Octahedral clusters in transition element chemistry

Journal of

ELSEVIER

Journal of Alloys and Compounds 21~2-21~3(1997) 10-21

Octahedral clusters in transition element chemistry Christiane Perrin* Lalu~ratoire d¢ Chimie d, Solide t'! lnorganiqtw Molt'.'c,laire. UMR CNRS 6511. AL~viueth~ Gt;nt;ralLeclerc. 35042 Rennes Cedex. France

Ab~t~¢t

The variou~ ways of condcn.~ation of the two different units ha~d on Me, clusters, the face-capped Me~.L~.~ or the edgeocapl~d Me~.L=,. arc pre~nted and illustrated by significant examples of the Mo¢. and Nb¢. chemistry, respectively. During the co~densation pr¢~ess, the molecular character of the compounds ba~d on Me¢.L~.=units decreases, while it is maintained in the compounds ba~d on Me,L =, units due to the important steric effects of the 18 Iigands in the latter case. The physical properties of compounds ba~d on such units are strongly de~ndent on the degree of their molecular character and on the filling of the electronic levels: deciron deficiently may produc=~'either metallic conductivity for compounds ba~d on strongly inicraciing unify, or p aramagnetic behaviour due to unpaired electrons when the units are isolated. ~ 1997 Elsevier

!, l ~ l ~ ¢ t t o n A chester can be d¢ltned as an a@regate of metallic atoms with finite dimensions and high symmetry, in which the metallic atoms are linked together by m~tal~m~tal bonds. In the cluster core, the distances ~tween the metallic atoms are close to the distances

in the metal hulk. In transition element chemistry, the clusters are formed when the non-metal/metal ratio is b e t ~ that needed for the preferred coordination h u m o r of the metal, when the d~rbitals are rclao tively large and when ~me valence electrons are available for the metal=metal ~nds. The cluster is encapsulated in a shell of surrounding ligands, thus forming a unit: e.g~ a Mo4 tetrahedral cluster is bounded to t~tur sulfur atoms, giving the Mo~S4 unit pre~nt in Mo4S~Bt~ [I], T h e ~ units are neutral or charged and |hey constitute the building bh~ks of variety of crystal structures. The cohos ion of the structur~s is m~vlc either by ¢oulomhic interactions

i~2{4.,~43~419"i1517,iill ,~ 1~? Elsevier ~l
betwcew~ the couiltcfcations and the anionic units, or by ligands shared between adjacenl units, or by Van dcr Waals inleractions between tire uuils when they are neutral, in a given structure, several of these various connections can occur simultaneously. All of the~ features lead to a large variety of different unit stackings in the cluster-based materials, giving original structure types with ~metimes anisotropic char° acler.

In the organomelailic chemistry of transition elements the clusters are mainly formed with electronrich elements and the ligands are frequently carbony! or phosphine groups [2]. These clusters can be triangles, telrahcdra, ociahedra or other a~regaies of more complex geometry. A spherical condensation of clusters can occur giving large clusters. In contrast, in solid°state chemistry of transition elements, the clusters are formed with e l e c t r o n = ~ r elements and the ligands are frequently halogens or chalcogens: they can be a l ~ triangles, tetrahedra, ociahedra or other types of relatively small clusters. The octahedra can be condensed by their apexes, edges or faces, thus giving frequently anisotropic compounds [2].

c Perdn/Journalof AIIoysand Cmnpounds262-263 0997) 10-21 Following the pioneering works of Brosset [3] and Pauling [4] on the octahedral clusters, their study has been extensively developed over the past 50 years. In this review we have chosen to focus on octahedrai clusters of early transition elements in solid-state chemistry., which appear in two different types of units, and we will compare the progressive condensation of these units and the evolution of the related physical properties. For this purpose we will detail the Mo~,, Nb~, and Ta~ duster chemistry and we will give some short comparisons with W~, Re6 and Zr6 clusters, including both old results and more recent developments. 2. The octahedral clusters in Me6Lt4 or Me6Lts units

2.1. Differences in the formation of Me~ L lq or Me6 Lss units In the solid state chemistry of transition elements, the octahedral clusters appear in two different units, i.e. Mec,Ln~ ',mit and Me, LIe unit. In the Met,Li4 unit, the cluster is face-capped by eight inner ligands (L~): six additional apical ligands (13) are located on the quaternaw axes of the octahedron (Fig. la). Such a unit is formed with Mo, W, Re and Tc. The highest levels occupied by the electrons are represented on the Fig. lb; the HOMO level exhibits an e~ symmetry, These levels have a metal~metal bonding character and are fully occupied for 24 valence electrons per chister, which corrc-

11

sponds to 12 two-electrons-two-centers bonds. The filling of these metal-metal bonding levels by the electrons is called VEC (Valence Flectron Concentration) per Me6 cluster. It corresponds to the number of electrons involved in the metal-metal bonding states after the charge transfer from the cations to the cluster and from the cluster to the ligands has been accounted for. In the Me6L~s unit, the cluster is edge-capped by 12 inner ligands; six additional a~ical ligands are located hke in the Me~L~4 unit (Fig. 2a). These units are favored with lower d-electron concentration than for Me6Ln4 units and they are formed ~ t h rare earths, Zr, Nb, Ta, Tb. The highest levels occupied by the electrons in this unit are represented in Fig. 2b; the HOMO level has an a2u symmetry. These levels are fully occupied by 16 valence electrons only, which corresponds to eight two-electron--three-center bonds. It most be pointed out that in the case of very poor valence electron transition elements like the rare earths, Zr and Th, an interstitial element which participates to the metal-metal bonding states, is necessary in order to stabilize the octahedral cluster. hi this review, we will not consider these octahedrai clusters filled with interstitial elements, except for a few significant examples taken from the zirconium chemistry. 2.2 Condensation of the Meo Lla or Me, Lt,~ traits

The Mc, L,.~ or Met,L,, units can condense by

? ~lgg~

-6

-?

B

Z

[

g

~

wsm

MD l I B I m

(eV)

- -

~8

imam"

~~

~lU

~ i J B

-9

Me @

Me6L8

00 a}

Me6Ls|,~,~

b)

Fig. I. (a) The Me, Ll~ unit: representation of the inner and apical ligands. (b) Energy level diagram for the metal-centered orbitals of Me~,Li4unit: influence of the apical ligands.The metal-metal bondingorbitals are labelled. Accordingto Simon el al. [5].

|2

C Pcmn /J.um.l

of+A #,,~+,+++o.,+d Comlm+,und:s ~+.+ . ++~+, "+.+~.+~+~+.+i I ~ 7 + I 0 ,.+21

+6

-7

..m

7"

m (ev)

..

a2U

- 8 II~

Im

---.----

mm--

---

tomb

t2g flu

-9

'= +

m+ ®

o

I0

Me6h 12

alg

Me61'l 2 h6a

0 a)

b)

Fig. 2+ ~a) The Mc, L,, unit: reprc~¢nta+ion o[ the inner and apical lig~mds. (h) Energy level diagram for the mclal centered orhitals of Me, Lt~ unit: influence o[ the apical ligand~. The mclal:-mctal l)onding orhilal.~ arc lal~llcd. According to Simon ctal. [5]. &

i

Fig+ 0++ ~hcmm(+¢rcprc+cntathm ol+ the unit ¢ondcnsali+m a¢~+++rd+

iog ¢o H, ~h,~fcr ++mlHG+ vo.~ ~hncrlng notatltm [¢~1+

s:haring ligands (Fig, 3), A ligand in apical position or in inner F~sition for two adjacent clusters is noted L; + or U +, res~ctively: a ligand in inner position for a cl~:¢tcr and in apical position ['or the adjacent cluster is noted L' tb+' the fi~t cluster and L'~=' for |he a~|ja~cii{ one, This notation has hcen prolapsed !k~r the first time by ~L!Jfcr and yon ~hnering [{~]and is ,¢e~ u~.fu] for~schematic description of all the strut+ tur¢~ o:~t them two ty~s of units, 3. ~ s s h + e

t~ ~y~.am

c|mden:$altoa of the Me+L/~ units m ~tstry

In the Me~,Lt~ units, the steric effect of the inner

ligand shell is not too important and it is possible to easily conden~ them at first by tne apical ligands and then by the inner iigands. This condensation can be illustrated by significant examples taken in the molybdet:um h:dides, chalcohalides and chalcogcnidcs chemistry [7], It is controlled by progressively changing the total number of anions per Mo,~ cluster (n). During this process the VEC ,+ 24 is usually main+ t~Jined, For instance, in the halides, any decrease in t+ is related to a dccrcam in the cationic charge; in the chalcohalides a decrease in u is obtained hy pmgrcs+ sively replacing two halogens by a chalcogcn, But, as d e , r i n d above, the Mo~ cluster remains always surrounded by i4 ligands, Then, during the condcnsa. lion, an increasing num~r of ligands will be shared ~twcen adjacent units: at fi~t apical ligands and then, when all the apical ligands are shared, the inner ligands start to ~ involved in the condenm~tion. 3. I. Comlum.d,s m which the molecular eh.s~¢ter is n,laim.d An example of the first term of this ~ries is represented in Fig. 4: PhMo~CIt~ (. ~ 14)[8], in which the units arc di~v:tc. ~ c next steps (Figs. 4b+~d) illustrate the condeusation of the Me~L+~ unils by apical ligands (L ~'+) in one, two and finally three directions of the space for the typical repr~ntativ¢ com~unds AgMo,,CIt~ O+ ~ 13) [SL Mo~CII2 (n + 12) [9] and Mo,,Clt~+~ (~+ + 11)[10], respectively. When the sL~ api~! ligands are shared between adjacent units, the inner ligands (L' ' ) start to ~ involved in lhe condensation giving the one-dimensional Mo,,IsSe 2 (n = 10)

C Pemn/Joamal of Alloys cmd Compounds 202-263 ¢i997) !0~21

13

a~ b)

e)

a)

Fig. 4° incrcasitlgcondcasatioa of the Mo~,L 14 uoi|.~via [/ ' ligands ia Mo, halid¢~ and chalcoh~lidcs. Examplc~ of: (~ t~ ~ [~. PbM,h,(~ls~ (Mo,C.I,CI,CI4/:): and (d) n ~ II ~.4o,Cls,,Sc m i,~) , ' " ' ~'' (c) n -i , C[KCI4CI2/:). -~ 1 2 . M o c , C i I: (Ij bMo~,ClsC ~,, (b) n ~ 13. AgMoc, Clj~ (AgMoc, o ~, ~ ! " i a ~J (Mo~,CI~S¢ CIf,/, ).

C f'crr#i f Jolirtlat o]',,It/ott~
14

10-21 H

E #

5 .....

t

I

#

ge

Te

#/

200 N: I

-

~I:QHt

I Kl'h~ .I

I

E ° /'%1 / .~. ~I~Hz Z" ..~-

,

/t e

/ ,

,'

O,t.

I

/

. •

/ ,';

5 l '

.

/

,I . ~ ;...* 2 g . - . -

-~ ~,-/

= ,, t..yr ~'I'- \ / ~ ~ " 2 7 " ~

t.-

I,,. '

'"--X//

d

1/

,

~

I,'/,,'/,'

• ,I " ,;' / , , . , .I,

b)

• o.I #

till

{1

1 {I

ft ~it

I [llvit

,/,,/,," i,

I,'/

compounds, ~ t h based on chains of Met,L,, units, at first isolated and then connected by inner-apical ligands tL'=") i~, one direction of tl~e s,..,.'., res~ctively In all of i h , ~ compounds, th~ dusters are far from each oih¢r anti cannot interact. The nlol~cular chili °o ~ctcr is retained and th~sc compounds ;ire of course

in~u!ating materials. Their VEC is always 24. so the clusters arc nonomagnctic. When It halogcn/chalcoo gen statistical distribution occurs, like in Mo~,CIt.S¢ (one S~ occupying statistically one of the eight inner ~ i t i o n s ) or in Mo01~Se~ (one Se statistically located on the six available inner positions), the unit acts as a permanent dipole due to the local dissymmetry of the anionic charges around toe cluster. In such ca~s, dielectric reluations can be observed as illustrated in Fig. 6 for the Mo,gll.Y (Y ,-S, ~ or T¢) mries. ~her¢ peaks of ," ~]'(T) appear clearly [10]. 3,2, (?OmlU>t~n,d~ tn w h i c h ~ith' redt~(~ed

¢e m o b c t d a r

c h a m c t ( ' r is

A further condensation (n ~ 8) is obtained for Mo.i]r:S~ [13]. i~typic with Chloral pha~s [14], in which the units are condensed by six L'~'~ extending in lh¢ three d~r¢~i~ons of the space (Fig. 7a). In this st~cific slacking of the units, the clusters are close to cac:h other and ih 9, can interact. Then, the molecular character is greatly reduced and actual ener~ bands

,I

l

!

k.,

17 ..... / I itEM, I

0,2

"-~.

s

i//

"i t

- 60

~i).

[11] and the two.dimensional Mo, Br~,So~ (n =9)[12]

[#

0,3

".. 7

.I,'i i/,,' l'k.J ! ,'i ,,' .,,,, d . I

Fig. ~. Increasing condensation of the Mo~,Lt~ units via L" ", L' ' ~ d L ~ ligands in Moo chalcohalides. Examples of: (a) n .~ 10. Mool~S¢. (Mo~I~iSc[/~I~/~); and (h) n -9. Mo~,BI'.Sj

"

#

/

I

o,s

,. ,..

,

l

a)

Il

,i

i ,.c.

-sb

-

Fig. 6. Rc ,l ( e ' ) and imaginary (e" ) part of the dielectric constant vs. tcmi:-'ratttrc for the Mo~,CIt,Sc series, for various frequencies.

are formed (Fig. 7b) [15]. The VEC can then be varied ,rum 20 to 24 corresponding to the filling of the e~ hand. Electron deficient compounds are metallic and most of them exhibit a superconducting behaviour. 111 this class of compounds, there are two possibilities In cllange the VEC and thus influence the '1:: (i) by halogen/chalcogen substitution, i.e. Mo~,Br~S~, and Mo~,|~S~, with VEC ~ 22 exhibit a T, ~ 13.8 K and 14 K, respectively, clo~:e to the ~ (15 K) of PbMo~S~ with the same VEC, while Mo~,S~ with VEC = 20 has a T, = 1.7 K [14]. In fact, in the sulfide ~ries. VEC 22 corres~nds to a maximum of density of states which can explain the highest T< ob~rved for the corresponding ~mpounds. With ~ and Te the solidsolutions M o ~ o , X ~ and Mo~Te~.,X, (X ~CI, Br or I and 0
[141, Finally. in Cs~,,Mo~S~ (~ ~ 7) based on chains of units connected together by six L'°'~ (Fig. 9), all the ligands are involved in the condensation and this communal constitutes the last l~.)ssible step of the

C Perrin/loumal o/Aitovs and Compounds262-263(199D 10-21

15

? I0 ~

(f

-12. ~

•i ('/" -16.~

a)

/. ~

,

......

partial

d 14o

.....

partial

p L

-----.

total

b) . (Mo,,Bt,S~,/2S,/:). Fig. 7. (a) Condensation of the Mo~,Lt4units via L*" ligar~dsin Mof,Br,S~, "* *" * (b) Densityof states (DOS)for the same compoundwith representation of the Fermilevel,accordingto Certain and Lissillour1151. TM

Me,,Lt.t unit condensation. Like in Chevrei phases, the molecular character is greatly reduced and this compound is a superconductor with "I~ ~ 7.7 K [171. 3,3. Comparison between trait condensation in Mo~, IVo and Rt'. chemistO, A similar condensation of the Me.Lt4 units has been obtained wilh tungsten and rhenium. However, for tungsten, only the first terms of the condensation. i.e. the W. halides, have been obtained, probably due to a limitation of the W. cluster stability. Indeed, it has been recently shown that the syntheses conditions have to be drastically precise so as not to decompose the W. cluster and a relatively low temperature of reaction has to be used [18]. Up to now, no Chevrel phase analogue based on W. cluster has been isolated. For the rhenium, a great number of compounds isostructural with molybdenum compounds have been obtained [19]. However, due to the fact that Re has seven valence electrons compared to Mo which has six, for the same n-value the chalcogen/halogen ratio is more important with rhenium than with molybdenum. This gives several compounds with original unit condensation, not found with molybdenum. For instance, in Re.SesCI 2 (n ~ 10) connections by four L ' " appear in two directions giving a two-dimensional compound [20]; in Re~,Se4CIto (n ~ 14) the

units are discrete without countercation giving a pure molecular compound, the cohesion of which being a,,'~ured by Van der Waal,~ interactions be~een neutral units [21], This feature explains why Chevrd phases cannot be obtained with Re, cluster. Moreover, in rhenium chemistry, a number of chalo cohalides ~xhibit Y=Y and/or Yobridges, e,g, "'~,~'N~'4t~'~"'~/ [221, 4. Progressive condensation of eh, MeaLt~ units in the niobium ehemistry In the M',~Lt~ unit the steric effect of the inner ligand shell is more important than in the Me,Lt~ one. Only a few examples of condensation involving the inner ligands have been reportecl and the molecuo lar character of the compounds based on this unit is retained. The n factor cannot be changed greatly, but the VEC can be varied from 16 to 14, which significantly differs from the Mot, cluster chemistry for which the VEC was always 24 for the molecular compounds. As a result, the magnetic behaviour of the cluster can be observed. Indeed, for a VEC ~ 15, the cluster is paramagnetic due to one unpaired e!ec~ tron on the a,~, level, in contrast, for a VEC ~ 14 or 16 the HOMO level is empty or filled, respectively and the cluster is non-magnetic. Illustrations of the Me,,Lt, unit condensatton will be taken from the

~

C, l Y m n / Jotmt. i o[ A Ilov~ . m l ('mnpo.ml,~ ~2h2 203 ¢1~ 7) I 0 2 !

× Fig, ~ /i evolution v~, halogen substitution l\u' the Mo.Sc~..Br.. Mo.~ , !, and Mo.Tc~ ~I, ~lid ~ohttions,

niobium halides and o~'halides chemistry: tantalum gives very similar compounds.

4o!. Unit condensation in the niobium halid,;'i~ Several niobium halides exist with Oi~rete units l~t ~ 18) and monovalent or divalent countercations. lmpo~:l~:tnl examples ~ t r e K~Nb.CI~, [23]. !n:Li~Nb.C!t, [24!. Ba~Nb, CIIM [25], Cs:EuNh,,Brl, [26i ~ i i h V ~ C ~ lfl, Wilh lfivalent ¢ountercation.~ like

rare earths an important series, MRENb~,X~s (M = monovalent cation, RE = rare earth. X = CI, Br) has been obtained. An additional monovalent countercation is present in the structure, which again gives a VEC of 16 [27,28]. These compounds crystallize in two structure-types (R3 or P31c) depending on the size of the monovalent cation. For a large cation like Cs + the units form a pseudo-hexagonal stacking (P31c) (Fig. 10a) and for smaller monovalent cations they form a fc.c. stacking ~R.3)(Fig. 10b). In the latter case the monovalent site is only half occupied; this site can be emptied without destroying the structural skeleton giving the ternary compounds RENb~,X~s with the quite unusual V E C = 15. This example is of first interest because it is possible to control the VEC in exactly the same structure~type (R3) and to compare the magnetism of the cluster in quaternary compounds with VEC = 16 and ternary compounds with a VEC--15. With a non-magnetic rare earth like Lu the magnetism of the cluster can be clearly quantified by comparing KLuNb~,CI~ (VEC = 16) and LuNb,Cl~s (VEC = 15): it corresponds to a paramagnetism due to one unpaired electron in the latter case. while a nearly temperature independent behaviour is observed for the quaternary compound (Fig. 11). For LuNb,Ci~s. a maximum of susceptibility at 2.5 K has been attributed to interactions hetwcen magnetic clusters [29]. The first step of unit condensation is obtained lk~r , Ic~ in L i : N I , , ~ ( I I ,~' , [311] in which the units arc

d o

O

O¢s

Mo

$

17

C P e r r i n / J o u m a l o f A l l o y s and Compounds .~6.;-. ~ ~ ~6~ . (1997) 10-21

Q o

~

RE o

4

-

eo

a}

e

%ft = i . ~ o +i

~20

°

8~

a}

x

-g I o

A,

2

@

°°

3

E

2

2

rcg:

O( Q

~c3

30

20

&O

0° °°° O

O

0 @

@

°

•

@

~0o=1=0o=, Vb) , o1,, ,, ~ 10 20

,

i

~

b 30

~,

°h t,O

• SO

T(K) Fig. 1I. Magnetic susceptibility vs. temperature for: (a) LuNb,Cl~s single crystals; and (b) KLuNbf,Clts powder. Insert: inverse of the magnetic susceptibility of LuNb.Clts.

a)

A

7

RE

by inner ligands is difficult, which explains that n cannot be decreased greatly in the Me~Lts based compounds.

4. 2. Unit comhmsation in the niobium o.9,halides ! I

@

,

¢

I

(

I I _

!

.

I I

i

I !

D

b) Fig. III. Unit-ccU representation of 111c Iwo MRENb, XL, ~MRENb,,X~:X~,) structure-types (t~ ~ 18): (a) I)31c: and (b) R3, For clarity the halogellS are 11OIreprcsentetl.

condensed in two directions by four apical-apical ligands (Fig. 12a). No intermediate step corresponding to n ~ 17 has been reported up to now. For n ~- 15, the units are connected in the three directions by six apical=apical ligands forming halogen bridges which arc linear in Nb, Ft~ (Fig. 12b) [31] and bent in Ta,CIt5 (Fig. 12c)[32] or NaNb, CIt5 [33]. In the last step, Nb, Cl~a (n ~ 14), the hmer ligands start to be involved in the condensation: the units are condensed by four CI" " and two CI ~'' (Fig. 12d), but due to the important steric effect of the 18 lig~mds, the Nb=CI * " distance is large (Nb=CI ~ 2.41 A, Nb=CI" " ~ 2.58 A, Nb~CI '~''~ 3.01 /~) and the clusters cannot get closer [34]. It appears that the M e , L ~ condensation

An alternative possibility to produce interactions between the clusters is to reduce the size of the Me.Lts units by introducing a small ligand like oxygen around the cluster. Then. three series of oxyhal!des have been recently isolated and they consti° lute it unique example of oxThalides based on Men cluster obtained tip 1o now in solid stale chemistry, For n ~ 18 two series have been obtained. M:RENI~(,II~O ( g 3 l (VEC ~ 16) [35] and M~RENb.CII~O ~ (P.~lc)(VEC ~ 14)[36], respectively isotypical or strongly related to the two structures of the MRENb,CIt~ chlorides (Rg and P~lc) demfibed above, in the first compound the oxygen is statistically distributed on the 12 inner positions which does not influence greatly the size of the unit. while in the second one three oxygens are ordered on inner l~si° tions. In the latter case tile size of the unit is signifi° cantly reduced, but no interaction can occur betwet:n the clu,"s~'ster, because the units are still discrete (Fig. 13a). In the third series gENboCIt~O~ (VEC~ 14) [37], three oxygens are again ordered in inner position like in the previous compound, but the main differ° ence is that now they are in cis=position with respect to the Nb, cluster, producing now a strong distortion. The units are condensed by tour L" '* and, due to the great deformation of the units, this connection pro° ducts helices of units (Fig. 13b). Now the clusters are relatively close to each other glue to the smaller size of the units induced by the three oxygens and to the type of their condensation. However, these three series of oxyhalides remain

18

C. tYrrin /Jo+~mal~l'Alloy.~ aml ('oml~mmls 202++2~3(1907) 10-21

b)

a)

a e) d) |:ig~ 1 ~ il~°re~in~ ¢ond¢ll~tli~n of the NbnLl~ tlnit~ ill Nb~, halidcs, t:]xamplcs of: la~ ~ + Ii% l°|: Nb++("l+,, ILl: Nh+,( I+ : ( 1 : ( I+ : ); (hi el N|+,++++++++(Nb,+F~+E + +)+ (¢) P+~+ 15++T+++,CI~++IT,+++,.,CI~:.... .+:.+ and (d) + 14, Nb++CI++ (NI ( | u , ( I , ,(,1, ,(1~) +

++

~

¢1o~ to the h~dides° In fact, three oxygens around the ¢lu~er are not sufficient enough to strongly reduce the si~,¢ of the units ~md to induce ~ctual interactio|as ~ t ~ e e n the dusters+ Moreover, up to now, it has not ~ e n ~ s i b t e to substitute more than three o~gens o,r~ the units+ ~e~el~,~e, a large gap exists between the~ o~h~dides ~,nd the next step formed by the m~ium oxides, The~ Nb. based oxides constitute a v~E¢ different cta~ of com~unds ~cause an im~r-

t+

..

+;+,'+:,+

+

+

`

~ant mlionic charge surrounds the cluster and. in a same com~und+ coexist frequently oetahedral clusters. triangles, i~laled element and sometimes conden~d cluste~ 138].

5. The Zr~ duste~: an e~mple of in|er'stitial octahedt~l cluster Zirconium has only four valence electrons and an

C Perrin ~Journal of AUoys and Compomuis 262-263 ~1997)10-21

19

0

a)

o

o

b) '~ ' core representations h~r M.RENb,,CII~O~ o o -~, ( n ~ 18). ( b ) unit°ceil, mb,,,,.~0.--~r't' c~, core Fig° 13~ ( a ) U n i t - c e l l and N b ~,(I,~0~ (Mo.RENb.CI,~O~CI.)' and helix of units representations for RENbc,Cll~Oa (RENb.CI,~OaCI~CI~/z) ~ ' " " ~ (it -~ 16). Only C I " ~ atoms are represented,

interstitial element (Z) is necessary o,1 the center of the zirconium octahedral cluster in order to stabilize it [39]. Such a situation is also encountered for octahedrai clusters of electron-poor elements like rare earths. This interstitial element participates in the

metal-metal bonding states and increases the number of electrons involved in the metal=metal bonds. A large variety of elements can be inserted like Be to N, AI to P, K, Cr to Ni and Ge. The crystallochemistry of these Zr~,(Z)based compounds is very rich and a great

20

C. P,n.in/Journal o#'AIl,~vxatt,I Comtm,tlnds 202'-2fi3 flt/97j 10-21

[21 G. Schmid tEd.). Clusters and Colloids. From Theory

Fig,

14. A [1(10i ~CCliOn of rhiimbohcdrai

Zr, X;:Be

humor of compounds are isotypic with the Nb,, and Ta. ones. Condensation of these Zr.(Z)L~s units by apical-apical ligands appears in a lot of compounds, but only few examples of condensation are known which involve inner ligandn. Indeed, the important stcric effect of the eighteen [igands makes difficult such a condensation and usually the molecular character in retained for these compounds. A particular zirconium compound is s~cially intercsting in the frame of this review: Zr,,X~:(Z) in which the units are condensed in the three directions by six inner:apical ligands (Fig. 14) [40], exactly as in Chevrel pha~c~, In fact this compound is the analogue of Chevrcl phases based ell Me,,LI~ units, instead of Mc,~L~0 But duc to the important stcriceffect of the lig~_md~, lhc ZroX ~ ' ' distiirlc¢~ arc large ~= like Nh=CI' ~' disl~i.c¢,~ h'i Nh~,CIi, ~ described ahov¢ ........ ~lld the ¢lo~lcr~ ~alriiOl ~¢I cloncr, Then in lhis corn. pOtllrd the molcctlhtr character is relahlcd in contrail io the ~ituation observed in ('llcvrcl pha."~'c~.

6. C~.¢ladlnil ~marlis Although the Me,~[.l~ and Me~,LI~ units look very similar, their condensation acts very differently due to the stro.g influc.ce of the steric effect of the ligand shell, ~yoild the condensation of the uilils, a further slop will involve the clusters by themselves. For in,lance, the Me,~L~ unii~ can bc condensed in one direction of lhe space giving the one-dimensional Mo,.I~S¢~ r¢~lrlcd above. A further condensation in the sanl¢ direction gives infinite molybdenum clunters c~,.de.~.~'d by lhcy faces like i. TI+Mo,,Sc<, I~2l, A htr~¢ vsrieiies of Stlch Colldcnsa!iOliS of cluster art:

w¢it [ . o w n giving a great number of n~aterials which con.~iiiule ml,olhcr h~lmri~mi ti¢ld of cluster chcnlislry [43l,

Re~e¢e~¢s [II (! Pcmm R, Chcvrci, M, Scrgcnt, C R Acad 5ci, 281 (lqT5)

to Applications, VCH, Wcinheim, 1994. [31 C. Brosset, Ark. Kern., Mineralog. Ser, A 20 (1945) 16. [4] P.A. Vaughan, ].H. Sturdivant, L. Pauling, J. Amer. Chem. Sty. 72 (1950) 5477. [5] A. Simon, H.J. Mattausch, G.J. Miller, W. Bauhofer, R.K. Krcmcr. in: K.A. Gschneidner, Jr., L Eyring (Eds.), Handbook on the Physics and Chemistry. of Rare Earths, vol. 15. Elsevier. 1991. [t,l H. Sch~fer. H.G. yon Schnering. Angew Chem. 76 (1964) 833. 171 C. Perrin. M. Sergent, J. Less Common Metals 123 (1987) 117. Is] M. Polcl, C. Pert'in, A. Perrin, M. Sergent, Mat. Res. Bull. 21 (1986) 1239. [,)1 H. Schiller, H.G. yon Schnering, J. Tillack, F. Kuhnen, H. W6hrl¢. H, Baumann. Z. Anorg. Allg. Chem. 353 (!%7)281. ll()l C. Perrin, M. Sergent, F. Lc Traon. A. Le Tra,m, J. Solid Slate Chem. 25 (1978) 197, fill C. Pcrrin. M. Sergent, J, Chore. Res, (M) (1083) 440. [12l C, Pcrrin, M. Polcl. M, Sergcnt. Acta Ctyst. 39C (IqS3)415. [131 C. Perrin, R. Chevrcl, M. Scrgcnl. O, Fischer. Mat. Res, Bull. 14 (1979) 1505. [141 R. Chevrcl, M. Sergent, Topics in Curren! Physics. in: O. Fischer. M.P. Maple (Eds.). Superconductivily in Tcrna~ Compounds, Springer-Verhlg, Berlin, llcidclherg, New York, 1082. [I.~l |). Certain, R. Lissillour, Z. Phys. D 3 (!q86)41 l. II~,l M. 5¢rgent, O. Fischer, M. Decroux, C. Perrin, R. Chewel, J. 5olid Slate Chem, 22 (1978) 8% P. Goug¢on, M. Polel, J. Padiou, M. Scrgenl, C.R. Acad. Set. Paris 5er. 2~7 (1983) 339. ll~l S. lhma'fne, C. Perrin, M. Sergent. Europ. J. Inorg. Solid Stale Chem 34 (1~)7) 16'k A, Pcrrin, M, 5crgenh New J, Chem, 12 (l~SS) 337,

[~il] L, Lcdoc, A, Perrm, M, 5crgenl, Acla Civsl C 3~i(19143)I~03, 1211 L, [.~di_i¢,A, Pcrrin, M, Sergcnl, F, Lc Traon, J,C, Pi!el,A,

i~1 W lh,mgcr, 11,4, Mi¢,~elL R Ncugl~0~cltcl, D, Schnlile, M, A Sinmm | | G , vo. 5¢hllclillg.II, Schiil¢i, Z, A,olg, AIIgcm~ ('hem, ,ln7 (19(~P) I 12~1 A, La,chgar, lt,oj, M~yer, J, Solid Slalc Chem, l iO (1~4) 15, 12sl A, Broil, I|, 5chllfcr,L l~ ('omiilollMelah 22 (It}Tll}3~?, [2i~l 5, Cordicr, C, Pctti., M, Scrgenl. Z, Aaor 8, Allgenl. Chcnl,

1271 C, Pertin, S, lhma'ine. M, Sergcni. New J, Chem, 12 (l~)n~) 321. C, Pcrri., 5. Cordicr, 5. lhma'i'nc.M, 5crgenl. J, Alloy,s(~mp,

12<,~l & lhma~nc, C, Pcma, O, Pc6a, M, Scr~cal, ,I,loc~ (7ommLm Mclah 13711~I 32J,

I.~o! B, Bajan, 11,4, Meyer. Z Anorg, Allgcm, Chore, ¢~23 II~7) 7~I.

{Sl] It, ~hiilct. It,G, vo. ~:hnering, 14,,,,I,Niehue,~, II,G, Nieder. vahrenhole,. L D.'~ C~.~imon Mehd~ ~ i l~li~)t~, D, Haunt, I|,G, yon Schncrmg, Z, An~,-g. Allg~.,m.('hem, 3~I tS31 M.E Sagcbalth, A. Simon, II, Imolo, W, Wcppi~ch G, [lithe, Z, Aaorg, Allge.~. Chelw, t~!l (1~t~};~1I~S~t.

tJ4l A. Silllon. H.G. ~'o, ~im.ering. li. WShrlc. II, Sch{ifei. Z. Ai~oig, Allgcnl, (!hem, ,I,I~I i I~I$) 155,

lS$l S, t'o~'dict,(7, P~riiil. M, Scrgeal. Mat. Rcs. Bull, 31 (l~if0 f~&! S, Cordier, (7, P~rm~, M, Scrgenl, ], Solid State (Them, 120 (199~) 4,k I37t S Cordi,er. C. Pcrlin. M. Scrlcni. Euix~p.I. &flid Stale lnorg. Chore, 31 (11!4) llt4~k

C Pen'in ~Journal of AIIoys and Compounds 262-263 (1997) 10-21 [38] J. Kiihler, G. Svensson, A. Simon, Angew. Chem. Int. Ed. Engi. 31 (1992) 1437. [39] J.D. Corbett, J. Alloys Comp. 229 (1995) 10, [40] J.D. Smith, J.D. Corbett, J. Am. Chem. Soc. 107 (1985) 5704.

21

[41] J.D. Corbett, J. Chem. Soc. Dalton Trans. (1996) 575. [42] M. Potel, R. Chevrel, M. Sergent, Acta Cq~st. B36 (1980) 1545. [43] A. Simon, Angew. Chem. Int. Ed. Engl. 27 (1988) t59.