Crystal structure studies of a new series of molybdovanadate polyanions and some related vanadates

*

www

^

Krippl m ^ Uhm -i 3 J-X

Umeå

University

CRYSTAL STRUCTURE STUDIES OF A NEW S ERIES OF MOLYBDOVANADATE POLYANIONS AND SOME RELATED VANADATES

by

A r n e B j ö r n b e r g Department of Inorganic Chemistry

CRYSTAL'STRUCTURE STUDIES OF A NEW SERIES OF MOLYBDOVANADATE POLYANIONS AND SOME RELATED VANADATES

by

A r n e B j ö r n b e r g Department of Inorganic Chemistry

AKADEMISK AVHANDLING

som med tillstånd av rektorsämbetet vid Umeå Universitet för erhållande av filoso­ fie doktorsexamen framlägges till offent­

lig granskning vid Kemiska institutionen, Sal C , LuO, fredagen den 25 april kl 10.00

Author : Arne Björnberg

Address : Department of Inorganic Chemistry, University of Umeå, S-SOV

87

Umeå, Sweden

Abstract : The thesis is a summary and discussion of six papers.

The determination of complexes formed in weakly acidic aqueous solutions containing pentavalent vanadium as well as hexavalent molybdenum has proved diffi cui t due to slow equilibria and 1 imi ted sol ubi 1 i ty of especially the vanadium species. The formation of several different polynuclear complexes with a very varied molybdenum/vanadium ratio also complicates the interpre­

tation of Potentiometrie data.

In order to clarify the picture of complexes formed and p rovide starting points for equilibrium calculations single-crystal X-ray studies were made on crystals obtained from âqueous solutions. In addition, these studies can provide information on bonding conditions and possibly f ormation mechanisms for molybdovanadate polyanions.

Crystals were synthesized by slow evaporation of aqueous solutions. Solu­ tions with varied molybdenum/vanadium ratios and also varied pH values were prepared and used in the synthesis experiments.

The X-ray measurements were performed with Philips PAILRED, Syntex P2^ and Syntex R3 automatic di ffTactometers. All data sets were corrected for ab­ sorption. Five of the structures were solved with heavy-atom methods and one by direct methods. The structures were refined by computer-performed

least-squares methods.

The following crystals were obtained and structurally determined: NaV03•1.89H2O, which contains chains of VO5 trigonal b ipyramids.

Nat^Oy (H2O) 1 e , containing discrete V2O74" anions which are completely

surrounded by sodium-coordinated water molecules. Discrete molybdovanadate polyanions were found in the structures of the compounds Na6Mo6\/2026 (H2O) 16 ,

KyMosVsO^o-SH^O, K8Moi,V8036 - 12H20 and K6 (V2 , Moi 0 ) VOi, 0 • 1 3H20. The last sub­

stance belongs to a class of compounds named 'h eteropoly blues', which con­ tain metal a toms in mixed-valence states, and has one unpaired electron on the polyanion. This compound was also investigated with electron spin re­ sonance spectroscopy.

The bonding configurations of oxygen atoms coordinated to molybdenum or va­ nadium are described and discussed. As the Moj+VsOa68~» MoeVsOi+o7" (which is

an isomer of the Keggin anion but has a quite different structure) and M06V2O266"" anions all cont ain remnants of mononuclear molybdate and vana­

date anions, it seems likely that these polyanions are formed mai nly through the condensation of mononuclear species.

An electrostatic model for the s imulation of bond distances in polyions, starting with perfectly regular idealized model s, is presented.

Key words and phrases: isopolyvanadates, molybdovanadate polyanions, single-crystal X-ray investigations, ESR spectroscopy, heteropoly blues.

The purpose of crystallography in chemistry is

to throw a radiant and glorious light upon that

darkness of matter, which is otherwise punctuated

merely by faintly glowing constants and spectra.

POLYANIONS AND SOME RELATED VANADATES.

ARNE BJÖRNBERG

Department of Inorganic Chemistry, University of Umeå.,

S-901 87 Umeå, Sweden.

This thesis presents a review of the results presented in papers

I-VI. In the text they will be referred to by their roman numerals.

The Crystal Structure of NaVO^* 1.89^0.

Björnberg, A. 6 Hedman, B.

Aota Chem. Soand.

A 31 (1977) 579-584.

Mul t (component Polyanions. 24. The Crystal Structure of Nâj^O^ (H^O)

Björnberg, A.

Aota Chem. Soand.

A 33 (1979) 539-546.

Multicomponent Polyanions. 22. The Molecular and Crystal Structure

of KgMo^VgO^'12^0» a Compound Containing a Structurally New

Hete-ropolyanion.

Björnberg, A.

Aota Cryet.

B 35 (1979)

1989-1995-Multicomponent Polyanions. 28. The Crystal Stru cture of K^MogV^O^Q"*

8H2O, a Compound Containing a Structurally New Potassium-Coordinated

Heteropol yan i on'.

V.

Mul11 component Polyanions. 26. The Crystal Struct ure of

NègMo^V202^(H20)a Compound Containing Sodium-Coordinated

Hexa-molybdodi vanadate Anions.

Björnberg, A.

Aota Cryst.

B 35 (1979)

1995-1999-VI.

Muiticomponent Polyanions. 27. The Crystal Struc ture and ESR Spect­

rum of K^^.MO

jqWOjjq

-13^0, a New "One-Electron Heteropoly Blue".

INTRODUCTION 1

ON PREVIOUS WORK 2

Complexes in aqueous solution 2

Complexes in crystals 5

Isopolymolybdates 5

Isopolyvanadates 6

Vanadates not containing ions included in

the explanation of solution equilibria 9 Heteropolyanions containing Mo or V 10 Reduced polyanions,

viz.

'heteropoly blues' 16

EXPERIMENTAL 17

Syntheses 17

Determination of crystal symmetry, space group

and cell parameters 19

Data collection and reduction 20

Computation methods and structure solution 22

DESCRIPTION OF THE STRUCTURES 24

The metavanadate, (VO, ) 24

4-The di vanadate anion , (V?07 )

8_ '

The Mo^VgO^k anion 25

The MogVj.Oj^ anion 27

The Mo6V2®26^ anion

The (^2'^°10^®40^ a n'o n 29

24

28

DISCUSSION 30

Mo-0 and V-O coordination and bonding 30 The coordination of Na+ and K+ ions to anions 36

ON THE FORMATION MECHANISMS FOR THE ANIONS INVESTIGATED 40 AN E LECTROSTATIC MODEL FOR THE SIMULATION OF BOND

DISTANCES AND ANGLES IN POLYANIONS 41

FUTURE PLANS 45

ACKNOWLEDGEMENTS 46

Appendix 1: The computer program KNUFF 48

1

INTRODUCTION

During the last twelve yea rs a project, initiated by P rofessor Nils Ingri of this department, has been in progress aimed at in­ vestigating three- and four-component equilibria in aqueous solu­ tion. The object has been to determine compositions, equilibrium constants and structures of the complexes formed. Among the

sys-» 2» O- X 2- 2" terns investigated have been H -MoO^ "HPO^ , H -MoO^ -HAsO^ ,

+ 2 - 1 2 3

H -MoO^ -(D-)manni toi (Pettersson , Strandberg , Hedman ), H+-MoOit2"-Si (OH)^, H+-Mo0ij2"-Ge(0H)/t and H^MoO^-HVO^2"

(Pettersson) as well as metal complexes with mixed ligands such as imidazole, THAM and OH , CIO^ (Bruno Lundberg, Staffan

Sjöberg, Willis Forsling, Gun Ivarsson, Inger Granberg, Lars-Olof Öhman).

The initial method employed for these investigations has been emf titrations, the results of which were evaluated by computer. These results consist of the compositions

(p,q,r)

and formation constants for equilibria of the type pH+ + qhoOj^ + rHVO^ £

+ 2- 2

-(H ) (MoO. ) (HVO. ) , but contain no information about th e

p

h

q

H

r

structures of the complexes formed. The equilibrium determina­ tions were therefore completed by spectroscopic measurements (UV,

1+-6

IR, Raman; Lyhamn, Pettersson ) and by crystal s tructure stu­ dies. In the latter X-ray and neutron single-crystal structure determination methods were employed. To determine whether the structures of complexes found in crystals were identical with those of complexes in aqueous solution , large-angle X-ray scat­ tering stu dies were also performed (Johansson, Pettersson,

7- 1 0

Ingri , Lyxell). As well as giving 'fingerprints' o f the complexes, Raman spectra from solutions and crystals further con­

firmed these results, in addition t o supplying information on 11-1 6

vibrational modes in the comple xes (Lyhamn

et al.

). It was shown that the structures of complexes in crystals were indeed

identical with those in aqueous solution. In order to obtain in­ formation on the mechanism of formation and decomposition of complexes, kinetic measurements using stopped-flow techniques were also included in the project (Mellström, Wennerholm,

17-18

Rehnberg, Ingri ). In the project, calorimetric measurements have also recently bee n included (Bo Danielsson).

+ 2-

2-The system H -MoO^ "HVO^ is among those investigated in the project, but, as opposed to the phosphate- and

arsenate-containing systems, results from the emf measurements have not been unambiguous 1 yexpla i ned . To a certain extent this also

applies to the system H+-HV0^ in near-neutral solution (see be­

low). In these syste ms, the aim of the present crystal structure investigation has therefore been to help evaluate the Potentiomet­ rie data and to aid in finding startin g p oints for equilibrium calculat ions.

ON PREVIOUS WORK

Complexes in aqueous solution.

In the studies of complex formation in aqueous solution the

method yielding the most rewarding res ults has been Potentiometrie titration in a NaCl o r NaClO^ ionic medium. Other methods, such as spectrophotometric, ultracent ri fuge and rapid-flow techniques have also been employed. The equilibria of interest to this project were pH+ + rHVO^2" t (H^pCHVQ^2")^, pH+ + çHoO^2" * (H+)p(MoO^2')^ and

M Mo (V I) 0.1 M Mo(V I) l,0r3'4/19,0 » .1 -pH (a) 4 M /

Mo 0 / /

_H9

_U

_59/

_U / / P ! ff 1 » _i in 1 1 1 / *o7

o

2

r

/

~ • ' X

o

(H

O]

* /

/Mo0

2)

(0

H

f

CVi <\J O r-o 2 fO

o~

r-o

S

/

— —

M

O

0

42

~

~

H2Mo04.

1

1

' *

O

_o

2

X J •

' . . . I - ' ' 1 •

1

2

3 '

₄

_{5 6 7} p« (b)

Fig. 1. Distribution (a) and predominance diagrams (b) for the system H+-MoO^ at ionic strength 1= 3 M and 25° C.19 The solid lines in

diagram (b) represent conditions under which the predominant species in adjacent r egions contain equal amounts of Mo(Vl).

pH+ + qMoO^2" + rHVO^2" z (H+)p(MoO^2") (HVO^2")r. The complexes

obtained are thus defined by the integers

p, q

and r3, and to de­

note a specific complex the notation

(p,q,r)

- possibly with

q

or

r

equal to zero - will subsequently be used.

+

2-The system ti -MoO'

. Investigations on this system have been 1 9 - 2 0

performed by Sasaki S Sillen. They reported m ononuclear (1,1,0) and (2,1,0) complexes and heptanuclear

(p,

7,0) complexes with

p

= 8-11 and its one-, two- and three-protonated forms). A large complex, suggested as (3^,19,0), was also repor­ ted at low pH values. In addition, indications for

(p

,8,0) comp­ lexes were found. These results are largely consistent with those

2 1 2 2

obtained by Byé S Schwing and Aveston

et al.

To confirm or disprove the existence of octamolybdates in aqueous solution a redetermination of this system was undertaken by Lage Pettersson

2 3

of this department. Results obtained are largely consistent with the previous model, and octamolybdates were indeed detect ed. The ratio of heptamolybdate to octamolybdate seems to be strong­

ly dependent on the concent rat i on of the ionic medium.

Fig. 1 shows distribution and p redominance diagrams for the hyd-2 3 - 2 4

rolysis of Mo(VI).

The system ït-RVO^

. The hydrolysis of vanadium(V) has been

subject to a number of investigations employing different methods. For the acid region the generally accepted model is

2 5

that of Rossotti S Rossott I who explained their data by the formation of a (3,0,1) complex (V0.2+) and a series of

(p

,0,10)

6-complexes with

p

= 14- 16 (V ^ 2 g and its one- and two-protona-ted forms). A subsequent investigation indicates the existence

1CT/»V(£) ' ' I ' I 1 L ' ( 1 I ' I 'VC 026(0H^-\ /V ^0^27^*"* '5

W£\

rVw026(0< >V!0O27(OH?' V03(0H)i

Wur VO(OH) \JvQ?(W r\—\^Cfe(OH )v2°7j 3

(al

0

- 1 - 2 > e

-3

o» o -4

-5

-6

0

2

4

6

8

10

12

14

pH rij

2. Distribution (a) and predominance diagrams (fc) for the system

+ o 21+

H -HVO^ at ionic strength 2= 1 m and 25 C. The solid lines in diagram (b) represent conditions under which the predominant species

in adjacent regions contain equal amounts of V(V), and the dashed line represents the solubility of ^0^ in terms of V(V) concentration.

4- 3-VO, 2-VO* (OH) VO

Q - 2 3

of the (13>0,10) complex (H^ QQ28 ^ re9'o n °f pH =

2 e

= 7.5-II was investigated by Ingri & Brito, who reported the ^J^pecies (1,0,2), , and (0,0,1) to be dominant on the more

basic side of the region. In still more basic solutions Newman

2 7

3-et al

. reported (-1,0,1), VO^ . These results have been con­

firmed by the more recent investi gat i on by Borgen, Mahmoud S

2 8

Skauvik. Newman

et al

. also reported th e (0,0,2) complex, ^2^7^ ' 'n m o r e concentrated ([v] > 0.1 M) solutions. Neutral

solutions, with an average charge per vanadium, s, between -0.6 and -1.0 (by Ingri & Brito named the 1 instab i 1 i ty range1) show

very slow equilibria. Depending upon the V concentration, and upon th e concentrât ion of the ionic medium, the dominant species are generally believed to be (1,0,1), (3,0,3) or (4,0,4)

/ - 3- 4- X 2 7

(H^VO^ , V^Og anc^ ^4^12 respectively). Distribution and

predominance diagrams for the hydrolysis of V(V) are shown in Fig. 2.

-

+

2

—

2"

The' system H -MoO^ -EVO^

. For this system extensive

poten-2 3 \"t i ométr i c data have been measured in the region pH = 1.5-7.5.

Owing ta.sol ubi 1 i ty conditions, especially for an excess of va­ nadium, the concentration range available is rather narro w. In addition, the emf effects are not very large. There are also

two areas having very slow equilibria. Of these, one can be explained by the slow equilibria in the 'instability range', -0.6

> z

> -1.0 (5 < pH < 7), of the H+-HV0^ system. In this

area, however, formation of three-component complexes is rapid, which makes ti trations at an excess of molybdenum (Mo/V > 1) feasible. The other instability area (at 2 < pH < 3.5) concerns the formation of acidic three-component completes.

(a) ( b )

(o)

(dì

(

e)

< f > ( g J

Fig. 3. Isopolymolybdates found in crystal struc tures. Apart from the Structures of

Figs, b

and

i

(From melts) and

g

(water-ethanol) the anions

2 - 1 h 6 2 - 1

shown all originate from aqueous solution,

a)

MoO^

^ b )

*

2- 1 U 3 6- 11+2 li- 11+8 6 2 9

o)

HO

6

0,

9

>

d)

Mo^ > e) MOG0

26

f)

MO

8

0

26

(CH)2

4-

7

6

g) MOgC^g • Rings or black dots on the corners of coordination polyhedra forthwith indicate the positions of OH ions or H^O molecules (except in Figs. 4a-d).

The conditions described have caused great d i ff i eu 11 i es in ex­ plaining the data from titrations. The formation Qf

three-component complexes starts at pH ~ 7, and there are a number of complex combinations that give almost equally good explanations for data at pH = '5"7- The se all have a nucleari ty Mo + V ~ 7 and a Mo/V ratio < 1. The combination of complexes that has best fitted data s o far contains the species (9,3,4), (10,3,4) and possibly traces of (11,4,4). In more acidic solutions (pH > 3-5) data indicate the presence of a series of complexes with Mo/V = = 3, the spec i es h itherto giving the best f it being (11,6,2) and i 12,6,2).

For the region pH < 3-5 calculations on the available data have not been completed.

Complexes -in c rystals.

Isopolymolybdates

. The different mono- and polynuclear molybda-tes have been given much attention by structural chemists. As

2-could be expected, the dominant species in solution - the MoO^ ion and the hepta- and octamolybdates - have been fou nd also in crystals. For a further discussion on isopolymolybdates in crystals, the interested reader is referred to the re view given

3

by Hedman. After the publication of this review, two new isopo-lymolybdate anions have been fou nd in crystals:

The compound (C^H ^N) ^^MogC^g * 2^0, which was crystallized from an aqueous solution of i sopropy1 ami ne and ammonium heptamolybda-te, was structurally determined by Isobe, Marumo, Yamase &

2 9 6

-Ikawa. It contains discrete MogC^^OH)^ anions (Fig.

3f)

( h )

(i)

M0I2 Mo13j

(k)

(i-

30

li-

31

.

O I

1

»

9

•

Fig. Z oont. h) {MOg02^)n , i) (MOg02^)n ' ^ Moio03it ' ' ^

8- 32

Mo

36°112<

H

2

0)

16 . Throughout this thesis,

the symbol t in the legend

of a figure indicates that the

polyhedra of the figure are

not

idealized.

3 0

chain structures of (NH^) ^MogC^y * 4^0 (Fig. J>h) , where the octamolybdate units are linked by sha ring corners. The same

oc-3 1

tamolybdate units are found in K^MogC^ß (Fig. 3i) where they form chains by sharin g edges.

From a 0.2 M K^MoO^ solution acidified with an equal a mount of 3 2

O.k M HNO^ Krebs and Paulat-Böschen recently isolated and d e­ termined the structure of the compound KgfMo^O^^ (H20)16]-36H2°.

The anion of this compound, which would have the

(p,q,r)

nota­ tion (64,36,0) is a discrete globular unit consisting of two 18-molybdate subunits which are connected

via

four common 0 atoms (Fig. 3k). It contains sixteen Mo-coordinated water molecules, of which four are bonded to two Mo atoms. Another unusual fea­ ture is the four seven -coordinated Mo atoms in the anion. The MoOy polyhedra are fairly regular pentagonal bipyramids, with all five Mo-0 bonds in the penta gon close to 1.98 Å. The apical Mo-0 bonds are - 1.67 Å (terminal 0 atom) and ~ 2.38 Å (0 atom shared between three Mo atoms). This anion is closely related t o

19

the (3^,19,0) complex suggested by Sasaki S Sillen from Poten­ tiometrie data.

Isopolyvanadates.

The isopolyvanadates have also had their

fair share of attention from structural chemists. The 19501 s and

601 s saw many structure determinations of these compounds, but

comparatively few novel vanadate structures have been determined more recently. Although not all of the compounds described below have been obtained from aqueous solution, they will be taken in an order of increasing z, which corresponds to starting at a high pH value and going towards more acidic solutions.

( b )

(e)

(o)

(d)

(f)

(a)

Fig, 4, Isopolyvanadates found in crystal structures, all except c origi 3 3 Z 4 -3 6 - 3 8

,

N

3-nating from aqueous solutions,

a)

V'O^

, b) . v

2

0

7

_<

_V0

_3»„

(from oxide melt),

d)

(\l

0,)

, e)

HV.0-,_ 3 n M i l vo2(c2oJt)|" 52.

.3-

3 9

Mononuclear species (orthovanadates). From basic solutions

2-V(V) precipitates as salts containing the VO^ anion (Fig. ka) . This anion is tetrahedral a nd has been found in a large number

3 3 of crystal s tructures. The most wel 1 -hydrated is Na^VO^* 12^0* which contains a small variable amount of OH anions (0-0.25 per vanadate ion). This anion is also found in orthovanadates crystal-1 i zed from melts.

Divanadates (pyrovanadates). As the pH in a vanadate solution

3-

k-is lowered t o 9~12, the VO^ anions dimerize to form 0^ 2 7

anions, (At low concentrations emf data indicate the presence

2-of HVOjj anions, but no compounds containing that anion have

3 -

k-been found in a solid ph ase). Like the VO^ anion, ^2^7 ^as 3 5

also been found in many crystal structures from melts as well 3 6

as from aqueous solutions. The anion consists of two corner-sharing tetrahedra (Fig. kb), usually with an eclipsed configuration.

(V0~ ) (metavanadates). Near-neutral solutions yield crystals 0 n

of so-called metavanadates (2 = -1). These are usually found to contain chain structures. Two different kinds of chains have been found: From aqueous solutions VO^ compounds usually have chain structures built up by VOj. trigonal bipyramids sharing

3 7

edges (Fig. kd)- The z = -1 vanadates crystallized from melts 3 8

contain chains of corner-sharing tetrahedra (Fig. ko). The 3-discrete polyanions predicted from emf measurements, ant*

4-V^0^2 > a r e UP t o this date represented in solid phas e only by

the compound which was crystallized from a

3 9

water-ethanol solution by Fuchs, Mahjour & Pickardt. The anion (Fig. kei consi sts of a ring of four tetrahedra with apices

8

pointing a 1ternatingly up and down from the plane of the ring.

ß_

V-irPoo -

_{1U Zo}

This anion, as well as its one- and two-protonated

2 5

forms, was postulated by Rossotti & Rossotti from Potentio­ metrie data. It has been found in three crystal structures:

kO

C03^1 0^28*1 ^2^ (Swallow, Ahmed £ Barnes ), K2^n2^l0^28

k

1

i+2

(Evans ) and 1®H2® (P u , , m a n )• The anion is built up

from VO^ octahedra, six of which share edges in a 2x3 array. Two additional octahedra are joined from above and two from below this array, sharing their equatorial oxygens with those of the apices of the octahedra in the rect angle (Fig.

kf)

.

The protonated forms of the an»on have not been found

in crystals. However, solid non-crysta11 i ne precipitates of (acridine) (ch i noi i ne) ^2^] 0^28 anc' (tetraP'neny 1

phos-k 3 - ^ 4

phon i um) gC^g have been prepared , but no structure de­ termination of these compounds has been perf ormed.

^2^5' s t r u c t u r e vanadium(V) oxide (2 = 0) has been

re-k 5

peatedly determined, most recently by Bachmann, Ahmed & Barnes. The same chains of VO^ trigonal bipyramids as in the hydrated metavanadate are found in the oxide, but in the ox ide the VO^ polyhedra also share corners in directions perpendicular to the chains, thus forming a three-dimensional framework.

The phases which can be obtained as a sluggish dark brown pre cipitate upon acidification of a vanadate(V)

solu-1+7-1+8

tion^ have not been structurally determined. The powder <+ 9

pattern of V^O^'^O has been recorded, but the pat tern was not i ndexed.

Large dark circles are Cs ions. V atoms are indicated by b lack dots.

Fig. 4if. Tfie layer structure of

Fig. 4j. The unit ^ounc' 'n the

6°

9

VO *. This species, which is found in strongly acidic

solu-5 0

t ions, has been foun d in crystals mainly with organic anions

5 1

and also as the compound VC^ 10^*2^0. Scheidt, Hoard

et

5 2 - 5 H

al.

determined the structures of compounds containing VC^ groups obtained from aqueous solutions with oxalate and EDTA anions. The VO2 group was found t o have a bent configuration with an O-V-O angle of 103.8 0 (oxalate compound; Fig.

bg)

or

IO7

.O

-IO7

.I ° (EDTA compounds). The V-0 distances in the group varied between 1.623 and 1.657 Å.

In solvents other than water, vanadyl(V) complexes with only one terminal oxy gen atom can be obtained,

e.g.

VOfOCH^)^, which

55

was crystallized from CH^OH by Caughlan

et al.

The vanadium atoms in this compound still co ordinate six oxygen atoms, with bond lengt hs 1.51*1-57 Å to the terminal vanadyl oxy gen atom.

Vanadates not containing ions included in the explana­

tion of solution equilibria.

Apart from the ions mentioned ab ove, which are all detected in emf measurements, or closely related t o such species, other va­ nadates have been obtained from aqueous solution, typically at temperatures above 25 °C.

Trivanadates

. Compounds with the formula AV^Og (A = K, Rb, Cs,

56

NH^) were described by Kel mers and the structures of KV^Og and

57

CsVjOg were determined by Evans £ Block. They prepared triva­ nadates by adding 0.4 equivalents of acid t o a metavanadate so­ lution, and crystals were obtained by evaporati on at

6O-8O

°C. The structure consists of layers of edge-sharing VO^ octahedra (Fig.

kh)

, some of which have an extremely long sixth V-0 bond

(2.97 Å) and might be regarded as VO^ square pyramids.

If less than the above mentioned amount of acid i s added to a potassium metavanadate solution, crystals of can be ob­ tained. The crystal structure of this compound was determined by

58

Byström S Evans (Fig.

hi)

. It has an unusual sheet structure with V atoms in VO^ tetrahedra and V0<- square pyramids which

link togeth er by shar ing corners, thus forming pentagonal sub-units within the sheets.

59

Block reported the structure of ^2^6^16' w^'c^ P^c i p i tated as

orange crystals from a 'slightly acid solution of potassium va­ nadate '. The structure was described as a layer structure con­ taining five- and six-coordinated V atoms. The mineral carnotite,

6 0 which was synthesized and investigated by Appleman S Evans

con-tains ^Og units consisting of two VO^ square pyramids sharing a basal e dge with apices pointing in opposite directions (Fig. kj).

Hetevopolyanions containing Mo or V.

There are a number of heteropolyanions where vanadium atoms have replaced other metal ato ms, usually Mo or W, in structures which are also found for vanadium-free compounds. In order to keep the gross weight of this publication at a reasonable level, the dis­ cussion on such compounds will be fairly brief. A more extensive treatise on these and other heteropolyanions more distantly re­ lated to the str uctures described in this work can be found in

6 1 6 2

the reviews by Evans and Weakley. Molybdophosphates are very well described in Refs. 2 and 3/ arid will c onsequently be given a rather superf i c i a 1 t reatment here .

(a) (b)

7_ 6<t Fig. 5.

Heteropol yvanadate ions found in crystals,

a)

NiV^.Oig

\k-

67

t

b) h

6aS6v4

0j

0

Heteropolyvanadates

. In earlier single-crystal studies only two different types of heteropolyanions, where vanadium is more than a replacement for

e.g.

some Mo or W atoms in polymolybdates or -tungstates, have been found:

Flynn £ Pope reported t he salts [MnV^^O^g] #^SH^O and

Ky i V J 3 0 3 gj • 16 H 2 0. The structure of the latter compound (Fig.

5a) was determined by Kobayashi & Sasaki. The anion consists of two units of five edge-sharing VO^ octahedra which are joi­ ned by sharing the four apical oxygen atoms at the 'bottom1 of

the units in the same way as in the W-jqO^ anion. Three more V0g octahedra are attached by sharing ed ges at the 'waist' of the 10-uni t. The heteroatom coordinates six oxygen atoms in an octahedral arrangement in the center of the anion. The same structure type is also found in the mineral s herwoodite, the

6 6 9- r

anion of which Evans and Konnert formulated A I V^O^q [with two vanadium atoms reduced t o V(IV)]. The sherwoodite crystal, being of natura 1;òr i g in, was not chemically pure, and the iden­

tification of this structure as a 14:1 complex is not altogether umambiguous. The apparent four-fold symmetry could, as was the case in the determinat ion by Kobayashi & Sasaki, be the result of rotational dis order of an anion with only three octahedra around the 'waist' instead of the apparent four.

6 7

Durif £ Averbuch-Pouchot recently determined the structure of a compound which they formulate d (NH^) (As^V^C^q) • ^^0. The anion of t h i s structure contains two pairs of edge-sharing VO^ octahedra. These pairs are connected to each other by four AsO^

tet rahedra wi th one more AsO^ tetrahedron attached on the out­ side of each pair (Fig.

5b)-

This is an interesting c ompound:

Fig. 6a.

The

a-

and ß-forms of

the

Mo

^2*®i»o (

Ke

99'

n

) anions.

Fig. 6b.

The MOgXO^(HjO)^ anion.

Fig.

6ö+. The a- and ß-forms of the

Mo

)8

X

2°

^2 anions.

Owing to the low water content, at least some of the six pro­ tons have to be situated on some of the anion oxygen atoms. The fact that the number of AsO^ tetrahedra coincides with the num­ ber of protons is suggestive.

There are, in additi on to these two types of heteropolyvanada-tes, a number of other compounds which have been described in the literature,

e.g.

vanad ophosphates, but very little has been done in the way of structure determination of such compounds.

Heteropolymolybdates.

The heteropolymolybdates, along with the

-tungstates, comprise the most varied and abundant species among the heteropolyanions. Many of them are found t o have si­ milar structures differing only in the kind of hetero atom. Since the oxygen coordination of vanadium represents an inter­ mediate between that of metals and non-metals, both types of he-teropolyanions will be described here. Some frequently encounte­ red structur e types with non-metal heteroatoms are:

n- 6 8

The anion (The 'Keggin structure1; Fig. 6a) has

6 9 7 0 been found and structurally determined with X = Si , Ge ,

2 6 9

P , and A s .It consists of twelve MoO^ octahedra sharing edges and corners forming a 'cage' around a central XO^ tetra­ hedron, which shares all fo ur corners with three Mo atoms.

o 2 ? 7 1

The MOgXO^ (OH^) ^ anion (X = P or As ' ) has an open 'cage1

structure that can be described as a Keggin structure (Fig. 6b\ having had three Mo atoms with their terminal and mut ually sha­ red oxygen atoms removed. The high concentration of negative charge that would have resulted from six terminal oxyg ens pointing in the same direction has been avoided by the

protona-If- 7 8 Fig. 6e. The H^As^Mo^O^q anion. The terminal 0 atoms of the AsO^ tetra-hedra are probably protonized.

ii- 80 Fig. 6ft. The (PhAs)2Mo602ltH20 anion.

2- 79

Fig. 6gf. The CH^AsMo^Q^ an'on«

Fig. 6hf. The structure of Fig. 6i. The ^2^°5®21 an'on-82 The

(6-n)~ 2 , 3 , 8 1

the H P.Mor0<,_ anions. hatched atoms are sulphur,

tion of three of them; the anion contains three Mo-coordinated terminal molecules.

If two such units are joined, disposing of their molecules, the anion ^0^X20^2^ 's obtained. This anion has also been

2 , 6 9

found with X = P or As. The structure type was first

encoun-

6-tered for the W^gP20^2 anion and is consequently often

refer-7 2

red to as the 'Dawson structure1 (Fig. 6c).

6 8 7 2

Both the Keggin and Da wson structures exist in two forms; the a forms, and 8 forms with one Mo^O^ unit or the PMo^O^ half-unit twisted 60 0 compared to the a structure. A structure

it-determination of ß-SiMoW^O^g , where Mo has substituted a W

7 3

atom, was recently reported.

2-With organic ligands on the As atoms, F^AsMo^O^OH anions have been found, with R = CH^, ^Hj- or C^H^. The tetramolybdoarsonate

ion (Fig. 6c?) consists of two pairs of face-sharing octahedra, the pairs being connected by sharing edges. The proton of the

7 - 7 5

anion is not acidic.

With As as heteroatom, two different structure types with a Mo/X ratio of 3 have been determined. The Mo^As2Û2^^ anion is

Z,_ 76

isotypic with the MogC^ anion described by Fuchs & Hartl (Fig. 3g) and consists of six MoO^ octahedra forming a ring by

sharing edges. This ring is capped o n each side by an AsO^

tet-7 tet-7

rahedron. This anion also exists in protonated for ms. The

k-H^As^MOi2^50 a n'o n (Fig. 6e) consists of twelve MoO^ octahedra

sharing edges to form four Mo^O^ units which are attached to

each other by sharing corners, turned inside out compared with those of the Keggin anion. The four AsO^ tetrahedra share corners with

1*»

7 8

three Mo^O^ units each. The protons were not located, but are very pr obably situated on the terminal 0 atoms of thè AsO^ tetra-hedra.

If the As atoms have organic ligands, e.g. methyl or phenyl groups replacing the terminal 0 atom of the AsO^ tetrahedra, other varieties of six-membered rings of MoO^ octahedra appear, with octahedra sharing corners and/or faces instead of edges

7 9 - 8 0

(Figs. 6f-g).

When the AsO^ tètrahedra are replaced by the smaller P0^ tetra­ hedra, a five-membered ring of MoO^ octahedra becomes the more stable configuration (Fig. 6h) . The P2^°5^23 9r ouP w't'1 0, 1, or

2 protons on the terminal 0 atoms of the PO^ tetrahedra has been

2 - 3 , 8 i

found in crystals. The S2M°5®21 anion (Fig. 6i) is very

similar, except that the coordination of the heteroatom is

tri-8 2

gonal pyramidal instead of tetrahedral.

Many of the above structures also exist with vanadium atoms sub­ stituting one or more of the Mo atoms. Anions with a Keggin

structure and formula PV Mo10 0.,n (x = 1-3) have been

descri-X 12-x

h0

6 2

bed, although none of these has been subject to a

single-crystal X-ray investigation. The Dawson anion has also been found with V re placing Mo or with V or Mo replacing W; compositions P0V Wi 0 0,, • „ and P0Mo W1 Q 0/-o with x values 1, 2, 3 and 6

2 x 1 8- x 6 1 - 6 2 2 x 1 0-x 6 2

6 2 8 3

have been reported.

Heteropolymolybdates with metal heteroatoms (in which category Te and I are included becau se of their oxygen coordination) dis­

play a rather more varied series of configurations than those with tetrahedral1 y coordinated hetero atoms. Face-sharing MoO^

8 7 8 8

Fig. 7a.

The TeMo^C^ anion.

Fig. 7d.

The H^Cc^MOjgO^g anion,

9 i

A

\j

n \/Éi

A

\j

n \/Éi

M

M

9 2

Fig. 7e

t. The ^°2

'2^16

an

'

oru octa

'

iec

'

ra

are hatched.

15

octahedra, which in the preceding text have been found onl y in connection with organic ligands, are also found.

8_

The anion CeMo^

2®k2

(F Î9-

Ja)

consists of six pairs of

face-sharing MoO^ octahedra, the pairs being joined by sharing corners. The central icosahedral cav ity of the anion harbours a

8 k

12-coordinated Ce atom. Isomorphous anions containing Th(lV)

6 2

and U(IV) are known.

The structure of the MnMoQ0oo^ anion has been determined by

y

ï>£-8 5 8 6 ,

Allman S d'Amour (NH^ salt) and Weakley (K salt). It con­ tains a central MnO^ octahedron which shares edges with two Mo^O-j^ un'ts (like those found in the Keggin m olecule), one on

top and one below. The remaining three Mo atoms are found in MoO^ octahedra that share one edge with the central MnO^ octa­ hedron and one edge with each of the 'top' and 'bottom' Mo^O^ units. The idealized anion possesses

32

symmetry (Fig.

lb)

.

The six-membered ring of edge-sharing octahedra found in molyb-doarsenates and -arsonates has been found also with octahedrally

8 7 - 8 8

coordinated h etero atoms. Evans determined the structure of

6- 8 9

the anion TeMo^C^ (Fig.

lo)

and Perloff demonstrated the

3-same structure for H^CrMo^C^ , with the hydrogen atoms pre­ sumably bonded t o the oxygen atoms of the central CrO^ octahed­

ron. The corresponding Co complex can also be readily prepared from a solution of cobaltous ion, Mo^O^^ and an oxidizing

9 0

agent such as bromine. If this reaction is carried out in the presence of active charcoal a dimer is formed, the structure of

9 1 6

-which was determined by Evans S Showell. The 0^38 anion (Fig.

Id)

consists of two CoMogC^/j rings having had one

MoO^ octahedron each removed and then joined with a mutual twist so that the CoO^ octahedra share edges. The hydrogen atoms are probably located on the four CoO^ oxygen atoms that only coordi­ nate two Mo atoms.

With \^+ as heteroatom yet another structural type of polyanion

has been found in the ^°2^2^16^ an'On> which was described by

9 2

Mattes, Matz S Sicking. This anion (Fig.

le)

consists of two mutually edge-sharing MoO^ octahedra. The 10^ octahedra each share edges with both MoO^ octahedra. The high formal charge of the iodine atom results in an unusual f eature: the 10^ octahedra have three terminal ox ygen atoms. This would otherwise result in a very high negative charge of the anion, and is not found in polyanions containing 5+ or 6+ central atoms.

Reduced poly anions

3

viz.

r

heteropoly blues'.

The reducibility of certain polyanions to mixed-valence comp­ lexes is well kn own and has been s tudied by a number of

wor-93-91+

kers. The ability to form mixed-valence anions appears to be restricted to polyanions where the metal ato ms are

coordina-9 5

ted to

one

unshared oxygen atom ('Type I' structures) . The two most intensely studied structure types of this kind are the

6 8 7 2

Keggin and Dawson structures. Among the Keggin structures, the anion has been shown to be able to undergo a

9 6

six-electron reduction while still mai ntaining the Keggin

6-structure, as has the anion ^2^12^40 ^'n there is no

9 7

central tetrahedrally coordinated atom). For various Keggin

4-(e.g. the S i Mo^ 2^40 an'on^ anc' ^a w s o n structures even further

9 3

17

unchanged anion structures in such compounds has not been con-f i rmed.

Most of the heteropoly blues reported are formed by the reduc­ tion of Mo(Vl) or W(VI). Molybdates or tungstates containing V(V) in place of Mo or W atoms form 'blues' as readily as the unsubstituted compounds, often with the extra electron(s)

trap-9 8

ped to a certain extent on the V atom(s).

The decavanadate ion, ^ ^ 0^28^* being a Type I struc­ ture, is also capable of reduction. ^x^^0-x^28 9r o uPs w'th

X = 2-7 have been observed, with the x = 3 and x = 7 species

9 9

as the most stable. The anion of the above mentioned mineral

6 6 IV V

9-sherwoodite, AIV2 ^12^0 ' 'S a n o t^e r e x a mple °f a mixed-9 5

valence vanadate. This last anion is what Pope named a Type III str ucture,

i.e.

there are metal atoms in the anion with one as well as metal atoms with two terminal ox ygen atoms.

EXPERIMENTAL

Syntheses.

The preparation of crystals was carried out by slow evaporation of aqueous solutions at room temperature. The compositions of the starting solutions were adjusted to (p,c?,^) values corre­ sponding with those of the complexes proposed from Potentiomet­ rie studies. Series of syntheses were also performed, where the

r

value (for pure vanadates) or

qfv

(= Mo/V) ratio was held constant and the p value (H+ concentration) varied. In addition,

solutions with varied

q/r

ratios were prepared, since it was found that the composi t ion of crystals more often than not dif­

fered f rom that of the starting solution. Thus phase V (with

q/r

= 6:2) could be obtained from solutions with 0.75 <

q/r <

10. Phase III was cr ystallized starting with a solution having a Mo/V ratio of 1:2, which is the same as in the crystals. However, phase III was obtai ned only after recrystal 1 ization ; the first product from the 1:2 solution was phase IV

{q/r

= 8:5).

As could be anticipated, the choice of cation has a large effect upon which complex is crystal 1îzed. A 1:2 solution with Li+ as

cation will not yield crystals containing either of the anions of phases I II or IV. From such a solution crystals containing

Mo-(6-x)- 10

substituted decavanadate anions, Mo Vln 0o O , are obtained.

X 10-x 2o

Compounds of this kind with 1 < x < 2.5 have been isolated from solutions with Mo/V ratios between 1:^4 and 3:^. The three com­ pounds found so far are isostructural with a slight increase in cell d imensions with increasing M o content. The structure deter­ mination is currently under way.

The solutions used for Potentiometrie studies usually could not be utilized f or crystallization experiments because of their high NaClOjj content (- 3.0 M). In these solutions the metal ion concentration is rarely above 0.1 M. Among the phases investiga­ ted only phase I has a solubility low enough to be crystallized from such a solution.

Hitherto, no protonated forms of phases I-VI have been obtained. When t he H+ concentration is raised, molybdovanadate solutions

yield yellow amorphous precipitates. The same kind of precipi­ tate is also formed if solutions containing crystals of phases

cryst-19

a l s h a v e f o r m e d . I n t h i s p r o c e s s p h a s e s i l l a n d I V w i l l d e c o m p ­ ose; phase V is slightly mo re stable and can still be present in crystalline form a few weeks after crystallization.

The phase VI anion is a species fundamentally different from those described in papers l-V. The numerous rep orts of Keggin phosphomolybdovanadates, and the existence of the WgV^O^

i o i

anion, where a V atom has taken the position of the central tetrahedral atom, led to the expectancy of finding the corre­ sponding Mo-V compounds. However, the only crystal containing a Mo-V Keggin anion found so far was produced f rom the solutions of the phase III and IV syntheses, which had been left sealed for a number of weeks. These solutions contain NH^+ ions, which

probably are the reducing agent involved in the reduct ion of V^ that led to the production of phase VI. It seems highly p robable that unredu ced species with the Keggin structure exist in solu­ tion. As long as the V content is low, howev er, the charge of these anions, and consequently the number of cations per anion, is probably to o low for stable crystalline phases to form. The yellow amorphous precipitates mentioned above might well contain unreduced Mo-V Keggin anions, probably with a high Mo/V ratio.

D e t e r m i n a t i o n o f c r y s t a l s y m me t r y > s p a c e g r o u p a n d c e l l p a r a r n e t e r s .

Each crystalline phase was examined by single-crystal Weissenberg and precession methods. Powder film data were also recorded, using a Guinier-Hägg camera and CuKa^ radiation (X = 1.5^051 Å) . All six phases are unstable in air (see below), and powder exposures had to be performed with the sample sealed between two layers of

adhe-(/) -o c 3

o

CL E

O

O

TD

_<D 4-» <D 0> (/) <D > C L. o t/> <D i. <D CL O -o C OJ CO "O fD in >-Q CM o> vO o ^ CD • • • O >» ' ' CM *- CM CM CM ro — _vO ^ • ' t/> • _{00 0\ -cr <Tk ro OO} C Q. _{<D X} CM vO r^. OO v£> <n • O <D CM «— CM CM CM CM

...

u O i. co — U L. •— CO C *0 fö •— O m m-— • •— ro no •M 3 OÌ E 3 3 E 1 0 ) •— U (0 t/> O U V) 0) Q-X) •— _X _ •— -Û ro ro _{O <D L. O O} L. 3 -C X CO -C CL *0 ro CL O </l 0) </) _c in cn Q) 0) o •— L. 1. u u. t _Q 3 _O 3 3 o o ^0 O O *—

_

u o o *0 a> 0) <D ro o _{u u >- >- >- •ö} O) ÛL _ö \

£

O 3 \ ö s \ CD O m \ ro O. u CM vO CM CM 1 — Mr to o> PX, O •O. «S3 -a- CM -3- •— ,,—„ irO O << > ^ CM OO <T\ -d- CM — <D _{E vO OO} LA oo CM LA ta CM OO vO vO <T\ 0) 3 -=r _{CM LA cn} O *•— -3" -d-*— CM CM CM CM CM oo «— cn cn -iT ro *— *— CM r— LA • • • • • • O t— _{-4- 00} ro LA

,

— *— o '— <T\ o •< —' II II II Ii II II <n ca QÛ QQ Ö oa >-L. s <D ro LA *—•«s. V- V LA •M CM ^ *— W' CM CM *— 00 LA ro CM <r CM Q) _OO >•»«' vw •>w' ^r E vO <Tv OO CM «— LA OO ro LA s» <T\ vO vO CM CM CO en ro CM OO -à- cn LA o LA ro ro vO CA L. vo O CM un CM -3- -d- CM *— -^r CM M3 * • • • • • » -• • • t • • • * Q. vO ro OO _{CPk vO} ro OO o CM o o O O *— CM *— +— CM T~" r— v— r— *—» d) II II II II II II II II II II II II II II II O a b ö $ o 3 <3 .o S O _P-» U u •— .— u c CO C c •~" c •— c ro E •— o •— 4-» a ) _{y CD o o u} tn 4 -» 0 co o ó u

>. </>

c _{X c c !o} o <D o 0 u 3 O (/> X X X X h- O o CM X vû ro o o • 00 CM CM O O •«—> X X CM -3" o CM _O CM 7 X O > X _CTN CM • 7 vO X vO o CM O CO ro -d- O T) ' • • O o CM cT C f—' OO LA > r 2 • o > > vO 0 _{Q. o} ro CM -3" OO O CM > 0 o X > E > -d- 2E X vO O _-2 co 9° r^. 05 vO o z z Z

20

s i ve tape. Phase II, which is very rapidl y transformed to solu­ t i o n i f e v e n a c o m p a r a t i v e l y l a r g e c r y s t a l i s l e f t i n o p e n a i r , had to be ground to powder with the sample immersed in carbon tetrachloride, and the X-ray powder exposures were made with the powder still drenched in CCI

For phases I, II I and IV the space group was uniquely determined from the systematic extinctions. For the others, the final choice of space group was determined during the procedure of solving the structure.

The powder data could be used t o accurately determine the unit cell parameters of phases I, II and VI, us ing internal reference substances (Si, a =.5.43088 Å or Pb(N03)2, a = 7.8575 Â; 25 °C)

mixed with the samples for

0

calibration. The powder films for phases III and IV could not be indexed w ith the program used

1 0 2

(PI RUM ) due to their extreme richness of lines. This was also the case with phase V, which is triclinio with all unit cell edges approximately equal in length. The unit cell parameters of p h a s e s I V a n d V w e r e t h e r e f o r e o b t a i n e d i n c o n n e c t i o n w i t h t h e i n ­ tensity measurements on the SYNTEX P2^ di ffTactometer. Unit cell parameters of phase III wer e determined using t he Enraf-Nonius CAD-4 di ffTactometer at the Dept. of Inorganic Chemistry, Univer­ si ty of Uppsala.

During all single-crystal investi gat ions, including intensity mea­ surements, the crystals of all six phases had to be enclosed in a Lindemann glass capillary together with part of the mother liquor.

D a t a c o l l e c t i o n a n d r e d u c t i o n .

0 _— "i 1

•

1 1 i 1 1 t/) O) *— cn _{vO sO -d"} c CD CM O r-* LA CM 03 L> c OJ vO LA LA \0 h- i_ o o O o ₀ 0 '—' "D L. U) -3-vD OO -cr OO C X OO fö X X X X X CM </) i/i 0° *— vO vO -3- 1 D X X X X X n> CD -3* -3- -4- OO *— la LA _{00 CM} CA 1 .«—• r^ _{r^. cn r^ CM} E ^E OA <T\ OA v£> -cr vO ;a. CM O CA OA CM rr\ cn c •— r—•i i—-» r—-1 4J O o v 4-> t— r— _{O i 1 1} Ò) O O o • CO I—J U—J k—-J c o _c/) o 4-» c O o X <ü •— L- in L. c O <D X LA *— _{o O LA r^} o E ö *— CM CM CA •— _ö CM C "O X X X X X X O "—^ — _{4-J nj} E _E OO _*— O oo OO r^. CA LA CM r— CM Q. 4-> CSI CM L. U) i o X X X X X X O i o </> v_ T— o CM LA _{o CM} 1—-X) OJ o *- OA CM 0A T3 C * C C O o •— •M -i-j m u •— _{Ö Ö Ö Ö ö ö} CU "O HJ O 0 O o O O o û£ r 2:

*:

£ 2! O flj 1 4-» (V _O x> CM X vO 0A +-» _o •— _{O CM X—v ' '}

.

c/n CM X 0 O C oo X OO CM -3* O) o CM X O 4-* CM rmm C X o • \0 •— UV CM \D _{o CM , O} oo X . CA -d" O • TD • _{O O CM o""} cva C *— _{00 LA} > 3 • o > > _\ö O oa CM OO 0 CM CL O _> o o 21. > £ > -3- X 2: vO ö o fö fD OO •. r^ <u vO Bs o z Z z

21

phases. This was achieved for phases l-lll with a Philips PAILRED two-circle diffTactometer, having equi - i nel inati on Weissenberg geometry with oo-scan and a fixed scintillation counter detector. These measurements were made with scan speed 1 ° min ^ and back­ ground time kO s at the beginning and end of each scan. Weak re­ flections were scanned twice. Data for phases IV and V were col­

lected with the SYNTEX P2^ au tomatic four-circle di ffTactometer of the Dept. of Inorganic Chemistry, Chalmers University of Techno­ logy and University of Göteborg, Göteborg. The 0/29 scan method was employed and the 20 scan speed was allowed to vary between 2 and 8 ° min ^ depending on the intensity of the measured reflexion. No separate background measurements were made (see below).

For phase VI, data were collected with a SYNTEX R3 automatic four-circle di ffTactometer, also employing the 0/20 scan method. The 20 scan speed was allowed to vary between 0.75 and 6 ° min \ and b a c k g r o u n d i n t e n s i t i e s w e r e m e a s u r e d f o r h a l f t h e s c a n t i m e a t each side of the interval. Test reflexions were recorded before and after each layer with the PAILRED di ffTactometer and between every kO or 50 reflexions with the SYNTEX di ff Tactometers.

Graph ite-monochromati zed MoKa radiation (X = 0.71069 Å) was used for all intensity measurements.

The crystals of phases II and V d eteriorated during th e X-ray exposure, and the intensities were therefore calibrated according to an intensity-time function calculated from the standard re­ flexions measured.

Intensities were corrected for Lorentz and polarization effects in the usual way . Absorption correct ion was applied on all data sets.

CSI CM CC4 3 CSI O fei i 3 w </> co

"O

_c <T3 fei O Pu I o fei

in

fü -o <D C <4-0) 13 </> 05 (/) D V) <D u <D C 0) 1. >-L. 0J D (/) 0Q r-C>

g

PO wn o vO CM sO LO r*- PO \0 O O o o o o et O o o o o o •ai . C. E O fü OO <y\ PO r^s. -3" o o f 1 e x pa vO OO CM CNI CM PO • < +- u o a> o;

_•z.

u. a. X !t> E <•<' \ CP — _{vO o -3-} LO sO o •< OO ao 00 v£> o to _{O o O O o} -u—

'

a> c 3 — • CT — X) *-» c a; c 3 t/) <U r- PO OO ao vX> un -i 1= o vO cn PO -3~ of . t ner rO CO -cr <T\ tn of . t ner v£> r^. LA . "4— L o «d a> 2 u s-<r u_ a) O 4-i a> 5- E a> <o _o L-E <T> 3 a 00 r-~ cn PO LO CM "O -- cu CTN *— PO vO -3" "O -- cu PO sO CM fü — O 03 H- > in a> o <3 U U

rtj

- — *-PO l/> V4_ C O *J l/l U L-C O

_u

•M O 1 X fc LU U- >-O C t/> fC »-a o 3 4J u o

_

1 OJ CM » 1 O f0 o <±-u 1

n CL

5- O !U i-<D O CO 1 * 1 • E tf> T5 — L. TO O CL — a — o ro i-E +-• i_ O <U i/i CM -3" -3" vO CM O Csl OO NO cn CM 1— c Csl -3" *— <TJ — t/> 03 i-C <D O ^ *U a> a> JJ E c _{(g O CM -T O -3"} CSI

m

i- u- PO ro O »— 00 . * O "3

o

' CM G- O- u o "Û c ^ — 0) tn g P» O vD O ao o O — _C CM PO r~- PM <r -O _CM . X-vO fv

"

i O _CM o CM 3 o 00 X X CM "o _(Ni CM

?

X o > ' X 3

?

«»o (TY Csl vO o CM o

OO

X ' -T O

OO

_O O CM Q £ 'T— \.r-~ 00 jun > • X" O > >

\0

o PO CM '.-3- oo

0-

CM 'a _•p O _> > _X 0

_i

x:

_Nf> > • •. Q _-2 f n> 2 OO

r^.

•«c SP-' .

yc

v£>

22

1 O 3- 1 Ok

For phases I-V the Gaussian integration method was employed. This correction was introduced at a stage when the structure de­ termination was almost completed, which made it possible t o use changes in values of temperature factors as a check on the result of the correction. For structures l-lll t he program ABSOT descri­ bed in Ref. 105 was used, while for structures IV-V the absorp­ tion correction was calculated with program DATAPH (Coppens, Leiserowitz S Rabinovich). Absorption correction for phase VI w as performed in quite a different way, using the empirical absorp­ tion correction available in the SYNTEX R3 system. For this cor­ rection , 20 reflexions with 29 values distributed over the 20 range investigated are measured 36 times each at 10-degree inter­ vals during a full rotation around the diffraction vector. Each reflexion so collected defines an absorption curve. The curve nea­ rest in 20 to a regular reflexion is then interpol ated t o correct this reflexion for absorption.

For equivalent reflexions a weighted mean value was calculated. Such reflexions were present only in the data sets for phases I (±/z00) and I I I (00+&) and for phase I I , where between two and

six equivalents of a reflexion could be present in the data.

C o m p u t a t i o n m e t h o d s a n d s t r u c t u r e s o l u t i o n .

Patterson syntheses were calculated for all structures, and were used as the basis of structure solution for all phases except phase IV, which was solved by direct methods using the program MULTAN (Main, Germain S Woolfson). Refinements were performed with full-matrix least-squares methods for all s tructures except

refi-ned with a 9x9 block-diagonal matrix.

For structures III and V the core memory capacity of the CYBER 172 computer used did not allow the simultaneous refinement of more than 140 parameters, which made a segmentation of refine­ ments necessary. This segmentation was carried out with at least some of the Mo atoms overlapping. The e.s.d.'s of these two struc­ tures may consequently be s 1ightly underestimated. Refinements of the hydrogen atomic positions of phases I and II were m ade on data sets with (sin0)/A < 0.50 and 0.52 respectively.

From the F , - F , list after isotropic refinement of structure

obs ca le r

V, it was evident that the calibration for degradation of the crystal was not quite successful; layers measured early or late d u r i n g d a t a c o l l e c t i o n h a d a n a v e r a g e F / F r a t i o o f ~ 1 . 0 2 ; i n ­ t e r m e d i a t e l a y e r s h a d a n a v e r a g e Fq/F^ ra t i o ~ 0 . 9 8 . I n d i v i d u a l

l a y e r s c a l e f a c t o r s w e r e t h e r e f o r e a p p l i e d . T h e s e w e r e h e l d i n ­ ternally fixed during refinements with anisotropic temperature factors.

The only structure seriously affected by extincti on was phase I, for which a secondary extinction parameter according to Coppens

106

S Hamilton was included in th e refinements.

The computer programs used, apart from those mentioned above, were those described in Ref. 105 with addition of the following:

Data reduction programs TAPER, AVEX and GECOR provided with the SYNTEX R3 crystallographic system.

FOUR, the Fourier summation program and BLOCK, the block-matrix least-squares refinement program from the R3 system.

Fig. 8. The (v0^n chain in phase I.

4 _

Fig. 9. A stereoscopic view formed by the (V0-) chains

' 3 n

NaOj^O)^ octahedra in pha

showing the layer structure and the double chains of

ORTEP-1 I, a thermal ellipsoid plot program for stereoscopic

illu-1 0 7

stration of crystal structures.

DESCRIPTION OF THE STRUCTURES

The metavanadate, (V07 ) The structure of NaVO,*1.89H,0

con-Ó YL J

T-tains chains of V0^ trigonal b i pyramids joined by common edges (Figs. 8 and 9)• These chains are connected by double chains of

NaOjAq^ octahedra to form a layer structure. These layers are con­ nected by hydrogen bonds.

The vanadium atoms coordinate two terminal oxyg en atoms at distan­ ces of 1.643(1) and 1.653(1) Â (of. th e 'V02+ anion1 in Refs.

50-54). The distances to the bridging 0 atoms are 1.882(3), 1.928(3) and 1.988(1) Å. The shortest V-V distances in the chain are 3-096(1) and 3.134(1) Â.

3 7 ^ 1 0 8 — 1 0 9

This structure agrees well with those previously reported with regard to the shape of the anion chain, as well as in the

respect that it gives little information about the metavanadates in

3-aqueous solution . The anion HV^O^ crystallized by Fuchs et 3 9

al. from a water-ethanol solution is in agreement with complexes

2 6 - 2 8

suggested from solution studies, but metavanadates precipita­ ted from aqueous solution seem rather remotely connected with the dominant species in solution.

4— II

The divanadate anion, (^2^7 This anion consists of two VO^ tetrahedra sharing a corner (Fig. 10). As can be seen from the

re-

4-ferences given in the paper , the ^2^7 anion has been found in a large number of structures, and the general outline and bonding pattern in the anion have proved very s imilar in all cas es. The

Fig. 10. The unit cell contents of Nai,V2°7^H20^ 18'

4-w i t h t h e ^ 2 ® - } a n i o n s h o w n i n d e t a i l a b o v e . I n t h e

chains of octahedra running along the s-axis the Na atomic positions with ztQ or s#1/2 are only half occupied.