reactions of nine-atom germanium clusters in solutions

2 downloads 0 Views 3MB Size Report
of the vertices are removed, 2n + 4 and 2n + 6 for the corresponding nido- and .... cluster is nido-species, and the 22 bonding electrons are in agreement with.
REACTIONS OF NINE-ATOM GERMANIUM CLUSTERS IN SOLUTIONS

A Dissertation

Submitted to the Graduate School of the University of Notre Dame in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

by

Angel Georgiev Ugrinov, M.S.

______________________________ Slavi C. Sevov, Director

Graduate Program in Chemistry and Biochemistry Notre Dame, Indiana November 2004

REACTIONS OF NINE-ATOM GERMANIUM CLUSTERS IN SOLUTIONS

Abstract

by

Angel Georgiev Ugrinov

The discovery of dimers of nine-atom germanium clusters, [Ge9–Ge9]6–, showed that 2-center–2-electron exo-bonds can exist between such clusters. This formally “oxidative” coupling of two clusters was the first redox reaction involving these species. This finding stimulated the systematic study of reactions of nine-atom germanium clusters with soft oxidizing agents, presented here. This study revealed that these clusters can bond to each other in different modes to form larger oligomers such as trimers, [Ge9=Ge9=Ge9]6–, and tetramers, [Ge9=Ge9=Ge9=Ge9]8–. Reactions of the clusters with Ph3Sb and Ph3Bi produced the first functionalized Zintl ions [Ph2Bi–Ge9–BiPh2]2–, [Ph2Sb–Ge9– SbPh2]2–, [Ph–Ge9–SbPh2]2–, and [Ph2Sb–Ge9–Ge9–SbPh2]4–. Initially the reaction mechanism for these additions was not clear but further detailed studies suggested that these reactions were in fact nucleophilic additions of the anions Ph2E– (E= Sb and Bi) to the naked clusters Ge9n-. The reaction mechanism is widely discussed in the thesis. These general ideas were explored further for possible use in more rational synthesis of similar derivatives of deltahedral Zintl

Angel Georgiev Ugrinov

ions by addition of other groups. Both mono- and di-substituted germanium clusters with –SnPh3, –SnMe3, –GePh3 and –CMe3 were obtained by similar reactions and were characterized by single crystal X-ray diffraction and in solution by NMR spectroscopy.

CONTENTS

TABLES ………………………………………………………………………………… v FIGURES …………………………………...…………………………………………. vi SYMBOLS …………………………………………………………………………….. ix PREFACE …………………………………………………………………………….... x ACKNOWLDGEMENTS ……...………..……………………………………….... xxvii CHAPTER 1 [Ph2Bi–(Ge9)–BiPh2]2-: A DELTAHEDRAL ZINTL ION FUNCTIONALIZED BY EXO-BONDED LIGANDS..................................................................................... 1 1.1. Abstract………………………………………………………………….……. 1 1.2. Introduction…………………………………………………………….…….. 1 1.3. Experimental Section………………………………………………….……. 3 1.4. Results and Discussions……………………………………………….…… 4 1.5. References…………………………………………………………….……... 8 CHAPTER 2 [Ge9=Ge9=Ge9]6-: A LINEAR TRIMER OF 27 GERMANIUM ATOMS……….. 10 2.1. Abstract………………………………………………………………….….. 10 2.2. Introduction…………………………………………………………….…… 10 2.3. Experimental Section………………………………………………….…... 12 2.4. Results and Discussions……………………………………………….…. 14 2.5. References…………………………………………………………….……. 19

ii

CHAPTER 3 [Ge9=Ge9=Ge9=Ge9]8-: A LINEAR TETRAMER OF NINE-ATOM GERMANIUM CLUSTERS, A NANOROD………………………………………………………….. 21 3.1. Abstract…………………………………………………………………..…. 21 3.2. Introduction…………………………………………………………….…… 21 3.3. Experimental Section………………………………………………….…... 22 3.4. Results and Discussions……………………………………………….…. 24 3.5. References…………………………………………………………….……. 29 CHAPTER 4 DERIVATIZATION OF DELTAHEDRAL ZINTL IONS BY NUCLEOPHILIC ADDITION: [Ph–Ge9–SbPh2]2- & [Ph2Sb–Ge9–Ge9–SbPh2]4-………………. 31 4.1. Abstract………………………………………………………………….….. 31 4.2. Introduction…………………………………………………………….…… 32 4.3. Experimental Section………………………………………………….…... 33 4.4. Results and Discussions………………………………………………..… 37 4.5. References…………………………………………………………….……. 51 CHAPTER 5 RATIONALLY FUNCTIONALIZED DELTAHEDRAL ZINTL IONS: SYNTHESIS AND CHARACTERIZATION OF [Ge9–ER3]3-, [R3E–Ge9–ER3]2-, and [R3E– Ge9–Ge9–ER3]4- (E = Ge, Sn; R = Me, Ph)……………………………………… 53 5.1. Abstract………………………………………………………………….….. 53 5.2. Introduction…………………………………………………………….…… 53 5.3. Experimental Section………………………………………………….…... 57 5.4. Results and Discussions……………………………………………….…. 66 5.5. References…………………………………………………………….……. 77

iii

CHAPTER 6 SYNTHESIS AND CHARACTERIZATION OF [Me3C-Ge9-Ge9-CMe3]4-……….. 79 6.1. Abstract………………………………………………………………….….. 79 6.2. Introduction…………………………………………………………….…… 79 6.3. Experimental Section………………………………………………….…... 81 6.4. Results and discussions……………………………………………….….. 85 6.5. References…………………………………………………………….……. 89 CHAPTER 7 SYNTHESIS OF A CHAIN OF NINE-ATOM GERMANIUM CLUSTERS ACCOMPANIED WITH DIMERIZATION OF THE SEQUESTERING AGENT .. 91 7.1. Abstract………………………………………………………………….….. 91 7.2. Introduction…………………………………………………………….…… 91 7.3. Experimental Section………………………………………………….…... 93 7.4. Results and Discussions……………………………………………….…. 96 7.5. References…………………………………………………………….…. 101 CHAPTER 8 SYNTHESIS AND CRYSTAL STRUCTURE OF [K-(12c4)2]2[K-12c4]2[Sn9]•4en ………………………………………………………………………………………… 103 8.1. Abstract………………………………………………………………….… 103 8.2. Introduction…………………………………………………………….….. 104 8.3. Experimental Section………………………………………………….…. 106 8.4. Results and Discussion……………………………………………….…. 107 8.5. References…………………………………………………………….….. 112 SUMMARY AND FUTURE PROSPECTS …………………………………...….. 114 REFERENCES …………………………………………………………..…………. 117

iv

TABLES

Table 4.1 Crystallographic data for (K-crypt)2[Ph–Ge9–SbPh2]•tol and (K-crypt)4[Ph2Sb– Ge9–Ge9–SbPh2]•2.5en…………………………………………………………..... 36 Table 5.1 Crystallographic data for all new compounds from Chapter 5 ………………..… 65 Table 5.2 Ge–Ge distances in the Ge9 clusters of all new compounds from Chapter 5…. 72 Table 6.1 Crystallographic data for (Rb–2,2,2-crypt)4[Me3C-Ge9-Ge9-CMe3]·7en ……….. 84 Table 7.1 Crystallographic data for [Rb2(4,2,1,1-crypt)]Ge9•en………………………………95 Table 8.1 Crystallographic data for [K-(12C4)2]2 [K-12C4]2[Sn9]•4en…………………… 108

v

FIGURES

Figure 1.1 ORTEP drawing of [Ph2Bi–(Ge9)–BiPh2]2- (70 % probability thermal ellipsoids). The numbering of the germanium atoms is shown………………………...………. 6 Figure 2.1 Comparison between the spectra of the trimer in solution and the trimer in solid state ………………………………………………………………………………..….. 13 Figure 2.2 ORTEP drawing of the trimer [Ge9=Ge9=Ge9]6- (50 % probability thermal ellipsoids). Each cluster is a tricapped trigonal prism with two elongated prismatic edges (shown as open bonds)…………………………………..……………..…… 15 Figure 2.3 Correlation diagram between two specific molecular orbitals in a square of germanium with four radially-pointing exo-bonds (left) and the same square with the exo-bonds parallel to a pair of Ge–Ge edges (right)………….......................18 Figure 3.1 Two views of the tetramer of [Ge9=Ge9=Ge9=Ge9]8- surrounded with eight rubidium cations (purple) "crowned" by 18-crown-6 polyether..………………… 25 Figure 3.2 ORTEP drawing of the tetramer [Ge9=Ge9=Ge9=Ge9]8- (90 % probability thermal ellipsoids). Each cluster is a tricapped trigonal prism with two elongated prismatic edges (shown as open bonds)…………………………...……………… 27 Figure 3.3 The HOMO of one of the nine-atom clusters of the tetramer………….………… 28 Figure 4.1 ORTEP drawing (thermal ellipsoids at the 50 % probability level) of [Ph–Ge9– SbPh2]2- with the distances [Å] to the two substituents shown……………….... 38 vi

Figure 4.2 ORTEP drawing (thermal ellipsoids at the 50 % probability level) of [Ph2Sb– Ge9–Ge9–SbPh2]4-........................................................................................... 39 Figure 4.3 MOLDEN plots of orbitals calculated for the core cluster Ge92- in [Ph–Ge9– SbPh2]2-: a) LUMO, b) HOMO, c) HOMO-1, and d) HOMO-2........................... 42 Figure 4.4 MOLDEN plots of orbitals calculated for hypothetical intermediate [Ge9–Ph]3with geometry taken from [Ph–Ge9–SbPh2]2-: a) HOMO, b) HOMO-1, c) HOMO2, and d) HOMO-9………………………………………………………………….... 46 Figure 5.1 119 Sn-NMR spectrum of [K-(2,2,2-crypt)]2[Ph3SnGe9SnPh3] dissolved in pyr…. 59 Figure 5.2 Sn-NMR spectrum of (K-18c8)2[Ph3SnGe9SnPh3]·0.25(18c6)·2en dissolved in pyr…………………………………………………………………………………….... 59 119

Figure 5.3 119 Sn-NMR spectrum of [K-(2,2,2-crypt)]2[Me3SnGe9SnMe3]·3.5tol dissolved in pyr ……………………………………………………………………………………... 60 Figure 5.4 119 Sn-NMR spectrum of [K-(2,2,2-crypt)]3[Ph3SnGe9]·en dissolved in pyr……... 62 Figure 5.5 119 Sn-NMR spectrum of [K-(2,2,2-crypt)]3[Me3SnGe9] dissolved in pyr…….…... 62 Figure 5.6 119 Sn-NMR spectrum of (K-18c6)3[Me3SnGe9]·thf·2en dissolved in pyr………... 63 Figure 5.7 ORTEP drawings of all new compounds from Chapter 5………………………... 68 Figure 6.1 13 C-NMR spectrum of [K-(2,2,2-crypt)]4[Me3CGe9-Ge9CMe3]·7en dissolved in pyridine ……………………………………………………………………………….. 83 Figure 6.2 ORTEP drawing of [Me3C-Ge9-Ge9-CMe3]4- (30% probab. thermal ellipsoids)... 87

vii

Figure 7.1 The infinite chain of Ge92--clusters in [Rb2(4,2,1,1-crypt)]Ge9•en……………….. 96 Figure 7.2 Two views of the new molecule named 4,2,1,1-crypt that is of two 2,1,1-crypt molecules……………………………………………………………………………… 98 Figure 7.3 Shown are the interactions between the sequestered rubidium cations and the chain of clusters…………………………………………………………………..….. 99 Figure 8.1 The deltahedral cluster Sn94- in [K-(12C4)2]2[K-12C4]2[Sn9]•4en is shown with the two [K-(12C4)] countercations that coordinate to two of its edges………... 110 Figure 8.2 Shown are the two cations K1 and K2 in [K-(12C4)2]2[K-12C4]2 [Sn9]•4en that are sandwiched by two molecules of 12-crown-4 each. The latter are staggered and tilted with respect to each other around K1, but are eclipsed and parallel around K2…………………………………………………………….……………… 110

viii

SYMBOLS

12c4

12-crown-4 ether; 1,4,7,10-tetraoxacyclododecane

15c5

15-crown-5 ether; 1,4,7,10,13-pentaoxacyclopentadecane

18c6

18-crown-6 ether; 1,4,7,10,13,16-hexaoxacyclooctadecane

2,1,1-crypt

4,7,13,18-tetraoxa-1,10-diazabicyclo[8.5.5]eicosane

2,2,1-crypt

4,7,13,16,21-Pentaoxa-1,10-diazabicyclo[8.8.5]tricosane

2,2,2-crypt

4,7,13,16,21,24-hexaoxa-1,10-diazabicyclo-[8.8.8]-hexacosane

2c-2e bond two-centre-two-electron bond en

ethylenediamine

mes

1,3,5-Trimethylbenzene

monoglyme ethylene glycol dimethyl ether Pn

pnictogens (the elements from group 15)

THF

tetrahydrofuran

tol

toluene

Tt

tetrels (the elements from group 14)

ix

PREFACE

Organization of the dissertation

The subject of this dissertation is reactions of nine-atom germanium clusters in solutions. Most results obtained over the period of nearly five years have already been published or submitted for publication in well-respected and esteemed US and international journals. This dissertation is organized as a collection of seven separate papers, ordered as individual chapters, and one chapter with unpublished results. Each chapter describes the experimental details of the particular subject, contains tables and figures of its own and is followed by applicable list of references. This information is accommodated in 113 pages of original research work, which comprise 8 chapters, 6 tables and 17 figures. Additionally, the dissertation contains a section with brief summary and future prospects. The publication status of each of 7 papers is as follows:

x

CHAPTER 1 Paper 1 (communication)

J. Am. Chem. Soc. 2002, 124, 2442-2443

Highlighted in: Chemistry & Industry, 2003, August 4, p.29

CHAPTER 2 Paper 2 (communication)

J. Am. Chem. Soc. 2002, 124, 10990-10991

Highlighted in: Nachrichten aus der Chemie, 2002, December, p.1332

CHAPTER 3 Paper 3 (communication)

Inorg. Chem. 2003, 42, 5789-5791

Highlighted in: CHEMICAL & Engineering News, 2003, September 15, p.21 Highlighted in: Chemistry & Industry, 2003, October 20, p.28.

CHAPTER 4 Paper 4

J. Am. Chem. Soc. 2003, 125, 14059-14064

CHAPTER 5 Paper 5

Chem. Eur. J. 2004, 10, 3727-3733

CHAPTER 6 Unpublished results

xi

CHAPTER 7 Paper 6

C. R. Chimie 2004, submitted

CHAPTER 8 Paper 7

Appl. Organometal. Chem. 2003, 17, 373-376

The dissertation ends with an alphabetically arranged total list of references that are cited in all of the eight chapters and an extensive introduction, which follows.

Introduction Background. The interest in clusters and their bonding have been the subject of intensive investigation for several decades. One of the first examples of clusters of main-group element were the boranes, BnHn2-. Their structural uniqueness caught the interest of many scientists of that time. The shapes of the clusters match some of the so-called Platonic solids, geometric objects of high symmetry built of equilateral triangles. Some examples of these are the tetrahedron, the octahedron, the tricapped trigonal prism, the icosahedron etc. Each boron atom is bonded to 3, 4 or 5 other boron atoms from the cluster and is exo bonded to a hydrogen atom. Empirical rules for counting the number of electrons needed for cluster bonding were developed by Wade1,

2

and quickly

became very useful tools for understanding bonding in the boranes and the carboranes, C2Bn-2Hn. The rules state that the number of skeletal electrons

xii

needed for bonding in a closo-cluster depends only on the number of vertices (n) of the cluster and is 2n+2. Moreover, the number remains the same even if some of the vertices are removed, 2n + 4 and 2n + 6 for the corresponding nido- and arachno-derivatives, respectively, where n is the number of remaining vertices. The logical next step was to explore the remaining main-group elements (M) and to try to synthesize similar clusters. It was realized that the corresponding M-H bonds are not likely to be favorable for the heavier maingroup elements because of the large energy difference between the H 1s and M np orbitals. Syntheses of alternative “naked” clusters (no ligands) or interconnected (each cluster has other clusters for exo-ligands) have been successful. The story of homoatomic “naked” clusters traces back to the observations made by Joannis in 1890.3 He observed that the reaction of the characteristic blue solution of sodium in liquid ammonia with excess lead colored the solution in intense green. At that stage more attention was given to the precipitates which were found to have the compositions NaPb4 or NaPb2 depending on conditions.4 Kraus was the first to study the green sodium-lead solution, showing that it was an electrolyte and the green color was associated with the anion.5 He also observed that the lead solute would plate out on an anode and form at a lead cathode, that the lead solute could be precipitated on addition of normal lead(II) salts. The electo-chemistry observations led Kraus to speculate that the polyanionic salt in equilibrium with the metal (assuming single solute species) was Na4Pb·Pb8 and that as electrolyte it contained the anion Pb94-.6,

xiii

7

Similar

observations with antimony and tellurium led him to predict Sb73- and Te42-, respectively. Further information on these anions came from extensive systematic studies by Zintl and co-workers published in the 1930s and the anions were labeled Zintl anions.8-11 Zintl recognized that such solutions formed from very “polar” intermetallics where the more electronegative element was a metalloid. Zintl was also a pioneer in determining the structures and properties of such polar intermetallics subsequently labeled “Zintl phases”.12, 13 A breakthrough in structural characterization of Zintl anions was in 1976 when Kummer and Diehl isolated the compound Na4Sn9·7en with Sn94clusters.14,

15

It was made from an alloy with nominal composition NaSn2.4-2.5

which was dissolved in ethylenediamine, and the compound was precipitated by the addition of monoglyme or THF. John Corbett discovered more general solution to the isolation of diverse examples of Zintl anions by using a sequestering agent for the cation, the compound known as 2,2,2-crypt.16, 17 The selection of this macrocyclic ligand was guided by reports by Dye and co-workers that this octadentate agent not only greatly enhances the solubility of the alkali metals,18 but also allows the isolation of the remarkable (Na+-2,2,2crypt)Na-.19 The radius of 2,2,2-crypt is approximately 1.35 – 1.40Å which makes it ideal for K+ (ionic radius 1.33-1.44Å) and also works quite well for Na+ (0.95-1.12Å) and Rb+ (1.40-150Å).20 The first characterized polyatomic anion of a metallic element with cryptated countercation is Sb73- in the salt (Na-2,2,2-crypt)3Sb7.16,

17

This

was followed by Sn94- and Pb52- in (Na-2,2,2-crypt)4Sn9 and (Na-2,2,2-crypt)2Pb9,

xiv

respectively.21-23 For a long time the addition of sequestering 2,2,2-crypt to solutions of Zintl anions was the only method to obtain crystals of good quality for single crystal structure determination. Recently, it was shown that the use of 18crown-6 ether also leads to single crystals with ordered clusters.24, 25 The cavity radius of 18-crown-6, 1.34-1.43Å,20 is ideal for K+, but works well for Rb+ also. It should be pointed out that the crown ether coordinates usually on one side of the cation. This makes it very different from cryptands which wrap the whole cation and do not leave any opening in its coordination sphere. Therefore, a cation sequestered in cryptand is completely incapable for any interactions with the anionic part of the structure. In contrast, crown ethers leave the cations quite exposed for additional interactions with the anions or the molecules of the solvents. Synthetic explorations with 2,2,1-crypt, 2,1,1-crypt, 15-crown-5 or 12crown-4 have generally been less productive, but not impossible. The list of polyatomic anions isolated from solutions and characterized by single crystal X-ray diffraction prior to this thesis included Tt52-, Tt9n- (n=3, 4 Td = Ge, Sn, Pb),26 [Ge9-Ge9]6-,27 chains of [-(Ge92-)-]1∞, 28 TlSn93-,29 Pn42-, Pn73-, Pn113(Pn = As, Sb, Bi),30 In4Bi53-, and InBi32-,31. In addition in 1997 Sevov and co-workers showed that the nine atom clusters Tt94- exist also in “neat” solids, i.e. solids made by solids-state reactions. Thus A4Tt9 where A = alkali metal contains Tt94- while A12Tt17 contains both Tt44and Tt94-.32-35 The clusters in these compounds are well separated from each other by the alkali-metal atoms. They can be viewed as salt-like compounds with alkali-metal cations and polyatomic polyanions. All they are soluble in

xv

ethylenediamine while 2,2,2-crypt or 18-crown-6 increase their solubility significantly. Introduction to group 14 nine-atom clusters. The elements of the carbon group themselves are especially prominent with respect to technical applications, chemical reactivity and chemical bonding. Going down the periodic table in group 14, the transition from insulating elements to metals occurs. Under standard conditions the elements above Sn are insulators or semiconductors. Sn has the outstanding property that the semiconducting low temperature α-form and the metallic β-form are almost equal in energy. Even the energy difference of 1 kcal/mol is small, chemical bonding changes from localized 2c-2e bonds and coordination number four for all atoms (diamond structure) to metallic bonding with a coordination number higher than the valence of Sn. The phase transition from α- to β-modification of Si and Ge are achieved by applying high pressures. The heaviest element Pb is a typical metal with delocalized bonds, but electrons form localized chemical bonds when Pb is alloyed with electropositive metals. The change in chemical bonding and thus the ability to form structures in which atoms have higher coordination numbers originate from the central position in the periodic table. This enables these elements to form homoatomic clusters in solid state, in solution, and in gas phase. The properties of the elements from group 14 and the properties of their clusters make them highly interesting candidates for building blocks in materials. The shape of the nine-atom clusters can be described generally as tricapped trigonal prisms elongated along one, two or three of the prismatic

xvi

edges (parallel to the three fold axes). Shown in Figure 1 are four idealized shapes for such clusters and the corresponding schematic molecular orbital diagrams. The cluster to the left (Fig. 1a) is the classical tricapped trigonal prism where the length of the vertical prismatic edges are similar to that of the remaining edges. The corresponding MO diagram has a HOMO-LUMO gap at 20 bonding electrons in agreement with the Wade’s rules for closo-species such as Tt92-, 2n + 2=20 for n=9. The classical example of such cluster is [B9H9]2-.36 For this particular geometry there is also a relatively large gap above the LUMO that makes the orbital quite attractive for additional electrons. Thus, the known paramagnetic Ge93-, Sn93-, and Pb93- clusters have 21 electrons where the extra electron occupies this particular orbital.37-39 The latter is made predominantly of pz orbitals of the atoms forming the trigonal prism and is bonding within each of the triangular bases but antibonding between them. This makes it is energy very sensitive to the height of the prism. Elongation of one vertical prismatic edge (shown in Fig. 1b) lower the orbital significantly. The resulting cluster is actually the classical monocapped square antiprism, C4v (shown lying on its side). This cluster is nido-species, and the 22 bonding electrons are in agreement with Wade’s rules, 2n+4 = 22 for n=9. The cluster shown in Fig. 1c is the result of elongation of one more vertical edge of the prism. The orbital in question drops further in energy because the antibonding interactions are lowered even more, but the HOMO-LUMO gap is still at 22 electrons. It should be pointed out that the degree of elongation of the two edges is typically smaller than that of only one edge. For example the one long edge in one of the Pb94- clusters in K4Pb9 is

xvii

4.376 Å while the two long edges of another Pb94- cluster in the same compound are both 3.895 Å.34 After elongation of all tree edges of the prism, we get the cluster shown in Fig. 1d which is again a tricapped trigonal prism, D3h, but with longer vertical edges. The molecular diagram does not change much and the cluster is stable with 22 electrons. An example of this geometry is Bi95+ found in Bi[Bi9](HfCl6)3.40. In general, the usual simple classification of nine-atom clusters into just two geometrical classes, tricapped trigonal prism and monocapped square antiprism, is not sufficient for an accurate description. It is perhaps better to discuss each particular cluster more in the context of the stoichiometry of the host compound, the possible charges based on this stoichiometry, and the corresponding molecular orbital diagram.

D3h

C4v

C2v

D3h

c

d

22 electrons 20 electrons

a

b

Figure1. Electronic structure of 9-atom deltahedral clusters

xviii

The first 3 geometries in Fig. 1 have two types of vertices, four- and fivebonded. The four-bonded vertices are the capping atoms and the atoms of the open square face(s). According to several theoretical calculations on nine-atom clusters the five-bonded positions carry lower negative charges than the fourbonded ones. For example, extended Hückel calculations on Ge94- and Sn94- give charges of -0.57, -0.45, -0.32 for the open square based, the capping atom, and five-bonded atoms.41 The distances in the clusters also depend strongly on the bonding of the corresponding vertices. Shortest are the distances between four-bonded atoms, while the longest bonds are formed between five-bonded atoms. For example the ranges of distances for Ge93- clusters are 2.54-2.62 Å and 2.78-2-91 Å, respectively.37 Clusters with charge 3-, Tt93-, are with odd number of delocalized electrons, i.e. delocalized radicals, and are paramagnetic. Several compounds have been investigated by EPR and magnetization measurements. The EPR signals of powder samples and of single crystals of compounds containing Tt93have been analyzed in detail for Tt = Ge, Sn and Pb.42 The simulations using three g tensor components of the anisotropic signals of powdered samples measured at 77K led to at least two different tensor components and indicated low symmetry Tt93- clusters. The g values of the powders were confirmed by EPR measurements of single crystals of (K-2,2,2-crypt)[Ge93-]·0.5en, (K-2,2,2crypt)6[Ge93-][Ge93-]·0.5en, (K-2,2,2-crypt)6[Sn93-][Sn93-]·1.5en·0.5tol and (K-2,2,2crypt)3[Pb93-]·1.5en.37, 42 Such measurements are very helpful when assigning the

xix

charge of the cluster, especially in compounds containing two independent clusters. For example, Corbett and co-workers assumed that the two clusters in (K-2,2,2-crypt)6[(Ge9)(Ge)9]·2.5en are Ge92- and Ge94- due to their different shapes.43 However, no measurements were done to confirm this. Very similar compound (K-2,2,2-crypt)6[Ge93-][Ge93-]·0.5en reported by Fassler is EPR active and suggests that perhaps the clusters in Corbett’s compound are also each of charge 3-.37 Tin and lead clusters in solution were studied by Rudolph and co-workers using

119

Sn and

207

Pb NMR.44 They showed that the Sn94- and Pb94- anions in

solution have only one NMR signal and therefore all atoms are equal on the NMR time scale and the polyhedral skeleton is apparently very flexible. The clusters have been studied also by Raman spectroscopy.45-49 For a nine-atom species with C4v symmetry one expects 21 fundamentals, Γvib = 4A1 (R, IR) + A2 (-,-) + 3B1 (R) + 3B2 (R) + 5E (R, IR), with 20 Raman active modes. Generally four to five broad bands are observed, a phenomenon that was attributed to accidental coincides and low intensities. Therefore Raman spectroscopy of solid samples cannot be used for analysis of small distortions, but it can prove the presence of the clusters in single crystals, powder, in noncrystalline samples and in solution.46, 48 Gas phase studies of the binary systems Cs/Sn and Cs/Pb show intensity maxima for the cationic molecules [Cs3Sn5]+, [Cs3Sn9]+, and [Cs3Pb9]+.50 These results support the existence of the polyanions Sn52-, Sn92-, and Pb92- in the gas phase. Time-of-flight mass spectrometric investigations showed that the gas

xx

phase anions can be easily achieved by laser desorption.51 The ion formation strongly depends on the nature of the starting materials used for the desorption experiments and can lead to rather large anionic clusters of the elements Ge, Sn, and Pb without using special gas phase clustering conditions or additional ionization process. Cluster anions from pure elements were observed with a maximum of 6 atoms and their relative intensities decreased exponentially with increasing cluster size. Larger clusters were observed if alkali metals were present. Prior to this thesis there were only several reports treating reactions with nine-atom Zintl anions. The reaction of Tt94- (Tt = Sn and Pb) with M(CO)3(mes) (M = Cr, Mo, W) leads to the addition of M(CO)3 to the open square of the nidocluster.52-54 Applying the isolobal analogy the resulting [Tt9Cr(CO)3]4- can be regarded as ten-atom closo-clusters in which the “zero-skeletal electron” fragment Cr(CO)3 occupies one vertex. The reactions of Ni(CO)2(PPh3)2 and Tt94(Tt = Ge, Sn) in ethylenediamine results in [Ge9(µ10-Ge)Ni-PPh3]2- and [Sn9Ni2(CO)]3- respectively.55,

56

The anions were described as a ten-atom

clusters [Tt9Ni] centered by one Ge or Ni atoms respectively. In fact, it is difficult to distinguish Ge and Ni atoms by X-ray diffraction. It is very likely that the central atom in the former is also Ni, and not Ge. Similar reaction between Pt(PPh3)4 and Sn94- leads to [Sn9Pt2PPh3]2-.56 The much desired functionalization of the nine-atom clusters by covalently exobonded groups was elusive until 1999.27 The first example of exo-bond to such a cluster was the synthesis of a dimer nine-atom clusters [Ge9-Ge9]6-. Two

xxi

Ge9 clusters are connected via an exo bond of 2.488 Å that corresponds to a GeGe single bond. This proved possible attaching main-group substituents to such clusters. After all, in [Ge9-Ge9]6- each cluster is a substituent to the other one. Furthermore, later it was shown that these clusters can have two exo bonds as in the chains of [-(Ge92-)-]1∞,28 The research described in this thesis is related to reactions of the nine-atom clusters with various reagents in order to add substituent groups. Furthermore, based on the existence of the dimer of clusters it was only natural to also look for larger oligomers.

xxii

References [1]

Wade, K. J. Chem. Soc. D 1971, 792.

[2]

Wade, K. Adv. Inorg. Chem. Radiochem. 1976, 18, 1.

[3]

Joannis, A. C. R. Hebd. Seances Acad. Sci 1891, 113, 795.

[4]

Joannis, A. C. Ann. Chi. Phys. 1906, 7, 75.

[5]

Kraus, C. A. J. Am. Chem. Soc. 1907, 29, 1557.

[6]

Kraus, C. A. J. Am. Chem. Soc. 1922, 44, 1216.

[7]

Kraus, C. A. Trans. Am. Electrochem. Soc. 1924, 45, 175.

[8]

Zintl, E.; Goubeau, J.; Dullenkopf, W. Z. Phys. Chem., Abt. A 1931, 154, 1.

[9]

Zintl, E.; Harder, A. Phys. Chem., Abt. A 1931, 154, 47.

[10]

Zintl, E.; Dullenkopf, W. Z. Phys. Chem., Abt. B 1932, 16, 183.

[11]

Zintl, E.; Kaiser, H. Z. Anorg. Allg. Chem. 1933, 211, 113.

[12]

Zintl, E. Angew. Chem. 1939, 52, 1.

[13]

Laves, F. Naturwissenschaften 1941, 29, 244.

[14]

Kummer, D.; Diehl, L. Angew. Chem., Int. Edit. 1970, 9, 895.

[15]

Diehl, L.; Khodadadeh, K.; Kummer, D.; Strahle, J. Chem. Ber. 1976, 109, 3404.

[16]

Corbett, J. D.; Adolphson, D. G.; Merryman, D. J.; Edwards, P. A.; Armatis, F. J. J. Am. Chem. Soc. 1975, 97, 6267.

[17]

Adolphson, D. G.; Corbett, J. D.; Merryman, D. J. J. Am. Chem. Soc. 1976, 98, 7234.

[18]

Lok, M. T.; Tehan, F. J.; Dye, J. L. J. Phys. Chem. 1972, 72, 2975.

xxiii

[19]

Tehan, F. J.; Barnett, B. L.; Dye, J. L. J. Am. Chem. Soc. 1974, 96, 7203.

[20]

Dillon, R. E. A.; Shriver, D. F. Chem. Mater. 1999, 11, 3296.

[21]

Corbett, J. D.; Edwards, P. A. J. Chem. Soc., Chem. Commun. 1975, 984.

[22]

Corbett, J. D.; Edwards, P. A. J. Am. Chem. Soc. 1977, 99, 3313.

[23]

Edwards, P. A.; Corbett, J. D. Inorg. Chem. 1977, 16, 903.

[24]

Fassler, T. F.; Hoffmann, R. J. Chem. Soc., Dalton 1999, 3339.

[25]

Fassler, T. F.; Hoffmann, R. Angew. Chem., Int. Edit. 1999, 38, 543.

[26]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[27]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[28]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[29]

Burns, R. C.; Corbett, J. D. J. Am. Chem. Soc. 1982, 104, 2804.

[30]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[31]

Xu, L.; Sevov, S. C. Inorg. Chem. 2000, 39, 5383.

[32]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

[33]

Queneau, V.; Todorov, E.; Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[34]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[35]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[36]

Guggenberger, L. J. Inorg. Chem. 1968, 7, 2260.

[37]

Fassler, T. F.; Schutz, U. Inorg. Chem. 1999, 38, 1866.

[38]

Fassler, T. F.; Hunziker, M. Z. Anorg. Allg. Chem. 1996, 622, 837. xxiv

[39]

Fassler, T. F.; Hunziker, M. Inorg. Chem. 1994, 33, 5380.

[40]

Friedman, R. M.; Corbett, J. D. Inorg. Chem. 1973, 12, 1134.

[41]

Lohr, L. L. Inorg. Chem. 1981, 20, 4229.

[42]

Fassler, T. F.; Hunziker, M.; Spahr, M. E.; Lueken, H.; Schilder, H. Z. Anorg. Allg. Chem. 2000, 626, 692.

[43]

Belin, C. H. E.; Corbett, J. D.; Cisar, A. J. Am. Chem. Soc. 1977, 99, 7163.

[44]

Rudolph, R. W.; Wilson, W. L.; Parker, F.; Taylor, R. C.; Young, D. C. J. Am. Chem. Soc. 1978, 100, 4629.

[45]

Campbell, J.; Schrobilgen, G. J. Inorg. Chem. 1997, 36, 4078.

[46]

vonSchnering, H. G.; Baitinger, M.; Bolle, U.; Carrillo Cabrera, W.; Curda, J.; Grin, Y.; Heinemann, F.; Llanos, J.; Peters, K.; Schmeding, A.; Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[47]

Somer, M.; Carrillo-Cabrera, W.; Peters, E. M.; Peters, K.; von Schnering, H. G. Z. Anorg. Allg. Chem. 1998, 624, 1915.

[48]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[49]

Somer, M.; Carrillo-Cabrera, W.; Peters, E. M.; Peters, K.; Kaupp, M.; von Schnering, H. G. Z. Anorg. Allg. Chem. 1999, 625, 37.

[50]

Martin, T. P. Angew. Chem., Int. Edit. 1986, 25, 197.

[51]

Fassler, T. F.; Muhr, H. J.; Hunziker, M. Eur. J. Inorg. Chem. 1998, 1433.

[52]

Eichhorn, B. W.; Haushalter, R. C.; Pennington, W. T. J. Am. Chem. Soc. 1988, 110, 8704.

[53]

Eichhorn, B. W.; Haushalter, R. C. J. Chem. Soc., Chem. Commun. 1990, 937.

[54]

Kesanli, B.; Fettinger, J.; Eichhorn, B. Chem. Eur. J. 2001, 7, 5277.

xxv

[55]

Gardner, D. R.; Fettinger, J. C.; Eichhorn, B. W. Angew. Chem., Int. Edit. 1996, 35, 2852.

[56]

Kesanli, B.; Fettinger, J.; Gardner, D. R.; Eichhorn, B. J. Am. Chem. Soc. 2002, 124, 4779.

xxvi

ACKNOWLDGEMENTS

I sincerely thank my advisor Professor Slavi C. Sevov for the chance he gave me to work in his group on this project. Thank you Slavi for your suggestions, advices and support! Thank you for the freedom which you gave me to explore new fields in solution chemistry of Zintl ions! I appreciate all these very much! I would like to thank my friends, Svilen Bobev and Franck Gascoin, who helped me during the difficult times, especially in the beginning of the PhD study. I thank my friend and colleague Iliya Todorov for making my life easier. He made most of the precursors for my reactions. I thank J. Zajicek and B. Milosavljevic for helping with Sn-NMR and solid UV-VIS spectra, respectably. I thank my dissertation committee Professor Lappin, Professor Lieberman and Professor Henderson for their suggestions and considerations. I am also grateful to the department of Chemistry and Biochemistry, all faculty members and staff, for accepting me as part of it, for their help and, of course, for giving me the honor of the Rohm & Haas Outstanding Graduate Student Award (2004) and of the Reilly Fellowship (2003 and 2004).

xxvii

Also, I would like to thank my family. I thank my wife Vesela for her love patience, understanding, and the most important for our beautiful 1-year-old son Alexander. I extend my thanks to my mother Evdokiya and to my father Georgi for making me who I am today and for their support for the choices that I made.

xxviii

CHAPTER 1

[Ph2Bi–(Ge9)–BiPh2]2-: A DELTAHEDRAL ZINTL ION FUNCTIONALIZED BY EXO-BONDED LIGANDS

1.1. Abstract A

new

compound,

(K-2,2,2-crypt)2[Ph2Bi-Ge9-BiPh2]·en

(1),

was

synthesized, isolated and characterized. It crystallized in triclinic space group P-1 (a = 13.522Å, b = 13.603Å, c = 24.498Å, α= 79.70o, β = 77.00o, γ = 65.88o, V = 3988.1Å3, z = 2). The compound is the first deltahedral Zintl ion functionalized by exo-bonded ligands.

1.2. Introduction The first nine-atom deltahedral Zintl ion of the carbon group, Sn94-, was structurally characterized 28 years ago.1 Since then the corresponding silicon, germanium and lead analogs have been also characterized, both in the solidstate and in solution (except for silicon).2-9 These clusters are distorted monocapped square antiprisms or tricapped trigonal prisms elongated along one or more of the prismatic edges (parallel to the three-fold axis). The bonding in the clusters is achieved by delocalized electrons as in the cage-like boranes. The

1

numbers of electrons needed for bonding follow the Wade's rules rather than the octet rule.10 The clusters are usually made by dissolving a binary compound of the corresponding element and alkali metal in ethylenediamine or liquid ammonia, and a compound containing the cluster is crystallized by addition of cation-sequestering agents such as 2,2,2-crypt (4,7,13,16,21,24-hexaoxa-1,10diazabicyclo- [8.8.8]-hexacosane) or crown ethers.8, 9 Ever since the discovery of these clusters, numerous attempts to use them in various reactions have been made, some successful and some not. Capping the open square face of the cluster by transition metals to form closo-clusters of [E9M(CO)3]4- for E = Sn, Pb and M = Cr, Mo, W has been successful.11-13 However, the much desired functionalization of the clusters by covalently exo-bonded groups has been elusive until now. The closest to this goal is the recently reported dimerization and polymerization of such Ge9 clusters to form [Ge9–Ge9]6- and [-(Ge92-)-]1∞ respectively, where the clusters are exo-bonded to each other by normal 2c–2e bonds.14, 15 Thus, in a way, this is functionalization of the clusters by each other. This achievement indicated that perhaps other main-group substituents could be exo-bonded in a similar way. We report here the first such species of a nine-atom germanium cluster ligated by two diphenylbismuth groups that are exo-bonded to two opposite germanium vertices of the open face of the cluster, [Ph2Bi–(Ge9)– BiPh2]2-.

2

1.3. Experimental Section Synthesis. All operations were carried out in inert atmosphere or under vacuum. The precursor K4Ge9 was made from stoichiometric mixture of the elements (Alfa-Aesar) heated at 900 °C for 2 days in sealed (by arc-welding) niobium containers that were jacketed in evacuated fused-silica ampules. The major phase of the product was K12Ge17, isostructural with Rb12Si17,3 which is known to contain both nine- and four-atom clusters, E94- and E44-. Inside a drybox, approximately 0.0272 mmol of K4Ge9 were dissolved in 1 ml of ethylenediamine (redistilled and packaged under nitrogen, 99.5+%, Aldrich) and 0.127 mmol of 2,2,2-crypt (98%, Acros) in a test-tube (brown-red solution). BiPh3 (Aldrich), 0.068 mmol, was then added, and the resulting bright-red solution was layered with 3 ml of toluene (dried and kept over sodium, 99.5%, Fisher). Brightred plates formed on the walls and the bottom of the test-tube. The mother liquor was decanted after a week and the crystals were dried (ca. 8 mg product, about 22 % yield based on the amount of dissolved precursor). IR (in KBr): 694, 720, and 730 cm-1 (out of plane C-H bending); 990, 1010, and 1052 cm-1 (in plane C–H bending); 1466 and 1560 cm-1 (C–C stretching); 3020 and 3051 cm-1 (CH stretching). Structure Determination. X-Ray diffraction data were collected from a red, plate-like crystal of 1 (0.35 x 0.28 x 0.06 mm) with graphite-monochromated Mo Kα radiation on a Bruker APEX diffractometer with a CCD area detector at 100 °C. The structure was solved by direct methods in P-1 and refined on F2 3

(full matrix, absorption corrections with SADABS) using the SHELXTL V5.1 package. Crystal data: triclinic, P-1, a = 13.5216(9), b = 13.6028(9) and c = 24.498(1) Å, α = 79.700(1), β = 77.003(1), and γ = 65.881(1)°, V = 3988.1(5) Å3, Z = 2, µ = 78.94 cm-1, dcalc = 1.891 g/cm3, R1/wR2 = 4.57/11.33% for 11680 unique observed reflections (I ≥ 2σI) and 826 variables (R1/wR2 = 5.64/11.75% for all data). An X-ray crystallographic file, in CIF format, is available free of charge via the Internet at: http://pubs3.acs.org/acs/journals/supporting_information.page?in_manuscript=ja 017813n 1

H-NMR and

13

C-NMR spectra were taken in ethylenediamine on a Varian

300 MHz spectrometer at room temperature. 1H-NMR (THF-d8): δ 7.07–7.17 (m; 12H) and 7.17–7.24 (m; 8H).

13

C NMR: δ 126.14 (p-C6H5), 129.03 (m-C6H5),

129.74 (o-C6H5).

1.4. Results and Discussions The new compound (K-2,2,2-crypt)2[Ge9(BiPh2)2]·en was synthesized by reacting ethylenediamine solution of a precursor of nominal composition K4Ge9 with BiPh3. The major phase of the precursor was K12Ge17, isostructural with Rb12Si17, which contains both Ge94- and Ge44- clusters.3 The color of the solution changed quickly from brown-red (for the dissolved precursor) to extremely deep-red upon reacting with BiPh3. The solution was carefully layered

4

with toluene and bright red crystals of the new compound grew to sizes suitable for structure determination in a week. The new compound contains nine-atom germanium clusters with two exobonded diphenylbismuth groups (Figure 1.1). The two Bi–Ge distances, 2.7327(8) and 2.7332(8) Å, are virtually identical, and compare well with other single-bond

distances

such

as

those

in

Bi{µ-Ge(C6F5)2}3Bi

and

[{(C6F5)2Ge}2Bi–Bi{Ge(C6F5)2}2] with distances in the range of 2.73–2.77 Å, as well as with Pauling's single-bond distance of 2.752 Å.16-18 The Bi–C distances in [Ph2Bi–(Ge9)–BiPh2]2-, 2.257(7)–2.277(8) Å, and the two C–Bi–C angles, 93.6(2) and 93.2(2)°, compare well with those of BiPh3, 2.237–2.273 Å and 92.7–94.7°, respectively.19 The effect of the inert s-pair of electrons at the bismuth makes the Ge–Bi-C angles also very close to orthogonal, 94.7(1)– 98.0(1)°. The Ge–Ge distances in the cluster follow the general pattern observed for other nine-atom clusters, and are quite similar to those in the various Ge9 clusters with different charges and connectivity.2,

9

Thus, they increase in the

following order: distances between four-bonded atoms (Ge6, 7, 8, 9), 2.553(1)– 2.585(1) Å (Figure 1.1), distances from four-bonded (Ge1, 6, 7, 8, 9) to fivebonded atoms (Ge2, 3, 4, 5), 2.578(1)– 2.653(1) Å, and distances between fivebonded atoms (Ge2, 3, 4, 5), 2.717(1)–2.826(1) Å. The shape of the Ge9 cluster is reminiscent of both a monocapped square antiprism with distorted open face (rhombic rather than square) and a tricapped trigonal prism with one elongated

5

1 2 5

Bi1

9

4

6

8

3

7

Bi2

Ge 9

Figure 1.1. ORTEP drawing of [Ph2Bi–(Ge9)–BiPh2]2- (70 % probability thermal ellipsoids). The numbering of the germanium atoms is shown.

prismatic edge.

The diphenylbismuth ligands are bonded to two opposite

vertices of the distorted open face, specifically the two vertices Ge7 and Ge9 (Figure 1.1) along the short diagonal of the rhombus. This diagonal is 2.985(1) while the corresponding long diagonal is 4.169(1) Å. Exactly the same type of distortion is observed in the chains of (-Ge9-) where the same two vertices of the open face are exo-bonded to other clusters and the corresponding short and long diagonals of the rhombus are 3.194 and 3.942 Å, respectively.15 In the dimer of Ge9–Ge9, where only one vertex is exo-bonded, the distortion is somewhat less pronounced with distances of 3.433 and 3.848 Å for the two diagonals.14 Thus, the two exo-bonded vertices are closer than in typical nido-species but are

6

farther than in typical closo-species. However, there is clearly no interactions between these two germanium atoms and, therefore, the reason for the distortion should be sought in effects induced by the exo-bonding. The replacement of a lone pair of electrons with a bonding one, the exo-bonds at Ge7 and Ge9, leads naturally to lesser repulsion with the bonds of the cluster, and this translates into larger endo-cluster angles at those atoms. Fenske-Hall (self-consistent field) and extended-Hückel molecular orbital calculations carried out for [Ph2Bi–(Ge9)– BiPh2]2- and the naked Ge94- cluster with the same shape confirmed the nidocharacter of both,20, 21 i.e. 22 cluster-bonding electrons (2n + 4 for n = 9), despite the shorter diagonal of the open face. It should be pointed out that although the charge of the former is 2- the bonding within the cluster and its stability have not changed since the number of bonding electrons is the same. What have changed are the two lone pairs of electrons that would have existed at Ge7 and Ge9 had the cluster been naked. They have been simply replaced by bonding pairs of electrons. As it has been discussed elsewhere,15 the highest occupied orbitals in the naked nido-Ge94- clusters are exactly the lone pairs of electrons at the atoms of the open face. This is why these particular atoms are most reactive and the first to form exo-bonds. The diphenylbismuth ligands have their phenyl rings positioned away from the cluster in the solid state (Figure 1.1) but are very likely fluctional in solution by rotation around the Bi–Ge bond. Crystals of the new compound can be redissolved in ethylenediamine and THF. The solubility of the compound opens

7

venues for further studies of the reactivity of the functionalized clusters towards eventual coordination to transition metals via the lone pairs of the bismuth atoms.

1.5. References [1]

Diehl, L.; Khodadadeh, K.; Kummer, D.; Strahle, J. Chem. Ber. 1976, 109, 3404.

[2]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Ed. 1997, 36, 1754.

[3]

Queneau, V.; Todorov, E.; Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[4]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[5]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[6]

vonSchnering, H. G.; Baitinger, M.; Bolle, U.; Carrillo Cabrera, W.; Curda, J.; Grin, Y.; Heinemann, F.; Llanos, J.; Peters, K.; Schmeding, A.; Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[7]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[8]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[9]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[10]

Wade, K. Adv. Inorg. Chem. Radiochem. 1976, 18, 1.

[11]

Eichhorn, B. W.; Haushalter, R. C.; Pennington, W. T. J. Am. Chem. Soc. 1988, 110, 8704.

8

[12]

Eichhorn, B. W.; Haushalter, R. C. J. Chem. Soc., Chem. Commun. 1990, 937.

[13]

Kesanli, B.; Fettinger, J.; Eichhorn, B. Chem. Eur. J. 2001, 7, 5277.

[14]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[15]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[16]

Bochkarev, M. N.; Razuvaev, G. A.; Zakharov, L. N.; Struchkov, Y. T. J. Organomet. Chem. 1980, 199, 205.

[17]

Pankratov, L. V.; Zakharov, L. N.; Bochkova, L. I.; Fukin, G. K.; Struchkov, Y. T. Izv. Akad. Nauk SSSR, Ser. Khim. 1994, 921.

[18]

Pauling, L. In Nature of the chemical bond, 3rd ed.; Cornell University Press: Ithaca, 1960; 'Vol.' p 403.

[19]

Jones, P. G.; Blaschette, A.; Henschel, D.; Weitze, A. Z. Kristallogr. 1995, 210, 377.

[20]

Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768.

[21]

Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.

9

CHAPTER 2

[Ge9=Ge9=Ge9]6-: A LINEAR TRIMER OF 27 GERMANIUM ATOMS

2.1. Abstract The new compound (Rb-2,2,2-crypt)6[Ge9=Ge9=Ge9]·3en (1) was isolated from the reactions of Ph3As and Ph3P with ethylenediamine solution of the precursor compound Rb4Ge9. The compound crystallized in monoclinic space group P21/n (a = 15.246Å, b = 47.223Å, c = 26.657Å, β = 98.37o, V = 18987Å3, z = 4). The trimer of the clusters is linear, and the nine-atom germanium clusters are bonded with two exo-bonds to each other.

2.2. Introduction Nine-atom deltahedral clusters of the heavier analogs of the carbon group exist in solutions of ethylenediamine and liquid ammonia, and can be crystallized from them with various countercations.1-3 More recently they have been found in neat solids as well.4-8 The clusters carry charge of 4- or 3- and have the shape of tricapped trigonal prisms elongated along one, two, or three of the prismatic edges (those parallel to the three-fold axis of the prism). As in the well-known cage-like boranes, the bonding in these clusters is achieved by delocalized

10

electrons. Recent studies have shown that upon oxidation these clusters couple in different modes and various species can be crystallized with countercations of appropriate sizes and shapes. Thus, the first dimer [Ge9-Ge9]6- of such clusters was stabilized with a combination of naked cesium and cryptated potassium cations in Cs2(K-2,2,2-crypt)4[Ge9–Ge9] where 2,2,2-crypt is 4,7,13,16,21,24hexaoxa-1,10-diazabicyclo-[8.8.8]-hexacosane.9 The tricapped trigonal prisms in this case are elongated along one edge (equivalent to a monocapped square antiprism) and are connected by a simple 2c–2e exo-bond between vertices of the elongated edges in each cluster. Another example of oxidative coupling is the chain of [-(Ge92-)-]1∞ stabilized in (K-18c6])9[Ge9]•en with potassium cations sequestered by 18-crown-6 ether.10 The shape of the clusters in the chain is the same as in the dimer but with both vertices of the elongated edge in each cluster involved in exo-bonding with two neighboring clusters in the chain. The existence of the dimer and the chain showed that, contrary to earlier assumptions, exobonds to such deltahedral ions are possible. Furthermore, it showed that the oxidation of the cluster does not affect the bonding within the cluster and its stability, but rather replaces lone pairs of electrons with bonding pairs. We followed on this idea and studied reactions of controlled mild oxidation of these clusters with Ph3E where E = P, As, Sb, and Bi. The reactions with the two heavier analogues, Ph3Sb or Ph3Bi, provided the first functionalized deltahedral Zintl ions, [Ph2E–Ge9–EPh2]2-, with two exo-bonds at the vertices of the elongated edge of the Ge9-prism.11 Here we present the product of the oxidation

11

of the germanium deltahedral clusters with Ph3P and Ph3As, a trimer of 9-atom clusters, [Ge9=Ge9=Ge9]6-.

2.3. Experimental Section Synthesis. All operations were carried out in inert atmosphere or under vacuum. The precursor compound Rb4Ge9 (isostructural with Cs4Ge95) was made from the elements heated at 900 °C for 2 days in sealed (by arc-welding) niobium containers that were jacketed in evacuated fused-silica ampules. Inside a drybox 108 mg of it were placed in a test tube and were covered with 2 ml of ethylenediamine and 200 mg of 2,2,2-crypt. After stirring for a few minutes the color of the solution became intense brown-red (UV-VIS: 368, broad peak from 450-500 with a broad shoulder extending to 650-700), indicative of Ge94- and Ge93-. Next, Ph3As was added and the solution was heated to 70 °C. After a few hours the color changed to dark-green (UV-VIS: 494, 641 and 825 nm). After cooling to room temperature the solution was filtered and carefully layered with 3 ml of monoglyme. Needle- and wedge-like crystals of exactly the same darkgreen color (UV-VIS: somewhat red-shifted to 503, 655 and 838 nm) formed after a few days (ca. 32 mg product, which is approximately 61 % yield based on the amount of dissolved precursor). The virtually identical absorption spectra prove the existence of the trimers in the solution (Figure 2.1). The compound can be synthesized also by oxidation of an ethylenediamine solution of the same precursor with Ph3P or elemental As and Sb at the same temperature followed

12

by layering with toluene. The elemental As and Sb in the latter two reactions are reduced to the well-known anions E73-: 3Ge94- + 14E —> [Ge9=Ge9=Ge9]6- + 2E73-.

Figure 2.1. Comparison between the spectra of the trimer in solution and the trimer in solid state.

Structure Determination. X-ray diffraction data were collected from a dark-green brick of 1 (0.39 x 0.30 x 0.19 mm) with graphite-monochromated Mo Kα radiation on a Bruker APEX diffractometer with a CCD area detector at 100 K. The structure was solved by direct methods in P21/n and refined on F2 (full matrix, absorption corrections with SADABS) using the SHELXTL V5.1 package. Crystal data: a = 15.246(2), b = 47.223(7) and c = 26.657(4) Å, β = 98.373(3)°, V

13

= 18987(5) Å3, Z = 4, µ = 57.90 cm-1, dcalc = 1.72 g/cm3, R1/wR2 = 7.28/14.48 % for the observed data (I ≥ 2σI). All 27 Ge-atoms of the trimer are unique and not disordered. The six 2,2,2-crypt molecules around the Rb-cations, on the other hand, show different degrees of disorder around the central rubidium. An X-ray crystallographic file, in CIF format, is available free of charge via the Internet at: http://pubs3.acs.org/acs/journals/supporting_information.page?in_manuscript=ja 026679j Properties. Magnetic measurements were carried out on a Quantum Design MPMS SQUID magnetometer. The molar magnetic susceptibility of the compound (measured on 12 mg) is negative and temperature ion dependent within 50–250 K: -0.035 emu/mol.

2.4. Results and Discussions The new compound with this trimer, (Rb-2,2,2-crypt)6[Ge9=Ge9=Ge9]· 3en, was isolated from reactions of Ph3As or Ph3P with ethylenediamine solution of the precursor compound Rb4Ge9. The latter is an ionic compound made of isolated Ge94- clusters and Rb-cations. Upon heating to 70 °C the color of the solution changes from brown-red (for the dissolved precursor) to extremely darkgreen. After cooling and careful layering with toluene or monoglyme needles of exactly the same dark-green color were recovered after a few days. The trimer of clusters is linear, and the nine-atom germanium clusters are bonded with two exo-bonds to each other (Figure 2.2). Each cluster is a tricapped trigonal prism elongated along two prismatic edges. The trigonal prism 14

of each cluster is formed of atoms 2, 3, 4, 5, 6, and 8 (Figure 2.2) with elongated edges of 2–5 and 6–8 while the capping atoms are 1, 7, and 9. Thus, the lengths of the edges 2–5 and 6–8 are in the range 2.953–3.085 Å while those of the edge 3–4 are within 2.753–2.772 Å. Like in other deltahedral nine-atom clusters,3,

4

the distances around the atoms that are four-bonded within the cluster (all atoms except 3 and 4, see Figure 2.2), 2.484–2.745 Å, are shorter than those that are five-bonded (atoms 3 and 4), 2.523–2.772 Å. 1

7 8

6 4

2

7 5

8

6

9 3

9

4

3

4

9 5

2

6

8 7

1

3

2

5 1

Figure 2.2. ORTEP drawing of the trimer [Ge9=Ge9=Ge9]6- (50 % probability thermal ellipsoids). Each cluster is a tricapped trigonal prism with two elongated prismatic edges (shown as open bonds).

Most interesting and nontrivial to rationalize in the new species is the character of the bonds between the clusters and the resulting charge of 6-. The end clusters have two exo-bonds each while the middle cluster has four such bonds (Figure 2.2). This type of connectivity with two bonds between the clusters is quite unexpected, almost as unexpected as a single exo-bond between such clusters was prior to the discovery of [Ge9-Ge9]6-.9 As in the latter, the oxidation of these clusters does not involve cluster-bonding electrons but rather the higherenergy lone-pair electrons, specifically the lone pairs at the atoms of the

15

elongated prismatic edges. The four intercluster bonds are nearly equidistant in the narrow range 2.579–2.601 Å. These bond lengths are significantly longer than the 2-center–2-electron intercluster bonds of 2.488(2) and 2.486(1) Å found in the dimer [Ge9–Ge9]6- and the chain [-(Ge92-)-]1∞, respectively,9,

10

and

indicate that the bond order for the bonds in the trimer is lower than one. Furthermore, a bond order of one for each exo-bond would have resulted in a charge of 4- for the trimer and not the observed 6- since the central cluster would have been neutral and the end ones would have carried charges of 2- each. Fenske-Hall and extended-Hückel molecular orbital calculations were carried out for the trimer in order to find the origin of the 6- charge.12,

13

The

results showed that the clusters needed 22 bonding electrons each (the same as for clusters with one elongated edge), confirmed the overall charge, and showed that it was distributed almost evenly among the three clusters resulting in a diamagnetic trimer (in agreement with measurements – see experimental section). It was quite clear that this “unexpected” charge originated from the unusual way the clusters were bonded. Thus, the exo-bonds are not “radial” to the clusters as they are in the boranes, the dimers [Ge9–Ge9]6-,9 the chains[(Ge92-)-]1∞10 and the ligated clusters [Ph2Bi–(Ge9)–BiPh2]2-,11 but are rather “parallel” to the elongated edges of the prisms. This changes quite drastically the intercluster interactions between the outward-pointing hybrid orbitals. Focusing on the central cluster, the effect of this is that one molecular orbital of lone-pair type that would have been empty for a cluster with four radially-positioned exobonds becomes stabilized and filled by mixing with a cluster-type orbital for the

16

observed arrangement of four parallel exo-bonds. The same effect was observed for a modeled trimer of octahedral clusters connected in the same way. In order to study the effect in more detail the model was simplified even further to a planar cluster of four germanium atoms with four hydrogen ligands (Figure 2.3). The cluster with radial exo-bonds is a typical arachno-species [Ge4H4]2- (D4h) with 2n + 6 = 14 cluster-bonding and 4 x 2 = 8 exo-bonding electrons. However, the Walsh diagram, part of which is shown in Figure 2.3, shows that upon bending of the exo-bonds to become parallel to two Ge–Ge edges (D2h), two orbitals of different symmetry, b2g and a2g, in the former, become of the same symmetry, b1g, and mix in the latter. Naturally, one of the combinations is stabilized and the other destabilized. The stabilized combination, basically a combination of lone pairs pointing at the available space vacated by moving the exo-bonds away from the radial positions, is filled and the species become [Ge4H4]4- with two additional electrons. This is exactly the same effect that is observed in many systems upon symmetry reduction such as linear to bent AH2, or planar to pyramidal AH3, etc., systems where the higher symmetry prevents mixing of orbitals that are otherwise allowed to mix in the lower symmetry to produce similarly a stabilized lone pair in the vacated space.

17

[Ge4H4]2H

[Ge4H4]4-

H Ge

Ge

Ge

Ge H

H

H

Ge

Ge

H

H

Ge

Ge

H

* b1g

a2g

b2g b1g

Figure 2.3. Correlation (Walsh) diagram between two specific molecular orbitals in a square of germanium with four radially-pointing exo-bonds (left, D4h) and the same square with the exo-bonds parallel to a pair of Ge–Ge edges (right, D4h). The two orbitals on the left are of different symmetry, b2g and a2g, and can not mix, but are of the same symmetry, b1g, and mix to form stabilized and destabilized combinations.

Any deductions about the mechanism of formation of this trimer should account for the fact that: a) the trimers exist in solution, i.e. they are not assembled upon crystallization; b) the reaction with the heavier triphenylbismuth and antimony

derivatives provides the disubstituted species [Ph2E–Ge9–

EPh2]2- instead.11 The different outcome of the latter reactions is perhaps due to the lower stability of the heavier diphenylpnictides [Ph2Sb]- and [Ph2Bi]- as free anions and their apparent preference to bond as ligands to the clusters.14 The same species, on the other hand, are more stable as free anions for the lighter phosphorus and arsenic. Thus, in analogy with the reaction of lithium with

18

triphenylphosphine to produce phenyllithium and lithium diphenylphosphide, a proposed scenario for the reaction at this preliminary stage is: 3Ge94- + 3Ph3E —> [Ge9=Ge9=Ge9]6- + 3[Ph2E]- + 3Ph-

2.5. References [1]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[2]

Corbett, J. D. Struct. Bond. 1997, 87, 157.

[3]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[4]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

[5]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[6]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[7]

vonSchnering, H. G.; Baitinger, M.; Bolle, U.; Carrillo Cabrera, W.; Curda, J.; Grin, Y.; Heinemann, F.; Llanos, J.; Peters, K.; Schmeding, A.; Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[8]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[9]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[10]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[11]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

19

[12]

Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768.

[13]

Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.

[14]

Doak, G. O.; Freedman, L. D., Organometallic Compounds of Arsenic, Antimony, and Bismuth. ed.; Wiley: New York, 1970.

20

CHAPTER 3

[Ge9=Ge9=Ge9=Ge9]8-: A LINEAR TETRAMER OF NINE-ATOM GERMANIUM CLUSTERS, A NANOROD

3.1. Abstract A tetramer of nine-atom deltahedral germanium clusters and charge 8-, [Ge9=Ge9=Ge9=Ge9]8--, has been characterized as a (Rb-18C6)+ salt. The compound crystallized in triclinic space group P-1. The clusters are connected by pairs of parallel bonds, and the electrons are delocalized over the whole anion. The size of the tetramer is of nanorod dimensions, ca. 2 nm.

3.2. Introduction Recent developments in the area of deltahedral Zintl ions, i.e. borane-like anions of the heavier carbon-group elements, have brought new light to their chemistry and have changed major concepts about their reactivity. First, it was shown that nine-atom ions E94- of group 14 (E = Ge, Sn, Pb), previously known only as crystallized from solutions, exist also in the precursor Zintl phases A4E9 (A = alkali metal).1-5 This was unexpected6 and, for the first time, established a connection between Zintl phases and Zintl ions.1 Second, contrary to the belief

21

that exo-bonds to these deltahedral clusters are not viable, dimers and chains of germanium

clusters,

discovered.7,

8

[Ge9-Ge9]6-

and

[-(Ge92-)-]1∞,

respectively,

were

Next came the discovery that these clusters can not only exo-

bond between themselves, but also to other groups such as Ph2E where E = Sb or Bi and form [Ph2E-Ge9-EPh2]2-.9 These findings show that the deltahedral Zintl ions are in many respects similar to the cage-like boranes as they can also connect to each other and tolerate various substituents. The boranes are also known to fuse via edges or faces, but this has not been observed yet for the deltahedral Zintl ions. Closest to fused clusters is the trimer [Ge9=Ge9=Ge9]6where each pair of clusters is bonded by two bonds from two neighboring vertices on each cluster.10 Here we report the synthesis and characterization of a tetramer of such clusters, [Ge9=Ge9=Ge9=Ge9]8-, that is long enough to qualify for a nanorod.

3.3. Experimental Section Synthesis. All manipulations were performed in a nitrogen-filled glove box with a moisture level below 1 ppm. The precursor of Rb4Ge9 was made from a stoichiometric mixture of the elements (Rb: Strem, 99+ %; Ge: Acros, 99.999 %) heated at 900 °C for 2 days in niobium containers that were sealed by arcwelding and were then jacketed and evacuated in fused-silica ampules. 97 mg precursor (~ 9.7 x 10-5 mol) were dissolved in 1 ml ethylenediamine (Aldrich, 99.5+ %, redistilled and packaged under nitrogen). The solution was stirred for

22

10 min at 50 °C and after cooling quickly to room temperature, orange-brown suspension of the known Rb4Ge9•1.1en formed.11

Most of this suspension

dissolved after adding 106 mg of 18-crown-6 (4 x 10-4 mol, Acros, 99 %) and stirring it for 15 min at 50 °C. It was left for 1 hour and was then centrifuged and filtered.

The resulting black-green solution (this color is associated with

oligomerization) was split in two test tubes and one of them was layered with THF and the other with toluene.

Three days later crystals of (Rb-

18C6)8[Ge9=Ge9=Ge9=Ge9]•2en (1) and (Rb-18C6)8[Ge9=Ge9=Ge9=Ge9]• 6en (2) were recovered from the containers, respectively. Structure Determination. Several crystals of 1 and 2 were placed on a micro slide and were covered with oil (Paratone–N) inside the glove box. The slide was taken out and, under a microscope crystals were picked with thin glass fibers. For the best ones, X-ray diffraction data were collected with graphitemonochromated Mo Kα radiation at 100 K on a Bruker APEX diffractometer with a CCD area detector. The structures were solved by direct methods and were refined on F2 using the SHELXTL V5.1 package. Crystal data for compound 1: plate-like; dichroic; 0.18 x 0.08 x 0.02 mm; triclinic; P-1; a = 15.120(1), b = 16.250(1), and c = 18.717(1) Å, α = 83.30(1), β = 83.01(1), γ = 78.00(3)°; V = 4445.2(5) Å3; Z = 1; µ = 82.2 cm-1; ρcalc = 2.066 g.cm-3; R1/wR2 = 5.76/13.15 % for 10334 observed reflections (I ≥ 2σI). Crystal data for compound 2: platelike; dichroic; 0.11 x 0.10 x 0.03 mm; triclinic; P-1; a = 13.571(1), b = 15.885(2), and c = 22.645(3) Å, α = 99.355(3), β = 97.813(3), γ = 101.257(3)°; V = 4652(1) Å3; Z = 1; µ = 78.6 cm-1; ρcalc = 2.06 g.cm-3; R1/wR2 = 6.49/12.01 % for 6622 23

observed reflections (I ≥ 2σI). X-ray crystallographic files, in CIF format, are available free of charge via the Internet at: http://pubs3.acs.org/acs/journals/supporting_information.page?in_manuscript=ic0 34677%2b Calculations. DFT calculations were performed on the tetramer (with the observed geometry) using the B3LYP functional with 3-21G basis set.12

3.4. Results and Discussions The

new

oligomer

is

found

in

the

compounds

(Rb-18C6)8

[Ge9=Ge9=Ge9=Ge9]•2en (1) and (Rb-18C6)8[Ge9=Ge9=Ge9=Ge9]•6en (2) where 18C6 and en stand for 18-crown-6 polyether and ethylenediamine, respectively.

Both compounds crystallize from a solution of Rb4Ge9

(isostructural with Cs4Ge9;1 contains isolated Ge94- clusters) dissolved to saturation in ethylenediamine.

The structures of the two compounds were

determined by single crystal X-ray diffraction. They are virtually the same and differ only by the number of solvent molecules involved in the crystallization. Both

structures

are

made

of

the

novel

discrete

tetramers

of

[Ge9=Ge9=Ge9=Ge9]8- and rubidium cations "crowned" by 18-crown-6, one ether molecule per cation (Figure 3.1). Since the cations do not reside within the planes of the crown-ether molecules but are rather removed quite a way from them, only half of their coordination spheres are occupied by the sequestering agent. Thus, they are quite open for interactions and coordinate around the germanium tetramers, capping faces and edges (Figure 3.1). Each cluster has

24

a)

b)

Figure 3.1. Two views of the tetramer of [Ge9=Ge9=Ge9=Ge9]8- surrounded with eight rubidium cations (purple) "crowned" by 18-crown-6 polyether: a) shown are the Rb–Ge interactions; b) the view along the nanorod's axis clearly shows how the crown-ether molecules "protect" the tetramer on the outside.

25

two capped faces and, in addition, the two inside clusters also have bridged edges. The Rb–Ge distances fall in the range 3.5 - 4.2 Å. The coordination of eight (Rb-18C6) cations around the tetramer provides a positively-charged and protective outer shell around the negatively-charged core of 36 germanium atoms. The tetramer is a rod-like formation with a length of about 2 nm and diameter of ca. 0.4 nm, i.e. a nanorod with an aspect ratio of about 5. These nanorods, in both compounds, are made of two crystallographically-different Ge9 clusters and their images generated by inversion centers (Figure 3.2). The four Ge9 clusters are quite similar and can be viewed as tricapped trigonal prisms where the triangular bases are made of atoms 4-5-8 and 2-3-6 while the capping atoms are 1, 7, and 9. The prisms are elongated along two of the three prismatic edges parallel to the three-fold axis, in this case edges 2–5 and 6–8.

The

average distance for these edges is 2.982 Å while it is 2.793 Å for the "normal" edges 3–4. It is exactly the atoms of the elongated edges, i.e. atoms 2, 5, 6, and 8 that are exo-bonded to the neighboring clusters. Furthermore, the exo-bonds are not radial to the clusters, as might have been expected, but are virtually collinear with the elongated edges (Figures 3.1 and 3.2). The reason for this is the geometry of the monomer and, subsequently, the shape of the HOMO, i. e. the orbital that is accessible for oxidative coupling. Studies of the molecular orbitals of a similar cluster but with three elongated edges, Bi95+, have already revealed the shape and character of this orbital.13 It is made predominantly of pz orbitals (z along the three-fold axis) with larger participation of the atoms at the

26

elongated edges and the one capping them, and almost no contribution of px,py (Figure 3.3). Thus, large parts of the orbital are found outside the cluster as extentions to the elongated edges and, therefore, the observed coupling along z is easily understood. It should be pointed out that coupled in the same way but along one edge are the clusters in the dimers of [Ge9–Ge9]6- and the infinite chains of ∞[–Ge9–]2-.7, 8 Also, the Ph2E groups in [Ph2E–Ge9–EPh2]2- (E = Sb, Bi) are attached along one elongated edge and the Ge–E exo-bonds are also collinear with that edge.9

7 6

5 2.546(1) 8

2.673(1)

1 5

2

3 4 9

8

2.546(1) 5

9 4

3

1

2

2.634(1)

6

7

8

Figure 3.2. ORTEP drawing of the tetramer [Ge9=Ge9=Ge9=Ge9]8- (90 % probability thermal ellipsoids). Each cluster is a tricapped trigonal prism with two elongated prismatic edges (shown as open bonds).

The clusters in the tetramer are stacked in such a way that the "normal" edges of the trigonal prisms, edges 3–4, alternate with respect of the plane defined by the exo-bonds. Had they all been on one side, the clusters would have been automatically three-connected along all three edges forming a columnar tetramer of stacked tricapped trigonal prisms elongated presumably

27

along all three edges. Such a structure might be accessible judging from the fact that already known are examples of clusters coupled along one and two edges, [Ge9–Ge9]6- and [Ge9=Ge9=Ge9=Ge9]8-, respectively, and based on the specifics of the HOMO as discussed above. One problem with such a columnar tetramer might be the relatively low negative charge of 6- calculated for it. Formations with more reasonable charge-to-size ratios would be columnar pentamers which would have a charge of 8- and higher columnar oligomers.

Figure 3.3. The HOMO of one of the nine-atom clusters of the tetramer. It is formed explicitly of pz orbitals (left, from extended-Hückel MO calculations). The largest contribution is from the atoms of the elongated edges and the corresponding vertex that caps those edges. Notice that the orbital is: a) πbonding within the triangular bases of the trigonal prism, b) σ-antibonding between the bases, and c) σ-bonding between the capping atoms and those at the bases. Most importantly, the orbital extends outside the cluster as extensions of the elongated edges. All this is confirmed by the shape of the orbital calculated by DFT (right).

28

The charge of the tetramer is easily verified by calculations. As in the trimer,10 this charge is equally distributed among the clusters. The intercluster bonding by two parallel bonds was initially considered to be unusual, and detailed electronic structure calculations and Wilson plots were performed for the trimer and various models of it.10 The conclusions for the latter are valid for the tetramer as well. It should be pointed out that the exo-bond distances, 2.546– 2.673 Å, are comparable with the distances within the clusters, 2.517–2.747 Å. This and the collinear exo-bonds and prismatic edges lead to delocalization of the bonding electrons over the whole tetramer and not only in each monomer. Thus, the exo-bonds can not be described as localized 2c–2e bonds but rather as participating in a system of delocalized bonding. The tetramer should not be viewed as made of four clusters of delocalized bonding that are bonded via localized bonds but rather as one monolithic formation where the delocalization is over the whole tetramer.

3.5. References [1]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

[2]

Queneau, V.; Todorov, E.; Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[3]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[4]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

29

[5]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[6]

Corbett, J. D. Struct. Bond. 1997, 87, 157.

[7]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[8]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[9]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

[10]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[11]

Somer, M.; Carrillo-Cabrera, W.; Peters, E. M.; Peters, K.; von Schnering, H. G. Z. Anorg. Allg. Chem. 1998, 624, 1915.

[12]

Becke, A. D. Journal of Chemical Physics 1996, 104, 1040.

[13]

Corbett, J. D.; Rundle, R. E. Inorg. Chem. 1964, 3, 1408.

30

CHAPTER 4

DERIVATIZATION OF DELTAHEDRAL ZINTL IONS BY NUCLEOPHILIC ADDITION: [Ph–Ge9–SbPh2]2- & [Ph2Sb–Ge9–Ge9–SbPh2]4-

4.1. Abstract The addition of exo-bonded groups to deltahedral Zintl ions such as Ge9nhas been established to proceed by the nucleophilic addition of anionic nucleophiles. Various nucleophiles such as Ph2Bi–, Ph2Sb–, Ph– interact with the relatively low-lying LUMO of Ge92- and/or the half filled HOMO of Ge93- and bond to the clusters. The title anions, characterized in their (K-2,2,2-crypt) salts, and the previously characterized [Ph2Sb–Ge9–SbPh2]2- are made by the reaction of K4Ge9 with SbPh3 in ethylenediamine. [Ph–Ge9–SbPh2]2- is the first organically-functionalized deltahedral Zintl ion, i.e. a deltahedral ion with a direct carbon-cluster covalent bond, that can exists without the substituents as well. The Ge9 clusters resemble tricapped trigonal prisms with one elongated edge (one of the three edges parallel to the pseudo three-fold axis). The two substituents are always bonded to the vertices of such an elongated edge. The same is true for the intercluster bond in [Ph2Sb–Ge9–Ge9–SbPh2]4-.

31

4.2. Introduction Zintl ions are polyatomic anions of the heavier main-group elements that can be stabilized in liquid ammonia or ethylenediamine and can be eventually crystallized from such solutions with appropriate countercations. A special group of such ions are the deltahedral Zintl ions of the heavier members in the carbon group, Ge, Sn, and Pb.1, 2 These are five- and nine-atom cage-like clusters with triangular faces and delocalized bonding as in the deltahedral boranes. The reactivity of these species, expected to be different from Zintl ions with normal 2center–2-electron bonds, has not been sufficiently studied.

Actually, until

recently, the clusters were assumed to be so highly reduced that any attempt for oxidation was expected to destroy them. Thus, all studies of their chemistry were confined to Lewis acid-base reactions where the clusters were used as ligands to coordinate to transition metals via their lone pairs of electrons.3-7

The first

breakthrough in the redox chemistry of these species was the successful “oxidative” coupling of nine-atom germanium clusters into dimers [Ge9–Ge9]6-.8 This was quickly followed by the synthesis of a chain of clusters [–(Ge92-)–]1∞,9 as well as smaller oligomers such as trimers and tetramers, [Ge9=Ge9=Ge9]6and [Ge9=Ge9=Ge9=Ge9]8-, respectively.10,

11

These results indicated that it

should be possible to covalently bond various groups to the clusters and hence synthesize functionalized clusters. Reported recently was the first such derivatized deltahedral cluster [Ph2Sb–Ge9–SbPh2]2- (1) made by a reaction of ethylenediamine solution of K4Ge9 with SbPh3.12 Here we report the structures

32

of two more products of the same reaction, the first organically functionalized deltahedral Zintl ion, [Ph–Ge9–SbPh2]2- (2), and a dimer of clusters functionalized with two diphenylantimony groups, [Ph2Sb–Ge9–Ge9–SbPh2]4(3). Discussed at length is also the probable course of the reactions of addition to the clusters. It is determined that they are nucleophilic addition reactions where an anionic nucleophile Nu– interacts with an empty or half-filled orbital of the cluster and forms an exo-bond.

4.3. Experimental Section Synthesis.

All manipulations were performed inside a nitrogen-filled

glove box with moisture level below 1 ppm. A precursor of nominal composition K4Ge9 was synthesized by heating (900 °C, 2 days) a stoichiometric mixture of the elements in niobium containers (arc-welded in Ar at low pressure) enclosed in quartz ampules under vacuum. Ethylenediamine solutions of this precursor were reacted with SbPh3 in various molar ratios and at different temperatures. Nevertheless, these different conditions did not seem to significantly affect the outcome of the reactions, and in many cases two or three different solid phases were crystallized. A typical reaction for obtaining (K–crypt)22•tol as one of the crystalline phases starts with 94 mg of K4Ge9 dissolved in 1 ml en mixed with 216 mg 2,2,2-crypt (tol = toluene, en = ethylenediamine, 2,2,2-crypt = 4,7,13,16,21,24-hexaoxa-1,10-diazabicyclo-[8.8.8]-hexacosane).

The resulting

brown-red solution is then reacted with 49 mg of SbPh3 while stirring for 6 hours.

33

The solution becomes blood-red. After filtration it is carefully layered with toluene and is then left undisturbed until large enough crystals are formed, 12 days in this case. The mother liquor is decanted and the crystals, orange plates of (K-2,2,2crypt)22•tol (yield of up to 55 % with respect to Ge) and brown-red polyhedra of (K-2,2,2-crypt)2Ge9, are collected. (This latter compound was initially reported as (K-2,2,2-crypt)2Ge10 but is more likely (K-2,2,2-crypt)2Ge9, although the structure has not been properly determined.13)

A typical reaction for

crystallization of (K-2,2,2-crypt)21•en and (K-2,2,2-crypt)43•2.5en starts with 35 mg of K4Ge9 and 72 mg crypt dissolved in 1 ml en. This solution is then reacted with 17 mg SbPh3 at 65 °C for 4 hours while stirring. The resulting red solution is cooled to room temperature, filtered, layered with toluene, and after a few days crystals are recovered. In this case orange-red plates of (K-2,2,2-crypt)21•en (up to 70 % yield), red plates of (K-2,2,2-crypt)43•2.5en (usually 10-15 % yield), and brown-red polyhedra of (K–crypt)2Ge9 were obtained.

Other reactions have

produced various combinations of these four compounds and some times also small amounts of (K-2,2,2-crypt)6Ge9Ge9•0.5en with isolated Ge93-.14 It should be pointed out that often the precursor does not dissolve completely, most likely because of presence of impurities, and this limits the control over the amounts of dissolved precursor.

Also, these systems are so highly reduced that often

manipulations that might be routine for many other systems easily destroy the clusters and result in precipitation of elemental germanium. 1H-NMR spectra of the ethylenediamine solutions before crystallization were taken on a Varian VXR 300 MHz spectrometer (a sealed capillary with 34

C6D12 + 1 % TMS for a reference). They show multiple sets of monosubstituted phenyl rings (6.8 – 8.0 ppm) due to 1, 2, 3, and various amounts of unreacted SbPh3 and SbPh2–.

This indicates that the three anions are present in all

reactions but different sets of them crystallize depending on the specifics of the reaction. In addition, a relatively strong signal of benzene is always observed. Structure determination. X-ray diffraction data were collected from an orange plate of (K-2,2,2-crypt)22•tol (0.23 x 0.17 x 0.02 mm) and a red plate of (K-2,2,2-crypt)43•2.5en (0.18 x 0.12 x 0.02 mm) with graphite-monochromated Mo Kα radiation at 100 K on a Bruker APEX diffractometer with a CCD area detector. The structures were solved by direct methods and refined on F2 using the SHELXTL V5.1 package (after absorption corrections with SADABS). Details of the data collections and refinements are given in Table 4.1. X-ray crystallographic files, in CIF format, are available free of charge via the Internet at: http://pubs3.acs.org/acs/journals/supporting_information.page?in_manuscript=ja 037007b Electronic-structure calculations. Single-point DFT calculations were carried out using the geometry from the structure determination. Used in the calculations was the Becke 3-parameter density functional with the Lee-YangParr correlation functional (B3LYP) in conjunction with the 3-21G basis set.15 The calculations were performed with the Gaussian 98 package, revision A.11.3 on Notre Dame's "Bunch-o-Boxes" Beowulf cluster. Orbital pictures were plotted

35

with Molden. Calculated HOMO-LUMO gaps: 3.09, 3.04, 2.45 eV for 1, 2, and 3, respectively.

TABLE 4.1 CRYSTALLOGRAPHIC DATA FOR (K-2,2,2-crypt)22•tol AND (K-2,2,2crypt)43•2.5en

formula

(K-2,2,2-crypt)22•tol

(K-2,2,2-crypt)43•2.5en

fw

1929.67

3671.13

space group, Z

Pna21, 4

P1, 4

a

23.584(4) Å

14.333(4) Å

b

24.664(4) Å

26.108(8) Å

c

12.910(2) Å

38.64(1) Å

α

80.472(7) °

β

82.169(6) °

γ

84.332(8) °

V

7510(2) Å3

14085(8) Å3

ρcalc

1.707 g•cm-3

1.731 g•cm-3

radiation, λ

Mo Kα, 0.71073 Å

Mo Kα, 0.71073 Å

µ

40.68 cm-1

43.33 cm-1

R1/wR2 (I ≥ 2σI)a

6.46/16.42 %

8.40/19.28 %

R1/wR2 (all data)

9.66/18.59 %

17.34/24.04 %

[a] R1 = ∑║Fo│ - │Fc║/∑│Fo│, wR2 = {[∑w[(Fo)2 - (Fc)2]2]/[ ∑w(F2o)2]}1/2 for Fo2 > 2 σ(Fo2), w = [σ2(Fo)2 + (AP)2]-1 where P = [(Fo)2 + 2(Fc)2]/3; A = 0.1 for (K-2,2,2-crypt)22•tol; A = 0.0891 for (K-2,2,2-crypt)43•2.5en.

36

4.4. Results and Discussions Structure Description. Crystallographically, there is one type of anion 2 (Figure 4.1) but two types, A and B, of anion 3 (Figure 4.2) in the corresponding structures. However, as it can be seen from Figure 4.2 as well as based on distances and angles, A and B are almost identical.

The cores of 2 and 3

contain the well-known nine-atom deltahedral clusters of germanium.

Such

naked clusters have been characterized in numerous compounds, and their geometry and the charge seem to be quite flexible.1,

2

They can be Ge92-,

Ge93-, or Ge94-, while the geometry is based on a tricapped trigonal prism distorted in various ways and extent. The distortions are always associated with elongation of one, two or all three edges of the trigonal prism that are parallel to the three-fold axis. For example, clusters with two elongated edges form the trimers and tetramers of [Ge9=Ge9=Ge9]6- and [Ge9=Ge9=Ge9=Ge9]8-, respectively.10, 11 On the other hand, only one edge is elongated in 1, 2 and 3, the edge 7–9 of the trigonal prism formed by the triangular bases 3–4–7 and 2– 5–9 (Figures 4.1 and 4.2). This edge in 2 and the pairs of such edges in 3A and 3B are 3.098(2), 3.044(2), 3.071(2), 3.054(2), and 3.074(2) Å, respectively, while the other two edges, 2–3 and 4–5, are rather normal, 2.705(2) and 2.881(2) Å in 2, 2.768(2), 2.718(2), 2.726(2), and 2.736(2) Å in 3A, and 2.765(2), 2.722(2), 2.747(2), and 2.728(2) Å in 3B. The remaining distances in the clusters are typical for these species and, in general, are longer for atoms with higher coordination.

37

2.658 3

4

Sb

7

8

2

6

5

9

1.94

1

Figure 4.1. ORTEP drawing (thermal ellipsoids at the 50 % probability level) of 2 with the distances [Å] to the two substituents shown. Other important distances are: Ge1-Ge2 2.577(2), Ge1-Ge3 2.606(2), Ge1-Ge4 2.606(2), Ge1-Ge5 2.595(2), Ge2-Ge3 2.705(2), Ge2-Ge5 2.927(2), Ge2-Ge8 2.642(2), Ge2-Ge9 2.602(2), Ge3-Ge4 2.881(2), Ge3-Ge7 2.626(2), Ge3-Ge8 2.643(2), Ge4-Ge5 2.722(2), Ge4-Ge6 2.642(2), Ge4-Ge7 2.634(2), Ge5-Ge6 2.667(2), Ge5-Ge9 2.607(2), Ge6-Ge7 2.572(2), Ge6-Ge9 2.562(2), Ge7-Ge8 2.543(2), Ge7-Ge9 3.098(2), Ge8-Ge9 2.567(2). The angles (°) at Sb are 93.1(5) for C-Sb-C and 98.2(4) and 96.9(3) for C-Sb-Ge.

38

2.658

2

2.661

Sb 5

9

1

6 8

2.482

4

7

2.486

3

7

3

6

4

9

2.649

2

5

2.651

8

1

Sb

B

A

Figure 4.2. ORTEP drawing (thermal ellipsoids at the 50 % probability level) of the two crystallographically different anions 3A and 3B.

39

The substituents, phenyl and diphenylantimony groups, are attached along the elongated edges of the clusters, i.e. to atoms 7 and 9. This is also the case for the intercluster bond in 3 where each Ge9 cluster acts as a substituent to the other one. The Ge–C distance of 1.94(1) Å in 2 and the average Ge–Sb distance of 2.655 Å in 2 and 3 compare well with similar bonds in other compounds and correspond to single bonds. For example, 1.957(4) Å is the Ge– C distance in Ph4Ge,16 while 2.636 and 2.624 Å are observed for Ge–Sb in ((CH3)3Ge)2Sb–Sb(Ge(CH3)3)2.17

Also, Ge–Sb distances of 2.6483(5) and

2.6505(5) Å are found in [Ph2Sb–Ge9–SbPh2]2-.12

The Ge–Ge distances

between the pairs of clusters in 3, 2.482(2) Å in 3A and 2.486(2) Å in 3B, also correspond to a single-bond distance and compare well with the distance of 2.488(1) Å in [Ge9–Ge9]6- and 2.486(1) Å in [-(Ge92-)-]1∞.8, 9 Electronic Structure.

The observed mode of exo-bonding where the

substituents are attached along an elongated edge of the cluster is the only mode observed so far for di-substituted clusters, and the preference seems to originate from the particular shapes of the HOMO and LUMO of the core clusters (below). The same mode of exo-bonding is observed also for 1 and the chains of [-(Ge92-)-]1∞.9,

12

Similarly, the single exo-bond for the dimer [Ge9–Ge9]6- is

along the elongated edges of the two clusters.8

Also, the clusters in

[Ge9=Ge9=Ge9]6- and [Ge9=Ge9=Ge9=Ge9]8- are connected to each other with pairs of bonds that are along the two elongated edges of each cluster.10, 11 Thus, based on these observations, a more general statement can be made that naked nine-atom clusters add substituents always along the elongated prismatic 40

edges of the clusters. Distinction should be made, however, from various ligated main-group clusters that are directly assembled from atoms that are already bonded to the substituents. For example, the three Si(SiMe3)3 substituents in [Ge9(Si(SiMe3))3]– are bonded to the three capping atoms of the nine atom cluster with three elongated prismatic edges.18

However, these species are

made by a reaction between GeIBr and Li(Si(SiMe3)3) that does not involve preformed Ge9 clusters. This positioning of the substituents in 1, 2, and 3 can be explained with the specifics of the electronic structure of the naked Ge9 clusters and their ability to carry different charges. DFT calculations on the core nine-atom clusters for all three anions provided very similar results as can be expected from their very similar geometries, and can be generalized for any such cluster with one elongated edge. The LUMO for a charge of 2- (HOMO for Ge93- and Ge94-) is in a relatively large energy gap, about 2.5 eV from the next higher orbital (LUMO+1) and 1.1 eV from the HOMO, and is quite indifferent of whether it is occupied or not. The calculated energy difference between Ge94- and Ge92- is only 0.0001 eV.

It has been pointed out, some time ago, that this orbital

(calculated by the extended-Hückel approach for Bi95+ with three elongated edges) is almost explicitly composed of the pz orbitals of the nine atoms (Figure 4.3a).19 Also similarly made of pz orbitals but with almost no participation from the capping atoms are the underlying HOMO and HOMO-1 (Figures 4.3b and 4.3c).

Going further down in energy we find that the next orbital, HOMO-2

41

(Figure 4.3d) is made of only pxpy of the six atoms forming the prism. The LUMO is π-bonding within the bases of the prism and σ-bonding to the capping atoms, but it is σ-antibonding between the bases, i.e. along the prismatic edges parallel to z. The HOMO and HOMO-1 (degenerate for ideal D3h), on the other hand, are π-antibonding within the bases but σ-bonding between them, along the prismatic edges.

Finally, HOMO-2 is σ-bonding within each base but π-

antibonding between them. Therefore, the positions of the LUMO, HOMO and HOMO-1 will depend very strongly on the number of elongated edges and the extent of the elongation. Larger elongations and a greater number of elongated edges will stabilize the LUMO and destabilize the HOMO and HOMO-1 while compression will act in the opposite direction. At the same time, HOMO-2 will be affected very slightly by such distortions along z.

a)

c)

b)

d)

Figure 4.3. MOLDEN plots of orbitals calculated for the core cluster Ge92- in 2: a) LUMO, b) HOMO, c) HOMO-1, and d) HOMO-2. The long edge 7–9 (see Figure 4.1) of the trigonal prism is in front (vertical) and is not drawn.

42

Clearly, both the LUMO and the HOMO are very suitable for interactions with substituents: they are frontier orbitals, both protrude outside the cluster, and both are mostly concentrated at the same two vertices, those along the long edge.

The HOMO is symmetrical while the LUMO is antisymmetrical with

respect to the two vertices of the open edge. These two orbitals, together with the underlying two lone-pair type orbitals at the two vertices, interact with the corresponding in-phase and out-of-phase combinations of the σ-orbitals of the two substituents.

This leads to formation of two strongly bonding, two less

bonding, and two antibonding molecular orbitals. The second pair of bonding orbitals is the former LUMO and HOMO of the core cluster being stabilized by the interactions with the substituents. They remain predominantly skeletal bonding orbitals for the cluster but with some mixing from the substituents. This mixing brings them from being the two cluster-bonding orbitals with the highest energy in the naked cluster to the lowest skeletal orbitals in the substituted species. More importantly, these two orbitals, together with HOMO-2 of the naked cluster (Figure 4.3d), define the positioning of the exo-bonds with respect to the cluster. The two low-lying strongly-bonding orbitals that replace the lone pairs involve predominantly germanium s-orbitals and are stereoinactive, i.e. their energies do not dependent of the position of the exo-bond. On the other hand, the LUMO and HOMO of the naked cluster point outward along the elongated edge and maximize their overlap when the substituents are positioned along the same direction, i.e. as extensions of the elongated edge. Also, notice that after the stabilization of the former LUMO and HOMO of the naked cluster the orbitals that

43

become frontier are the HOMO-1 and HOMO-2 of the naked cluster.

While

HOMO-1 has no contribution from the vertices of the elongated edge and can not be stereoactive for the two exo-bonds, HOMO-2 protrudes outside the cluster at these atoms pointing radially with respect to the bases of the prism and will resist bending of these exo-bonds along that direction. Similar sets of orbitals are found for clusters with two elongated edges where two high-lying orbitals are positioned outside these edges and the interactions between such clusters are maximized also along these edges as observed in [Ge9=Ge9=Ge9]6- and [Ge9=Ge9=Ge9=Ge9]8-.10, 11 Each cluster in these oligomers participates in the exo-bonding with exactly these two orbitals resulting in a charge of 2independent of whether the cluster is an end or a middle cluster. The process of replacing two lone pairs (the two antibonding orbitals) with two exo-bonds (the two pairs of bonding orbitals) seems to be directly related to the course of the reaction. In order to study it in more detail, calculations were carried out also on the cluster with one substituent, the phenyl group (chosen over SbPh2 for the simpler drawings; the results with SbPh2 are identical). Shown in Figure 4.4 are the HOMO, HOMO-1, HOMO-2, and HOMO-9 of such a hypothetical mono-substituted cluster, [Ge9–Ph]3-. Clearly the phenyl group has stabilized the HOMO of the naked Ge92- cluster (Figure 4.3b) lowering its energy to HOMO-9 (Figure 4.4d).

Nevertheless, this orbital remains predominantly

skeletal for the cluster as most of it is distributed over the cluster. There is another, much lower-lying orbital that is predominantly responsible for the exobond (not shown). Notice that the HOMO of [Ge9–Ph]3- (Figure 4.4a) originates 44

from the LUMO of Ge92- (Figure 4.3a). It is somewhat perturbed, of course, by the interactions with the phenyl group, but is still concentrated outside the second vertex of the elongated edge (bottom) and is still along the same direction as in the naked cluster. Furthermore, exactly as the LUMO of the naked Ge92- cluster (Figure 4.3a), this orbital is in a large energy gap, 1.98 eV from the next higher orbital and 1.06 eV from the one below it. (It is the LUMO for [Ge9–Ph]–.) Also, similarly to Ge94- and Ge92-, the energy difference between the species with occupied and empty orbital, [Ge9–Ph]3- and [Ge9–Ph]–, respectively, is only 5.8 x 10-5 eV (0.00188 eV for [Ge9–SbPh2]3- and [Ge9–SbPh2]–). Thus, in analogy with the existence of nine-atom clusters of germanium with different charges, Ge94-, Ge93-, and Ge92-, it is very likely that species such as [Ge9–Ph]3-, [Ge9–Ph]2-, and [Ge9–Ph]- can exist as well. This orbital, the HOMO of [Ge9– Ph]3-, in the di-substituted species is also stabilized, and the top two orbitals are the HOMO-1 and HOMO-2 of [Ge9–Ph]3- (Figures 4.4b and 4.4c). Again, while HOMO-2 has no contribution from the exo-bonded vertices, HOMO-1 does and plays stereoactive role for the positioning of the exo-bonds. Reaction Description. The addition of various groups to the Ge9 cluster can be described as a replacement of a lone pair of electrons at a germanium vertex by an exo-bond between the substituent and that vertex. The course of the reaction is most likely very similar to SN1 and/or SRN1 nucleophilic substitutions where the nucleophiles, SbPh2– and Ph– in our case, interact with

45

an empty or half-filled orbital of the substrate. The difference with the classical mechanisms is that the leaving group is one or more electrons. Similarly to the classical mechanisms this step results in an empty or half-filled low-lying orbital available for the nucleophilic attack. Another difference is that the leaving electrons in our case are also the electrons that reduce SbPh3 and generate the nucleophiles SbPh2– and Ph–. In classical reactions the nucleophiles are often generated by separate reduction with alkali metals.

a)

b)

c)

d)

Figure 4.4. MOLDEN plots of orbitals calculated for hypothetical intermediate [Ge9–Ph]3- with geometry taken from 2: a) HOMO, b) HOMO-1, c) HOMO-2, and d) HOMO-9. The long edge 7–9 (see Figure 4.1) of the trigonal prism is in front (vertical) and is not drawn.

46

The "release" of electrons by the clusters needs some more elaboration and discussion.

We have proposed in the past that the Ge9 clusters with

different charges, i.e. Ge94-, Ge93- and Ge92-, coexist in the ethylenediamine solutions and that depending on the available countercations different species crystallize.

We have now carried out two experiments that prove this very

convincingly. First, a saturated dark brown-green solution of K4Ge9 was divided equally in four test tubes labeled from #1 to #4. Added to these four containers were the following reagents (molar ratio to potassium is in parentheses): #1 – 2,2,2-crypt (1.5 : 1), #2 – 2,2,2-crypt (0.3 : 1), #3 - 18-crown-6 ether (1 : 1) and heated at 50 °C for 2 hours, #4 - nothing. They were all then layered carefully with toluene except that the toluene for test tube #4 was saturated with 18-crown6 ether. After a few days various compounds crystallized from these test tubes: #1 - (K–2,2,2-crypt)6(Ge9)(Ge9)•0.5en with monomers of Ge93-,9 #2 - both (K– 2,2,2-crypt)6(Ge9)(Ge9)•0.5en14 and (K–2,2,2-crypt)2K4[Ge9–Ge9]•8en with dimers of [Ge9–Ge9]6-,8 #3 - (K–18C6)6[Ge9=Ge9=Ge9]•2en with trimers of [Ge9=Ge9=Ge9]6-,10 #4 - both (K–18C6)2K2Ge9 with monomers of Ge94- and (K–18C6)3Ge9•3en with monomers of Ge93-. Thus, these compounds contain nine-atom clusters with different charges: Ge94- as monomers, Ge92- in the trimer and Ge93- as monomers and in the dimer. The second experiment is even more telling.

A saturated dark red-brown solution of Rb4Ge9 in

ethylenediamine was prepared. Added to this solution were both crypt and 18crown-6 ether, each in a molar ratio of 2 : 1 with respect to Ge9. After layering

47

with toluene and waiting for a few days, the following phases were found in the solid

product:

(Rb–crypt)6[Ge9=Ge9=Ge9]•3en,10

[Ge9=Ge9=Ge9=Ge9]•2en,11

(Rb–crypt)2Rb4(Ge9)(Ge9)•7en,8

(Rb–18C6)8 and

(Rb–

18C6)2Rb2Ge9, which contains known Ge94- clusters. Thus, nine-atom clusters with all three charges can co-crystallize from the same solution. How is it possible that Ge94-, Ge93- and Ge92- coexist in these solutions?

Most likely there are equilibria between themselves and free

electrons, i. e. Ge94- ↔ Ge93- + e- and Ge93- ↔ Ge92- + e-. This is exactly the same phenomenon as observed for elemental alkali metals when dissolved in ethylenediamine or liquid ammonia, A ↔ A+ + e- (A = alkali metal).20, 21 The free electrons are solvated by the ethylenediamine and, although they react with the solvent to produce amide and hydrogen, the reaction is very slow in the absence of catalysts or strong nucleophiles. These electrons, however, react readily with SbPh3 to produce nucleophiles, Ph– and initially a radical of •SbPh2 which becomes SbPh2–.

This is exactly the same reaction as the well known

generation of nucleophiles using elemental alkali metals dissolved in liquid ammonia, i.e. 2A + SbPh3 —> 2A+ + SbPh2– + Ph–. The phenyl anion is quite reactive and extracts a proton from the ethylenediamine to form benzene and amide. The presence of large amounts of benzene was confirmed by proton NMR in all of the reactions. Due to the simultaneous presence of numerous differently-bonded phenyl groups it was impossible to unequivocally assign a set of phenyl hydrogens in the 1H NMR to the other nucleophile, SbPh2–. Thus, it is

48

clear that all of these addition reactions, start with the nucleophile donating electrons to the LUMO of Ge92- (SN1-like), Ge92– + SbPh2– —> [Ge9– SbPh2]3–, or to the half-occupied HOMO of the radical •Ge93- (SRN1-like), •Ge93– + SbPh2– —> {[•Ge9–SbPh2]4–} —> [Ge9–SbPh2]3– + e-.

The

resulting monosubstituted species, either [Ge9–SbPh2]3– or [Ge9–Ph]3-, have not been isolated yet but should be relatively stable. As already discussed, their HOMOs are of exactly the same type as the HOMO of Ge94- (Figure 4.3a), which is the LUMO of Ge92-. Their character is also the same, i.e. they are in large energy gaps and there is very small energy differences between when occupied and empty. Thus, we propose that monosubsituted species of different charges exist in the solutions and they form equilibria similar to the naked clusters, i.e. [Ge9–SbPh2]3– ↔ [•Ge9–SbPh2]2– + e- and [•Ge9–SbPh2]2– ↔ [Ge9–SbPh2]– + e-. The final products are formed when another nucleophile attacks the empty LUMO of [Ge9–SbPh2]– or the half-occupied HOMO of [•Ge9– SbPh2]2– as follows: [Ge9–SbPh2]– + SbPh2– —> 1 [•Ge9–SbPh2]2– + SbPh2– —> {[Ph2Sb–Ge9–SbPh2•]3-} —> 1 + e– [Ge9–SbPh2]– + Ph– —> 2 [•Ge9–SbPh2]2– + Ph– —> {[Ph–Ge9–SbPh2•]3-} —> 2 + e– [Ge9–SbPh2]– + [Ge9–SbPh2]3– —> 3 [•Ge9–SbPh2]2– + [Ge9–SbPh2]3– —> 3 + e– 49

[•Ge9–SbPh2]2– + [•Ge9–SbPh2]2– —> 3 These multiple possibilities for reactions of different species with different nucleophiles are clearly manifested in the multiphase crystalline products obtained from such reactions. It is possible that other anions form as well but can not be crystallized with the available countercations. Thus, species such as [Ph–Ge9–Ph]2-, [Ph–Ge9–Ge9–SbPh2]4-, [Ph–Ge9–Ge9–Ph]4-, etc., should not be ruled out. Also, as mentioned in the experimental part, often found in the product are the naked Ge93- and Ge92- coexisting with 1, 2, and/or 3. This clearly indicates that not only Ph– undergoes side reactions but perhaps SbPh2– does that as well. Furthermore, anion 1 exists also with bismuth, i. e. [Ph2Bi– Ge9–BiPh2]2-, and perhaps 2 and 3 are possible.

However, none of the

corresponding arsenic and phosphorus analogs of these species has been made.

The analogous reactions with AsPh3 and PPh3 result in the more

oxidized clusters of Ge92- found as oligomers ([Ge9=Ge9=Ge9]6- or [Ge9=Ge9=Ge9=Ge9]8-) and the corresponding free nucleophiles AsPh2– and PPh2–. It is possible that phenyl-disubstituted species are formed but, as in the reactions with SbPh3 and BiPh3, they do not crystallize.

The failure to

synthesize phosphorus and arsenic analogs of 1, 2, and/or 3, is most likely due to the less nucleophilic nature of the corresponding anions PPh2– and AsPh2–. Summary. The reaction between Ge9n- where n = 2, 3, or 4, and SbPh3 produces

[Ph2Sb–Ge9–SbPh2]2-,

[Ph–Ge9–SbPh2]2-,

50

[Ph2Sb–Ge9–Ge9–

SbPh2]4-, and perhaps other species that have not been characterized yet. The reaction proceeds via formation of negatively-charged nucleophiles Ph– and SbPh2– from SbPh3 and the solvated free electrons that are released from the nine-atom clusters.

High-energy cluster orbitals with significant distribution

outside the cluster are available for the nucleophilic attack and formation of bonds. Different products are formed depending upon the available nucleophile in the immediate vicinity of the cluster. This knowledge can be used to envision other, interesting cluster-based species.

4.5. References [1]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[2]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[3]

Eichhorn, B. W.; Haushalter, R. C.; Pennington, W. T. J. Am. Chem. Soc. 1988, 110, 8704.

[4]

Eichhorn, B. W.; Haushalter, R. C. J. Chem. Soc., Chem. Commun. 1990, 937.

[5]

Gardner, D. R.; Fettinger, J. C.; Eichhorn, B. W. Angew. Chem., Int. Edit. 1996, 35, 2852.

[6]

Kesanli, B.; Fettinger, J.; Eichhorn, B. Chem. Eur. J. 2001, 7, 5277.

[7]

Kesanli, B.; Fettinger, J.; Gardner, D. R.; Eichhorn, B. J. Am. Chem. Soc. 2002, 124, 4779.

[8]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

51

[9]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[10]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[11]

Ugrinov, A.; Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

[12]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

[13]

Belin, C.; Mercier, H.; Angilella, V. New Journal of Chemistry 1991, 15, 931.

[14]

Fassler, T. F.; Schutz, U. Inorg. Chem. 1999, 38, 1866.

[15]

Becke, A. D. Physical Review A 1988, 38, 3098.

[16]

Karipides, A.; Haller, D. A. Acta Crystallogr. 1972, B28, 2889.

[17]

Roller, S.; Drager, M.; Breunig, H. J.; Astes, M.; Gulec, S. J. Organomet. Chem. 1989, 378, 327.

[18]

Schnepf, A. Angew. Chem. Int. Edit. 2003, 42, 2624.

[19]

Corbett, J. D.; Rundle, R. E. Inorg. Chem. 1964, 3, 1408.

[20]

Kraus, C. A. J. Am. Chem. Soc. 1908, 30, 1323.

[21]

Thompson, J. C., Electrons in liquid Ammonia. ed.; Clarendon Press: Oxford, 1976.

52

CHAPTER 5

RATIONALLY FUNCTIONALIZED DELTAHEDRAL ZINTL IONS: SYNTHESIS AND CHARACTERIZATION OF [Ge9–ER3]3-, [R3E–Ge9–ER3]2-, and [R3E– Ge9–Ge9–ER3]4- (E = Ge, Sn; R = Me, Ph)

5.1. Abstract Six new derivatized deltahedral Zintl ions have been synthesized by reactions between the known Zintl ions Ge9n– with the halides R3EX and/or the corresponding anions R3E– for E = Ge or Sn. This rational approach is based on our previous discovery that the nature of these derivatization reactions is of nucleophilic addition to the clusters. All species were structurally characterized as their salts with potassium counteractions sequestered in 2,2,2-crypt or 18crown-6 ether. The tin-containing anions were characterized also in solutions with 119Sn NMR. Discussed are the reaction types for such substitutions and the structures of the new anions.

5.2. Introduction Despite their more than 100 year history the Zintl ions have remained a remarkably little understood area of inorganic chemistry.1-3 These are negatively 53

charged polyatomic clusters of main-group metallic and/or semimetallic elements that can be crystallized from solutions with alkali metal countercations (often sequestered in various crypts or crown ethers). Although they have received some more attention recently, the efforts, until a few years ago, involved mostly reproduction and better characterization of the already known or slightly modified species.4, 5 One of the more recent developments on this front was the discovery that, contrary to the expectations, such Zintl ions exist also in Zintl phases. The latter are polar intermetallic compounds of the heavier main-group p-elements with alkali or alkaline-earth metals that are electronically-balanced (or can be considered as such based on their structures). The Zintl phases can be viewed as salts where the more electropositive element transfers its valence-shell electrons to the more electronegative element and both achieve closed-shell formations. Despite their name, Zintl phases were considered unrelated to Zintl ions until the discovery of the Zintl phases A4E9 (A = alkali metal, E = group 14 element) which contain the known deltahedral Zintl ions E94-.6-10 The latter were previously characterized in solids crystallized from ethylenediamine or ammonia solutions of the precursors AxEy.4,

5

These nine-atom clusters are cage-like

species with the shape of a distorted tricapped trigonal prism (the distortions are elongations along one or more of the prismatic edges parallel to the three-fold axis). Bonding in such clusters is achieved by delocalized electrons as in the well known cage-like boranes. More recent studies revealed that, despite earlier beliefs, such deltahedral E9n- clusters can bond to each other in different modes and form dimers [Ge9–

54

Ge9]6-,11 trimers [Ge9=Ge9=Ge9]6–,12 tetramers [Ge9=Ge9=Ge9=Ge9]8–,13 and infinite chains [–(Ge92–)-]1∞.14 This indicated that perhaps clusters can bond not only to other clusters but also to other groups. Following on this idea we studied reactions of ethylenediamine solutions of K4Ge9 with mild oxidizing agents such as Ph3Sb and Ph3Bi and, indeed, these reactions produced the first functionalized Zintl ions [Ph2Sb–Ge9–SbPh2]2–, [Ph2Bi–Ge9–BiPh2]2–, [Ph– Ge9–BiPh2]2–, and [Ph2Sb–Ge9–Ge9–SbPh2]4–.15, 16 However, at this stage it was unclear how the reactions proceeded and what the reacting species might be. More detailed studies were carried out and they suggested that these reactions are in fact nucleophilic additions of the anions Ph2Sb– and Ph2Bi– to the naked Zintl ions Ge9n-.16 The charges of the latter are, most likely, lower than 4-, i.e. Ge92- and/or the radicals •Ge93-. These naked clusters are well known: •Ge93- has been characterized in a few compounds,5 while Ge92- has been also crystallized from solutions but the structure shows great disorder and was erroneously reported as Ge102-.17

(We and others have many times

collected single-crystal X-ray diffraction data on compounds containing this cluster, Ge92-, but all structure determinations have shown the same disorder, and no additional results have been published.) In ethylenediamine, these Zintl ions of lower charges are most likely in equilibria between themselves and with Ge94- and free solvated electrons, i.e. Ge94- ↔ •Ge93- + e– ↔ Ge92- + 2e– analogous to alkali metals dissolved in ethylenediamine. Both •Ge93- and Ge92-

55

have a low-lying molecular orbital available for electron donation, it is half-filled for the former and empty for the latter. Thus, nucleophiles such as Ph2Sb– and Ph2Bi– (labeled R2Pn–) donate electrons to this orbital and form bonds to the clusters. We proposed that such reactions should produce the monosubstituted species [Ge9–PnR2]3-, but these were not observed at that time. They, in turn, would be in equilibria with [•Ge9–PnR2]2-, [Ge9–PnR2]–, and free electrons. Such less-reduced species have similarly a half-filled or empty orbital and can add one more anion PnR2– to produce [R2Pn–Ge9–PnR2]2-.

They can also

react between themselves either as two radicals or as a donor [Ge9–PnR2]3- and an acceptor of either [•Ge9–PnR2]2- or [Ge9–PnR2]–, and produce the observed dimers of [R2Pn–Ge9–Ge9–PnR2]4-. These general ideas were explored further for possible use in more rational synthesis of similar derivatives of deltahedral Zintl ions by addition of other groups. Reported here are the rational syntheses of both mono- and di-substituted germanium clusters by reactions of the clusters with either the organometallic compounds ER4, their halides ER3X, or the corresponding anion ER3– prepared separately (E = Sn, Ge and R = Me, Ph). The new species are characterized structurally as their [K–2,2,2-crypt]+ or [K– 18c6)]+ salts in the solid state, and the tin-containing species are also characterized by 119Sn NMR in solution.

56

5.3. Experimental Section Materials and Techniques: All manipulations were performed in a nitrogen-filled glove box with moisture level below 1 ppm. A precursor of nominal composition K4Ge9 was synthesized by heating (900 °C, 2 days) a stoichiometric mixture of the elements (K: 99+ %, Strem; Ge: 99.999 %, Alfa-Aesar) in sealed niobium containers (arc-welded in Ar at low pressure) that were in turn sealed in quartz

ampoules

under

vacuum.

The

reactions

were

performed

in

ethylenediamine (99.5+%, Aldrich) or pyridine (extra dry, Acros). The reagents 2,2,2-crypt (98%, Acros), Ph4Sn (95%, Acros), Ph3SnCl (95%, Strem), Me3SnCl (99%, Acros), Ph4Ge (97%, Aldrich), and Ph3GeCl (99%, Strem) were used after careful drying under vacuum. 119Sn NMR spectra were taken in ethylenediamine or pyridine on a Bruker 400MHz spectrometer.

Sealed capillaries with C6D12 were used as

deuterted solvent while Me4Sn in CDCl3 was used as an external reference. Synthesis of (K–2,2,2-crypt)2[Ph3Sn–Ge9–SnPh3]: K4Ge9 (133 mg, 0.16 mmol) and 2,2,2-crypt (247 mg, 0.65 mmol) were dissolved in ethylenediamine (1 mL) and the solution was reacted with SnPh4 (118 mg, 0.28 mmol) while stirring at 50 °C for 24h.

The resulting red-brown solution was

filtered, carefully layered with toluene, and left undisturbed for a few days. After decanting the mother liquor, yellow-orange plates of [K–(2,2,2-crypt)]2[Ph3Sn– Ge9–SnPh3] were recovered (~45% yield based on the precursor). The same

57

reaction performed in pyridine instead of ethylenediamine gave exactly the same result. (It should be pointed out that the yields in all the reactions have very large uncertainties due to the uncertainty of the purity and the degree of decomposition of the precursors.) The compound was also prepared by reacting the precursor with solution containing Ph3Sn–. The latter was made by reduction of Ph3SnCl (100 mg, 0.23 mmol) dissolved in ethylenediamine (1mL) with elemental K (18 mg, 0.46 mmol). This solution was stirred until all potassium was dissolved and then the resulting pale orange suspension was centrifuged and the solid KCl was filtered out. 119Sn NMR: -113.0 ppm (s, SnPh3-). The solution was reacted with K4Ge9 (95 mg, 0.117 mmol) and stirred for 24h at 50 °C. After filtration, 2,2,2-crypt (264 mg, 0.7 mmol) was added and the solution was carefully layered with toluene. After a few days the mother liquor was decanted and crystals of [K–(2,2,2crypt)]21 (~ 20% yield) were collected.

119Sn NMR of crystals dissolved in

pyridine: 15.1 ppm (s, [Ph3Sn–Ge9–SnPh3]2-), –108.1 ppm (s, SnPh3-) (Figure 5.1). Synthesis of (K-18c6)2[Ph3Sn–Ge9–SnPh3]•0.25(18c-6)•2en:

K4Ge9

(43.5 mg, 0.05mmol) was dissolved in ethylenediamine (0.5 mL) to form intensely colored red-brown solution. Mixed in a separate test tube were Ph3SnCl (20 mg, 0.05 mmol) and ethylenediamine (1 mL), and this solution was added slowly to the precursor solution. 18-crown-6 ether (105 mg, 0.39 mmol, molar ratio of 2 : 1 with respect to potassium) was added to the resulting very dark red-brown solution and this was stirred for 15 min. After filtration the solution was layered

58

with toluene and was left undisturbed until crystals of good size formed. Yelloworange to orange-red crystals of (K–18c6)2[Ph3Sn–Ge9–SnPh3]•0.25(18c6)•2en were obtained (~ 40-50 % yield). 119Sn NMR of crystals dissolved in pyridine: 14.9 ppm (s, [Ph3Sn–Ge9–SnPh3]2-) (Figure 5.2).

Figure 5.1. in pyr.

119

Sn-NMR spectrum of [K-(2,2,2-crypt)]2[Ph3SnGe9SnPh3] dissolved

Figure 5.2. 119Sn-NMR spectrum of (K-18c6)2[Ph3SnGe9SnPh3] •0.25(18c6)•2en dissolved in pyr.

59

Synthesis of (K–2,2,2-crypt2[Me3Sn–Ge9–SnMe3]•3.5tol:

Me3SnCl

(88 mg, 0.44 mmol) and K (35 mg, 0.88 mmol) were added to ethylenediamine (1mL) and the solution was stirred until all K dissolved.

The solution was

centrifuged and filtered. 119Sn NMR: -176.9 ppm (s, Me3Sn-). K4Ge9 (44 mg, 0.05 mmol) and 2,2,2-crypt (250 mg, 0.65 mmol) were added and the solution was stirred for 24 h at 50 °C. The resulting very dark red-brown solution was filtered (119Sn NMR: -9.2 ppm (s, [Me3Sn–Ge9–SnMe3]2-)) and layered with toluene. It was left undisturbed, and when crystals of good size were visible, the mother liquor was decanted and the yellow-orange plates of [K–(2,2,2crypt)]2[Me3Sn–Ge9–SnMe3]•3.5tol (ca. 10% yield) were collected. 119Sn-NMR of crystals dissolved in pyridine: 21.9 ppm (s, [Me3Sn–Ge9–SnMe3]2-) (Figure 5.3).

Figure 5.3. 119Sn-NMR spectrum of [(K-(2,2,2crypt)]2[Me3SnGe9SnMe3]·3.5tol dissolved in pyr.

60

Synthesis

of

(K–2,2,2-crypt)2[Ph3Ge–Ge9–GePh3]•tol•0.5en:

Ph3GeCl (71 mg, 0.21 mmol) in ethylenediamine (1 mL) was reacted with K (16 mg, 0.42 mmol), and the mixture was stirred until everything was dissolved. The yellow-orange solution was centrifuged and filtered. K4Ge9 (55 mg, 0.07 mmol) and 2,2,2-crypt (160 mg, 0.42 mmol) were added to the solution and stirred for 24 h at 50 °C.

The resulting very dark brown-red solution was filtered and

carefully layered with toluene. After a few days the mother liquor was decanted and orange-red crystals of (K–2,2,2-crypt)2[Ph3Ge–Ge9–GePh3]•3.5tol were recovered (ca. 50% yield). Synthesis of (K–2,2,2-crypt)3[Ge9–SnPh3]•en: Ph3SnCl ( 20 mg, 0.05 mmol) in ethylenediamine (1 mL) was reacted with K (3.8 mg, 0.1 mmol), and the mixture was stirred until everything dissolved. The resulting pale orange solution was separated from the KCl precipitation and was reacted with K4Ge9 (105 mg, 0.13 mmol) and 2,2,2-crypt ( 233 mg, 0.62 mmol) while stirring for 1 hour at 50 °C. After filtration the orange-red solution was layered with toluene, and after a few days orange-red plates of (K–2,2,2-crypt)3[Ge9–SnPh3]•en were recovered (ca. 20 % yield). 119Sn NMR of crystals dissolved in pyridine: 260.0 ppm (s, (Ph3Sn–Ge9]3-), –109.8ppm (s, SnPh3-) (Figure 5.4). Synthesis of (K–2,2,2-crypt)3[Ge9–SnMe3]:

Me3SnCl (10 mg, 0.05

mmol) in ethylenediamine (1 mL) was reacted with K (3.8 mg, 0.1 mmol), and the mixture was stirred until everything dissolved.

Added to this solution were

K4Ge9 (90 mg, 0.11 mmol) and 2,2,2-crypt (203 mg, 0.54 mmol) and the mixture was stirred for 1 hour at 50 °C. After filtration the red-brown solution was layered 61

with toluene and left undisturbed for a few days. Large red-orange crystals of (K–2,2,2-crypt)3[Ge9–SnMe3] were recovered (ca. 55 % yield). 119Sn NMR of crystals dissolved in pyridine: 130.3ppm (s, [Me3Sn–Ge9]3-) (Figure 5.5). The solution in pyridine was then layered with toluene and [Me3Sn–Ge9]3- was recrystallized in (K–2,2,2-crypt)3[Ge9–SnMe3]•2tol•py with a different unit cell due to the extra solvent molecules.

Figure 5.4. pyr.

119

Sn-NMR spectrum of [K-(2,2,2-crypt)]3[Ph3SnGe9]·en dissolved in

Figure 5.5. 119Sn-NMR spectrum of [K-(2,2,2-crypt)]3[Me3SnGe9]·dissolved in pyr. 62

Synthesis of (K–18c6)3[Me3Sn–Ge9]•thf•2en:

Following the same

procedure as for (K–2,2,2-crypt)3[Ge9–SnMe3] used was 18-crown-6 ether (143 mg, 0.54 mmol) instead of 2,2,2-crypt. The resulting solution was divided in two reaction tubes and one was layered with toluene and the other with THF. Red crystals of (K–18-crown-6)3[Ge9–SnMe3]•thf•2en were recovered from the latter (ca. 30-40 %). 119Sn-NMR of crystals dissolved in pyridine: 116.2 ppm (s, [Me3Sn–Ge9]3-).

Figure 5.6. pyr.

119

Sn-NMR spectrum of (K-18c6)3[Me3SnGe9] •thf•2en dissolved in

Synthesis of (K–2,2,2-crypt)3[Ph3Sn–Ge9–Ge9–SnPh3]•2en:

K4Ge9

(43.5 mg, 0.05 mmol) was dissolved in ethylenediamine (0.5 mL) and formed intensely colored red-brown solution. Separately, Ph3SnCl (40 mg, 0.1 mmol) was dissolved in ethylenediamine (1 mL), and 0.2 ml of this solution (0.02 mmol Ph3SnCl) were added to the solution of the precursor. The red color became 63

darker and after stirring the solution for 15 min it was layered with 2 ml of a solution of 2,2,2-crypt (83 mg, 0.22 mmol) in toluene. After a few days crystals of two phases were recovered: orange plates of (K-2,2,2-crypt)2[Ph3Sn–Ge9– SnPh3] and orange plates of (K-2,2,2-crypt)4[Ph3Sn–Ge9–Ge9–SnPh3]•2en. Structure determination: Single crystal X-ray diffraction data sets were collected at 100 K on a Bruker APEX diffractometer with a CCD area detector (graphite-monochromated Mo Kα radiation, crystals protected with Parathone-N oil). The structures were solved by direct methods and refined on F2 using the SHELXTL V5.1 package (after absorption corrections with SADABS). Details of the data collections and refinements are given in Table 5.1. Further details of the crystal structure investigation are available free upon request from the Cambridge Crystallographic Data Centre (12 Union Road, Cambridge

CB2

1EZ,

UK;

www.ccdc.cam.ac.uk/conts/retrieving.html;

fax: or

(+44)

1223-336-033;

[email protected])

upon

quoting depository numbers CCDC 230728 - 230735. Some disorder was observed in [Me3Sn–Ge9]3- in both its (K–2,2,2-crypt) and the (K–18c6)] salts. Each germanium atom in the former has an image, i.e. the Ge9 core has two positions. The disorder in the latter is at the tin atom with two close positions. This makes somewhat less certain the distances in these species. However, such a disorder is quite understandable in light of the very weak interactions between cluster and sequestering agents, and has been observed quite often before.

64

TABLE 5.1 CRYSTALLOGRAPHIC DATA FOR [K–2,2,2-crypt)21, [K–18c6)]21·0.25(18c6)· 2en, (K–2,2,2-crypt)22·3.5tol, (K–2,2,2-crypt)23·tol·en, (K–2,2,2-crypt)34·en, (K– 2,2,2-crypt)35, (K–18c6)35·2en·thf, and (K–2,2,2-crypt)46·2en. (K-crypt)22· 3.5tol 2042.40 Cc, 8

(K-crypt)23· tol·0.5en 2214.46 P21/c, 4

13.732(1) 24.0378(1) 47.707(3)

16.768(1) 17.949(1) 30.423(2)

97.392(1)

93.259(2)

16135.7(2) 1.681 88978

9141.9(1) 1.606 76833

28557 20759 3971 50 5.92/13.71

39202 34835 4.064 56 4.10/8.86

16085 14402 3.708 50 3.31/8.56

8.67/15.14

4.88/9.18

3.83/8.88

formula

(K-crypt)21

(K-18C6)21· (18c6).25·2en

fw space group, Z

2184.47 P-1, 2

2142.47 P-1, 4

a [Å] b [Å] c [Å] α[°] β [°] γ[°] V [Å3 ] ρcalc [g•cm-3] No. of measured refl. No. of indep. refl. No. of used refl. µ [cm-1] 2θmax [°] R1/wR2 (I≥2σ1)* [%] R1/wR2 (all data) [%]

15.714(1) 16.693(1) 17.361(2) 106.347(2) 98.984(2) 99.595(2) 4208.4(6) 1.724 45190

16.056(1) 16.850(1) 32.336(3) 78.280(3) 82.054(2) 75.809(10) 8268.8(11) 1.721 49422

20749 16248 3.902 56 3.58/7.68 5.16/8.27

[*] R1 = ∑║Fo│ - │Fc║/∑│Fo│, wR2 = {[∑w[(Fo)2 - (Fc)2]2]/[ ∑w(F2o)2]}1/2 for Fo2 > 2 σ(Fo2), w = [σ2(Fo)2 + (AP)2]-1 where P = [(Fo)2 + 2(Fc)2] / 3; A (B) = 0.0347 (0.000) for (K-crypt)21, A (B) = 0.0599 (58.2235) for (K18c6)31·0.25(18c6)·2en, A (B) = 0.0325 (12.5834) for (K-crypt)22·3.5tol, A (B) = 0.0394 (19.6685) for (K-crypt)23·tol·0.5en, A (B) = 0.0495 (115.9095) for (Kcrypt)34·en, A (B) = 0.0688 (135.2153) for (K-crypt)35, A(B) = 0.0501 (33.4451) for (K-18c6)35·2en·thf, and A(B) = 0.1208 (0.0) for (K-crypt)46·2en 65

TABLE 5.1 (contd.)

formula

(K-crypt)34• (K-crypt)35 en

fw space group, Z a [Å] b [Å] c [Å] α[°] β [°] γ[°] V [Å3 ] ρcalc [g•cm-3] No. of measured refl. No. of indep. refl. No. of used refl. µ [cm-1] 2θmax [°] R1/wR2 (I≥2σ1)* [%] R1/wR2 (all data) [%]

2310.17 P21/c, 4 27.192(2) 20.457(2) 17.151(2)

2063.87 P21/c, 4 16.834(1) 20.904(1) 23.587(1)

(K18C6)35• 2en•thf 1909.57 P21/c, 4 14.378(1) 30.157(2) 17.541(1)

93.560(2)

94.837(1)

95.492(1)

9524.0(1) 1.611 78720 16783 12703 3.249 50 9.19/18.41 11.92/19.32

8270.6(7) 1.658 48770 14439 13661 3.729 50 7.45/19.69 7.71/19.84

7571.1(9) 1.675 44778 13278 10282 4.057 50 4.82/11.25 6.65/12.24

(K-crypt)46• 2en 3789.16 P-1, 1 12.618(3) 14.961(3) 21.589(4) 90.097(4) 102.601(3) 111.103(3) 3694.9(13) 1.703 21422 12622 7102 4.105 50 7.70/18.93 14.38/23.01

5.4. Results and Discussions Synthesis. The reaction path for addition of functional groups to the Ge9 clusters that was proposed before (above) can be described in more details by the following reactions:16 Equilibria 1: Ge94– ↔•Ge93– +e– ↔Ge92– + 2e–

66

Reaction steps 1: Ge92– + SbPh2– ↔ [Ge9–SbPh2]3– (intermediate) and/or •Ge93– + SbPh2– ↔ [•Ge9–SbPh2]4– ↔ [Ge9–SbPh2]3– (intermediate) + e– Side reactions of Ge9n-: •Ge93- + •Ge93- ↔ [Ge9–Ge9]6- and/or Ge92- + Ge94- ↔[Ge9–Ge9]6- and/or •Ge93- + Ge94- ↔[Ge9–Ge9]6- + e– Equilibria 2: [Ge9–SbPh2]3– ↔[•Ge9–SbPh2]2– + e– ↔ [Ge9–SbPh2]– + 2e– Reaction steps 2: [Ge9–SbPh2]– + SbPh2– ↔ [Ph2Sb–Ge9–SbPh2]2- and/or [•Ge9–SbPh2]2– + SbPh2– ↔ [•Ph2Sb–Ge9–SbPh2]3- ↔ [Ph2Sb–Ge9–SbPh2]2+ e– Side reactions of [Ge9–SbPh2]n-: [•Ge9–SbPh2]2– + [•Ge9–SbPh2]2– ↔ [Ph2Sb–Ge9–Ge9–SbPh2]4- and/or [Ge9–SbPh2]– + [Ge9–SbPh2]3– ↔[Ph2Sb–Ge9–Ge9–SbPh2]4- and/or [•Ge9–SbPh2]2– + [Ge9–SbPh2]3– ↔ [Ph2Sb–Ge9–Ge9–SbPh2]4- + e– Based intermediates

on

this,

we

[Ge9–SbPh2]3-

expected that,

relatively

perhaps,

stable

might

be

mono-substituted susceptible

for

crystallization at appropriate conditions. However, all attempts to crystallize such species from solution were unsuccessful. Also, it was impossible to prove their existence in solution because neither Ge nor Sb are good nuclei for NMR

67

spectroscopy, and therefore their existence, and ultimately the reaction path, could not be unequivocally established. This led to shifting our attention to tinbased substituents for which the corresponding reactions can be followed in solutions by 119Sn NMR. Furthermore, it is relatively easy to prepare simple R3E– anions (rather than various aggregates) of the carbon group by reduction of the corresponding halide with alkali metal. Although this reaction can proceed along different pathways with different intermediates, its final product with small excess of alkali metal is always the anion R3E–.18 Initially, solutions of K4Ge9 were reacted with GePh4 and SnPh4 in order to find out whether these react in a similar way to SbPh3 and BiPh3. The tin reaction was successful and di-substituted [Ph3Sn–Ge9–SnPh3]2– (1) was obtained and characterized as its (K–2,2,2-crypt)+ salt.

The reaction with

GePh4, on the other hand, produced only •Ge93- and dimers of [Ge9-Ge9]6- as crystalline products (depending on the amount of 2,2,2-crypt used).11, 19 We have shown before, however, that these same products can be obtained directly from solutions of K4Ge9, i.e. without reacting them with GePh4.16

This indicates,

therefore, that the Ge9–clusters do not react with GePh4, most likely because of the stronger Ph–Ge bond that prevents generation of Ph3Ge– anions in the solution. Next, we tested reactions with the much more ionic Ph3GeCl (as well as Ph3SnCl and Me3SnCl) which can be reduced very easily to anions Ph3Ge– and Cl–. As expected, the reaction was very vigorous and produced plenty of gray-black precipitation of elemental germanium. This particular experiment was 68

performed with excess of Ph3GeCl, and clearly the following reaction must have occurred: 2Ph3GeCl + K4Ge9 → 2KCl + 2K+ + 2Ph3Ge– + 9Ge0. Therefore, Ph3Ge– anions were generated for the expense of all the available Ge94clusters which became fully oxidized to Ge0 and precipitated.

If excess of

clusters were used the additional amount would have been able to react with the generated Ph3Ge– anions. This was tested and, indeed, reactions with excess of K4Ge9 with respect to R3EX produced various R3E–substituted clusters including a di-substituted dimer of germanium clusters [Ph3Sn–Ge9–Ge9– SnPh3]4- (6). This confirmed again that needed for successful addition are the corresponding anions of the substituents.

The direct reactions between the

clusters and the halides are somewhat difficult to control and also involve "sacrifice" of a fraction of the clusters for the reduction of the halide. For these reasons it was decided to generate the R3E– anions separately by reduction of the halides with alkali metals.

Thus, reactions of R3EX with potassium in

ethylenediamine were used for this purpose.

The resulting solutions were

reacted with solutions of K4Ge9 and, depending on the molar ratios, the following substituted species were crystallized with (K–2,2,2-crypt)+ as countercations (Figure 5.7): [Ph3Sn–Ge9–SnPh3]2– (1), [Me3Sn–Ge9–SnMe3]2- (2), [Ph3Ge– Ge9–GePh3]2- (3), [Ge9–SnPh3]3- (4), and [Ge9–SnMe3]3- (5) (1 and 5 were also characterized with (K–18c6)]+ as countercations). The rational synthesis of these species by this approach proves that the reaction type is of nucleophilic addition of R3E– to the germanium clusters. 69

One more advantage of using pre-

formed anions is that the reaction is independent of the strengths of the E–R and E–X bonds. The successful synthesis of Ph3Ge–substituted clusters using this approach confirms this. Furthermore, the use of the anions allows for better control of the reaction, and this made possible the synthesis of the monosubstituted species 4 and 5 when smaller amounts of R3Sn– were used (R3E– : K4Ge9 = 1 : 1). Their synthesis and characterization prove the proposed reaction path described above. In addition to the structural characterization, the tin-containing anions 1, 2, 4, and 5 were characterized also in solution with 119Sn NMR. Crystals of (K– 2,2,2-crypt)21, [K–18c6)]21•0.25(18c6)•2en, (K–2,2,2-crypt)22•3.5tol, (K–2,2,2crypt)34•en,

(K–2,2,2-crypt)35, and (K–18c6)]35•thf•2en were dissolved in

pyridine and their spectra showed peaks with chemical shifts (with respect to Me4Sn in CDCl3 as external standard) at 15.1, 14.9, 21.9, 260.0, 130.3, 116.2 ppm, respectively. For comparison, solutions of Ph3Sn– and Me3Sn– prepared by reduction of the corresponding chlorides with excess potassium in ethylenediamine showed chemical shifts of -113.0 ppm and -176.9 ppm, respectively. It should be pointed out that Ph3Sn– anions were observed in the spectra of 1 and 4 (in pyridine) at -108.1 and -109.8 ppm, respectively, but no Me3Sn– was observed in the solutions of 2 and 5. It may be that Ph3Sn– has higher stability in pyridine, most likely because of better solvation by the aromatic solvent.

70

b)

a) Sn 2.608

2.621

Sn

Sn

2.618

2.617

2.607 (2.594)

2.617 (2.605)

Sn

Sn

Sn

B

A

e)

d)

c)

Sn

7)

2.434

85

2.9

(2.69

2.9

40

Sn

2.695

2.435

Ge

Ge

f)

2.650 2.481

Sn

Sn

2.650

Figure 5.7. ORTEP drawings (50 % thermal ellipsoids) of: a) 1 in (K–2,2,2crypt)21 (and in (K–18c6)]21·0.25(18c6)·2en), b) 2 in (K-2,2,2-crypt)22·3.5tol, c) 3 in (K–2,2,2-crypt)23·tol·0.5en, d) 4 in (K–2,2,2-crypt)34·en, e) 5 in (K–2,2,2crypt)35 (and in (K–18c6)35·THF·2en), and f) 6 in (K–2,2,2-crypt)46·2en. Anion 1 in (K–18c6)21·0.25(18-crown-6)·2en and anion 2 are represented by two crystallographically slightly different geometries in the corresponding structures, A and B. The Ge–Sn distances are shown (the distances in parentheses in a) and e) are for (K–18c6)21·0.25(18c6)·2en and (K–18c6)35·THF·2en, respectively) while the Ge–Ge distances are listed in Table 5.2. Open bonds indicate elongated edges. The numbering of the germanium atoms of the clusters is shown in a) and is the same for all clusters.

71

TABLE 5.2 Ge–Ge distances [Å] in the Ge9 clusters of the new species 1 – 3 (1' and 1" are the anions in (K–2,2,2-crypt)21 and [K-18c6)]21•0.25(18c6)•2en, respectively).[a] Atoms[b] 1-2 1-3 1-4 1-5 2-3 2-5 2-6 2-7 3-4 3-7 3-8 4-5 4-8 4-9 5-6 5-9 6-7 6-9 7-8 7-9 8-9

1' 2.588(1) 2.573(1) 2.587(1) 2.576(1) 2.823(1) 2.735(1) 2.636(1) 2.617(1) 2.717(1) 2.628(1) 2.613(1) 2.827(1) 2.625(1) 2.622(1) 2.615(1) 2.620(1) 2.559(1) 2.544(1) 2.578(1) 3.035(1) 2.586(1)

1"(A)

1"(B)

2(A)

2(B)

2.560(2) 2.573(1) 2.573(2) 2.571(2) 2.805(1) 2.713(1) 2.637(1) 2.612(1) 2.739(1) 2.593(1) 2..636(1) 2.852(1) 2.598(2) 2.642(2) 2.596(1) 2.644(1) 2.566(1) 2.559(1) 2.550(1) 3.018(1) 2.561(1)

2.565(2) 2.573(1) 2.562(2) 2.585(1) 2.821(1) 2.734(1) 2.605(1) 2.636(2) 2.729(1) 2.636(1) 2.603(1) 2.815(1) 2.634(1) 2.606(1) 2.628(1) 2.591(1) 2.560(1) 2.555(1) 2.575(1) 2.997(1) 2.560(1)

2.579(1) 2.597(1) 2.583(1) 2.590(1) 2.819(1) 2.744(1) 2.647(1) 2.646(1) 2.752(1) 2.660(1) 2.671(1) 2.865(1) 2.637(1) 2.634(1) 2.645(1) 2.643(1) 2.571(1) 2.587(1) 2.595(1) 3.096(1) 2.602(1)

2.611(1) 2.597(1) 2.590(1) 2.599(1) 2.866(1) 2.800(1) 2.676(1) 2.667(1) 2.752(1) 2.669(1) 2.639(1) 2.855(1) 2.661(1) 2.612(1) 2.682(1) 2.635(1) 2.638(1) 2.621(1) 2.568(1) 3.107(1) 2.565(1)

3 2.579(1) 2.595(1) 2.575(1) 2.577(1) 2.862(1) 2.719(1) 2.621(1) 2.612(1) 2.687(1) 2.587(1) 2.655(1) 2.887(1) 2.633(1) 2.609(1) 2.647(1) 2.600(1) 2.564(1) 2.566(1) 2.553(1) 3.089(1) 2.548(1)

[a] The underlined distances are the elongated prismatic edges of the Ge9clusters shown as open bonds in Figure 5.7. [b] The numbering of the atoms is shown in Figure 5.7a. [c] This cluster is disordered among two positions and the reported distances are averages.

72

TABLE 5.2 (contd.)

Atoms[b] 1-2 1-3 1-4 1-5 2-3 2-5 2-6 2-7 3-4 3-7 3-8 4-5 4-8 4-9 5-6 5-9 6-7 6-9 7-8 7-9 8-9

4 2.552(2) 2.527(2) 2.618(2) 2.591(2) 2.756(2) 2.850(2) 2.574(2) 2.734(2) 2.949(2) 2.687(2) 2.537(2) 2.659(2) 2.627(2) 2.613(2) 2.582(2) 2.659(2) 2.560(2) 2.595(2) 2.551(2) 2.985(2) 2.626(2)

5'[c] 2.435 2.533 2.608 2.504 2.811 2.899 2.651 2.723 2.791 2.694 2.622 2.817 2.619 2.615 2.584 2.596 2.525 2.609 2.505 3.212 2.609

5"

6

2.562(1) 2.538(1) 2.619(1) 2.603(1) 2.813(1) 2.855(1) 2.609(1) 2.660(1) 2.864(1) 2.657(1) 2.594(1) 2.794(1) 2.618(1) 2.657(1) 2.616(1) 2.637(1) 2.511(1) 2.666(1) 2.503(1) 3.218(1) 2.661(1)

2.560(2) 2.556(2) 2.583(2) 2.586(2) 2.868(2) 2.729(2) 2.588(2) 2.652(2) 2.737(2) 2.670(2) 2.574(2) 2.796(2) 2.623(2) 2.623(2) 2.675(2) 2.606(2) 2.553(2) 2.540(2) 2.579(2) 3.077(2) 2.570(2)

Structure. The six new species 1 – 6 are shown in Figure 5.7. Those of type 1 in the (K–18c6) salt and 2 have two crystallographically different geometries each, A and B, that are otherwise very similar. The cores of all these species are the well known deltahedral Zintl ions Ge9n- that can exist on their own, i.e. without substituents.

As already discussed, they have been

73

characterized both as Zintl ions in numerous compounds crystallized from solutions as well as in Zintl phases.5, 6 These clusters are quite flexible in shape and charge. Overall, they resemble tricapped trigonal prisms with variously elongated one, two, or three trigonal prismatic edges parallel to the three-fold axis and can carry charges of 2-, 3-, and 4-. Referring to the numbering of the Ge-atoms shown in Figure 5.7a, the trigonal prism is made of the triangular bases 4-5-9 and 2-3-7 (the three-fold axis is vertical) while the capping atoms are 1, 5, and 6.

It should be mentioned that tricapped trigonal prisms with one

elongated edge can be viewed also as monocapped (atom 1) square antiprisms (squares 2-3-4-5 and 6-7-8-9). The same type distortions are observed in the core clusters of the substituted species 1 – 6.

Thus, the edges 7–9 are

elongated in all of them, while also elongated in 4 is the edge 3–4 (Figure 5.7 and Table 5.2). As observed before, the elongations in such substituted species are related to the positioning of the substituents. It can be seen in Figure 5.7 that the exo-bonds are always to atoms of these elongated edges, i.e. atoms 7 and 9 for 1 – 3, atom 7 in 5, and atoms 3 and 7 in 4. Furthermore, with the exception of 4, the exo-bonds are almost parallel to the elongated edges and look as their outward extensions.

The subsittuent in 4 bonds to two atoms, and it is

geometrically impossible for the two exo-bonds to be parallel to these edges. The closest to such a position is for the subsittuent to be within the plane defined by these two edges, and this is exactly where Ph3Sn is located in this structure. The shapes of the Ge9 clusters and the positioning of the substituents are directly related to the cluster's electronic structure. This relationship has been

74

already discussed in some detail based on molecular orbital calculations for Ph3Sb-substituted clusters.16 The main points are that the HOMO for Ge94- is made of pz orbitals (z along the three-fold axis of the prism) and is within a relatively large energy gap, i.e. there are comparable gaps below and above it. This same orbital is half-filled for Ge93- and is empty and LUMO for Ge92-. Its energy depends very strongly on the elongation of the prismatic edges: greater elongation and more elongated edges lower the orbital's energy due to relief of antibonding interactions between the two triangular bases of the prism. Thus, Ge94- clusters (nido-species according to Wade's rules; 22 cluster-bonding electrons) typically have one or more elongated edges and, correspondingly, this orbital is at relatively low energy and is occupied. For example, the lengths of such edges are approximately in the ranges 3.61–3.70, 3.15-3.54, and 3.03–3.19 Å for clusters with one, two, and three elongated edges, respectively. It should be pointed out that, as these numbers show, the elongation is less pronounced when more edges are elongated.

Although the structure of Ge92- (closo-

species, 20 cluster-bonding electrons) has not been determined well, it is clear that none of the edges is elongated.17

This makes the molecular orbital in

question more antibonding and empty. The situation with Ge93- is somewhat in the middle, i.e. the elongation of the edges is not as pronounced as in Ge94-. The lengths of the elongated edges fall in the ranges 3.21–3.55 and 3.02–3.33 Å for the known examples of such clusters with one and two elongated edges, respectively.

75

The structures of the new species 1 – 6 show that the antibonding character of the same molecular orbital can be alleviated not only by edge elongations but also by the addition of substituents along the general direction of the pz orbitals at the atoms of the trigonal prism. Thus, although the Ge9–cluster cores in the new species carry the same number of cluster-bonding electrons, 22, as the naked Ge94- cluster, their single elongated edges are quite shorter: 2.997 – 3.107 Å in the di-substituted 1, 2, 3, and 6, and 3.212 Å in the monosubstituted 5 (3.218 Å with (K-18c6)). The latter is somewhat longer than the former but is shorter than in non-substituted species. The trend is observed also for the clusters in the dimer [Ge9–Ge9]6- with one exo-bond per cluster and the chain [–(Ge92–)-]1∞.with two such bonds per cluster where the lengths of the single-elongated edges are 3.433 and 3.194 Å, respectively.11,

14

The same is

true for clusters with two elongated edges, i.e. the elongation is more pronounced in naked clusters and less so in exo-bonded ones. Thus, these two edges in the monosubstituted 4 with bridging Ph3Sn, 2.949 and 2.985 Å, are much shorter than in the corresponding naked clusters (above). The exo-bond distances in 1 and 2, 2.594–2.617 Å, and in 3, 2.434 and 2.435 Å, correspond to single-bond Ge–Sn and Ge–Ge distances, respectively. The same is true for the intercluster distance of 2.481 Å in 6. They compare well with 2.599 Å for Ge–Sn in Me3GeSnPh3 and with 2.438 and 2.441 Å for Ge–Ge in Ph3Ge-Ge(Ph)2-GePh3.20,

21

The effect of the bridging in 4 and the

associated 3-center–2-electron interaction are clearly manifested in the longer Sn–Ge distances of 2.940 and 2.985 Å. Although the trimethyltin substituent in 5 76

is not exactly bridging, it is noticeably bent towards the three-fold axis of the prism and brings the tin atom close to the other two germanium atoms of the triangular base Ge2 and Ge3 (3.3 – 3.9 Å). The weak interactions with these two germanium atoms results in a somewhat longer Ge–Sn bond distance of 2.695 Å (2.697 Å for the (K–18c6) salt) that is somewhat longer than in 1 and 2 but shorter than in 4. Summary.

Deltahedral Zintl ions functionalized with various organic

derivatives of elements of groups 14 and 15, R3E– and R2E–, respectively, can be synthesized by design.

The rationality of the approach is based on the

understanding that the substituents are added as anions to the clusters. The nature of the reaction is nucleophilic addition to an empty or half-filled cluster orbital.

This knowledge opens doors for further exploration of many other

nucleophiles as potential substituents in such clusters.

5.5. References

[1]

Joannis, A. C. R. Hebd. Seances Acad. Sci 1891, 113, 795.

[2]

Zintl, E.; Harder, A. Phys. Chem., Abt. A 1931, 154, 47.

[3]

Zintl, E.; Dullenkopf, W. Z. Phys. Chem., Abt. B 1932, 16, 183.

[4]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[5]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[6]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

77

[7]

vonSchnering, H. G.; Baitinger, M.; Bolle, U.; CarrilloCabrera, W.; Curda, J.; Grin, Y.; Heinemann, F.; Llanos, J.; Peters, K.; Schmeding, A.; Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[8]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[9]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[10]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[11]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[12]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[13]

Ugrinov, A.; Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

[14]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[15]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

[16]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2003, 125, 14059.

[17]

Belin, C.; Mercier, H.; Angilella, V. New Journal of Chemistry 1991, 15, 931.

[18]

Dessy, R. E.; Kitching, W.; Chivers, T. J. Am. Chem. Soc. 1966, 88, 453.

[19]

Fassler, T. F.; Schutz, U. Inorg. Chem. 1999, 38, 1866.

[20]

Pannell, K. H.; Parkanyi, L.; Sharma, H.; Cervanteslee, F. Inorg. Chem. 1992, 31, 522.

[21]

Roller, S.; Simon, D.; Drager, M. J. Organomet. Chem. 1986, 301, 27.

78

CHAPTER 6

SYNTHESIS AND CHARACTERIZATION OF [Me3C-Ge9-Ge9-CMe3]4-

6.1. Abstract A new ligated Zintl anion was synthesized by a direct reaction between the known Zintl ions Ge9n- and the organic halide Me3CCl. The functionalized cluster was structurally characterized in a salt with rubidium countercations sequestered in 2,2,2-crypt. The new compound, (Rb-2,2,2-crypt)4[Me3C-Ge9-Ge9-CMe3]·7en crystallized in the monoclinic space group C2/c. The anion was also characterized with

13

C-NMR in solution. Discussed are the possible reaction type

and the structure of the anion.

6.2. Introduction Recent developments in the area of deltahedral Zintl ions have brought new light to their chemistry and have changed the major concepts about their reactivity. In the beginning, the Zintl anions were observed in solutions of liquid ammonia or ethylenediamine, and were isolated and characterized only from such solutions as salts with the alkali metals.1,

2

Later, Sevov and co-workers

showed that the same nine-atom clusters of the heavy elements of group IV exist

79

in “neat” solids, i.e. in solids made by solids state reactions.3-8 Also, in 1999 dimers of nine-atom germanium clusters [Ge9-Ge9]6- were discovered.9 This was the first example of a redox reaction with these species, and it was quickly followed by the synthesis of a chain of clusters [-(Ge92-)-]1∞,10 as well as larger oligomers

such

as

trimers

and

[Ge9=Ge9=Ge9=Ge9]8-, respectively.11,

12

tetramers,

[Ge9=Ge9=Ge9]6-

and

All these compounds showed that exo

bonds between the clusters can exist. Functionalization of the nine-atom clusters by covalently exo-bonded groups was achieved, for first time, by reacting ethylenediamine solution of a precursor of nominal composition K4Ge9 with BiPh3.13 In the resulting compound, (K-2,2,2-crypt)2[Ph2Bi-Ge9-BiPh2]·en, two diphenylbismuth groups are exobonded to two opposite germanium vertexes of the open face of the cluster. Shortly after that were also synthesized the following ligated anions, [Ph2Sb-Ge9SbPh2]2-, [Ph3E-Ge9-EPh3]2- where E=Sn, Ge, [Me3Sn-Ge9-SnMe3]2-, [Ph-Ge9SbPh2]2-, [Ph2Sb-Ge9-Ge9-SbPh2]4-, [Ph3Sn-Ge9-Ge9-SnPh3]4-, [Ge9SnR3]3- where R = Ph and Me.14, 15 Furthermore, for first time, the mechanism of such reactions was investigated and was found to be of nucleophilic addition of the anions RnEto the Zintl anions. Based on that success, we extended our investigations to reactions between nine-atom germanium clusters and organic halides. Here we report the first successful direct reaction between the Zintl anions and Me3CCl. The resulting

anion,

[Me3C-Ge9-Ge9-CMe3]4-

(1),

is

the

second

functionalized deltahedral Zintl anion after [Ph-Ge9-SbPh2]2-.14

80

organically

6.3. Experimental Section Materials and Techniques:

All manipulations were performed in a

nitrogen-filled glove box with moisture level below 1 ppm. All precursors were synthesized by heating (900 °C, 2 days) stoichiometric mixtures of the elements (K: 99+ %, Strem; Rb:99+%, Strem; Eu: 99.9%, ingot, Ames Laboratory DOE; Ge: 99.999 %, Alfa-Aesar) in sealed niobium containers (arc-welded in Ar at low pressure) that were in turn sealed in quartz ampoules under vacuum.

The

solvent ethylenediamine (99.5+%, Aldrich) and the reagents 2,2,2-crypt (98%, Acros), Me3CCl (99%, Acros) were used after careful drying under vacuum. 13

C-NMR spectra were taken in ethylenediamine or pyridine on a Varian

300MHz spectrometer. Sealed capillaries with C6D12 and TMS were used as deprotonated solvent and reference. Synthesis of (Rb-2,2,2-crypt)41·7en: RbEuGe (100mg, 0.3mmol) was dissolved in ethylenediamine (1.5 mL) and the resulting green-brown solution was filtrated and divided in three different test-tubes, which were labeled with 1, 2, and 3 respectively. These three solutions were treated with different amount of Me3CCl as follows: Me3CCl (25mg, 0.27mmol) was dissolved in 1mL ethylenediamine and 1, 3, and 5 drops of this solution were added to the testtubes 1, 2, and 3 respectively. Each of the three resulting black-green solutions was layered with 80 mg (0.2 mmol) 2,2,2-crypt dissolved in 2 mL toluene. Red crystals of (Rb-2,2,2-crypt)4 1·7en were grown on the walls of the test-tubes number 2 and number 3, i.e. when bigger amount of Me3CCl was used. In both test tubes were found several crystals and a lot grey-black precipitation.

81

Synthesis of (K-2,2,2-crypt)41·7en:

K4Ge9 (147mg, 0.18mmol) was

dissolved in 3mL ethylenediamine. The red-brown solution was filtrated and divided in three different test-tubes, which were labeled with 1, 2, and 3 respectively. Me3CCl (74 mg, 0.8 mmol) was dissolved in 1mL ethylenediamine and 0.1 mL, 0.2 mL, and 0.3 mL of this solution were added to the test-tubes 1, 2, and 3, respectively. Each of the three resulting black-green solutions was layered with 50 mg (0.13 mmol) 2,2,2-crypt dissolved in 3 mL toluene. Big blackred crystals of (K-2,2,2-crypt)41·7en were grown on the walls of the test-tubes 2 and 3, respectively. Again significant amount of black-grey precipitations were presented in test tubes 2 and 3. It should be mentioned that the amount of the crystals in test-tube number 2 were more than the amount in test-tube number 3, and the black-grey precipitation was less in test-tube number 2. Based on that we can conclude that the ratio between the clusters, Ge9n-, and the reactant, Me3CCl, in test-tube number 2 is the best used ratio for successful reaction at that stage. Several crystals of (K-2,2,2-crypt)41·7en were dissolved in pyridine and 13

C-NMR spectrum was run: δ = 21.567 ppm, s, -CH3 (Figure 6.1)

82

83

Structure determination: Single crystal X-ray diffraction data of (Rb2,2,2-crypt)4 1·7en were collected at 100K on a Bruker APEX diffractometer with a CCD area detector (graphite-monochromated Mo Kα radiation, crystals protected with Parathone-N oil). The structure was solved by direct methods and refined on F2 using the SHELXTL V5.1 package (after absorption corrections with SADABS). Details of the data collection and refinement are given in Table 6.1.

TABLE 6.1 Crystallographic data for (Rb–2,2,2-crypt)41·7en (Rb–2,2,2-crypt)41·4.5en 3502.86 C2/c, 4 34.776(5) 24.843(3) 16.755(2) 90.000 94.929(2) 90.000 14422(3) 1.613 37876 12665 6106

fw space group, Z a [Å] b [Å] c [Å] α [°] β [°] γ [°] V [Å3 ] ρcalc [g•cm-3] No. of measured refl. No. of indep. refl. No. of used refl. µ [cm-1]

5.089 50 8.97/23.69 18.83/26.81

2θmax [°] R1/wR2 (I≥2σ1)a [%] R1/wR2 (all data) [%]

[a] R1 = Σ||Fo| - |Fc|| / Σ|Fo|, wR2 = {[Σ[(Fo)2 – (Fc)2]2] / [Σw(Fo2)2]}1/2 for Fo2 > 2σ(Fo2), w = [σ2(Fo )2 + (AP)2 + BP]-1 where P = [(Fo)2 + 2(Fc)2] / 3; A (B) = 0.1498 (0.000).

84

6.4. Results and discussions Detailed studies of reactions of nine-atom germanium clusters with soft oxidizing agents were carried out and they suggested that these reactions were in fact nucleophilic additions of the anions Ph2E- (E = Sb, Bi) and R3E- (R = Me, Ph; E = Ge, Sn) to the naked Zintl ions Ge9n- (n = 2, 3, or 4).14, 15 Based on this mechanism, we predicted that mono-substituted intermediates, [Ge9-ER2]3- and [Ge9-ER3]3-, respectively, should be relatively stable and later they were isolated and characterized. Furthermore, the mechanism was proved by direct reaction between the clusters, Ge9n-, and corresponding anions, R2E- or R3E-. We extended our investigations in to reactions between nine-atom germanium clusters and organic compounds. Something more, it was already known at that time that organically functionalized deltahedral Zintl ions can form and [Ph-Ge9SbPh2]2- was an example for that.14 However, the attempts to attach organic species to nine-atom germanium clusters by using organic anions as reagents were unsuccessful at that stage. Based on our experience we can summarize the following important facts: (a) the Ph-ring only attached to the cluster in direct reaction between Ph3Sb and Ge9n-, but never in direct reaction between Ph- (or R-) and Ge9n-; (b) direct reactions between Ge9n- and R2E- or R3E- produce ligate clusters only for E = Bi, Sb, Sn and Ge, but not for E = As, P, Si and C; (c) disubstituted dimers such as [Ph3Sn-Ge9-Ge9-SnPh3]4- are synthesized only in direct reaction between Ph4Sn or

Ph3SnCl with Ge9n-, but never in reaction

between the anions Ph3Sn- and Ge9n-. Obviously the reactions between the Zintl

85

anions and the neutral species RnE and Rn-1ECl lead to more than one product. This suggested that parallel with the nucleophilic addition there might be other reaction mechanisms, most likely involving radical species. In other words, the phenyl groups were most likely attached to the cluster as radicals of Ph·. Similarly the disubstituted dimers of [RnE-Ge9-Ge9ERn]4- were probably formed by radical dimerizations between two monosubstituted radicals, [RnE-Ge9·]2-. In order to prove these assumptions we carried out several direct reactions between Me3CCl and Ge9n-. Me3CCl was chosen because (a) it forms a stable radical Me3C· (b) the anion Me3C- can not exist in ethylenediamine. The latter reacts readily with the solvent to form Me3CH. The reactions were performed both with RbEuGe and K4Ge9. Different stoichiometric ratios between clusters and Me3CCl were explored. Substituted dimers of [Me3C-Ge9-Ge9-CMe3]4(Figure 6.2), were synthesized with both precursors, but only when large excess of Me3CCl was used, larger than two times. Such excesses of the oxidizing agent (Me3CCl + e- → Me3C· + Cl-) leads to large amounts of elemental germanium. This is most likely due to the additional and non productive oxidation step of Me3C· + e- → Me3C- followed by reaction with H2NCH2CH2NH2 to form Me3CH and amide. All attempts to isolate mono- or di-substituted monomers such as [Me3CGe9]3- or [Me3C-Ge9-CMe3]2- were unsuccessful. This may be due to a radical mechanism such as Ge92- + Me3C· → [Me3C-Ge9·]2- followed by dimerization. Clearly reactions need more investigations in future.

86

1

4

3

5

2

8 C1

1.99

2

7

6

9

2.524

Figure 6.2. ORTEP drawing of [Me3C-Ge9-Ge9-CMe3]4- (30% probability thermal ellipsoids).

The two new compounds (Rb-2,2,2-crypt)41·7en and (K-2,2,2-crypt)41·7en are isostructural and only the former was structurally characterized by single crystal X-ray diffraction. It crystallized in monoclinic space group, C(2)/c. The second

compound,

with

potassium,

is

isostructural

(C2/c,

34.776(5)Å,

24.843(3)Å, 16.755(2)Å, 94.929(3)o, V = 14422(3)Å3, Z = 4). Details of the data collections and refinements for it are given in Table 6.1. The new disubstituted dimer is structurally similar to the known [Ph2SbGe9-Ge9-SbPh2]4- and [Ph3Sn-Ge9-Ge9-SnPh3]4-.14, 15 Two nine-atom germanium clusters are bonded to each other by a single Ge-Ge bond with distance 2.524Å which compares well with 2.488Å found in the naked dimer, [Ge9-Ge9]6-.9 The

87

distances inside the clusters are listed in Figure 1 and are in the expected range for such clusters. The substituents, t-butyl groups, are attached along the elongated edges of the clusters. The two exo-bonds are equivalent and with distance of 1.992Å. Such a distance is quite similar to other Ge-C 2c-2e bonds listed in the literature. For example, the range of distances between germanium atoms and t-butyl groups

observed

in

4,8-dibromoocta-tert-butyltetracyclo[3.3.0.02,7.03,6]

octagermane is 1.989 – 2.028Å. In addition to the structural characterization, the compound was also characterized with 13C-NMR in solution. The spectrum of crystals of (K–2,2,2crypt)41·7en, dissolved in pyridine showed a peak with chemical shift (with respect to TMS) at 21.57 ppm for CH3-groups. This spectrum was compared with the spectra of Me3CCl (δ = 34.82ppm, s, CH3-groups) in en and the reaction mixture, K4Ge9 + Me3CCl + en, (δ = 34.82ppm, s, CH3-groups of the reactant; δ = 24.57ppm, s, CH3-groups of the product) before layering it with solution of 2,2,2crypt in toluene. Obviously attaching t-butyl group to the dimer [Ge9-Ge9]6- shifts the

13

C signal for the CH3-groups to higher field, 34.82ppm to 24.57ppm, exactly

as it can be expected. Because of the solvent effects the chemical shifts of the CH3 groups in pyridine and in ethylenediamine, differ somewhat from shift each other. Conclusions. t-Butyl groups were attached to dimers of nine-atom germanium clusters. The new anion [Me3C-Ge9-Ge9-CMe3]4- was characterized by single crystal X-ray diffraction and by

88

13

C-NMR. This is the first successful

designed reaction between such clusters and organic substituents. We can speculate that the addition occurs via radical mechanism.

6.5. References [1]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[2]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[3]

Queneau, V.; Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

[4]

Queneau, V.; Todorov, E.; Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[5]

Queneau, V.; Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

[6]

Todorov, E.; Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[7]

vonSchnering, H. G.; Baitinger, M.; Bolle, U.; Carrillo Cabrera, W.; Curda, J.; Grin, Y.; Heinemann, F.; Llanos, J.; Peters, K.; Schmeding, A.; Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[8]

von Schnering, H. G.; Somer, M.; Kaupp, M.; Carrillo-Cabrera, W.; Baitinger, M.; Schmeding, A.; Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[9]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[10]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[11]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[12]

Ugrinov, A.; Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

89

[13]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

[14]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2003, 125, 14059.

[15]

Ugrinov, A.; Sevov, S. C. Chem. Eur. J. 2004, 10, 3727.

90

CHAPTER 7

SYNTHESIS OF A CHAIN OF NINE-ATOM GERMANIUM CLUSTERS ACCOMPANIED WITH DIMERIZATION OF THE SEQUESTERING AGENT

7.1. Abstract [Rb2(4,2,1,1-crypt)]Ge9•en was synthesized from a precursor of nominal composition Li2RbGe17 that was dissolved in ethylenediamine. Red crystals of the compound were obtained after addition of 2,1,1-crypt and layering with THF. The structure (triclinic, P-1, a = 11.1927(5), b = 15.0516(7), and c = 15.7847(7) Å, α = 57.671(1), β = 83.594(1), and γ = 80.217(1) °, Z = 2) features infinite chains of (–Ge92-–)∞ where each cluster of Ge92- is bonded to two other clusters via single Ge–Ge bonds. Most interestingly, the 2,1,1-crypt molecules form dimers by opening C–N bonds within the molecules and restoring them to another molecule.

Each dimer sequesters two rubidium

cations forming a complex dication that coordinates to the chain of clusters.

7.2. Introduction Significant progress in research on nine-atom clusters of germanium has been made in recent years.1, 2 Typically, the clusters are obtained by dissolving

91

alkali-metal/germanium precursors in ethylenediamine or liquid ammonia followed by crystallization with the aid of various sequestering agents for the alkali-metal countercations. The sequestering agents, typically crypts or crown ethers, are not necessary for the dissolution process. The precursors of stoichiometry A4Ge9 (A = Na, K, Rb) can be dissolved in ethylenediamine at room temperature without a sequestering additive. The latter is added later in order to crystallize a compound.

The shape of the Ge9-clusters is that of a

tricapped trigonal prism with elongated one, two or three prismatic edges, those parallel to the 3-fold axis of the prism, and the clusters can carry charges of 4-, 3or 2-.

We have already established that these differently charged clusters

coexist in the ethylenediamine solutions.3

Depending on the type of

sequestering agent and its concentration, compounds with clusters of different charges can be crystallized from such solutions. One of the most noticeable recent discoveries is that these clusters can couple

to

form

dimers,4

trimers,5

and

even

tetramers,6

[Ge9–Ge9]6-,

[Ge9=Ge9=Ge9]6-, and [Ge9=Ge9=Ge9=Ge9]8-, respectively. It was found that this depends to a great extent on the degree of concentration of clusters in the solution and also on the type of the sequestering agent. This and the observed infinite chains of such clusters, (–Ge92-–)∞ in [K(18-crown-6)]2Ge9 and [K(2,2diaza-18-crown-6)]KGe9·3en,7, 8 provide hopes for achieving controlled stepwise growth of even larger oligomers in the future. Furthermore, the fact that these clusters can carry different charges, i.e. 2-, 3-, or 4-, makes their redox chemistry

92

interesting and offers possibilities for their use as building blocks in materials with extended structures. Here we report the synthesis and characterization of a third compound with infinite chains of (–Ge92-–)∞, [Rb2(4,2,1,1-crypt)]Ge9•en (1). Furthermore, the sequestering agent used for capturing the alkali-metal cations, 2,1,1-crypt (4,7,13,18-tetraoxa-1,10-diazabicyclo[8.5.5]eicosane), undergoes a condensation reaction of dimerization and forms a new compound that can be named 4,10,13,19,25,29,33,38-octaoxa-1,7,16,22-tetraazatricyclo[20,8,5,5]tetracontane, or 4,2,1,1-crypt for short.

7.3. Experimental Section Synthesis. All manipulations and reactions were carried out under argon. A precursor of nominal composition Li2RbGe17 was synthesized from a stoichiometric mixture of the elements (Li, Acros, 99+%; Rb, Strem, 99+%; Ge, Acros, 99.999%). The mixture was heated at 900oC for 2 days in a niobium tubular container sealed by arc-welding at both ends and then jacketed by flamesealing under vacuum in a fused-silica ampoule. Approximately 30 mg (ca. 2.25x10-5 mol) of the precursor were dissolved in 1 ml of ethylenediamine (Aldrich, redistilled, 99.5+%, packaged under nitrogen) and formed brown solution. 32 mg of 2,1,1-crypt (11.0x10-5 mol) (liq.90%, Aldrich packaged in ampoules) was added and the resulting green-brown solution was layered with 3 ml THF (anhydrous, 99.9%, Aldrich, packaged under nitrogen) in a test tube.

93

Red crystals of [Rb2(4,2,1,1-crypt)]Ge9•en (1) grew on the walls and the bottom of the test tube and were collected 2 days later. Structure determination. X-ray diffraction data of a single crystal of 1 were collected at 100 K on a Bruker APEX diffractometer with a CCD area detector (graphite-monochromated Mo Kα radiation, crystals protected with Parathone-N oil). The structure was solved by direct methods and refined on F2 using the SHELXTL V5.1 package (after absorption correction with SADABS). Details of the data collection and refinement are given in Table 1. Further information on the crystal structure investigation is available free upon request from the Cambridge Crystallographic Data Centre (12 Union Road, Cambridge CB2 1EZ, UK; fax: (+44) 1223-336-033; www.ccdc.cam.ac.uk/conts/retrieving.html; or [email protected]) upon quoting depository number CCDC 248253.

94

TABLE 7.1 Crystallographic data for [Rb2(4,2,1,1-crypt)]Ge9•en

fw

1457.09

crystal color, size [mm]

red, 0.19 x 0.10 x 0.04

space group, Z

P-1, 2

a [Å]

11.1927(5)

b [Å]

15.0516(7)

c [Å]

15.7847(7)

α [°]

57.671(1)

β [°]

83.594(1)

γ [°]

80.217(1)

V [Å3]

2420.8(2)

ρcalc [g•cm-3]

1.999

µ [cm-1]

7.549

2θmax [°]

50

R1/wR2 (I≥2σ1)a [%]

4.72/9.74

R1/wR2 (all data) [%]

8.00/10.94

[a] R1 = Σ||Fo| - |Fc|| / Σ|Fo|, wR2 = {[Σ[(Fo)2 – (Fc)2]2] / [Σw(Fo2)2]}1/2 for Fo2 > 2σ(Fo2), w = [σ2(Fo )2 + (0.0469P)2]-1 where P = [(Fo)2 + 2(Fc)2] / 3.

95

7.4. Results and Discussions The new compound [Rb2(4,2,1,1-crypt)]Ge9•en crystallizes in the centric triclinic space group P-1. The asymmetric unit consists of one Ge9 deltahedral cluster, one complex dication [Rb2(4,2,1,1-crypt)]2+, and one ethylenediamine molecule. The cluster has the shape of a distorted tricapped trigonal prism with one elongated prismatic edge (one of the three edges parallel to the three-fold axis). The edge elongation creates an open square-like face, and the cluster can be viewed also as a monocapped square antiprism. The clusters are linked to each other by two exo-bonds at the two atoms of the elongated edge (the shorter diagonal of the open square-like face) and form an infinite chain (Figure 7.1). The clusters alternate “up” and “down” within the chain, exactly as observed before in [K(18-crown-6)]2Ge9 and [K(2,2-diaza-18-crown-6)]KGe9·3en.7,

8

The

intercluster bonds of 2.493 and 2.475 Å compare well with the corresponding bonds in the latter two compounds and correspond to a single Ge–Ge bond. 1

3 2

2.493 7

7

4 5

8

6

9

2.475 9

Figure 7.1. The infinite chain of Ge92--clusters in 1. The intercluster distances (shown) correspond to a Ge–Ge single bond distance. Intracluster distances (Å): 1-2 2.598, 1-3 2.577, 1-4 2.575, 1-5 2.581, 2-3 2.774, 2-5 2.724, 2-6 2.640, 2-7 2.642, 3-4 2.770, 3-7 2.620, 3-8 2.627, 4-5 2.805, 4-8 2.629, 4-9 2.627, 5-6 2.642, 5-9 2.635 6-7 2.548, 6-9 2.550, 7-8 2.554, 8-9 2.544 (standard deviations of 0.001 Å for all distances).

96

Perhaps the most interesting aspect of the new compound is the dimerization reaction of the sequestering agent 2,1,1-crypt that occurs during the synthesis. This involves the opening of a C–N bond in one of the two singleoxygen chains in each 2,1,1-crypt followed by restoration of the same type of bonds but with partners from different molecules. The resulting dimer has one large ring (Figure 7.2) that contains 4 nitrogen, 6 oxygen, and 20 carbon atoms. Although somewhat distorted, the ring can be considered as defining a plane (Figure 7.2b), similar to a crown polyether. Attached to pairs of nitrogen atoms in this ring are the two remaining single-oxygen links of the original 2,1,1-crypts. It is important to point out that the two linkages are on the same side of the plane defined by the large ring, and the dimer can occupy only a half of the coordination spheres of the sequestered rubidium cations. This is very similar to the way crown ethers coordinate to large cations, usually on one side of the cation. On the other hand, it is very different from cryptands which wrap the whole cation and do not leave any opening in its coordination sphere. Therefore, a cation sequestered in cryptand is completely incapable for any interactions with the anionic part of the structure. In contrast, crown ethers as well as the dimer of 2,1,1-crypt in the present case leaves the cations quite exposed for additional interactions and they coordinate to the chain (Figure 7.3) and stabilize it. It can be argued that, perhaps, it is this coordination of cations that causes the formation of the chain itself. It should be noted that similarly the tetramer of [Ge9=Ge9=Ge9=Ge9]8- is stabilized by eight rubidium cations sequestered by 18crown-6.6

97

a)

N

C N

Rb Rb

N

C N

b)

Rb Rb

Figure 7.2. Two views of the new molecule named 4,2,1,1-crypt that is of two 2,1,1-crypt molecules. A C–N bond is broken in each of the latter and two new intermolecular C–N bonds (yellow) are formed. The two original 2,1,1-crypt molecules can be distinguished by the colors of their bonds. The side view in b) shows that the Rb+ cations have open coordination hemispheres for further interactions.

98

1

Rb 3 4

8

7

5

6

9 9

6

Rb

Rb

2

8

7

5

4 2

3

Rb 1

Figure 7.3. Shown are the interactions between the sequestered rubidium cations and the chain of clusters (the view is approximately along the chain of cluster). One of the cations interacts with a triangular face of one cluster while the second interacts with two atoms of two different clusters and a nitrogen atom from the solvent ethylenediamine captured in the structure.

The reaction of dimerization of the 2,1,1-crypt needs to be discussed in some more detail.

First of all, it should be recognized that ethylenediamine

solutions of alkali-metal containing main-group intermetallics are highly reducing and contain solvated free electrons, the ultimate reducing agent. Thus, the first step in the dimerization reaction is a molecule of 2,1,1-crypt taking an electron and breaking a C–N bond as a result of that. The extra electron will most likely reside on the more electronegative nitrogen and provide a two-bonded N– and a very reactive radical –CH2• end (Scheme I). The process may be assisted by the

99

lithium cations present in the solution which would insert inside the crypt and add more strain to the C–N bonds of the single-oxygen chains.

An empty open

molecule would coordinate directly a rubidium cation, while a molecule with lithium inside would replace the cation with rubidium due to the expected better complexation of the latter. The radius of Rb+ is larger than the cavity of a closed 2,1,1-crypt, and therefore the cation would keep the molecule open (Scheme I). It should be pointed out that 2,1,1-crypt is a strained molecule to start with, especially at the single-oxygen rings, one of which is the ring that opens in this case. Molecules of this type most likely open and close constantly at the C–N bonds when in solutions with free electrons even when without cations larger than the molecule’s cavity.

However, in the presence of large cations the

molecule may not be able to close its original C–N bond and rather forms a bond to another molecule. How exactly this latter step is accomplished, i.e. the type and mechanism of the reaction, is not clear at all. One possibility is for the –CH2• radical to attack the nitrogen of another molecule and cause opening of a C–N bond in the latter with creation of the same radical end on the second molecule. Another possibility is for the nucleophilic N– of the first molecule to attack a carbon atom of the second molecule causing a bond opening and formation of another N– species. This process is also accompanied with insertion of another Rb+ in the second molecule. In both cases the last step must be recombination of N– and –CH2• with release of an electron.

100

One, perhaps, less important difference between compound 1 and the two other compounds with chains of germanium clusters is that while the color of the latter is reported as blue-green, the crystals of 1 are red. As it was already mentioned, both the inter- and intra-cluster distances in the three compounds are very similar, and the shapes of the clusters are almost identical. Therefore, the difference in color must be related to the cations captured in this unprecedented sequestering agent and their interactions with the chain. After all, it should be recognized that the chains do not exist in solution but rather assemble only in the solid product. Thus, the color of the solutions for all three compounds is the same, dark green to green-brown. In other words, they all contain the same or very similar species, most likely oligomers of three, four, or more clusters as we have shown before.5

7.5. References [1]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[2]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

101

[3]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2003, 125, 14059.

[4]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[5]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[6]

Ugrinov, A.; Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

[7]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[8]

Downie, C.; Mao, J. G.; Parmar, H.; Guloy, A. M. Inorg. Chem. 2004, 43, 1992.

102

CHAPTER 8

SYNTHESIS AND CRYSTAL STRUCTURE OF [K-(12c4)2]2[K-12c4]2[Sn9]•4en

8.1. Abstract The new compound [K-(12C4)2]2[K-12C4]2[Sn9]•4en (1), where 12c4 is 12-crown-4 polyether, was crystallized from an ethylenediamine solution of an intermetallic precursor with nominal composition K4Li2Sn8. The structure of the compound (orthorhombic, Cmca, a = 21.763(4), b = 16.030(3), c = 51.85(1) Å, Z = 8) contains isolated deltahedral clusters of [Sn9]4-. The clusters are of Cs symmetry and are nido- according to the number of bonding electrons. The potassium countercations are sequestered by the very small 12-crown-4 ether, half of them are sandwiched between a pair of molecules while the other half are coordinated by a single molecule.

The latter potassium cations also bridge

edges of the tin clusters, and each cluster has two such coordinated cations. Coordination of potassium cations by such small crown ethers is quite rare.

103

8.2. Introduction Deltahedral Zintl anions have fascinated scientists ever since their discovery in the 19th century, and many such species have been synthesized and characterized since then.1,

2

Typically, they are crystallized from

ethylenediamine or liquid-ammonia solutions with the help of various cationsequestering agents such as 2,2,2-crypt or crown ethers.

The most stable

species seem to be the nine-atom clusters with charges of 3- or 4- and shapes that can be considered as variously distorted tricapped trigonal prisms.3 As in the well-known cage-like boranes, bonding in these clusters is achieved via delocalization of the available electrons, and the numbers of the latter follow the Wade’s rules.

We have shown that upon soft oxidation such germanium

clusters, can couple in different modes to form various oligomeric species when crystallized with appropriate countercations.

The first dimer of deltahedral

clusters was stabilized with a combination of non-sequestered cesium and cryptated potassium cations in Cs4(K-crypt-2,2,2)2[Ge9-Ge9].4 Later, a trimer and

a

tetramer

of

two-bonded

clusters

were

found

in

(Rb-crypt-

2,2,2)6[Ge9=Ge9=Ge9]•2en as well as (K-18C6)6[Ge9=Ge9=Ge9]•2en and in (Rb-18C6)8[Ge9=Ge9=Ge9=Ge9]•4en, respectively.5,

6

Also, a compound, (K-

18C6)2[Ge9]•en, with infinite chains of [–(Ge92-)–]∞ was synthesized and characterized by Guloy's group.7 The sizes and the shapes of the available cations in the solution and the concentration of the precursor are apparently of great importance for the 104

crystallization of various species as it seems that differently-charged clusters coexist in such solutions, i.e. E92-, E93-, E94-, and various oligomers are in equilibria between themselves and free solvated electrons (E = Ge, Sn, Pb). Thus, the availability of only large and bulky cryptated cations (in excess of crypt) leads usually to crystallization of E93-, apparently because only three large cations can pack with one cluster.8, 9 The presence of both free and sequestered alkali-metal cations leads to either E94- or [E9–E9]6- depending on the ratio between the two types of cations.4 Flat sequestering agents, such as crown ethers, on the other hand, seem to have different effect depending on the element forming the cluster.

Thus, depending on the concentration of the

precursor, 18-crown-6 stabilizes either linear oligomers or polymers of germanium, but the same ether stabilizes just the E94- cluster for tin in [K18C6]3K[Sn9], [K-18C6]4[Sn9], and [Rb-18C6]2Rb2[Sn9]•1.5en.10, 11 Guided by these considerations we have started exploring the sequestering capabilities of other, less conventional crypts and crown ethers, and their capability to stabilize different cluster species for different elements. Here we present a nine-atom tin cluster, Sn94-, crystallized with potassium countercations that are sequestered by the very small 12-crown-4 ether (12c4). Apparently, the small size of the crown ether and its inability to coordinate spherically around the larger potassium stabilize the cluster with a charge of 4- instead of 3- observed with crypt-2,2,2 as a sequestering agent.

105

8.3. Experimental Section Materials and techniques. All operations were carried out in inert atmosphere or under vacuum.

The ethylenediamine ( Aldrich, 99.5+%,

redistilled, packaged under nitrogen), 12-crown-4 (Acros, 97%), and tin (Alfa, rod, 99.999%) were used as received. The surfaces of the potassium (Alfa, sticks, 98% metal basis, packaged in mineral oil) and lithium (Aldrich, rods, 99.9% metal basis, sealed under argon) were cleaned before use. The precursor solid K4Li2Sn8 was made from the elements. The mixture was enclosed in a niobium container and the latter was sealed by arc-welding. The container was then sealed in an evacuated fused-silica ampoule. This assembly was heated at 400 °C for 20 days. Synthesis of [K-(12C4)2]2[K-12C4]2[Sn9]•4en. Inside a glove box, 71 mg (~6 x 10-5 mol) of precursor were placed in a test tube and were covered with 1 mL ethylenediamine. Brown solution was obtained. Large excess of 12-crown4 (~300 mg; 1.7 x 10-3 mol) was added, the solution was heated to 45-50 °C, and was stirred at that temperature for about 40 min. The color of the solution changed to red-brown. Some remaining solid was filtered out and the solution was cooled to room temperature and left standing for about 20 hours. Red platelike crystals with sizes suitable for X-ray diffraction studies were formed on the walls and the bottom of the test tube. Structure Determination. Several crystals of 1 were placed on a micro slide and were covered with oil (Paratone–N) inside the glove box. The slide was

106

taken out and, under a microscope, crystals were picked with thin glass fiber. Xray diffraction data were collected of the best one at 100 K using a Bruker APEX diffractometer with a CCD area detector and graphite-monochromated Mo Kα radiation . The structure was solved by direct methods in Cmca and was refined on F2 (full matrix, absorption correction with XABS) using the SHELXTL V5.1 package. Details of the data collection and refinement are given in Table 8.1.

8.4. Results and Discussion The specific precursor with nominal composition K4Li2Sn8 that was used for the preparation of 1 was chosen because a compound with this stoichiometry was characterized recently.

This compound and the isostructural Rb4Li2Sn8

contain isolated arachno-clusters of Sn86-, square antiprisms where the square faces are capped by lithium atoms forming [Li2Sn8]4- closo-species.12

The

relatively small 12-crown-4 ether was intended to "exo-cap" the lithium and eventually extract the whole cluster as [(Li-12C4)2Sn8]4- in solution. However, the ether sequestered the potassium cations instead. It should be mentioned that the precursor made by this particular synthesis (above) contained more than one phase: K4Li2Sn8, K12Sn17, and one or more phases that could not be identified. K12Sn17 is isostructural with the known Rb12Si17 and contains isolated tetrahedra of Sn44- and nine-atom clusters of Sn94-.13 This compound is known to dissolve readily in ethylenediamine, even without the help of sequestering agents.

The solubility of K4Li2Sn8, on the other hand, is not

107

known, and therefore it is not clear whether the solution of the precursor had either phases dissolved or only K12Sn17.

TABLE 8.1 CRYSTALLOGRAPHIC DATA FOR [K-(12C4)2]2[K-12C4]2[Sn9]•4en Empirical formula

C56H128O24N8K4Sn9

Formula weight

2522.27 g•mol-1

Wavelength

0.71073 Å

Crystal system

Orthorhombing

Space group, Z

Cmca, 8

Unit cell dimensions

21.763(4) Å, 16.030(3) Å, 51.85(1) Å

Volume

18089(5) Å3

Density

1.852 g•cm3

Absorption coefficient

2.691 mm-1

Crystal size.

0.25 x 0.25 x 0.02 mm

Crystal color/shape

red/plate

2θ range

3.96 – 46.64 °

Collected/independent reflections

53642

Parameters/data (I ≥ 2σI)

459/5227

R1/wR2 (I ≥ 2σI)

8.8/17.27 %

R1/wR2 (all data)

11.03/18.00 %

108

The structure of compound 1 is made of isolated deltahedral clusters of Sn94- (Figure 8.1.) that are separated by sequestered potassium cations (Figures 8.1 and 8.2) and some solvent molecules. Similarly to many nine-atom clusters of the carbon group the clusters of this compound can be described as distorted tricapped trigonal prisms. This particular cluster has Cs symmetry with the mirror plane containing atoms 1, 4, and 6 (Figure 8.1). The capping atoms are 1, 5, 5 and the two triangular bases of the prism are made of 2, 2, 4 and 3, 3, 6.

Typically, the distortions are elongations of one or more of the trigonal

prismatic edges along the three-fold axis. A special case among these is an elongation of one prismatic edge to such extent that a monocapped square antiprism with C4v symmetry is achieved. However, more common are clusters of lower symmetry due to either the degree of elongation or the number of elongated edges. It has been shown that independent of the type of distortion all these clusters carry charges of 3- or 4-, but not 2-. The distortion observed for the clusters in 1 is clearly elongation along the three prismatic edges, i.e. along 4–6 and the two edges 2–3. The lengths of these edges, 3.734(2) Å for 4–6 and 3.419(1) Å for 2–3, are much longer than the rest of the distances in the cluster which fall in the range 2.927(1) – 3.059(2) Å.

Extended-Hückel calculations

carried out for [Sn9]4- confirmed the nido-character of the cluster with a HOMOLUMO gap of 3.69 eV at 22 bonding electrons (in addition to 18 for lone pairs).

109

1 2

2

3

4

5 N

3

5 6

O K3 O

N

O O

Figure 8.1 The deltahedral cluster Sn94- in 1 is shown with the two [K-(12C4)] countercations that coordinate to two of its edges (edges 4-5). Distances [Å]: Sn1-Sn2 2.929(2), Sn1-Sn3 2.937(2), Sn2-Sn2 3.056(2), Sn2-Sn3 3.419(2), Sn2Sn4 2.977(2), Sn2-Sn5 2.958(2), Sn3-Sn3 3.059(3), Sn3-Sn5 2.962(2), Sn3-Sn6 2.973(2), Sn4-Sn5 2.943(2), Sn5-Sn6 2.964(2).

K1

K2

Figure 8.2. Shown are the two cations K1 and K2 in [K-(12C4)2]2[K-12C4]2 [Sn9]•4en that are sandwiched by two molecules of 12-crown-4 each. The latter are staggered and tilted with respect to each other around K1, but are eclipsed and parallel around K2.

110

The four potassium countercations in the structure are sequestered in two different modes by the 12-crown-4 ether molecules, K1 and K2 are sandwiched by two ether molecules (Figure 8.2) while K3 is coordinated by only one such molecule (Figure 8.1). The potassium cations are removed quite far from the planes of the cycles. This is quite normal as a result of the small size of 12crown-4 ether and in order to achieve normal K–O distances, within 2.68(1) – 2.88(2) Å in this particular case. These distances are actually shorter than those observed between potassium and 12–crown-4 in the structure of the alkalide [K+–(18c6)(12C4)](Na-), 2.815-3.131 Å.14 The two sandwiched cations in 1 differ from each other in the way they are coordinated by the crown ethers.

The oxygen atoms of the two cycles are

staggered around K1 but are eclipsed around K2. Furthermore, while the planes of the rings are parallel in the latter they are substantially tilted with respect to each other in the former. One possible reason for the tilt is the presence of additional ligands such as solvent molecules coordinated to the open side of this potassium atom. However, no electron density was found in this vicinity and also, there is no room for an ethylenediamine molecule there. A closer look at the surrounding of this sandwich revealed the presence of two [K2-(12C4)2] cations nearby that have their crown ethers too close to those of K1. Thus, simple space issues seem to be the reason for the somewhat distorted coordination around K1. K3 is the cation coordinated with only one crown-ether molecule, and as a result of this half of its coordination sphere is exposed for other interactions.

111

Thus, it is found additionally coordinated by lone pairs of electrons from the tin cluster as well as the –NH2 groups of the ethylenediamine solvent captured in the structure. There are two such potassium cations per cluster, each bridging an edge. The two bridged edges are equivalent, between tin atoms 4 and 5, and the two K–Sn distances are 4.193(5) and 4.030(6) Å, respectively. distances are similar to those observed in [K-18c6]3K[Sn9].

These

The two

ethylenediamine molecules around K3 complete its coordination sphere with K–N distances of 2.96 and 2.84 Å. Summary. In conclusion, we have found that small crown ethers such as 12-crown-4 stabilize Sn94- by sandwiching some of the countercations but also by capping cations that can interact with the clusters. Only the latter type of coordination is similar to that observed for larger crown ethers such as 18c6.

8.5. References [1]

Joannis, A. C. R. Hebd. Seances Acad. Sci 1891, 113, 795.

[2]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[3]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[4]

Xu, L.; Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[5]

Ugrinov, A.; Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[6]

Ugrinov, A.; Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

112

[7]

Downie, C.; Tang, Z. J.; Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[8]

Fassler, T. F.; Hunziker, M. Inorg. Chem. 1994, 33, 5380.

[9]

Fassler, T. F.; Hunziker, M. Z. Anorg. Allg. Chem. 1996, 622, 837.

[10]

Fassler, T. F.; Hoffmann, R. Angew. Chem., Int. Edit. 1999, 38, 543.

[11]

Hauptmann, R.; Fassler, T. F. Z. Anorg. Allg. Chem. 2002, 628, 1500.

[12]

Bobev, S.; Sevov, S. C. Angew. Chem. Int. Edit. 2000, 39, 4108.

[13]

Queneau, V.; Todorov, E.; Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[14]

Huang, R. H.; Eglin, J. L.; Huang, S. Z.; Mcmills, L. E. H.; Dye, J. L. J. Am. Chem. Soc. 1993, 115, 9542.

113

SUMMARY AND FUTURE PROSPECTS

Despite their more than 100 year history the deltahedral Zintl ions have remained a remarkably little understood area of inorganic chemistry. The reactivity of these species, expected to be different from Zintl ions with normal 2c-2e bonds, has not been sufficiently studied. As a matter of fact, until 1999, the clusters were assumed to be so highly reduced that any attempt for oxidation was expected to destroy them. Thus, all studies of their chemistry were confined to Lewis acid-base reactions where the clusters were used as ligands to coordinate to transition metals via their lone pairs of electrons. The first example of a redox reaction with these species was the synthesis of a dimer of nine-atom clusters [Ge9-Ge9]6-. After that, extended and systematic studies of nine-atom germanium clusters were carried out and the results are presented in this thesis. The investigations revealed that deltahedral Ge9n- clusters can bond to each other in different modes and to larger oligomers as trimers [Ge9=Ge9=Ge9]6- and tetramers [Ge9=Ge9=Ge9=Ge9]6-. The clusters are bonded with two exo-bonds to each other. The tetramer is the largest discrete species in a growing class of compounds formed by oxidative condensation of the Zintl ion. It is long enough to qualify for a nanorod.

114

These results are highly promising for the future synthesis of larger oligomers, 2D-, 3D-frames or chains with different connections between the clusters. Furthermore, the clusters in the trimer and in the tetramer are stacked in such a way that the "normal" edges of the trigonal prisms alternate with respect of the plane defined by the exo-bonds. Had they all been on one side, the clusters would have been automatically three-connected along all three edges forming a columnar species of stacked tricapped trigonal prisms elongated presumably along all three edges. Such a structure might be accessible judging from the fact that already known are examples of clusters coupled along one and two edges and based on the specifics of the HOMO as discussed in the thesis. The fact that germanium nine-atom clusters can connect to each other by covalent exo-bonds indicated that perhaps clusters can bond to other groups. Thus, for the first time, deltahedral Zintl ions were functionalized with various exo bonded groups such as Ph2Bi- and Ph2Sb-. The first organically-functionalized deltahedral Zintl ion [Ph-Ge9-SbPh2]2- was also synthesized and structurally characterized. At that stage it was unclear how the reactions proceeded and what the reacting species might be. The detailed studies suggested that the reaction proceeds via formation of negatively-charged nucleophiles Ph2Bi- and Ph2Sb- from Ph3Bi and Ph3Sb, respectively, and the solvated free electrons that are released from the nine-atom clusters.

High-energy cluster orbitals with

significant distribution outside the cluster are available for the nucleophilic attack and formation of bonds.

Different products are formed depending upon the

available nucleophile in the immediate vicinity of the cluster. These general ideas

115

were explored further for possible use in more rational synthesis of similar derivatives of deltahedral Zintl ions by addition of other groups. Both mono- and di-substituted germanium clusters with –SnPh3, –SnMe3 and –GePh3 were obtained and characterized by such synthesis. Later, t-butyl groups were also attached to dimers of nine-atom clusters. This was the first successful designed reaction between such clusters and organic substituents. We speculate that the addition of organic species occurs via radical mechanism. This knowledge opens doors for further exploration of many other nucleophiles and radicals as potential substituents in such clusters. I will mention that the compounds of the elements from group 13 and 16 are not explored at that stage. The ranges of the anions and the radicals in organic chemistry are tremendous. Species with two or more active centers can be used to attach two or more cluster to such species. In other words, the achieved results can be used as an initial and a basic step for the synthesis of new compounds and materials.

116

REFERENCES

[1]

Adolphson, D. G.;Corbett, J. D.;Merryman, D. J. J. Am. Chem. Soc. 1976, 98, 7234.

[2]

Becke, A. D. Physical Review A 1988, 38, 3098.

[3]

Becke, A. D. Journal of Chemical Physics 1996, 104, 1040.

[4]

Belin, C.;Mercier, H.;Angilella, V. New Journal of Chemistry 1991, 15, 931.

[5]

Belin, C. H. E.;Corbett, J. D.;Cisar, A. J. Am. Chem. Soc. 1977, 99, 7163.

[6]

Bobev, S.;Sevov, S. C. Angew. Chem. Int. Edit. 2000, 39, 4108.

[7]

Bochkarev, M. N.;Razuvaev, G. A.;Zakharov, L. N.;Struchkov, Y. T. J. Organomet. Chem. 1980, 199, 205.

[8]

Burns, R. C.;Corbett, J. D. J. Am. Chem. Soc. 1982, 104, 2804.

[9]

Campbell, J.;Schrobilgen, G. J. Inorg. Chem. 1997, 36, 4078.

[10]

Corbett, J. D. Chem. Rev. 1985, 85, 383.

[11]

Corbett, J. D. Struct. Bond. 1997, 87, 157.

[12]

Corbett, J. D.;Adolphson, D. G.;Merryman, D. J.;Edwards, P. A.;Armatis, F. J. J. Am. Chem. Soc. 1975, 97, 6267.

117

[13]

Corbett, J. D.;Edwards, P. A. J. Chem. Soc., Chem. Commun. 1975, 984.

[14]

Corbett, J. D.;Edwards, P. A. J. Am. Chem. Soc. 1977, 99, 3313.

[15]

Corbett, J. D.;Rundle, R. E. Inorg. Chem. 1964, 3, 1408.

[16]

Dessy, R. E.;Kitching, W.;Chivers, T. J. Am. Chem. Soc. 1966, 88, 453.

[17]

Diehl, L.;Khodadadeh, K.;Kummer, D.;Strahle, J. Chem. Ber. 1976, 109, 3404.

[18]

Dillon, R. E. A.;Shriver, D. F. Chem. Mater. 1999, 11, 3296.

[19]

Doak, G. O.;Freedman, L. D., Organometallic Compounds of Arsenic, Antimony, and Bismuth, Wiley, New York, 1970.

[20]

Downie, C.;Mao, J. G.;Parmar, H.;Guloy, A. M. Inorg. Chem. 2004, 43, 1992.

[21]

Downie, C.;Tang, Z. J.;Guloy, A. M. Angew. Chem., Int. Edit. 2000, 39, 338.

[22]

Edwards, P. A.;Corbett, J. D. Inorg. Chem. 1977, 16, 903.

[23]

Eichhorn, B. W.;Haushalter, R. C. J. Chem. Soc., Chem. Commun. 1990, 937.

[24]

Eichhorn, B. W.;Haushalter, R. C.;Pennington, W. T. J. Am. Chem. Soc. 1988, 110, 8704.

[25]

Fassler, T. F. Coord. Chem. Rev. 2001, 215, 347.

[26]

Fassler, T. F.;Hoffmann, R. Angew. Chem., Int. Edit. 1999, 38, 543.

[27]

Fassler, T. F.;Hoffmann, R. J. Chem. Soc., Dalton 1999, 3339.

118

[28]

Fassler, T. F.;Hunziker, M. Inorg. Chem. 1994, 33, 5380.

[29]

Fassler, T. F.;Hunziker, M. Z. Anorg. Allg. Chem. 1996, 622, 837.

[30]

Fassler, T. F.;Hunziker, M.;Spahr, M. E.;Lueken, H.;Schilder, H. Z. Anorg. Allg. Chem. 2000, 626, 692.

[31]

Fassler, T. F.;Muhr, H. J.;Hunziker, M. Eur. J. Inorg. Chem. 1998, 1433.

[32]

Fassler, T. F.;Schutz, U. Inorg. Chem. 1999, 38, 1866.

[33]

Friedman, R. M.;Corbett, J. D. Inorg. Chem. 1973, 12, 1134.

[34]

Gardner, D. R.;Fettinger, J. C.;Eichhorn, B. W. Angew. Chem., Int. Edit. 1996, 35, 2852.

[35]

Guggenberger, L. J. Inorg. Chem. 1968, 7, 2260.

[36]

Hall, M. B.;Fenske, R. F. Inorg. Chem. 1972, 11, 768.

[37]

Hauptmann, R.;Fassler, T. F. Z. Anorg. Allg. Chem. 2002, 628, 1500.

[38]

Hoffmann, R. J. Chem. Phys. 1963, 39, 1397.

[39]

Huang, R. H.;Eglin, J. L.;Huang, S. Z.;Mcmills, L. E. H.;Dye, J. L. J. Am. Chem. Soc. 1993, 115, 9542.

[40]

Joannis, A. C. R. Hebd. Seances Acad. Sci 1891, 113, 795.

[41]

Joannis, A. C. Ann. Chi. Phys. 1906, 7, 75.

[42]

Jones, P. G.;Blaschette, A.;Henschel, D.;Weitze, A. Z. Kristallogr. 1995, 210, 377.

[43]

Karipides, A.;Haller, D. A. Acta Crystallogr. 1972, B28, 2889.

119

[44]

Kesanli, B.;Fettinger, J.;Eichhorn, B. Chem. Eur. J. 2001, 7, 5277.

[45]

Kesanli, B.;Fettinger, J.;Gardner, D. R.;Eichhorn, B. J. Am. Chem. Soc. 2002, 124, 4779.

[46]

Kraus, C. A. J. Am. Chem. Soc. 1907, 29, 1557.

[47]

Kraus, C. A. J. Am. Chem. Soc. 1908, 30, 1323.

[48]

Kraus, C. A. J. Am. Chem. Soc. 1922, 44, 1216.

[49]

Kraus, C. A. Trans. Am. Electrochem. Soc. 1924, 45, 175.

[50]

Kummer, D.;Diehl, L. Angew. Chem., Int. Edit. 1970, 9, 895.

[51]

Laves, F. Naturwissenschaften 1941, 29, 244.

[52]

Lohr, L. L. Inorg. Chem. 1981, 20, 4229.

[53]

Lok, M. T.;Tehan, F. J.;Dye, J. L. J. Phys. Chem. 1972, 72, 2975.

[54]

Martin, T. P. Angew. Chem., Int. Edit. 1986, 25, 197.

[55]

Pankratov, L. V.;Zakharov, L. N.;Bochkova, L. I.;Fukin, G. K.;Struchkov, Y. T. Izv. Akad. Nauk SSSR, Ser. Khim. 1994, 921.

[56]

Pannell, K. H.;Parkanyi, L.;Sharma, H.;Cervanteslee, F. Inorg. Chem. 1992, 31, 522.

[57]

Pauling, L., Cornell University Press, Ithaca, 1960.

[58]

Queneau, V.;Sevov, S. C. Angew. Chem., Int. Edit. 1997, 36, 1754.

[59]

Queneau, V.;Sevov, S. C. Inorg. Chem. 1998, 37, 1358.

120

[60]

Queneau, V.;Todorov, E.;Sevov, S. C. J. Am. Chem. Soc. 1998, 120, 3263.

[61]

Roller, S.;Drager, M.;Breunig, H. J.;Astes, M.;Gulec, S. J. Organomet. Chem. 1989, 378, 327.

[62]

Roller, S.;Simon, D.;Drager, M. J. Organomet. Chem. 1986, 301, 27.

[63]

Rudolph, R. W.;Wilson, W. L.;Parker, F.;Taylor, R. C.;Young, D. C. J. Am. Chem. Soc. 1978, 100, 4629.

[64]

Schnepf, A. Angew. Chem. Int. Edit. 2003, 42, 2624.

[65]

Somer, M.;Carrillo-Cabrera, W.;Peters, E. M.;Peters, K.;Kaupp, M.;von Schnering, H. G. Z. Anorg. Allg. Chem. 1999, 625, 37.

[66]

Somer, M.;Carrillo-Cabrera, W.;Peters, E. M.;Peters, K.;von Schnering, H. G. Z. Anorg. Allg. Chem. 1998, 624, 1915.

[67]

Tehan, F. J.;Barnett, B. L.;Dye, J. L. J. Am. Chem. Soc. 1974, 96, 7203.

[68]

Thompson, J. C., Electrons in liquid Ammonia, Clarendon Press, Oxford, 1976.

[69]

Todorov, E.;Sevov, S. C. Inorg. Chem. 1998, 37, 3889.

[70]

Ugrinov, A.;Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 10990.

[71]

Ugrinov, A.;Sevov, S. C. J. Am. Chem. Soc. 2002, 124, 2442.

[72]

Ugrinov, A.;Sevov, S. C. J. Am. Chem. Soc. 2003, 125, 14059.

[73]

Ugrinov, A.;Sevov, S. C. Inorg. Chem. 2003, 42, 5789.

[74]

Ugrinov, A.;Sevov, S. C. Chem. Eur. J. 2004, 10, 3727.

121

[75]

von Schnering, H. G.;Somer, M.;Kaupp, M.;Carrillo-Cabrera, W.;Baitinger, M.;Schmeding, A.;Grin, Y. Angew. Chem., Int. Edit. 1998, 37, 2359.

[76]

vonSchnering, H. G.;Baitinger, M.;Bolle, U.;CarrilloCabrera, W.;Curda, J.;Grin, Y.;Heinemann, F.;Llanos, J.;Peters, K.;Schmeding, A.;Somer, M. Z. Anorg. Allg. Chem. 1997, 623, 1037.

[77]

Wade, K. J. Chem. Soc. D 1971, 792.

[78]

Wade, K. Adv. Inorg. Chem. Radiochem. 1976, 18, 1.

[79]

Xu, L.;Sevov, S. C. J. Am. Chem. Soc. 1999, 121, 9245.

[80]

Xu, L.;Sevov, S. C. Inorg. Chem. 2000, 39, 5383.

[81]

Zintl, E. Angew. Chem. 1939, 52, 1.

[82]

Zintl, E.;Dullenkopf, W. Z. Phys. Chem., Abt. B 1932, 16, 183.

[83]

Zintl, E.;Goubeau, J.;Dullenkopf, W. Z. Phys. Chem., Abt. A 1931, 154, 1.

[84]

Zintl, E.;Harder, A. Phys. Chem., Abt. A 1931, 154, 47.

[85]

Zintl, E.;Kaiser, H. Z. Anorg. Allg. Chem. 1933, 211, 113.

122