf ntnc forces Hanic to distribute two copper atoms over four positions (4d), and two sets of eight water moleculesover two 16-foldpositions (l6h). He confirmsBeintema's proposed structure for the (UOrXOa)""- layer and also finds the copper atoms coordinated by four water molecules in a square-planar arrangement. Hanic also states that hydrogen-bondsare formed between the water moleculesand the uranyl oxygen atoms. He gives values of 1.94 and 1.78A for the uranium-uranyl-oxygenbonds; 2.55and 2.5SA for the copper-uranyloxygen bonds; 2.18 A for the uranium-arsenateoxygen distance; 2.14 h for the copper-waterbond; ancl1.77 A for the arsenic- arsenate oxygen bond. 1620 r11.ROSS, H, T, EVANS, JR. AND D. E. APPLEMAN

It is strongly suspectedthat meta-zeuneriteis isostructuralwith meta- torbernite and that the details of Hanic's structure are incorrect. rn order to check the crystallography found by Hanic, a number of so-calledmeta- zeuneriteswere examined. Only one sample, that from Majuba Hill, Nevada, appearedto contain significant amounts of arsenic. Semiquanti- tative c-rav fluorescence analysis (F. J. Flanagan, analyst) showed major amounts of uranium, copper and arsenic to be present. There is probably a minor amount of phosphorus substituting for arsenic in this sample. We did not have enough material to confirm this analvtically. The unit-cell data as determined by the precessiontechnique are compared to Hanic's data in Table 8. The crystalsexamined are dull-greenin color and showed cracks which may be evidence of an earlier deh_vdra- tion. The conditionlimiting the possiblereflections is: hh0:hlh:2n

The hkt photographsshow evidencef.or 4f m Laue symmetry; thus the probable space group is P4/n-the same as that found for metatorbernite. The present study suggeststhat meta-zeuneriteis isostructural with meta-torbernite, although until crystals of higher quality than those obtainedfor this studv are found the questionwill remain in doubt.

AcrwowlBnGMENTS We wish to thank Daphne R. Ross for preparing and analyzing the o-ray powder patterns; Francis J. Flanagan for the o-ray fluorescence analysis;Katherine V. Hazel for the spectrographicdeterminations; and Professorscliffort Frondel and cornelius s. Hurlbut, Jr. for their interest and advicegiven us during the progressof this study. Rrlrnnxcrs

BrrNtru.l, J. (1938) On the composition and the crystallography of autunite and the meta-autunites. Rec. lrauaur chim. Pays-Bas et Belgique,57, 155-175. BusrNc, W. R. axo H. A. Lrl'v (1959) A crystallographic function and error program for the IBM 7M. Oak Rid.geNatl. Lab. Central Files Metno.59-12-3, Dec. Cr.enr, Joex R., C. L. Cnnrsr aNo D. E. Apprnnneu (1962) Studies of borate minerals (X). The crystal structure of CaBaOa(OH). Acta Cryst.15,207-213. DoNN,l,v,G. eNo J. D. H. Doxn,lv (1955) Contribution to the crystallography of uranium minerals. U. S. Geol. Su.r.teyRept. TEI-507. Eveus, H. T., Jn., D. E. Appr.nlrll' ero D. S. HeNownnrrn (1963) The least squares refinement of crystal unit cells with powder difiraction data by an automatic computer indexing method. Am. Cryst. Assoc.Program, Ann. Meet. 1963 (abs.),4243- G.a.rtow, G. (1958) Die Kristallstruktur von CuSeOa.2H:O (chalkomenit). Acta Cryst. lt, 377-383. Grrlrn, S. (1961) Parameter interaction in least squaresstructure refinement. Acta Cr"tst. 14, 1026-1035. M ET A -TORB]iRN I TD, ST RU CTU RE 1621

W. L. BoNn (1958) Crystal structure of copper fluoride dihydrate, CuFz. 2HzO four. Cl.rem.Phys. 29, 925-930. HaNrc, F. (1960) The crystal structure of meta-zeunerite, Cu(UO)r(AsOs)z.8HzO. Czech. Jow. Phys. Bl0, 169-181. MAxanov, E. S. .q.Nl K. I. Tolnr.xo (1960) Crystal structure of metatorbernite. Dokl. Ahad. Nauk.SSSR, 131, 87-89 (in Russian). Pntrnsox, S. W. e:co H. A. LBw (1957) Proton positions in CuCl:.2HrO by neutron dif- fraction. f our. Chem. Phys. 26, 220-221. Ross, Mar.cor.u (1962) The crystallography, crystal structure, and crystal chemistry of various minerals and compounds belonging to the torbernite mineral group. Ph.D. thesis, Harvard University, Cambridge, Mass. and Ll. S. Geot. Surt:eyopen fle rept. --- (1963) The crl.stallography of meta-autunite (I) . Am. Mineral,.48, 1389-1393. -- AND H. T. Evaxs, ln. (196a) Studies of the torbernite minerals (I): The crystal structure of abernathyite and the structurally related compounds NHr(UO:AsOa) '3H:O and K(HgO)(UO:AsOa)r.6H:O. Am. Mineral.49, 1578 1602. --- ANDH. T. Evaxs, Jn. (1965) Studies of the torbernite minerals (III): Role of the interlayer oxonium, potassium, and ammonium ions, and \yater molecules Arn.. Mineral (in press). Tnolras, L H. arn K. Uuroe (1957) Atomic scattering factors calculated from the 'fhomas-Fermi-Dirac (TFD) atomic model. J our. Chem. P hys. 26, 293-303. Wesrn, J. (1951) Lorentz and polarization correction for the Buerger precessionmethod. Ilet. Sci. Inst 22,567-568.

Manu,scr'ipt receited.,IIarch8, 1961; accepted.Jorpublicotion, June 17, 1964.