ARTICLES PUBLISHED ONLINE: 16 FEBRUARY 2014 | DOI: 10.1038/NCHEM.1870
Stable GaX2 , InX2 and TlX2 radicals Andrey V. Protchenko1, Deepak Dange2, Jeffrey R. Harmer3,4, Christina Y. Tang1, Andrew D. Schwarz1, Michael J. Kelly1, Nicholas Phillips1, Remi Tirfoin1, Krishna Hassomal Birjkumar5, Cameron Jones2, Nikolas Kaltsoyannis5, Philip Mountford1 and Simon Aldridge1 * The chemistry of the Group 13 metals is dominated by the 11 and 13 oxidation states, and simple monomeric M(II) species are typically short-lived, highly reactive species. Here we report the first thermally robust monomeric MX2 radicals of gallium, indium and thallium. By making use of sterically demanding boryl substituents, compounds of the type M(II)(boryl)2 (M 5 Ga, In, Tl) can be synthesized. These decompose above 130 8 C and are amenable to structural characterization in the solid state by X-ray crystallography. Electron paramagnetic resonance and computational studies reveal a dominant metal-centred character for all three radicals (>70% spin density at the metal). M(II) species have been invoked as key short-lived intermediates in well-known electron-transfer processes; consistently, the chemical behaviour of these novel isolated species reveals facile one-electron shuttling processes at the metal centre.
T
he metals of Group 13 differ markedly in their terrestrial distribution and the scale of their technological exploitation, but the chemistry of all four elements—aluminium, gallium, indium and thallium—is dominated by two oxidation states, þ1 and þ3 (ref. 1). Therefore, compounds of the composition MX2 are typically mixed-valence species (for example, M(I)M(III)X4), and those that are not adopt diamagnetic M2X4 structures that feature M–M bonds2. Although mechanistically important, hitherto isolated examples of simple MX2 radical species that feature the metal in the þ2 oxidation state have been confined to trapping experiments at cryogenic temperatures2,3; controversial claims of room-temperature stable Ga(II) species, for example, have subsequently been shown to be erroneous4. By contrast, here we report the first thermally robust monomeric MX2 species of gallium, indium and thallium. By making use of sterically demanding boryl substituents, compounds of the type M(II)(boryl)2 (M ¼ Ga, In, Tl) that decompose above 130 8C can be synthesized. Electron paramagnetic resonance (EPR) and computational studies reveal a dominant metal-centred character for all three radicals (.70% spin density at the metal). M(II) species have been invoked as key short-lived intermediates in well-known electrontransfer processes5–7; consistently, the chemical behaviour of these novel isolated species reveals facile one-electron shuttling processes at the metal centre.
Results and discussion Syntheses. In recent work the lithium reagent (thf)2Li[B(NDippCH)2] (1, where Dipp ¼ 2,6-iPr2C6H3 and thf ¼ tetrahydrofuran) was utilized as a nucleophilic source of the boryl fragment in the synthesis of novel E–B chemical bonds8,9. In particular, the competence of this boryl substituent in stabilizing low-valent, low-coordinate main-group systems has been demonstrated through the synthesis of the Si(II) species Si[B(NDippCH)2][N(Dipp)SiMe3] (ref. 10). Herein, the reactivity of 1 towards M(I) precursors (M ¼ Ga, In or Tl) is shown to lead to the formation of the respective metal bis(boryl) complexes M[B(NDippCH)2]2 (M ¼ Ga (2-Ga), In (2-In) or Tl (2-Tl), Fig. 1) in yields of
38–75%. In the cases of the precursors TlCl, In[N(SiMe3)Dipp*] or Ga[N(SiMe3)Dipp*] (where Dipp* ¼ 2,6-(Ph2CH)2-4-Me-C6H2), apparent disproportionation of the putative M(I) intermediate M[B(NDippCH)2] is implied by the precipitation of elemental metal. In the case of Tl[N(SiMe3)2], however, the reaction stoichiometry can be evaluated explicitly through the isolation at 230 8C of the co-product Tl8[B(NDippCH)2]4 , a mixed-valence cluster that contains four Tl(0) and four Tl(I) centres (see Supplementary Fig. 1)11. Synthetic routes were also established from M(III) precursors; the reactions of ClIn[B(NDippCH)2]2 (3-In) with strong reducing agents provide alternative routes to the bis(boryl) system 2-In. Good yields (65–70%) of 2-In are obtained using Sm(h5-C5Me5)2(thf), but the use of potassium metal brings with it the possibility of over-reduction, which leads to the accompanying formation of aggregated systems that feature a lower average metal-oxidation state, such as In2[B(NDippCH)2]3 (4-In). Solid-state structures. All three bis(boryl) compounds, 2-Ga, 2-In and 2-Tl, can be isolated as analytically pure, thermally robust crystalline materials that decompose at 132–134 8C, 196–198 8C and 225–230 8C, respectively. The molecular structure of each compound in the solid state was determined to be monomeric by X-ray crystallography (see Fig. 2 and Supplementary Table 1). This contrasts markedly with previously reported M(II) compounds of the Group 13 metals, which feature multimetallic structures2; presumably, dimerization of 2-Ga, 2-In or 2-Tl to give an M–M-bonded M2X4 framework akin to related alkyl and silyl derivatives is prevented by the extremely sterically demanding nature of the boryl substituents. Accordingly, comparison of the structure of 2-Ga with those of related systems of the type XGa[B(NDippCH)2]2 (X ¼ Cl (3-Ga) and X ¼ Ga[B(NDippCH)2] (4-Ga), see Supplementary Fig. 2) reveals, as the bulk of the X substituent increases: (1) significant narrowing of the GaB2 angle (156.0(1)8, 152.0(1)8 and 146.4(1)8 for 2-Ga, 3-Ga and 4-Ga, respectively) and (2) lengthening of the Ga–B bonds within the GaB2 unit (mean distances 2.048(2) Å, 2.078(2) Å and 2.094(3) Å, respectively).
1
Inorganic Chemistry Laboratory, Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QR, UK, 2 School of Chemistry, PO Box 23, Monash University, Melbourne, Victoria 3800, Australia, 3 Centre for Advanced Electron Spin Resonance, Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QR, 4 Center for Advanced Imaging, University of Queensland, St Lucia, Queensland 4072, Australia, 5 Department of Chemistry, University College London, Christopher Ingold Laboratories, 20 Gordon Street, London WC1H 0AJ, UK. * e-mail: [email protected] NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
315
ARTICLES [M(NRR´)] n
i
or
NATURE CHEMISTRY
Dipp
Dipp
N
N
B
M
N
Dipp M = Ga, In, Tl
ii
B
N
TlCl
Cl
Dipp N
Dipp
M
B
N
N
Dipp
N
B
Dipp
Dipp
Dipp = 2,6- i Pr2C6H3
Dipp = 2,6- i Pr2C6H3
M = Ga: 2-Ga M = In: 2-In M = Tl: 2-Tl
M = In: 3-In
Figure 1 | Syntheses of M(II) complexes that feature a supporting bis(boryl) ligand set from either M(I) or M(III) precursors. i, For 2-Ga from Ga[N(SiMe3)Dipp*], (thf)2Li[B(NDippCH)2] (1, 2.0 equiv.), toluene, 278 8C to 0 8C over five hours, then 12-crown-4 (2.0 equiv.), room temperature, one hour, 64%; for 2-In from In[N(SiMe3)Dipp*], 1 (2.0 equiv.), toluene, 278 8C to 0 8C over five hours, then 12-crown-4 (2.0 equiv.), room temperature, one hour, 38%; for 2-Tl from Tl[N(SiMe3)2], 1 (1.0 equiv.), hexane, 245 8C, five minutes, 52%; for 2-Tl from TlCl, 1 (1.0 equiv.), hexane, 278 8C to room temperature, 12 hours, 75%. ii, For 2-In from 3-In, K metal (0.94 equiv.), benzene-d6/thf (12:1), room temperature, sonication for two hours, 75% (by NMR spectroscopy), or Sm(h5-C5Me5)2(thf) (1.0 equiv.), benzene, room temperature, 10 minutes, 67% (isolated). The syntheses of 2-Ga and 2-In can also be accomplished from M(I) precursors and 1 (1 equiv.) without the use of 12-crown-4, although slightly lower yields and greater contamination with HB(NDippCH)2 are observed. R and R′ are aryl or silyl groups.
The molecular structures of 2-Ga and 2-In differ notably from that (linear) of 2-Tl in that the B–M–B framework for the lighter congeners is bent (B–M–B angles of 156.0(1)8 and 145.4(1)8 versus 177.6(2)8, respectively). The presence of an additional metal-bound hydrogen atom (although not easily discounted on the basis of X-ray diffraction data) is at odds with spectroscopic and magnetic data, which indicate that all three molecules possess an unpaired electron. The observed trend in the B–M–B angles is consistent with that predicted by density functional theory (DFT) (160.38, 152.78 and 178.88, respectively (Supplementary Tables 2–5)), and with calculated metal ns orbital contributions to the respective singularly occupied molecular orbitals (SOMOs), which increase in the order Tl , Ga , In (2-Tl, ,1.0% 6s, 65.2% 6p; 2-Ga, 3.0% 4s, 62.7% 4p; 2-In, 5.9% 5s, 60.3% 5p). By analogy with the classical Walsh diagram for linear/bent dihydrides12, greater SOMO s character for a system of five valence electrons is linked intrinsically to increased bending at the metal, and leads to a lower SOMO energy (2-Tl, 22.740 eV; 2-Ga, 23.069 eV; 2-In, 23.113 eV). Moreover, a
N(1)
Ga(1) B(2)
N(4)
the relative degrees of orbital mixing for the three metal systems and hence the observed deviations from linearity can readily be rationalized in terms of the respective atomic ns/np energy separations (Tl, 7.109 eV; Ga, 6.319 eV; In, 5.652 eV, as calculated by DFT). The largest ns/np gap is thus associated with the most linear molecule (that is, 2-Tl). All three compounds are red– orange in colour, and the respective ultraviolet/visible (UV-vis) spectra reveal bands at 350 nm that tail off into the visible region (Supplementary Fig. 8). In the case of 2-Ga, the two strongest bands (at l ¼ 362 nm and 309 nm) are thought, on the basis of time-dependent DFT calculations, to feature significant contributions from transitions that originate in the SOMO (Supplementary Fig. 9 and Supplementary Tables 7 and 8). Solution-phase characterization. The paramagnetic nature of 2-Ga, 2-In and 2-Tl implied by their structures in the solid state can be shown spectroscopically to be retained in solution. Thus, broadened 1H NMR resonances for the boryl ligand and magnetic moments determined by the Evans method (meff at 293 K) of 2.0 (+0.1) mB , 2.0 (+0.1) mB and 1.2 (+0.1) mB , respectively, are consistent with the presence of an unpaired electron for each species. The magnetic moment measured for 2-Tl is reproducibly smaller than that expected on the basis of the spin-only formula (one unpaired electron, 1.73 mB), consistent with the linear geometry at thallium and with the occupation of a p-orbital manifold with high Tl 6p character (DFT-calculated SOMO, 65.2% Tl 6p; a SOMO þ 1, 86.0% Tl 6p). Related radicals featuring a non-zero orbital component that arises from a single electron occupying degenerate p orbitals (such as NO) have a (diamagnetic) j ¼ 1/2 ground state, which causes meff to approach zero at low values of kT/l (ref. 13). In the case of 2-Tl, the much larger magnitude of the spin-orbit coupling parameter l (ref. 14) implies near-exclusive occupancy of the j ¼ 1/2 state, and the fact that meff = 0 presumably reflects that the SOMO and SOMO þ 1 are not actually degenerate in the minimum energy calculated structure (DE ¼ 0.458 eV). No evidence for the formation in solution of a second (diamagnetic) component (for example, dimeric [(HCNDipp)2B]2Tl–Tl[B(NDippCH)2]2) was revealed by multinuclear variable-temperature NMR spectroscopy. X- and W-band field-sweep EPR spectra measured in frozen pentane/hexane solution (1:1) show that all three species have g values that deviate significantly from the free-electron value, and exhibit very large hyperfine couplings to their respective metals (Fig. 3 and Supplementary Table 6). The corresponding couplings (Ax , Ay , Az) measured for the matrix-isolated GaH2 radical are larger (1,957 MHz, 1,917 MHz and 2,379 MHz, respectively), which presumably reflects c
b
N(3)
N(4)
N(3)
N(2)
In(1) B(2)
B(1)
DOI: 10.1038/NCHEM.1870
N(3)
N(1)
Tl(1)
B(1) N(1)
N(2)
B(2) N(4)
B(1) N(2)
Figure 2 | Molecular structures. a, 2-Ga. b, 2-In. c, 2-Tl. Ellipsoids are at the 50% probability level, carbon-bound hydrogen atoms are omitted and iPr groups are shown in a wireframe format for clarity. Selected bond lengths: 2-Ga, Ga(1)–B(1) 2.045(2) Å, Ga(1)–B(2) 2.051(2) Å; 2-In, In(1)–B(1) 2.246(3) Å, In(1)–B(2) 2.242(3) Å; 2-Tl, Tl(1)–B(1) 2.173(6) Å, Tl(1)–B(2) 2.167(5) Å. 316
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1870
a
b Simulation Ga, experimental Simulation Ga, experimental Simulation In, experimental
Simulation In, experimental
Simulation
Simulation
Tl, experimental Tl, experimental
0
250
500
750
1,000
1,250
1,500
3,000
3,500
4,000
B0 (mT)
4,500
5,000
5,500
6,000
B0 (mT)
Figure 3 | Field-sweep EPR spectra for 2-Ga, 2-In and 2-Tl measured in frozen pentane/hexane solution (1:1), along with the corresponding simulations. a, X-band continuous-wave EPR spectra. b, W-band echo-detected EPR spectra, first-derivative representation. For 2-Tl the signal starts at 4.5 T (g1 ¼ 1.23) and could only be recorded to 6 T (the upper limit of our magnet). The principal g values gi and metal hyperfine couplings Ai derived from the simulation are as follows (the corresponding values calculated using DFT are italicized and in parentheses). 2-Ga, g1 , g2 , g3 ¼ 1.881, 1.998, 2.014 (1.8760, 1.9960, 2.0113); A(69Ga) ¼ 444, 1,032, 533 ¼ 670 þ (2226, 362, 2137) MHz (214 þ (2185, 312, 2127) MHz). 2-In, g1 , g2 , g3 ¼ 1.719, 1.954, 1.999 (1.7338, 1.9689, 2.0125); A(115In) ¼ 913, 1,936, 793 ¼ 1,214 þ (2301, 722, 2421) MHz (645 þ (2408, 621, 2214) MHz). 2-Tl, g1 g2, g3 ¼ 0.6, 0.7, 1.23 (1.0047, 1.4381, 1.697); A(205Tl) ¼ 28,000, 10,300, 28,100 ¼ 21,933 þ (26,067, 12,233, 26,167) MHz (22866 þ (26,105, 7,592, 21,478) MHz).
reduced possibilities for spin delocalization in the dihydride system3. The unusually small g values for 2-Tl determined from the simulations were confirmed by an X-band nutation experiment (see Supplementary Fig. 4). Smaller hyperfine couplings unresolved in the field-sweep EPR spectra from the 11B and 14N nuclei of the [B(NDippCH)2] ligands were characterized by electron nuclear double resonance (ENDOR) spectroscopy and hyperfine sublevel correlation spectroscopy (HYSCORE; see Supplementary Figs 5 and 6). For 2-Ga and 2-In, small hyperfine couplings from both 11 B (20–42 MHz) and 14N (1–7 MHZ) nuclei were detected that, by comparison with tabulated 2s and 2p atomic orbital hyperfine couplings for the 11B and 14N nuclei, allow spin densities on the ligands of 0.25 to be calculated. The summed ligand-based spin densities from DFT calculations are 0.266, 0.273 and 0.277 for 2-Ga, 2-In and 2-Tl, respectively. In the case of 2-Tl, only small 14 N hyperfine couplings could be detected (3.5–6.5 MHz) via HYSCORE, and no 11B ENDOR data could be measured, presumably as a result of fast electron relaxation. The experimental principal g values and metal hyperfine couplings, along with those from DFT calculations, are summarized in Fig. 3 (see Supplementary Table 6 for a full list). There is good overall agreement between g values and the anisotropic part of the metal hyperfine couplings (the isotropic part is very dependent on spin polarization of the core metal orbitals and is more difficult to calculate), with both experimental and DFT-calculated g values decreasing and metal hyperfine couplings increasing across the series Ga, In, Tl. The EPR data leave no doubt that all three radicals are characterized best as metal centred (.70% spin density at the metal). Although stable paramagnetic metal-centred radicals are common among the elements of the d block, in recent years welldefined radicals have also started to emerge for the heavier p-block elements15–19. Despite this, group 13 metal-based radicals are still very rare20–31, which in part reflects the ease of dimerization of low-coordinate species in the absence of significant steric shielding. Although such processes can be subverted for charged systems, and † the anionic trisilylmetalate species [M(SiMetBu2)3] 2 (M ¼ Al, Ga) 31 have been reported , to date these remain the only other monometallic group 13 radicals to have been structurally characterized.
Redox chemistry. Implicit in the descriptions of 2-Ga, 2-In and 2-Tl as monomeric M(II) species is relatively facile redox chemistry that leads to the formation of complexes featuring the more familiar M(III) and M(I) oxidation states. Tl(II) systems, for example, have been implicated as intermediate species in twoelectron self-exchange processes and non-complementary redox reactions with transition-metal cations6,7. Accordingly, taking 2-Tl as the exemplar, single-electron oxidation and reduction chemistry to generate tractable Tl(III) cations and Tl(I) anions can be achieved both electrochemically and chemically (Fig. 4 and Supplementary Fig. 7). Moreover, the reaction of [2-Tl]þ with [2-Tl]2 can be shown to proceed through comproportionation to regenerate the neutral Tl(II) compound 2-Tl. The reactions of 2-Tl with borate salts of the trityl cation [CPh3]þ led to the formation of diamagnetic [2-Tl]þ, together with the head-to-tail Gomberg dimer Ph3C(C6H5)CPh2 of the trityl radical (ferrocenium oxidants give the same thallium-containing product). The stoichiometry of this oxidation, and the implied transfer of one electron per molecule of 2-Tl, can be established by integration of the 1H NMR signals of the products. In addition, the identity of the [2-Tl]þ cation (as the [B(C6H3(CF3)2-3,5)]2 salt) has been confirmed crystallographically (see Supplementary Fig. 3), with the linear coordination geometry at thallium being consistent with related four-valence-electron systems12. Conversely, reduction of 2-Tl with [K(18-crown-6)][C10H8] led to the formation of diamagnetic [K(18-crown-6)][Tl[B(NDippCH)2]2 ], the crystal structure of which reveals a bent TlB2 core (with the angle B–Tl–B ¼ 109.8(1)8) and elongated Tl–B bonds (2.454(4) Å and 2.456(4) Å) (Fig. 4). The bent geometry of [2-Tl]2 is consistent with the greater central-atom ns character in the highest occupied molecular orbital for a system of six valence electrons (10.4% Tl 6s, compared with ,1.0% Tl 6s for five-electron charge-neutral 2-Tl)32. Consequently, there is a greater Tl 6p character in molecular orbitals with Tl–B bonding character (compared with 2-Tl), and significant (.13%) elongation of the Tl–B bonds therefore accompanies the reduction process. In conclusion, we report the syntheses of divalent bis(boryl)gallium, -indium and -thallium radicals from either M(I) or M(III)
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
317
ARTICLES a
NATURE CHEMISTRY
[BAr4] – Dipp
[K(18-crown-6)] +
Dipp
N
+
Tl
B
Dipp
N
i
B
N
iii
N
Dipp
ii
2-Tl
iii
N B N Dipp
Dipp [2-Tl]+[BAr4]–
Dipp Tl
–
N B
[K(18-crown-6)]+[2-Tl]–
K(60)
Tl(1) N(4) B(2) N(3)
Figure 4 | Redox chemistry of 2-Tl. a, One-electron chemical oxidation and reduction processes. b, Molecular structure of [K(18-crown-6)][2-Tl] . OEt2 (ellipsoids are shown at the 50% probability level, diethyl ether solvated molecule, C-bound hydrogen atoms are omitted and iPr groups are shown in a wireframe format for clarity). Selected bond lengths and angles: Tl(1)–B(1) 2.454(4) Å, Tl(1)–B(2) 2.456(4) Å, B(1)–Tl(1)–B(2) 109.8(1)8. i, [Fe(h5C5H5)2][B(C6H3(CF3)2-3,5)4] (1.0 equiv.), diethyl ether, room temperature, ten minutes, 98%, or [CPh3][B(C6F5)4] (1.0 equiv.), benzene-d6 , room temperature, NMR scale. ii, [K(18-crown-6)][C10H8 ], (1.33 equiv.), THF, 2196 8C to room temperature, two minutes, 26%. iii, [2-Tl][B{C6H3(CF3)2-3,5}4] plus [K(18-crown-6)][2-Tl] (1:1), THF/benzene-d6 , room temperature, two minutes, quantitative.
precursors. These species have unprecedented thermal stability, and decompose, in each case, at temperatures above 130 8C. The monomeric structures of these systems (enforced by the steric bulk of the boryl substituents) in the solid state were confirmed by X-ray crystallography. The description of each of these as predominantly metal-centred radical species is consistent with the results of EPR and DFT studies.
Methods Manipulations were carried out under a dry, oxygen-free argon or dinitrogen atmosphere, with reagents dissolved or suspended in aprotic solvents, and combined or isolated using cannula and glove-box techniques. Compounds M[B(NDippCH)2]2 (M ¼ Ga (2-Ga), In (2-In) or Tl (2-Tl)) were synthesized, respectively, via the reactions of Ga[N(SiMe3)Dipp*], In[N(SiMe3)Dipp*] or one of Tl[N(SiMe3)2] or TlCl, with (thf )2Li[B(NDippCH)2] (1) in toluene or hexane solution, and isolated by low-temperature crystallization. 2-In was also synthesized by the reduction of ClIn[B(NDippCH)2]2 (3-In) with Sm(h5-C5Me5)2(thf ) in benzene. The Tl(I) species [K(18-crown-6)][2-Tl] was synthesized by the reactions of 2-Tl with [K(18-crown-6)][C10H8] in THF, and the Tl(III) species [2-Tl][B{C6H3(CF3)2-3,5}4] from the reaction of 2-Tl with [Fe(h5-C5H5)2][B(C6H3 (CF3)2-3,5)4] in diethyl ether. Reaction of isolated samples of [K(18-crown-6)][2-Tl] with [2-Tl][B{C6H3(CF3)2-3,5}4] in THF or benzene led to the quantitative regeneration of 2-Tl. We characterized new compounds by elemental analysis, mass spectrometry, UV/vis spectroscopy, cyclic voltammetry (for 2-Tl), multinuclear NMR spectroscopy (for diamagnetic compounds) and EPR spectroscopy (for paramagnetic compounds). Solution-phase magnetic moments were determined for 2-Ga, 2-In and 2-Tl using the Evans method, and the structures of 2-Ga, 2-In, 2-Tl, 3-Ga, 3-In, 4-Ga, 4-In, Tl8[B(NDippCH)2]4 , [K(18-crown-6)][2-Tl] and [2-Tl][B{C6H3(CF3)2-3,5}4] in the solid state were determined by single-crystal X-ray diffraction studies. X- and W-band field-sweep EPR spectra were measured in frozen pentane/hexane solution (1:1) for 2-Ga, 2-In and 2-Tl. Comparisons of the 318
Received 13 September 2013; accepted 13 January 2014; published online 16 February 2014
References
b
N(2)
11 B and 14N hyperfine couplings determined by ENDOR and HYSCORE with tabulated 2s and 2p atomic orbital hyperfine couplings for the 11B and 14N nuclei allowed the ligand (and hence (by difference) the metal-centred) spin densities to be calculated for each species. DFT, as implemented in the Amsterdam Density Functional code, was employed to calculate molecular geometries, orbital energies and compositions, spin densities, EPR parameters and electron-transition energies.
N Dipp
Ar = C6F5 or C6H3(CF3)2-3,5
N(1) B(1)
DOI: 10.1038/NCHEM.1870
1. Downs, A. J. & Himmel, H-J. The Group 13 Metals Aluminium, Gallium, Indium and Thallium: Chemical Patterns and Peculiarities (eds Aldridge, S. & Downs, A. J.) Ch. 1 (Wiley, 2011). 2. Uhl, W. & Layh, M. The Group 13 Metals Aluminium, Gallium, Indium and Thallium: Chemical Patterns and Peculiarities (eds Aldridge, S. & Downs, A. J.) Ch. 4 (Wiley, 2011). 3. Knight, L. B. Jr, Banisaukas, J. J. III, Babb, R. & Davidson, E. R. Electron spin resonance matrix isolation and ab initio theoretical investigations of 69,71 GaH2 , 69,71GaD2 , H69,71GaCH3 , and D69,71GaCD3. J. Chem. Phys. 105, 6607–6615 (1996). 4. Kaim, W. & Matheis, W. Bis(1,4-di-tert-butyl-1,4-diazabutadiene)gallium is not a gallium(II) compound. Chem. Commun. 597–598 (1991). 5. Pardoe, J. A. J. & Downs, A. J. Development of the chemistry of indium in formal oxidation states lower than þ3. Chem. Rev. 107, 2–45 (2007). 6. Khoshtariya, D. E., Dolidze, T. D., Zusman, L., Lindberg, G. & Glaser, J. Two-electron transfer for Tl(aq)3þ/Tl(aq)þ revisited. Common virtual [Tl(II)– Tl(II)]4þ intermediate for homogeneous (superexchange) and electrode (sequential) mechanisms. Inorg. Chem. 41, 1728–1738 (2002). 7. Falcinella, B., Felgate, P. D. & Laurence, G. S. Aqueous chemistry of thallium(II). Part I. Kinetics of reaction of thallium(II) with cobalt(III) and iron(III) ions and oxidation–reduction potentials of thallium(II). J. Chem. Soc., Dalton Trans. 1367–1373 (1974). 8. Segawa, Y., Yamashita, M. & Nozaki, K. Boryllithium: isolation, characterization, and reactivity as a boryl anion. Science 314, 113–115 (2006). 9. Saleh, L. M. A. et al. Group 3 and lanthanide boryl compounds: syntheses, structures and bonding analysis of Sc–B, Y–B and Lu–B s-coordinated NHC analogues. J. Am. Chem. Soc. 133, 3836–3839 (2011). 10. Protchenko, A. V. et al. A stable two-coordinate acyclic silylene. J. Am. Chem. Soc. 134, 6500–6503 (2012). 11. Eichler, B. E., Hardman, N. J. & Power, P. P. In8(C6H3-2,6-Mes2)4 (Mes¼C6H22,4,6-Me3): a metal-rich main-group cluster with a distorted cubane structure. Angew. Chem. Int. Ed. 39, 383–385 (2000). 12. Albright, T. A., Burdett, J. K. & Whangbo, M. H. Orbital Interactions in Chemistry 2nd edn (Wiley, 2013). 13. Cooke, A. H. & Duffus, H. J. The magnetic susceptibility of nitric oxide in a clathrate compound. Proc. Phys. Soc. A 67, 525–527 (1954). 14. Leininger, T. et al. Spin–orbit interaction in heavy group 13 atoms and TlAr. Chem. Phys. 217, 19–27 (1997). 15. Power, P. P. Main group elements as transition metals. Nature 463, 171–177 (2010). 16. Davidson, P. J., Hudson, A., Lappert, M. F. & Lednor, P. W. † Tris[bis(trimethylsilyl)methyl]tin(III), R3Sn : an unusually stable stannyl radical, from photolysis of R2Sn. J. Chem. Soc., Chem. Commun. 829–830 (1973). 17. Power, P. P. Persistent and stable radicals of the heavier main group elements and related species. Chem. Rev. 103, 789–810 (2003). 18. Lee, V. Y. & Sekiguchi, A. Stable silyl, germyl, and stannyl cations, radicals, and anions: heavy versions of carbocations, carbon radicals, and carbanions. Acc. Chem. Res. 40, 410–419 (2007). 19. Rodriguez, A., Prasang, C., Gandon, V., Bourg, J-B. & Bertrand, G. Stable singlet diradicals based on boron and phosphorus. ACS Symp. Ser. 917, 81–93 (2007). 20. Sprague, E. D. & Williams, F. ESR spectrum of BH3†2 in gamma-irradiated tetramethylammonium borohydride. Mol. Phys. 20, 375–378 (1971). 21. Pluta, C., Po¨rschke, K-R., Kru¨ger, C. & Hildebrand, K. An Al–Al one-electron p bond. Angew. Chem. Int. Ed. Engl. 32, 388–390 (1993). 22. He, X. et al. Reduction of a tetraaryldigallane to afford a radical anion with Ga–Ga multiple bonding character. Angew. Chem. Int. Ed. Engl. 32, 717–719 (1993). 23. Dohmeier, C. et al. [AltBu]6†2: EPR spectroscopic evidence and ab initio calculations. Angew. Chem. Int. Ed. Engl. 32, 1428–1430 (1993). 24. Wehmschulte, R. J. et al. Reduction of a tetraaryldialane to generate aluminum– aluminum p-bonding. Inorg. Chem. 32, 2983–2984 (1993). 25. Uhl, W., Vester, A., Kaim, W. & Poppe, J. Dialan-radikalanionen [R2Al–AlR2]†2. J. Organomet. Chem. 454, 9–13 (1993). 26. Uhl, W., Schu¨tz, U., Kaim, W. & Waldho¨r, E. Das tetraalkyldigallan-radikalanion [R2Ga–GaR2]†2 [R¼CH(SiMe3)2] mit langer einelektron-p-bindung. J. Organomet. Chem. 501, 79–85 (1995). NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1870
27. Haaland, A. et al. Gas-phase structure of the monomeric alkylgallium(I) compound Ga[C(SiMe3)3] and the electrochemical behavior of Ga4[C(SiMe3)3]4 and In4[C(SiMe3)3]4 with EPR evidence for a Ga4R4 radical anion. Organometallics 15, 1146–1150 (1996). 28. Wiberg, N., Amelunxen, K., No¨th, H., Schmidt, M. & Schwenk, H. Tetrasupersilyldiindium (In–In) and tetrasupersilyldithallium (Tl–Tl): (tBu3Si)2M–M(SitBu3)2 (M¼In, Tl). Angew. Chem. Int. Ed. Engl. 35, 65–67 (1996). 29. Uhl, W., Cuypers, L., Kaim, W., Schwederski, B. & Koch, R. [Ga9(CMe3)9]†2 – a persistent cluster radical anion, boron-analogous chemistry with the heavier homologue gallium. Angew. Chem. Int. Ed. 42, 2422–2423 (2003). 30. Uson, R., Fornies, J., Tomas, M., Garde, R. & Alonso, P. J. Synthesis and structure of [nBu4N]2[Tl{Pt(C6F5)4}2], the first paramagnetic compound containing thallium(III). J. Am. Chem. Soc. 117, 1837–1838 (1995). 31. Nakamoto, M., Yamasaki, T. & Sekiguchi, A. Stable mononuclear radical anions of heavier group 13 elements: [(tBu2MeSi)3E†–][Kþ(2.2.2-cryptand)] (E¼Al, Ga). J. Am. Chem. Soc. 127, 6954–6955 (2005). 32. Hitchcock, P. B., Huang, Q., Lappert, M. F. & Zhou, M. The coordination chemistry of the C1-symmetric bis(silyl)methyl ligand [CH(SiMe3){SiMe(OMe)2}]2 revisited: Li/M- (M¼Zn, Tl, Ce), Li4- or Ce2methoxy-bridged alkyls. Dalton Trans. 2988–2993 (2005).
Acknowledgements We thank the Leverhulme Trust (F/08699/E), the Oxford University John Fell Fund, the Australian Research Council (DP120101300 and FT120100421) and the Engineering and
Physical Sciences Research Council (EPSRC) (EP/F019181/1, EP/F055412/1 and access to the NMSF, Swansea). We are grateful for computational resources from the EPSRC’s National Service for Computational Chemistry Software, http://www.nsccs.ac.uk), and also thank the University College London’s High Performance Computing Facility (Legion@UCL) and associated support services, and the e-Infrastructure South consortium’s Centre for Innovation for computing resources via its ‘Iridis’ facility.
Author contributions A.V.P. and D.D. synthesized and characterized the compounds. J.R.H. carried out the EPR studies. C.Y.T., A.D.S., M.J.K., N.P. and C.J. mounted the crystals, collected the single-crystal X-ray crystallographic data and solved the crystal structures. K.H.B. and N.K. carried out the DFT calculations. R.T. carried out the electrochemical measurements. N.K., P.M., C.J. and S.A. generated and managed the project and wrote the manuscript. All authors discussed the results and commented on the manuscript.
Additional information Supplementary information and chemical compound information are available in the online version of the paper. Reprints and permissions information is available online at www.nature.com/reprints. Correspondence and requests for materials should be addressed to S.A.
Competing financial interests The authors declare no competing financial interests.
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
319
Stable GaX2 , InX2 and TlX2 radicals Andrey V. Protchenko1, Deepak Dange2, Jeffrey R. Harmer3,4, Christina Y. Tang1, Andrew D. Schwarz1, Michael J. Kelly1, Nicholas Phillips1, Remi Tirfoin1, Krishna Hassomal Birjkumar5, Cameron Jones2, Nikolas Kaltsoyannis5, Philip Mountford1 and Simon Aldridge1 * The chemistry of the Group 13 metals is dominated by the 11 and 13 oxidation states, and simple monomeric M(II) species are typically short-lived, highly reactive species. Here we report the first thermally robust monomeric MX2 radicals of gallium, indium and thallium. By making use of sterically demanding boryl substituents, compounds of the type M(II)(boryl)2 (M 5 Ga, In, Tl) can be synthesized. These decompose above 130 8 C and are amenable to structural characterization in the solid state by X-ray crystallography. Electron paramagnetic resonance and computational studies reveal a dominant metal-centred character for all three radicals (>70% spin density at the metal). M(II) species have been invoked as key short-lived intermediates in well-known electron-transfer processes; consistently, the chemical behaviour of these novel isolated species reveals facile one-electron shuttling processes at the metal centre.
T
he metals of Group 13 differ markedly in their terrestrial distribution and the scale of their technological exploitation, but the chemistry of all four elements—aluminium, gallium, indium and thallium—is dominated by two oxidation states, þ1 and þ3 (ref. 1). Therefore, compounds of the composition MX2 are typically mixed-valence species (for example, M(I)M(III)X4), and those that are not adopt diamagnetic M2X4 structures that feature M–M bonds2. Although mechanistically important, hitherto isolated examples of simple MX2 radical species that feature the metal in the þ2 oxidation state have been confined to trapping experiments at cryogenic temperatures2,3; controversial claims of room-temperature stable Ga(II) species, for example, have subsequently been shown to be erroneous4. By contrast, here we report the first thermally robust monomeric MX2 species of gallium, indium and thallium. By making use of sterically demanding boryl substituents, compounds of the type M(II)(boryl)2 (M ¼ Ga, In, Tl) that decompose above 130 8C can be synthesized. Electron paramagnetic resonance (EPR) and computational studies reveal a dominant metal-centred character for all three radicals (.70% spin density at the metal). M(II) species have been invoked as key short-lived intermediates in well-known electrontransfer processes5–7; consistently, the chemical behaviour of these novel isolated species reveals facile one-electron shuttling processes at the metal centre.
Results and discussion Syntheses. In recent work the lithium reagent (thf)2Li[B(NDippCH)2] (1, where Dipp ¼ 2,6-iPr2C6H3 and thf ¼ tetrahydrofuran) was utilized as a nucleophilic source of the boryl fragment in the synthesis of novel E–B chemical bonds8,9. In particular, the competence of this boryl substituent in stabilizing low-valent, low-coordinate main-group systems has been demonstrated through the synthesis of the Si(II) species Si[B(NDippCH)2][N(Dipp)SiMe3] (ref. 10). Herein, the reactivity of 1 towards M(I) precursors (M ¼ Ga, In or Tl) is shown to lead to the formation of the respective metal bis(boryl) complexes M[B(NDippCH)2]2 (M ¼ Ga (2-Ga), In (2-In) or Tl (2-Tl), Fig. 1) in yields of
38–75%. In the cases of the precursors TlCl, In[N(SiMe3)Dipp*] or Ga[N(SiMe3)Dipp*] (where Dipp* ¼ 2,6-(Ph2CH)2-4-Me-C6H2), apparent disproportionation of the putative M(I) intermediate M[B(NDippCH)2] is implied by the precipitation of elemental metal. In the case of Tl[N(SiMe3)2], however, the reaction stoichiometry can be evaluated explicitly through the isolation at 230 8C of the co-product Tl8[B(NDippCH)2]4 , a mixed-valence cluster that contains four Tl(0) and four Tl(I) centres (see Supplementary Fig. 1)11. Synthetic routes were also established from M(III) precursors; the reactions of ClIn[B(NDippCH)2]2 (3-In) with strong reducing agents provide alternative routes to the bis(boryl) system 2-In. Good yields (65–70%) of 2-In are obtained using Sm(h5-C5Me5)2(thf), but the use of potassium metal brings with it the possibility of over-reduction, which leads to the accompanying formation of aggregated systems that feature a lower average metal-oxidation state, such as In2[B(NDippCH)2]3 (4-In). Solid-state structures. All three bis(boryl) compounds, 2-Ga, 2-In and 2-Tl, can be isolated as analytically pure, thermally robust crystalline materials that decompose at 132–134 8C, 196–198 8C and 225–230 8C, respectively. The molecular structure of each compound in the solid state was determined to be monomeric by X-ray crystallography (see Fig. 2 and Supplementary Table 1). This contrasts markedly with previously reported M(II) compounds of the Group 13 metals, which feature multimetallic structures2; presumably, dimerization of 2-Ga, 2-In or 2-Tl to give an M–M-bonded M2X4 framework akin to related alkyl and silyl derivatives is prevented by the extremely sterically demanding nature of the boryl substituents. Accordingly, comparison of the structure of 2-Ga with those of related systems of the type XGa[B(NDippCH)2]2 (X ¼ Cl (3-Ga) and X ¼ Ga[B(NDippCH)2] (4-Ga), see Supplementary Fig. 2) reveals, as the bulk of the X substituent increases: (1) significant narrowing of the GaB2 angle (156.0(1)8, 152.0(1)8 and 146.4(1)8 for 2-Ga, 3-Ga and 4-Ga, respectively) and (2) lengthening of the Ga–B bonds within the GaB2 unit (mean distances 2.048(2) Å, 2.078(2) Å and 2.094(3) Å, respectively).
1
Inorganic Chemistry Laboratory, Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QR, UK, 2 School of Chemistry, PO Box 23, Monash University, Melbourne, Victoria 3800, Australia, 3 Centre for Advanced Electron Spin Resonance, Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QR, 4 Center for Advanced Imaging, University of Queensland, St Lucia, Queensland 4072, Australia, 5 Department of Chemistry, University College London, Christopher Ingold Laboratories, 20 Gordon Street, London WC1H 0AJ, UK. * e-mail: [email protected] NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
315
ARTICLES [M(NRR´)] n
i
or
NATURE CHEMISTRY
Dipp
Dipp
N
N
B
M
N
Dipp M = Ga, In, Tl
ii
B
N
TlCl
Cl
Dipp N
Dipp
M
B
N
N
Dipp
N
B
Dipp
Dipp
Dipp = 2,6- i Pr2C6H3
Dipp = 2,6- i Pr2C6H3
M = Ga: 2-Ga M = In: 2-In M = Tl: 2-Tl
M = In: 3-In
Figure 1 | Syntheses of M(II) complexes that feature a supporting bis(boryl) ligand set from either M(I) or M(III) precursors. i, For 2-Ga from Ga[N(SiMe3)Dipp*], (thf)2Li[B(NDippCH)2] (1, 2.0 equiv.), toluene, 278 8C to 0 8C over five hours, then 12-crown-4 (2.0 equiv.), room temperature, one hour, 64%; for 2-In from In[N(SiMe3)Dipp*], 1 (2.0 equiv.), toluene, 278 8C to 0 8C over five hours, then 12-crown-4 (2.0 equiv.), room temperature, one hour, 38%; for 2-Tl from Tl[N(SiMe3)2], 1 (1.0 equiv.), hexane, 245 8C, five minutes, 52%; for 2-Tl from TlCl, 1 (1.0 equiv.), hexane, 278 8C to room temperature, 12 hours, 75%. ii, For 2-In from 3-In, K metal (0.94 equiv.), benzene-d6/thf (12:1), room temperature, sonication for two hours, 75% (by NMR spectroscopy), or Sm(h5-C5Me5)2(thf) (1.0 equiv.), benzene, room temperature, 10 minutes, 67% (isolated). The syntheses of 2-Ga and 2-In can also be accomplished from M(I) precursors and 1 (1 equiv.) without the use of 12-crown-4, although slightly lower yields and greater contamination with HB(NDippCH)2 are observed. R and R′ are aryl or silyl groups.
The molecular structures of 2-Ga and 2-In differ notably from that (linear) of 2-Tl in that the B–M–B framework for the lighter congeners is bent (B–M–B angles of 156.0(1)8 and 145.4(1)8 versus 177.6(2)8, respectively). The presence of an additional metal-bound hydrogen atom (although not easily discounted on the basis of X-ray diffraction data) is at odds with spectroscopic and magnetic data, which indicate that all three molecules possess an unpaired electron. The observed trend in the B–M–B angles is consistent with that predicted by density functional theory (DFT) (160.38, 152.78 and 178.88, respectively (Supplementary Tables 2–5)), and with calculated metal ns orbital contributions to the respective singularly occupied molecular orbitals (SOMOs), which increase in the order Tl , Ga , In (2-Tl, ,1.0% 6s, 65.2% 6p; 2-Ga, 3.0% 4s, 62.7% 4p; 2-In, 5.9% 5s, 60.3% 5p). By analogy with the classical Walsh diagram for linear/bent dihydrides12, greater SOMO s character for a system of five valence electrons is linked intrinsically to increased bending at the metal, and leads to a lower SOMO energy (2-Tl, 22.740 eV; 2-Ga, 23.069 eV; 2-In, 23.113 eV). Moreover, a
N(1)
Ga(1) B(2)
N(4)
the relative degrees of orbital mixing for the three metal systems and hence the observed deviations from linearity can readily be rationalized in terms of the respective atomic ns/np energy separations (Tl, 7.109 eV; Ga, 6.319 eV; In, 5.652 eV, as calculated by DFT). The largest ns/np gap is thus associated with the most linear molecule (that is, 2-Tl). All three compounds are red– orange in colour, and the respective ultraviolet/visible (UV-vis) spectra reveal bands at 350 nm that tail off into the visible region (Supplementary Fig. 8). In the case of 2-Ga, the two strongest bands (at l ¼ 362 nm and 309 nm) are thought, on the basis of time-dependent DFT calculations, to feature significant contributions from transitions that originate in the SOMO (Supplementary Fig. 9 and Supplementary Tables 7 and 8). Solution-phase characterization. The paramagnetic nature of 2-Ga, 2-In and 2-Tl implied by their structures in the solid state can be shown spectroscopically to be retained in solution. Thus, broadened 1H NMR resonances for the boryl ligand and magnetic moments determined by the Evans method (meff at 293 K) of 2.0 (+0.1) mB , 2.0 (+0.1) mB and 1.2 (+0.1) mB , respectively, are consistent with the presence of an unpaired electron for each species. The magnetic moment measured for 2-Tl is reproducibly smaller than that expected on the basis of the spin-only formula (one unpaired electron, 1.73 mB), consistent with the linear geometry at thallium and with the occupation of a p-orbital manifold with high Tl 6p character (DFT-calculated SOMO, 65.2% Tl 6p; a SOMO þ 1, 86.0% Tl 6p). Related radicals featuring a non-zero orbital component that arises from a single electron occupying degenerate p orbitals (such as NO) have a (diamagnetic) j ¼ 1/2 ground state, which causes meff to approach zero at low values of kT/l (ref. 13). In the case of 2-Tl, the much larger magnitude of the spin-orbit coupling parameter l (ref. 14) implies near-exclusive occupancy of the j ¼ 1/2 state, and the fact that meff = 0 presumably reflects that the SOMO and SOMO þ 1 are not actually degenerate in the minimum energy calculated structure (DE ¼ 0.458 eV). No evidence for the formation in solution of a second (diamagnetic) component (for example, dimeric [(HCNDipp)2B]2Tl–Tl[B(NDippCH)2]2) was revealed by multinuclear variable-temperature NMR spectroscopy. X- and W-band field-sweep EPR spectra measured in frozen pentane/hexane solution (1:1) show that all three species have g values that deviate significantly from the free-electron value, and exhibit very large hyperfine couplings to their respective metals (Fig. 3 and Supplementary Table 6). The corresponding couplings (Ax , Ay , Az) measured for the matrix-isolated GaH2 radical are larger (1,957 MHz, 1,917 MHz and 2,379 MHz, respectively), which presumably reflects c
b
N(3)
N(4)
N(3)
N(2)
In(1) B(2)
B(1)
DOI: 10.1038/NCHEM.1870
N(3)
N(1)
Tl(1)
B(1) N(1)
N(2)
B(2) N(4)
B(1) N(2)
Figure 2 | Molecular structures. a, 2-Ga. b, 2-In. c, 2-Tl. Ellipsoids are at the 50% probability level, carbon-bound hydrogen atoms are omitted and iPr groups are shown in a wireframe format for clarity. Selected bond lengths: 2-Ga, Ga(1)–B(1) 2.045(2) Å, Ga(1)–B(2) 2.051(2) Å; 2-In, In(1)–B(1) 2.246(3) Å, In(1)–B(2) 2.242(3) Å; 2-Tl, Tl(1)–B(1) 2.173(6) Å, Tl(1)–B(2) 2.167(5) Å. 316
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1870
a
b Simulation Ga, experimental Simulation Ga, experimental Simulation In, experimental
Simulation In, experimental
Simulation
Simulation
Tl, experimental Tl, experimental
0
250
500
750
1,000
1,250
1,500
3,000
3,500
4,000
B0 (mT)
4,500
5,000
5,500
6,000
B0 (mT)
Figure 3 | Field-sweep EPR spectra for 2-Ga, 2-In and 2-Tl measured in frozen pentane/hexane solution (1:1), along with the corresponding simulations. a, X-band continuous-wave EPR spectra. b, W-band echo-detected EPR spectra, first-derivative representation. For 2-Tl the signal starts at 4.5 T (g1 ¼ 1.23) and could only be recorded to 6 T (the upper limit of our magnet). The principal g values gi and metal hyperfine couplings Ai derived from the simulation are as follows (the corresponding values calculated using DFT are italicized and in parentheses). 2-Ga, g1 , g2 , g3 ¼ 1.881, 1.998, 2.014 (1.8760, 1.9960, 2.0113); A(69Ga) ¼ 444, 1,032, 533 ¼ 670 þ (2226, 362, 2137) MHz (214 þ (2185, 312, 2127) MHz). 2-In, g1 , g2 , g3 ¼ 1.719, 1.954, 1.999 (1.7338, 1.9689, 2.0125); A(115In) ¼ 913, 1,936, 793 ¼ 1,214 þ (2301, 722, 2421) MHz (645 þ (2408, 621, 2214) MHz). 2-Tl, g1 g2, g3 ¼ 0.6, 0.7, 1.23 (1.0047, 1.4381, 1.697); A(205Tl) ¼ 28,000, 10,300, 28,100 ¼ 21,933 þ (26,067, 12,233, 26,167) MHz (22866 þ (26,105, 7,592, 21,478) MHz).
reduced possibilities for spin delocalization in the dihydride system3. The unusually small g values for 2-Tl determined from the simulations were confirmed by an X-band nutation experiment (see Supplementary Fig. 4). Smaller hyperfine couplings unresolved in the field-sweep EPR spectra from the 11B and 14N nuclei of the [B(NDippCH)2] ligands were characterized by electron nuclear double resonance (ENDOR) spectroscopy and hyperfine sublevel correlation spectroscopy (HYSCORE; see Supplementary Figs 5 and 6). For 2-Ga and 2-In, small hyperfine couplings from both 11 B (20–42 MHz) and 14N (1–7 MHZ) nuclei were detected that, by comparison with tabulated 2s and 2p atomic orbital hyperfine couplings for the 11B and 14N nuclei, allow spin densities on the ligands of 0.25 to be calculated. The summed ligand-based spin densities from DFT calculations are 0.266, 0.273 and 0.277 for 2-Ga, 2-In and 2-Tl, respectively. In the case of 2-Tl, only small 14 N hyperfine couplings could be detected (3.5–6.5 MHz) via HYSCORE, and no 11B ENDOR data could be measured, presumably as a result of fast electron relaxation. The experimental principal g values and metal hyperfine couplings, along with those from DFT calculations, are summarized in Fig. 3 (see Supplementary Table 6 for a full list). There is good overall agreement between g values and the anisotropic part of the metal hyperfine couplings (the isotropic part is very dependent on spin polarization of the core metal orbitals and is more difficult to calculate), with both experimental and DFT-calculated g values decreasing and metal hyperfine couplings increasing across the series Ga, In, Tl. The EPR data leave no doubt that all three radicals are characterized best as metal centred (.70% spin density at the metal). Although stable paramagnetic metal-centred radicals are common among the elements of the d block, in recent years welldefined radicals have also started to emerge for the heavier p-block elements15–19. Despite this, group 13 metal-based radicals are still very rare20–31, which in part reflects the ease of dimerization of low-coordinate species in the absence of significant steric shielding. Although such processes can be subverted for charged systems, and † the anionic trisilylmetalate species [M(SiMetBu2)3] 2 (M ¼ Al, Ga) 31 have been reported , to date these remain the only other monometallic group 13 radicals to have been structurally characterized.
Redox chemistry. Implicit in the descriptions of 2-Ga, 2-In and 2-Tl as monomeric M(II) species is relatively facile redox chemistry that leads to the formation of complexes featuring the more familiar M(III) and M(I) oxidation states. Tl(II) systems, for example, have been implicated as intermediate species in twoelectron self-exchange processes and non-complementary redox reactions with transition-metal cations6,7. Accordingly, taking 2-Tl as the exemplar, single-electron oxidation and reduction chemistry to generate tractable Tl(III) cations and Tl(I) anions can be achieved both electrochemically and chemically (Fig. 4 and Supplementary Fig. 7). Moreover, the reaction of [2-Tl]þ with [2-Tl]2 can be shown to proceed through comproportionation to regenerate the neutral Tl(II) compound 2-Tl. The reactions of 2-Tl with borate salts of the trityl cation [CPh3]þ led to the formation of diamagnetic [2-Tl]þ, together with the head-to-tail Gomberg dimer Ph3C(C6H5)CPh2 of the trityl radical (ferrocenium oxidants give the same thallium-containing product). The stoichiometry of this oxidation, and the implied transfer of one electron per molecule of 2-Tl, can be established by integration of the 1H NMR signals of the products. In addition, the identity of the [2-Tl]þ cation (as the [B(C6H3(CF3)2-3,5)]2 salt) has been confirmed crystallographically (see Supplementary Fig. 3), with the linear coordination geometry at thallium being consistent with related four-valence-electron systems12. Conversely, reduction of 2-Tl with [K(18-crown-6)][C10H8] led to the formation of diamagnetic [K(18-crown-6)][Tl[B(NDippCH)2]2 ], the crystal structure of which reveals a bent TlB2 core (with the angle B–Tl–B ¼ 109.8(1)8) and elongated Tl–B bonds (2.454(4) Å and 2.456(4) Å) (Fig. 4). The bent geometry of [2-Tl]2 is consistent with the greater central-atom ns character in the highest occupied molecular orbital for a system of six valence electrons (10.4% Tl 6s, compared with ,1.0% Tl 6s for five-electron charge-neutral 2-Tl)32. Consequently, there is a greater Tl 6p character in molecular orbitals with Tl–B bonding character (compared with 2-Tl), and significant (.13%) elongation of the Tl–B bonds therefore accompanies the reduction process. In conclusion, we report the syntheses of divalent bis(boryl)gallium, -indium and -thallium radicals from either M(I) or M(III)
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
317
ARTICLES a
NATURE CHEMISTRY
[BAr4] – Dipp
[K(18-crown-6)] +
Dipp
N
+
Tl
B
Dipp
N
i
B
N
iii
N
Dipp
ii
2-Tl
iii
N B N Dipp
Dipp [2-Tl]+[BAr4]–
Dipp Tl
–
N B
[K(18-crown-6)]+[2-Tl]–
K(60)
Tl(1) N(4) B(2) N(3)
Figure 4 | Redox chemistry of 2-Tl. a, One-electron chemical oxidation and reduction processes. b, Molecular structure of [K(18-crown-6)][2-Tl] . OEt2 (ellipsoids are shown at the 50% probability level, diethyl ether solvated molecule, C-bound hydrogen atoms are omitted and iPr groups are shown in a wireframe format for clarity). Selected bond lengths and angles: Tl(1)–B(1) 2.454(4) Å, Tl(1)–B(2) 2.456(4) Å, B(1)–Tl(1)–B(2) 109.8(1)8. i, [Fe(h5C5H5)2][B(C6H3(CF3)2-3,5)4] (1.0 equiv.), diethyl ether, room temperature, ten minutes, 98%, or [CPh3][B(C6F5)4] (1.0 equiv.), benzene-d6 , room temperature, NMR scale. ii, [K(18-crown-6)][C10H8 ], (1.33 equiv.), THF, 2196 8C to room temperature, two minutes, 26%. iii, [2-Tl][B{C6H3(CF3)2-3,5}4] plus [K(18-crown-6)][2-Tl] (1:1), THF/benzene-d6 , room temperature, two minutes, quantitative.
precursors. These species have unprecedented thermal stability, and decompose, in each case, at temperatures above 130 8C. The monomeric structures of these systems (enforced by the steric bulk of the boryl substituents) in the solid state were confirmed by X-ray crystallography. The description of each of these as predominantly metal-centred radical species is consistent with the results of EPR and DFT studies.
Methods Manipulations were carried out under a dry, oxygen-free argon or dinitrogen atmosphere, with reagents dissolved or suspended in aprotic solvents, and combined or isolated using cannula and glove-box techniques. Compounds M[B(NDippCH)2]2 (M ¼ Ga (2-Ga), In (2-In) or Tl (2-Tl)) were synthesized, respectively, via the reactions of Ga[N(SiMe3)Dipp*], In[N(SiMe3)Dipp*] or one of Tl[N(SiMe3)2] or TlCl, with (thf )2Li[B(NDippCH)2] (1) in toluene or hexane solution, and isolated by low-temperature crystallization. 2-In was also synthesized by the reduction of ClIn[B(NDippCH)2]2 (3-In) with Sm(h5-C5Me5)2(thf ) in benzene. The Tl(I) species [K(18-crown-6)][2-Tl] was synthesized by the reactions of 2-Tl with [K(18-crown-6)][C10H8] in THF, and the Tl(III) species [2-Tl][B{C6H3(CF3)2-3,5}4] from the reaction of 2-Tl with [Fe(h5-C5H5)2][B(C6H3 (CF3)2-3,5)4] in diethyl ether. Reaction of isolated samples of [K(18-crown-6)][2-Tl] with [2-Tl][B{C6H3(CF3)2-3,5}4] in THF or benzene led to the quantitative regeneration of 2-Tl. We characterized new compounds by elemental analysis, mass spectrometry, UV/vis spectroscopy, cyclic voltammetry (for 2-Tl), multinuclear NMR spectroscopy (for diamagnetic compounds) and EPR spectroscopy (for paramagnetic compounds). Solution-phase magnetic moments were determined for 2-Ga, 2-In and 2-Tl using the Evans method, and the structures of 2-Ga, 2-In, 2-Tl, 3-Ga, 3-In, 4-Ga, 4-In, Tl8[B(NDippCH)2]4 , [K(18-crown-6)][2-Tl] and [2-Tl][B{C6H3(CF3)2-3,5}4] in the solid state were determined by single-crystal X-ray diffraction studies. X- and W-band field-sweep EPR spectra were measured in frozen pentane/hexane solution (1:1) for 2-Ga, 2-In and 2-Tl. Comparisons of the 318
Received 13 September 2013; accepted 13 January 2014; published online 16 February 2014
References
b
N(2)
11 B and 14N hyperfine couplings determined by ENDOR and HYSCORE with tabulated 2s and 2p atomic orbital hyperfine couplings for the 11B and 14N nuclei allowed the ligand (and hence (by difference) the metal-centred) spin densities to be calculated for each species. DFT, as implemented in the Amsterdam Density Functional code, was employed to calculate molecular geometries, orbital energies and compositions, spin densities, EPR parameters and electron-transition energies.
N Dipp
Ar = C6F5 or C6H3(CF3)2-3,5
N(1) B(1)
DOI: 10.1038/NCHEM.1870
1. Downs, A. J. & Himmel, H-J. The Group 13 Metals Aluminium, Gallium, Indium and Thallium: Chemical Patterns and Peculiarities (eds Aldridge, S. & Downs, A. J.) Ch. 1 (Wiley, 2011). 2. Uhl, W. & Layh, M. The Group 13 Metals Aluminium, Gallium, Indium and Thallium: Chemical Patterns and Peculiarities (eds Aldridge, S. & Downs, A. J.) Ch. 4 (Wiley, 2011). 3. Knight, L. B. Jr, Banisaukas, J. J. III, Babb, R. & Davidson, E. R. Electron spin resonance matrix isolation and ab initio theoretical investigations of 69,71 GaH2 , 69,71GaD2 , H69,71GaCH3 , and D69,71GaCD3. J. Chem. Phys. 105, 6607–6615 (1996). 4. Kaim, W. & Matheis, W. Bis(1,4-di-tert-butyl-1,4-diazabutadiene)gallium is not a gallium(II) compound. Chem. Commun. 597–598 (1991). 5. Pardoe, J. A. J. & Downs, A. J. Development of the chemistry of indium in formal oxidation states lower than þ3. Chem. Rev. 107, 2–45 (2007). 6. Khoshtariya, D. E., Dolidze, T. D., Zusman, L., Lindberg, G. & Glaser, J. Two-electron transfer for Tl(aq)3þ/Tl(aq)þ revisited. Common virtual [Tl(II)– Tl(II)]4þ intermediate for homogeneous (superexchange) and electrode (sequential) mechanisms. Inorg. Chem. 41, 1728–1738 (2002). 7. Falcinella, B., Felgate, P. D. & Laurence, G. S. Aqueous chemistry of thallium(II). Part I. Kinetics of reaction of thallium(II) with cobalt(III) and iron(III) ions and oxidation–reduction potentials of thallium(II). J. Chem. Soc., Dalton Trans. 1367–1373 (1974). 8. Segawa, Y., Yamashita, M. & Nozaki, K. Boryllithium: isolation, characterization, and reactivity as a boryl anion. Science 314, 113–115 (2006). 9. Saleh, L. M. A. et al. Group 3 and lanthanide boryl compounds: syntheses, structures and bonding analysis of Sc–B, Y–B and Lu–B s-coordinated NHC analogues. J. Am. Chem. Soc. 133, 3836–3839 (2011). 10. Protchenko, A. V. et al. A stable two-coordinate acyclic silylene. J. Am. Chem. Soc. 134, 6500–6503 (2012). 11. Eichler, B. E., Hardman, N. J. & Power, P. P. In8(C6H3-2,6-Mes2)4 (Mes¼C6H22,4,6-Me3): a metal-rich main-group cluster with a distorted cubane structure. Angew. Chem. Int. Ed. 39, 383–385 (2000). 12. Albright, T. A., Burdett, J. K. & Whangbo, M. H. Orbital Interactions in Chemistry 2nd edn (Wiley, 2013). 13. Cooke, A. H. & Duffus, H. J. The magnetic susceptibility of nitric oxide in a clathrate compound. Proc. Phys. Soc. A 67, 525–527 (1954). 14. Leininger, T. et al. Spin–orbit interaction in heavy group 13 atoms and TlAr. Chem. Phys. 217, 19–27 (1997). 15. Power, P. P. Main group elements as transition metals. Nature 463, 171–177 (2010). 16. Davidson, P. J., Hudson, A., Lappert, M. F. & Lednor, P. W. † Tris[bis(trimethylsilyl)methyl]tin(III), R3Sn : an unusually stable stannyl radical, from photolysis of R2Sn. J. Chem. Soc., Chem. Commun. 829–830 (1973). 17. Power, P. P. Persistent and stable radicals of the heavier main group elements and related species. Chem. Rev. 103, 789–810 (2003). 18. Lee, V. Y. & Sekiguchi, A. Stable silyl, germyl, and stannyl cations, radicals, and anions: heavy versions of carbocations, carbon radicals, and carbanions. Acc. Chem. Res. 40, 410–419 (2007). 19. Rodriguez, A., Prasang, C., Gandon, V., Bourg, J-B. & Bertrand, G. Stable singlet diradicals based on boron and phosphorus. ACS Symp. Ser. 917, 81–93 (2007). 20. Sprague, E. D. & Williams, F. ESR spectrum of BH3†2 in gamma-irradiated tetramethylammonium borohydride. Mol. Phys. 20, 375–378 (1971). 21. Pluta, C., Po¨rschke, K-R., Kru¨ger, C. & Hildebrand, K. An Al–Al one-electron p bond. Angew. Chem. Int. Ed. Engl. 32, 388–390 (1993). 22. He, X. et al. Reduction of a tetraaryldigallane to afford a radical anion with Ga–Ga multiple bonding character. Angew. Chem. Int. Ed. Engl. 32, 717–719 (1993). 23. Dohmeier, C. et al. [AltBu]6†2: EPR spectroscopic evidence and ab initio calculations. Angew. Chem. Int. Ed. Engl. 32, 1428–1430 (1993). 24. Wehmschulte, R. J. et al. Reduction of a tetraaryldialane to generate aluminum– aluminum p-bonding. Inorg. Chem. 32, 2983–2984 (1993). 25. Uhl, W., Vester, A., Kaim, W. & Poppe, J. Dialan-radikalanionen [R2Al–AlR2]†2. J. Organomet. Chem. 454, 9–13 (1993). 26. Uhl, W., Schu¨tz, U., Kaim, W. & Waldho¨r, E. Das tetraalkyldigallan-radikalanion [R2Ga–GaR2]†2 [R¼CH(SiMe3)2] mit langer einelektron-p-bindung. J. Organomet. Chem. 501, 79–85 (1995). NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1870
27. Haaland, A. et al. Gas-phase structure of the monomeric alkylgallium(I) compound Ga[C(SiMe3)3] and the electrochemical behavior of Ga4[C(SiMe3)3]4 and In4[C(SiMe3)3]4 with EPR evidence for a Ga4R4 radical anion. Organometallics 15, 1146–1150 (1996). 28. Wiberg, N., Amelunxen, K., No¨th, H., Schmidt, M. & Schwenk, H. Tetrasupersilyldiindium (In–In) and tetrasupersilyldithallium (Tl–Tl): (tBu3Si)2M–M(SitBu3)2 (M¼In, Tl). Angew. Chem. Int. Ed. Engl. 35, 65–67 (1996). 29. Uhl, W., Cuypers, L., Kaim, W., Schwederski, B. & Koch, R. [Ga9(CMe3)9]†2 – a persistent cluster radical anion, boron-analogous chemistry with the heavier homologue gallium. Angew. Chem. Int. Ed. 42, 2422–2423 (2003). 30. Uson, R., Fornies, J., Tomas, M., Garde, R. & Alonso, P. J. Synthesis and structure of [nBu4N]2[Tl{Pt(C6F5)4}2], the first paramagnetic compound containing thallium(III). J. Am. Chem. Soc. 117, 1837–1838 (1995). 31. Nakamoto, M., Yamasaki, T. & Sekiguchi, A. Stable mononuclear radical anions of heavier group 13 elements: [(tBu2MeSi)3E†–][Kþ(2.2.2-cryptand)] (E¼Al, Ga). J. Am. Chem. Soc. 127, 6954–6955 (2005). 32. Hitchcock, P. B., Huang, Q., Lappert, M. F. & Zhou, M. The coordination chemistry of the C1-symmetric bis(silyl)methyl ligand [CH(SiMe3){SiMe(OMe)2}]2 revisited: Li/M- (M¼Zn, Tl, Ce), Li4- or Ce2methoxy-bridged alkyls. Dalton Trans. 2988–2993 (2005).
Acknowledgements We thank the Leverhulme Trust (F/08699/E), the Oxford University John Fell Fund, the Australian Research Council (DP120101300 and FT120100421) and the Engineering and
Physical Sciences Research Council (EPSRC) (EP/F019181/1, EP/F055412/1 and access to the NMSF, Swansea). We are grateful for computational resources from the EPSRC’s National Service for Computational Chemistry Software, http://www.nsccs.ac.uk), and also thank the University College London’s High Performance Computing Facility (Legion@UCL) and associated support services, and the e-Infrastructure South consortium’s Centre for Innovation for computing resources via its ‘Iridis’ facility.
Author contributions A.V.P. and D.D. synthesized and characterized the compounds. J.R.H. carried out the EPR studies. C.Y.T., A.D.S., M.J.K., N.P. and C.J. mounted the crystals, collected the single-crystal X-ray crystallographic data and solved the crystal structures. K.H.B. and N.K. carried out the DFT calculations. R.T. carried out the electrochemical measurements. N.K., P.M., C.J. and S.A. generated and managed the project and wrote the manuscript. All authors discussed the results and commented on the manuscript.
Additional information Supplementary information and chemical compound information are available in the online version of the paper. Reprints and permissions information is available online at www.nature.com/reprints. Correspondence and requests for materials should be addressed to S.A.
Competing financial interests The authors declare no competing financial interests.
NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
319
