Stepwise isolation of low-valent, low-coordinate Sn and Pb mono- and dications in the coordination sphere of platinum† †Electronic supplementary information (ESI) available: Experimental and synthetic procedures, characterisation data, computational details, and crystallographic methods employed in this work are given. CCDC 1002025–1002028 and 1023363–1023368. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4sc02948h Click here for additional data file. Click here for additional data file.

ion of two chlorides from 2 or 4 may provide simple, monocoordinate, low-valent Sn or Pb dications if the strongly Lewis basic nature of the metal center can be harnessed to partly relieve the electron deciency of such a dication. The greater extent of Pt–M (M1⁄4 Sn, Pb) s donation in 7a and 8a over that in 4 and 2might provide enough electronic stabilization to attenuate the reactivity of the M dication and allow its isolation. Addition of two equivalents of AlBr3 to 4 and AlCl3 to 2 resulted in the formation of the loosely-associated salts [(Cy3P)2Pt(Sn)][AlBr4]2 (10) and [(Cy3P)2Pt(Pb)][AlCl4]2 (11) (Scheme 4). Alternatively, halide abstraction from 7a and 8a using stoichiometric amounts of AlBr3 or AlCl3 also led to the desired dicationic species 10 and 11, respectively. The P{H} NMR spectrum of compound 10 reveals a sharp singlet with Pt satellites at d 1⁄4 52.9 ppm (JP–Pt 1⁄4 2993 Hz), marginally downeld from that of 4 (d1⁄4 49.7 ppm, JP–Pt 1⁄4 3421 Hz). In contrast to 10, the low-valent Pb dicationic complex 11 exhibits a resonance (d1⁄4 47.5 ppm, JP–Pt1⁄4 2950 Hz) marginally upeld of that of 2. However, the decrease of the coupling 428 | Chem. Sci., 2015, 6, 425–435 constants from the neutral species as well as from 7a/8a indicates the formation of a dicationic species. The Al NMR spectra of 10 and 11 reveal broad singlets at d 1⁄4 80.8 and 103.6 ppm, respectively, indicating a tetrahedral aluminate. No Sn{H} or Pb{H} NMR resonances were detected despite several attempts. Single crystal X-ray diffraction studies of complex 10 conrmed the distorted T-shaped geometry around its platinum centre (Fig. 3) – a common feature of MOLPs involving [PtL2] Lewis bases. The Pt–Sn bond of this dicationic complex is even shorter than that of the monocationic complex (10: 2.502(1) Å; 7b: 2.524(1) Å). The tin center is weakly coordinated by three bromide atoms of two aluminate counteranions, with a Sn/Br(WCA)avg distance of 3.055(5) Å, markedly shorter than the sum of their van der Waals radii (4.0 Å). The shortest Sn–Br distance in 10 (2.956(1) Å) is signicantly longer than regular Sn– Br bond lengths (ca. 2.59 Å). The published arene-stabilized, chlorine-bridged tin monocations [h-(C6H6)2SnCl(AlCl4)]2 (12), and [h-(C6Me6)SnCl(AlCl4)]4 (13) also possess bonds that are shorter than sum of their van der Waals radii (3.496(2) Å vs. 3.92 Å). Similar structural features are also observed in the case of 11 (Fig. 3), which exhibits a distorted T-shaped geometry around its platinum centre. The Pt–Pb bond length of 11 (2.564(1) Å) is 0.04 Å shorter than that in 8a and is shorter than typical Pt–Pb bonds. For example, the Pt–Pb distances in the metallocryptate [Pt2(P2phen)3Pb][ClO4]2 are 2.747(1) and 2.733(1) Å, indicating an attractive metallophilic interaction rather than Lewis acid– base interactions. The Pt–Pb distance in the complex [(AcO) Pb(crown-P2)Pt(CN)2][(O2CCH3)] (H) (3.313(2) Å) is long enough to raise questions about the existence of bonding between these two entities. Another Pt–Pb compound, [nBu4N]2[{Pt(C6F5)4}2(Pb)] (I), shows interactions between the Pb center and the endo uorine atoms. The Pb center in 11 also exhibits weak interactions with two aluminate anions (Pb/Cl(WCA)avg 3.001(1) Å), leading to a pseudo-T-shaped geometry at the Pb atom. We cannot disregard two additional Pb/Cl interactions with the [AlCl4] anions (3.592(1) Å), which although long are nevertheless This journal is © The Royal Society of Chemistry 2015 Fig. 2 Molecular structures of 7b and 8a. Ellipsoids are drawn at the 50% probability level. Hydrogen atoms, counteranions (two [AlCl4] or [BAr4] units), and ellipsoids of the cyclohexyl rings are omitted for clarity. Selected bond lengths [Å] and angles [ ], calculated values are in parentheses: 7b: Pt1–Sn1 2.524(1), Sn1–Br1 2.780(1), Sn1–Br10 2.821(1), P1–Pt–P2 159.9(1), Br1–Sn1–Br10 84.1(1), Pt1–Sn1–Br1 93.9(1), Pt1– Sn1–Br10 114.0(1); 8a: Pt1–Pb1 2.603(1) (2.626), Pb1–Cl1 2.763(4) (2.720), Pb1–Cl2 2.829(3) (2.777), Pt2–Pb2 2.603(2) (2.622), Pb2–Cl2 2.719(4) (2.724), Pb2–Cl1 2.766(4) (2.786), Pb1/Cl(WCA) 3.420(1); P2– Pt1–P1 162.56(10) (159.9), Cl1–Pb1–Cl2 77.32(10) (80.4), P3–Pt2– Pb2–Cl2 112.28(10) (104.6). Scheme 4 Formation of the pseudo-monocoordinate dications of tin (10) and lead (11). Edge Article Chemical Science O pe n A cc es s A rt ic le . P ub lis he d on 1 5 O ct ob er 2 01 4. D ow nl oa de d on 0 7/ 09 /2 01 7 10 :2 0: 34 . T hi s ar tic le is li ce ns ed u nd er a C re at iv e C om m on s A ttr ib ut io nN on C om m er ci al 3 .0 U np or te d L ic en ce . View Article Online shorter than the sum of the van der Waals radii (3.77 Å). Our calculations also showed that these interactions persist even in the gas phase, provided non-constrained calculations were used. The Pb/Cl interaction is considerably shorter than those in 8a (3.420(1) Å) but signicantly longer than the Pb–Cl single bond length (2.300(4) Å) reported for LPbCl [L 1⁄4 HC(CMeNAr)2 (Ar 1⁄4 2,6-iPr2C6H3)] (14). The distance is in the same range as other weak Pb/Cl interactions, e.g. in the arene-stabilized lead(II) dications [(1,2-C6H4Me2)2Pb (AlCl4)2] (F) (2.969(1) Å) or [(C6H6)2Pb][AlCl4]2 (15) (2.854(8)–3.218(9) Å). We subsequently attempted exchange of the donor fragment [(Cy3P)2Pt] from the dication 11. Upon addition of aromatic donor ligands (benzene and uorobenzene), no exchange reaction was observed. Likewise, no exchange was observed upon addition of the strong N-donor ligand DMAP. This journal is © The Royal Society of Chemistry 2015 Synthesis of a stannyl anion Aer the isolation of low-valent tin monoand dications by halide abstraction from 4, we sought to conversely add another halide to form anionic tin compounds. The preparation of the anionic species [NnBu4][(Cy3P)2Pt(SnBr3)] (16) is shown in Scheme 5. The stoichiometric addition of [NnBu4]Br to 4 resulted in the formation of an anionic complex containing a formal Pt–Sn dative bond, along with a tetrabutylammonium countercation, [NnBu4][(Cy3P)2Pt(SnBr3)] (16), in good yields (80%). Such additions of bromide anions to tin(II) compounds, yielding anionic tin(II) salts, are known in cases where the tin atom is ligated by main group donors, but are unknown in cases where the Lewis base is a transition metal fragment. The P{H} NMR spectra of 16 reveals a sharp singlet at d 1⁄4 56.2 ppm (JP–Pt 1⁄4 4634 Hz), slightly upeld in comparison to that of precursor 4 d 1⁄4 49.7 ppm (JP–Pt 1⁄4 3421 Hz), while the coupling constant increased by 1200 Hz, indicating a signicantly modied chemical environment around the Sn atom. As in previous cases, all attempts to detect Sn{H} NMR resonances failed. Red crystals of 16 were analyzed by single-crystal X-ray diffraction (Fig. 4), displaying a Pt–Sn distance (2.604(1) Å) nearly identical to that of its precursor (2.605(2) Å). As expected, the Sn atom is strongly pyramidalized. The average Sn–Br distance in the stannyl anion (2.69 Å) is longer than the average Sn–Br distance (2.63 Å) in 4, presumably due to the introduction of another Br, which increases the steric demand at the Sn center. Reactivity of low-valent Pb cations Cationic species are known to form adducts with Lewis bases, prompting us to study the reaction of 8a with 4-picoline. This reaction afforded the 4-picoline-coordinated low-valent Pb cation [{(Cy3P)2Pt}PbCl(4-pic)][AlCl4] (17) (Scheme 6). Although all attempts to grow X-ray quality single crystals for diffraction analysis failed, the formation of 17 was clearly demonstrated Chem. Sci., 2015, 6, 425–435 | 429 Fig. 3 Molecular structures of 10 and 11. Thermal ellipsoids are drawn at the 50% probability level. Hydrogen atoms and ellipsoids of the cyclohexyl rings are omitted for clarity. Selected bond lengths [Å] and angles [ ]: 10: Pt1–Sn1 2.502(1), Sn1–Br1 3.155(1), Sn1–Br2 3.264 (1), Sn1–Br3 2.956(1), P1–Pt–P2 158.0(1), P1–Pt1–Sn1 93.6(1), P2–Pt1–Sn1 106.7(1); 11: Pt1–Pb1 2.564(1), Pb1/Cl(WCA) 3.001(1); P10–Pt1–P1 162.23(4), Cl1–Pb1–Cl10 177.90(4), P1–Pt1–Pb1–Cl1 77.65(3), Pt1– Pb1–Cl1–Al1 143.81(4). Scheme 5 Formation of the anionic complex 16. Fig. 4 Molecular structure of the anion of 16. Ellipsoids are drawn at the 50% probability level. Hydrogen atoms, countercations ([nBu4N] ), solventmolecules and ellipsoids of the cyclohexyl rings are omitted for clarity. Selected bond lengths [Å] and angles [ ]: Pt1–Sn1 2.604(1), Sn1– Br1 2.698(2), Sn1–Br2 2.675(2), Sn1–Br3 2.718(1), P1–Pt–P2 145.1(1), P1–Pt1–Sn1 111.4(2), P2–Pt1–Sn1 102.3(1). Scheme 6 Reactivity of the new cationic PtPb complexes. Chemical Science Edge Article O pe n A cc es s A rt ic le . P ub lis he d on 1 5 O ct ob er 2 01 4. D ow nl oa de d on 0 7/ 09 /2 01 7 10 :2 0: 34 . T hi s ar tic le is li ce ns ed u nd er a C re at iv e C om m on s A ttr ib ut io nN on C om m er ci al 3 .0 U np or te d L ic en ce . View Article Online from its P{H} NMR spectrum (17: d 1⁄4 47.1 ppm; JP–Pt 1⁄4 3370 Hz), which showed an increase of the coupling constant with respect to that of the starting material (8a: d 1⁄4 46.6 (JP–Pt 1⁄4 3100 Hz) ppm). Additionally, we were able to see 430 | Chem. Sci., 2015, 6, 425–435 the respective picoline resonances, with appropriate integration, in the H and C{H} NMR spectra, slightly shied with respect to free picoline. To further investigate the reactivity of our cationic plumbylene complexes we attempted halide exchange reactions. This journal is © The Royal Society of Chemistry 2015 Edge Article Chemical Science O pe n A cc es s A rt ic le . P ub lis he d on 1 5 O ct ob er 2 01 4. D ow nl oa de d on 0 7/ 09 /2 01 7 10 :2 0: 34 . T hi s ar tic le is li ce ns ed u nd er a C re at iv e C om m on s A ttr ib ut io nN on C om m er ci al 3 .0 U np or te d L ic en ce . View Article Online Addition of NaI to a THF solution of 8a led to a halide exchange reaction at the Pb atoms and isolation of the new dinuclear dicationic complex [(Cy3P)2Pt(PbI)]2[AlCl4]2 (18) (Scheme 6). The P{H} NMR spectrum of 18 reveals a sharp singlet at d 1⁄4 45.5 ppm (JP–Pt 1⁄4 3130 Hz). Single crystal X-ray determination showed the same constitution as compound 8a. The Pt–Pb and Pb–I bond lengths in 18 are 2.61(1) Å and 2.917(1)/3.092(1) Å, respectively (see ESI†). All bond lengths are slightly elongated with respect to the chloro precursor 8a, as expected due to the larger covalent radius of iodide. In contrast to 8a, the reaction between 8b and [NnBu4]Br in THF resulted in a MOLP containing a lead atom with mixed halogens [(Cy3P)2Pt–PbClBr] (19) instead of a putative halogen exchange product (Scheme 6). Despite the absence of structural evidence, the formation of 19 was apparent from the P{H} and Pt{H} NMR spectra, which show a singlet at d 1⁄4 48.6 ppm (JP–Pt 1⁄4 3520 Hz) and a triplet at d 1⁄4 3950 ppm, respectively. These data show very little difference to those of the dichloro analogue 2 (P: d 1⁄4 48.3 (JPt–P 1⁄4 3450 Hz) ppm, Pt: d 1⁄4 4025 ppm). During crystallization from toluene, a disproportionative halide exchange takes place, leading to the formation of two neutral species, [(Cy3P)2Pt–PbX2] (X 1⁄4 Cl, Br), which were conrmed by X-ray structure analysis as well as by P{H} NMR spectroscopy. This serendipitous nding serves as further evidence for the formation of complex 19. Fig. 5 Above: optimized geometries of 2, 8a, 11 and 11a at the RID3(BJ)-BP86/def2-TZVP + def2-QZVP level. Bond lengths are given in Å, angles in ( ) and energies in kcal mol ; below: plot of (a) HOMO-8 of 8a and (b) HOMO-5 of 11 (eV) at the D3(BJ) + BP86/TZ2P//RID3(BJ)-BP86/def2-TZVP + def2-QZVP level. DFT calculations on the PtPb complexes 2, 8a and 11 In order to gain a deeper understanding of the bonding situation in the PtPb complexes 2, 8a and 11, quantum chemical calculations were carried out at the RI-D3(BJ)-BP86 level using def2-TZVP for the non-metal atoms and def2-QZVP for the Pt and Pb atoms. The optimized geometries of the complexes are shown in Fig. 5, including the most important bond lengths and bond angles. The calculated bond lengths agree well with the experimental bond lengths when the intermolecular dispersion interactions, and especially the damping factor (BJ), have been taken into account, as its importance has been previously reported by Grimme. A constrained geometry optimization was performed for 11 to further analyze the Pt–Pb bond because its non-constrained optimization converged to a structure in which two chloride atoms Cl2 and Cl20 are bound to Pb1 (11a, Fig. 5). This is presumably a result of the absence of crystal packing effects, as the chloride atoms are presumably stabilized by the dielectric eld created by the positively charged protons on the cyclohexyl groups of the surrounding molecules. We discuss the interaction of Cl2 and Cl20 with Pb1 in detail in the bond analysis section below. NBO analysis shows that the Pt atom always carries a negative charge in these complexes ( 0.483, 0.420 and 0.485e in 2, 8a and 11, respectively), which means there is a signicant charge donation from the ligands to Pt, especially from the strong PCy3 donor ligands. However, the partial charges of the Pb atoms are always positive (1.174, 1.124 and 1.070e in 2, 8a and 11, respectively), and remain relatively constant upon successive halide abstraction and increase of net positive charge. This suggests that build-up of positive charge This journal is © The Royal Society of Chemistry 2015 created by the removal of halides from Pb is effectively compensated by the connected groups, which, given the relatively constant charge on Pt, presumably stems from the short contacts with nearby halides. We investigated the nature of the Pt–Pb bonding in the complexes 2, 8a and 11 with an energy decomposition analysis (EDA-NOCV) using two different fragmentation patterns, D and E. D describes donor–acceptor interactions whereas E describes an electron sharing bonding scheme between the Pt and Pb fragments. The EDA-NOCV data given in Table 2 suggest Chem. Sci., 2015, 6, 425–435 | 431 Table 2 EDA-NOCV results of the analysis of Pt–Pb bonds in 2, 8a and 11 at the D3(BJ) + BP86/TZ2P//RI-D3(BJ)-BP86/def2-TZVP + def2-QZVP level (kcal mol ) Compound 2 2 8a 8a 11 11 Fragmentation pattern D E D E D E Fragment 1 [(Cy3P)2Pt] [Cy3P]2Pt + [(Cy3P)2Pt] [(Cy3P)2Pt] + [(Cy3P)2Pt] [(Cy3P)2Pt] + Fragment 2 [PbCl2–PbCl2– Pt(PCy3)2] [PbCl2–PbCl2– Pt(PCy3)2] [PbCl–PbCl– Pt(PCy3)2] 2+ [PbCl–PbCl– Pt(PCy3)2] + [Cl4Al–Pb– AlCl4] [Cl4Al–Pb– AlCl4] DEint 68.3 176.0 128.1 77.1 120.5 163.9 DEPauli 146.9 169.5 173.7 192.9 200.5 249.7 DEelstat a 104.7 (48.7%) 183.1 (53.0%) 124.2 (41.2%) 112.7 (41.7%) 139.9 (43.6%) 246.5 (59.6%) DEdisp a 44.7 (20.8%) 44.7 (12.9%) 50.1 (16.6%) 50.1 (18.5%) 49.8 (15.5%) 49.8 (12.0%) DEorb a 65.7 (30.5%) 117.7 (34.1%) 127.5 (42.2%) 107.3 (39.7%) 131.4 (40.9%) 117.4 (28.4%) DEs b Pt / Pb 32.6 (49.6%) — 65.1 (51.1%) — 60.2 (45.8%) — DEs b Pt–Pb — 87.2 (74.1%) — 67.5 (62.9%) — 70.1 (59.7%) DEpk b Pt / Pb 7.1 (10.8%) 2.4 (2.0%) 9.7 (7.6%) 10.6 (9.9%) 9.7 (7.4%) 17.2 (14.7%) DEpt b Pt / Pb 4.8 (7.3%) 4.8 (4.1%) 15.4 (12.1%) 6.2 (5.8%) 25.4 (19.3%) 6.4 (5.5%) DEs b Pb / Pt 3.8 (5.8%) — 5.3 (4.2%) — 4.0 (3.0%) — Rest 17.4 (26.5%) 23.3 (19.8%) 32.0 (25.0%) 23.0 (21.4%) 32.1 (24.5%) 23.6 (20.2%) DEprep 12.9 5.1 37.5 17.3 52.9 12.3 De 55.4 170.9 90.6 59.8 67.6 151.6 a The values in parentheses give the percentage contribution to the total attractive interactions DEelstat + DEdisp + DEorb. b The values in parentheses give the percentage contribution to the total orbital interactions DEorb. c D: donor–acceptor bonding model; E: electron sharing bonding model. Chemical Science Edge Article O pe n A cc es s A rt ic le . P ub lis he d on 1 5 O ct ob er 2 01 4. D ow nl oa de d on 0 7/ 09 /2 01 7 10 :2 0: 34 . T hi s ar tic le is li ce ns ed u nd er a C re at iv e C om m on s A ttr ib ut io nN on C om m er ci al 3 .0 U np or te d L ic en ce . View Article Online that the Pt–Pb bond in 2 is dative in nature, whereas those of 8a and 11 are electron-sharing bonds. The DEorb value of 2 in model D ( 65.7 kcal mol ), where both the singlet state of [(Cy3P)2Pt] and [PbCl2–PbCl2–Pt(PCy3)2] were used as interacting fragments, has a weaker interaction than model E (DEorb 1⁄4 117.7 kcal mol ), where doublet states of [(Cy3P)2Pt] and [PbCl2–PbCl2–Pt(PCy3)2] were used as interacting fragments. Here we applied the rule that the best description of a chemical bond comes from the interacting fragments that give the weakest orbital interaction (DEorb), as suggested by Frenking based on his work on carbodiphosphoranes in which the same charge and spin states of the fragments were compared, in addition to his other work in which different charge and spin states of the fragments were compared for the analysis of M–E (M 1⁄4 Fe, Ru, Os; E 1⁄4 C–Sn) and E–C bonds (E 1⁄4 Be, B, C, N, O).73 The DEorb values of 8a ( 107.3 kcal mol ) and 11 ( 117.4 kcal mol ) in model E are signicantly lower than those of 8a ( 127.5 kcal mol ) and 11 ( 131.4 kcal mol ) in model D. Therefore, the Pt–Pb bonds in 8a and 11 are electronsharing in nature, whereas that of 2 is a dative bond. This difference in bonding upon halide abstraction suggests that the electron pair donated by Pt to Pb in 2 is converted to a fullycovalent bonding s orbital in 8a and 11. The analysis of the Kohn–Sham molecular orbitals reveals that the HOMO-8 of 8a (a, Fig. 5) and HOMO-5 of 11 (b, Fig. 5) both exhibit s orbitals formed from a d orbital of Pt and a p orbital of Pb. We have analyzed the Pb–Cl bonding in 2, 8a and 11 with the EDA-NOCVmethod using singlet [(Cy3P)2Pt–PbCl2] for 2, singlet [(Cy3P)2Pt–PbCl] + for 8a and singlet [(Cy3P)2Pt–Pb] 2+ and [AlCl4/AlCl4] 2 for 11 as interacting fragments. The EDA-NOCV 432 | Chem. Sci., 2015, 6, 425–435 results (Table 3) show that the orbital (covalent) interactions DEorb are stronger ( 67.7 kcal mol ) than the electrostatic (ionic) interactions DEelstat ( 48.7 kcal mol ) in 8a. However, the electrostatic interactions are stronger in 2 ( 43.0 kcal mol ) and 11 ( 257.6 kcal mol ) than the corresponding orbital interactions in 2 ( 26.6 kcal mol ) and 11 ( 69.9 kcal mol ) (Table 3). The breakdown of DEorb into contributions from orbitals having different symmetry reveals that the dominant contribution comes from Cl / Pb s-donation. The Pb / Cl s backdonation is very weak for 8a (11.1%) and almost negligible for 2 (4.9%) and 11 (1.7%). The crystal structure of 11 shows two additional Pb1/Cl2 and Pb1/Cl20 short contacts (3.592 Å) and non-constrained gas phase optimization (11a, Fig. 5) converged to a structure in which these short contacts strengthen (3.205 Å), the structure thereby becoming similar to the structure of [(1,2-C6H4Me2)2Pb][AlCl4]2 (F) reported by Frank et al. According to the EDA-NOCV method, the sum of the orbital interaction energies of the two Pb1/Cl2 and Pb1/Cl20 contacts is very weak (11.9 kcal mol ), indicating that each of these contacts has a roughly 6.0 kcal mol 1 orbital interaction energy. These two chloride atoms in the crystal are presumably stabilized by the positively charged protons on the cyclohexyl groups of the surrounding molecules. In the absence of crystal packing effects, Pb1/Cl2 and Pb1/Cl20 interactions strengthen and the non-constrained gas phase optimized geometry 11a becomes similar to the aforementioned structure reported by Frank et al. The bond dissociation energies decrease to 4.7 kcal mol 1 for 2 and +25.3 kcal mol 1 for 8a without including dispersion energies, which shows the importance of introducing the dispersion interactions. It is clear that This journal is © The Royal Society of Chemistry 2015 Table 3 EDA-NOCV results of the analysis of Pb–Cl bonding in 2, 8a and 11 at the D3(BJ) + BP86/TZ2P//RI-D3(BJ)-BP86/def2-TZVP + def2QZVP level (kcal mol )