ACS SYMPOSIUM SERIES 822 Group 13 Chemistry From Fundamentals to Applications Pamela J. Shapiro, Editor University of Idaho David A. Atwood, Editor University of Kentucky American Chemical Society, Washington, DC In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. Preface This volume is based on two symposia sponsored by the American Chemical Society (ACS) Division of Inorganic Chemistry, one at Pacifichem 2000 and the other at the 221st National ACS Meeting. The volume is different from traditional conference proceedings in that contributions were not limited to conference participants. Rather, a general invitation was sent out to chemists specializing in group 13 chemistry, and this collection of chapters represents the response to that invitation. As such, the content of this volume does not reflect any preconceived outline orchestrated by the editors. We had two objectives in organizing this volume. One was to offer a perspective on the present and future impact of group 13 chemistry on research in the 21st century. The other was to provide a resource for presenting modern main group chemistry to students at the college and graduate levels. Therefore, besides bringing the reader up to date on recent developments, some chapters offer in-depth reviews of their topic. The headings under which the chapters are organized (i.e., fundamentals of structures, bonding, and reactivity; organic synthesis and catalysis, new materials and clusters; and aluminum compounds: biological and environmental aspects) are a rough guide to help the reader navigate through the multitude of topics involving group 13 chemistry. Due to the increasing interdisciplinary nature of research, the different headings are not mutually exclusive and many of the chapters could have easily been listed under more than one heading. Because concepts of structure and bonding are the foundation upon which we build our understanding of chemistry, chapters dealing with these topics are presented first. In Chapters 1 and 2, Power and Jutzi describe the syntheses and properties of group 13 compounds in the +1 oxidation state, which is an unstable and rare oxidation state for the lighter group 13 elements B-Ga. These two chapters illustrate the synthetic ingenuity chemists use to prepare highly reactive species to uncover new modes of reactivity and bonding. Even compounds with B¬ -Ga in the more prevalent +3 oxidation state exhibit unusual and ix In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. interesting bonding modes, particularly when they are complexed by cyclopentadienyl ligands, as described by Shapiro in Chapter 3. Theories of bonding in electron-deficient molecules are the hallmark of group 13 compounds. These theories continue to evolve as new structures are discovered and as computational methods become more sophisticated. In Chapter 4, Fehlner develops a conceptual framework for explaining the structures of earlier transition metal-containing metalloboranes that are not easily accommodated by general electron-counting rules. The next section examines modern applications of group 13 compounds to organic synthesis and catalysis. In Chapter 5, Carter et al. describe their use of diboranes for the stereoselective coupling of diarylaldimines and the selective deoxygenation of non-metal oxides and organic carbonyl compounds. The Lewis-acidity of group 13 compounds is key to the acceleration of various organic reactions and the activation of transition metals for catalysis. Efforts to "juice-up" the Lewis acidity of these compounds by having multiple Lewis-acidic group 13 centers act in concert is a recent focus of several research groups. The number of contributions about multidentate Lewis acids, from Eisch, Jàkle, and Gabbaï, demonstrates the growing importance of this relatively new field. In Chapter 9, Harvey and Atwood review chelated aluminum anions, which are finding application as counter anions in catalysis. Recent reports of ethylene polymerization by aluminum compounds have caused a flurry of prospecting for aluminum compounds that could be the next generation of alkene polymerization catalysts. Is ethylene polymerization at alumi-num centers too good to be true? Budzelaar and Talarico take a theo-retical approach to examining this question in Chapter 10. Group 13 elements are components of many technologically important materials, such as heat-resistant ceramics, zeolites, and the semiconductors used in transistors and light-emitting devices. Current efforts in this area are at the molecular and materials interface— developing molecular precursors and methods for the controlled growth of nanostructural materials and materials with well-defined structures and compositions. In Chapters 11-13, Schnôckel, Sneddon, and Mason describe new types of fundamentally interesting and technologically relevant clusters, networks, and materials. In Chapter 14, Chivers reveals the unique oligomeric structures that result when trisamido derivatives of the group 13 elements are made anionic. In Chapter 15, King et al. use χ In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. density functional theory to understand the distorted structures of hypoelectronic six-atom clusters of B, In, and Tl. The final section of this volume takes a look at the biological and environmental properties of aluminum compounds. Despite its relative abundance, aluminum does not appear to play a supportive role in biological systems. On the contrary, as discussed in Chapter 16 by Anitha et al., aluminum appears to be involved in the etiology Alzheimer's Disease and other neurological disorders in humans. Understanding the role of aluminum in these age-related diseases becomes that much more important as the age-expectancy of the human population increases. For this reason, it is also important to understand the complex aqueous chemistry of aluminum and how it is released into aquatic environments, and this is the topic of Chapter 17 by Wang et al. The potential benefits of aluminum to biology in the form of AlF- are presented in Chapters 4 18 and 19. Conley et al. outline the fundamental chemistry of AlF¬ 4 -whereas Strunecka reveals how this anion behaves in biological systems. We are grateful to the contributing authors for the time and care they have given to this project. The diversity of topics reveals how widespread the influence of group 13 element chemistry is in science and technology. We also thank the Petroleum Research Fund, administered by the ACS, and the ACS Division of Inorganic Chemistry for their generous support of the two symposia from which this volume issued as well as the ACS Books Department for recognizing the timeliness of such a volume. Pamela J. Shapiro Department of Chemistry University of Idaho Moscow, ID 83844-2343 Email: [email protected] Telephone: 208-885-5785 Fax; 208-885-6173 David A. Atwood Department of Chemistry The University of Kentucky Lexington, KY 40506-0055 Email: [email protected] xi In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. Chapter 1 Bonding and Reactivity of a β-Diketiminate, Gallium(I), Carbene Analogue Ned J. Hardman, Andrew D. Phillips, and Philip P. Power* Department of Chemistry, University of California at Davis, One Shields Avenue, Davis, CA 95616 The gallium(I) diketiminate, :Ga[{N(Dipp)C(Me)}CH] (Dipp 2 = CH-2,6-i-Pr), is a carbene analogue featuring a planar 6 3 2 GaNC ring structure in which the gallium valence shell 2 3 contains two bond pairs, a lone pair and an empty orbital perpendicular to the ring plane. Computational studies for the model species :Ga[{N(H)C(H)}CH] showed that, although 2 the HOMO corresponds to the gallium lone pair, the LUMO does not involve the gallium p-orbital. The latter is represented by the LUMO+1 level, and this is separated from the HOMO by ca. 110 kcal mol-1. The low energy of the HOMO and its large energy separation from the acceptor orbital point to good σ-donor but poor π-acceptor characteristics for the :Ga[{N(Dipp)C(Me)}CH] ligand. 2 Calculations on the aluminum analogue showed that the HOMO-LUMO+1 separation is ca. 12 kcal mol-1 less than the gallium species which is due primarily to the higher energy of the aluminum lone pair. The reaction of :Ga[{N(Dipp)C(Me)}CH] with NO or S produced the novel 2 2 8 μ-chalcogenides (EGa[{N(Dipp)C(Me)}CH]) (Ε = Ο or S). 2 2 However, the reaction with NSiMe afforded the isomeric 3 3 tetrazole [HC{C(Me)N(Dipp)}]GaN(SiMe)NN(SiMe) 2 3 2 3 2 © 2002 American Chemical Society In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. 3 and amide/azide [HC{C(Me)N(Dipp)}]Ga(N)N(SiMe)2. In 2 3 3 contrast, the reaction with the bulkier azide NAr* (Ar* = 3 CH-2,6-Trip, Trip = CH-2,4,6-i-Pr) yielded the first 6 3 2 6 2 3 monomeric gallium imide [H{C(Me)N(Dipp)}]Ga=NAr*. 2 Computational data on model complexes for the latter showed that the π-bonding component in the Ga-N imide bond is 48.6 kJ mol-1 which is similar to the value calculated for HGaNH. 2 2 The relative weakness of the Ga-N π-bonding is consistent with the strong σ-donor/weak π-acceptor bonding character expected from the large HOMO-LUMO+1 gap in the model gallium(I) diketiminate precursor. Introduction The sterically crowded A1(I) and Ga(I) β-diketiminates :M(Dippnacnac), 2 ie. :M[{N(Dipp)C(Me)}CH], (M = Al, (/), or Ga, (2), Dipp = CH-2,6-*-Pr), 2 6 3 2 possess monomeric structures (see Figure 1) in which the metal valence shells contain two bond pairs, a lone pair and an empty p-orbital. Stable, neutral, monomers of this general type, ie. :MR (R = organic or related ligand, M = Al, Ga, In or Tl), are rare in the solid state, particularly in the case of Al and Ga. Structurally characterized examples include :M(TpBu 2) (M = Ga, (3), or In, («0; TpBu'2 = tris(3,5-di-tert-butylpyrazolyl)hydroborato), :In(Tpph), (5), :In(TpBu), (4), :In(Tp(CF)), (5), several Tl derivatives of the Tp ligands, (7), some Tl 3 2 derivatives of bis(pyrazolyl)hydroborato (Bp) ligands, in which Tl is bound to two nitrogens and the B-H hydrogen of the Bp ligand, (8), as well as the one- coordinate terphenyl derivatives :MCH-2,6-Trip (M = In, (10), orTl, (//); 6 3 2 Trip = -C<>H-2,4,6-i-Pr). In addition the anionic gallium carbene analogue 2 3 [Ga{N(But)CH}]" has been isolated although its chemistry has not been 2 explored (P). The presence of a pair of electrons and formally unoccupied p- orbital(s) in low (1 or 2) coordinate :MR species suggest that they may behave as both electron donors and acceptors. However, the acceptor capacity of the metal in the Tp derivatives is to some extent compromised by the fact that the Tp ligand, being tridentate, occupies three coordination sites and potentially engages three of the four available metal valence orbitals in bonding. Even for the formally two-coordinate Bp derivatives, the hydrogen of the B-H moiety occupies a site in the metal coordination sphere (8,9). The :M(Dippnacnac) (M 2 = Al or Ga) and MCH-2,6-Trip (M = In orTl) species differ in that they 6 3 2 contain only bi or monodentate ligands, in which case either one or two metal p- orbitals are formally available for bonding. They resemble singlet carbenes (72,73), both electronically and structurally, and are expected to have some In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. 4 similarity in their reactivity patterns. However, important differences can also be anticipated which may arise from several factors. For example, the properties of the central aluminum or gallium and carbon atoms differ markedly in terms of their electronegativities and sizes. In addition, the energy difference between lone pair and empty p-orbital(s) of the Al or Ga derivatives is expected to be larger than that in carbenes, and this difference should exert a strong influence on the bonding characteristics of these ligands. In this article, we focus on the structure, physical properties and some reaction chemistry of the two coordinate gallium species :Ga(Dippnacnac) (2) and show that it possesses some unique 2 and unexpected bonding and chemical characteristics. Structure and Bonding The structure of the :Ga(Dipp»2nacnac) is shown in Figure 1. The GaNC3 2 ring has planar geometry with an approximate C symmetry axis along the 2 Figure L Drawing of :Ga(Dippnacnac) illustrating the planar structure of the 2 β-diketiminate ring. Ga(l)—C(2) vector which indicates derealization of the C-C and C-N double bonds over the NC ring array. Some important structural parameters for this 2 3 compound are compared with those of its aluminum analogue in Table I. Calculated structural parameters for the unsubstituted model species :M(nacnac) (M = Al or Ga) are also provided in parenthesis (14). These were calculated using a hybrid method (15) of Hartree-Fock and density functional theories (DFT) with Beckefs exchange functional (B3), (16), and the Lee-Yang-Parr non­ local correlation functional (LYP), (17). A basis set of 6-31G was employed for all atoms. Calculations were carried out using the Gaussian 98 suite of programs (18). Both molecules represent energy minima and were optimized as In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. 5 having C symmetry. It is obvious from the data in Table I that there is good 2v agreement between the calculated and experimental results. The major Table I. Comparison of Experimental and Calculated Structural Data for :M(Dippnacnac) and :M(nacnac) (M = Al or Ga) 2 2 : Al(Dippnacnac)a : Ga(Dippnacnac)b 2 2 M-N(A) 1.957(2) (1.992)c 2.055(2) (2.063)c N-C(A)d 1.341(3) (1.329) 1.338(2) (1.326) C-C(A)d 1.391(3) (1.396) 1.400(3) (1.397) N-M-N° 89.86(8) (86.3) 87.53(5) (85.2) aRef. 1, bRef. 2, Calculated values for :M(nacnac), Ref. 14, dMNC ring. 2 3 difference between the two compounds involves the M-N bond lengths which were experimentally found to be ca. 0Λλ longer for the gallium compound. In low coordinate Al-N and Ga-N compounds, longer distances are usually observed in the case of gallium (19). However, in these compounds the difference is twice that normally observed. Both the Al-N and Ga-N distances are considerably longer (ca. 0.1Â) than those reported (20) for Al(III) and Ga(III), β-diketiminates and this is consistent with the larger effective ionic radii expected for the M(I) ions. The computational data provide further insight into the bonding through the ordering of various energy levels. These levels and their energies are listed in Table II. An illustration of the electron density surfaces for the :Ga(pacnac) species is given in Figure 2. It can be seen from Table II that the order and energies of the MO's are similar for the Al and Ga derivatives. In each compound, the HOMO is the lone pair orbital, and it is primarily s in character. However, the LUMO is not the Al or Ga valence p- orbital but a π* orbital localized at the C-N moieties. The LUMO+1 level corresponds to the metal p-orbital with a small component localized on the organic ring. The HOMO-LUMO and HOMO-LUMO+1 energy differences are significantly larger in gallium where the HOMO-LUMO+1 gap is calculated to be 109.81 kcal mol"1 (ca. 4.76 eV). It is notable that the HOMO also has σ*- antibonding character with respect to the M-N bond. The HOMO-1 has M-N π-bonding as well as some C-C π-bonding character. The HOMO-2 orbital involves M-N σ-bonding as well as C-C and C-H σ-bonding while the HOMO- 4 is a ring π-orbital with slight derealization onto the metal. In sum the HOMO-1 (π) and ΗΟΜΟ-2(σ) are the orbitals that are mainly In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. 6 Table II. Molecular Orbital Energies (kcal moll) for the Model Species :M(nacnac) (M = Al or Ga) at 6-31G Level8 Orbital Al Ga Orbital Character LUMO+1 -8.16 -10.43 Ga(4p) ΝΒ, C-C π* LUMO -29.73 -24.00 π* C-N HOMO -105.49 -120.24 Ga(4s-p) LP, Ga-N σ* HOMO-1 -142.57 -136.57 π Ga-N, C-N HOMO-2 -200.58 -190.94 σ* Ga-N, C-N, CH HOMO-3 -208.38 -203.80 π C-N HOMO-4 -242.85 -237.34 π Ga-N, C-N, C-C HOMO-5 -255.93 -254.07 σ C-H, C-N, C-C Δ HOMO-LUMO 75.76 96.24 Δ HOMO-LUMO+1 97.33 109.81 aPhillips, A, D., Power, P. P. Unpublished results. In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002. 7 Figure 2. Electron density surfaces and MO energies of :Ga(nacnac). associated with M-N bonding. Also, there is some M-N σ-antibonding character associated with the HOMO and some M-N π-bonding character from HOMO-4, but, to a first approximation, these effects cancel each other so that M-N bonding corresponds to the delocalized HOMO-1 and HOMO-2 orbitals. Even though HOMO-1 is a π-orbital which also extends to the carbon atoms of the ring, it does not give rise to short M-N bonds since it does not confer multiple character on the M-N interaction. In addition we note that the σ M-N bonding HOMO-2 orbital involves a considerable component localized on other atoms. Thus, the computational data suggest weakened Ga-N bonding which is in agreement with the relatively long Ga-N bonds seen in the crystal structure. Both the computational and structural data are also in harmony with Haaland's analysis (21) of related Al-N bonded compounds which regards the M-N bonds as having equal components of normal and dative character. The low energy and high s-character of the HOMO, and the large energy difference between the HOMO and the rather diffuse acceptor 4p-orbital (LUMO+1) suggest that :Ga(Dippnacnac) should be a good σ-donor but a poor π-acceptor. The π*- 2 LUMO orbital, which is ca. 14 kcal mof1 more stable than the LUMO+1 level, In Group 13 Chemistry; Shapiro, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.