Atom Exchange Versus Reconstruction: (GexAs4−x)x− (x=2, 3) as Building Blocks for the Supertetrahedral Zintl Cluster [Au6(Ge3As)(Ge2As2)3]3−

Abstract The Zintl anion (Ge2As2)2− represents an isostructural and isoelectronic binary counterpart of yellow arsenic, yet without being studied with the same intensity so far. Upon introducing [(PPh3)AuMe] into the 1,2‐diaminoethane (en) solution of (Ge2As2)2−, the heterometallic cluster anion [Au6(Ge3As)(Ge2As2)3]3− is obtained as its salt [K(crypt‐222)]3[Au6(Ge3As)(Ge2As2)3]⋅en⋅2 tol (1). The anion represents a rare example of a superpolyhedral Zintl cluster, and it comprises the largest number of Au atoms relative to main group (semi)metal atoms in such clusters. The overall supertetrahedral structure is based on a (non‐bonding) octahedron of six Au atoms that is face‐capped by four (GexAs4−x)x− (x=2, 3) units. The Au atoms bind to four main group atoms in a rectangular manner, and this way hold the four units together to form this unprecedented architecture. The presence of one (Ge3As)3− unit besides three (Ge2As2)2− units as a consequence of an exchange reaction in solution was verified by detailed quantum chemical (DFT) calculations, which ruled out all other compositions besides [Au6(Ge3As)(Ge2As2)3]3−. Reactions of the heavier homologues (Tt2Pn2)2− (Tt=Sn, Pb; Pn=Sb, Bi) did not yield clusters corresponding to that in 1, but dimers of ternary nine‐vertex clusters, {[AuTt5Pn3]2}4− (in 2–4; Tt/Pn=Sn/Sb, Sn/Bi, Pb/Sb), since the underlying pseudo‐tetrahedral units comprising heavier atoms do not tend to undergo the said exchange reactions as readily as (Ge2As2)2−, according to the DFT calculations.


General
All manipulations and reactions were performed under dry Ar atmosphere using standard Schlenk or glovebox techniques. All solvents were dried and freshly distilled prior to use, crypt-222 (Merck) 1 was dried in vacuo for at least 18 hours. [K(crypt-222)]2(Tt2Pn2)·en (Tt/Pn = Ge/As, Sn/Bi, Sn/Sb and Pb/Sb) were prepared according to literatures, respectively. 2 [(PPh3)AuMe] was commercially available from Aldrich. Samples were shielded from ambient light throughout cluster syntheses.

Single Crystal X-ray Diffraction
The data for the X-ray structural analyses were collected at T =100.0 K with Mo-Kα-radiation (λ = 0.71073 Å) on a Bruker D8Quest with a CMOS detector for 1 and 3, and on an area detector systems Stoe IPDS/2T for 4, and at T = 100(2) K with Cu Kα-radiation (λ = 1.5418 Å) on a STOE STADIVARI with an image plate detector for 2. The structure was solved by methods of SHELXT from SHELXL-2018/136, 3 and refined by full matrix least-squares methods against F 2 with the SHELXL program. 4 All hydrogen atoms were kept riding on calculated positions with isotropic displacement parameters U = 1.2 Ueq of the bonding partners. In the refinement of compounds 1 and 5, atoms Ge/As or Sn/Sb were assigned according to the result of the perturbation theory results based on the DFT investigations. The relatively large tendency for disorder of the crypt-222 molecules in compounds 2 and 3 required the application of restraints to this large number of atoms (the alternative application of the back-Fouriertransform method did not produce more reliable results   Figure S5. Unit cell of compound 1.  Figure S6. Unit cell of compound 2.  Figure S7. Unit cell of compound 3.  Figure S8. Unit cell of compound 4.

Energy Dispersive X-ray Spectroscopy (EDS) of 1
EDS analysis of was carried out using an EDS-device Voyager 4.0 of Noran Instruments coupled with an electron microscope CamScan CS 4DV. Data acquisition was performed with an acceleration voltage of 25 kV and 100 s accumulation time. Results are summarized in Table  S6 and illustrated in Figure S9. Note that the deviation of the K content is observed very frequently in EDS analyses of these very air-sensitive Zintl cluster compounds.  Figure S9. EDS analysis of 1.

Micro-X-ray Fluorescence Spectroscopy (µ-XFS) Analysis
All µ-XFS measurements were performed with a Bruker M4 Tornado, equipped with an Rhtarget X-ray tube and a Si drift detector. The emitted fluorescence photons are detected with an acquisition time of 100 s. Quantification of the elements is achieved through deconvolution of the spectra. Results are summarized in Table S7. Figures S10-S12 present the spectra for 2-4 along with the results of the deconvolution algorithm. Several measurements produced unreasonably large values for the % K. Removal of K from the calculations afforded excellent S10 agreement with the expected atomic ratio of close to Au2.00Sn10.00Pn6.00 in 2-4. We assume that accumulation of K at the crystal surface upon exposure to air is responsible for the anomalous results. This is observed commonly for very these very air-sensitive compounds.  Figure S10. Micro X-ray fluorescence spectrum of 2 with the results of the deconvolution algorithm. Colors are used as follows: K (pink), Au (red), Sn (light blue), Bi (green). S11 Figure S11. Micro X-ray fluorescence spectrum of compound 3 with the results of the deconvolution algorithm. Colors are used as follows: K (light blue), Au (red), Sn (green), Sb (yellow).

Methods
All mass spectra were recorded with a Thermo Fischer Scientific Finnigan LTQ-FT spectrometer in negative ion mode. We prepared a fresh reaction solution of the reaction leading to compound 1 in DMF, and solutions of single crystals of the compounds 2-4 in freshly distilled DMF inside a glovebox. The solutions were injected into the spectrometer with gastight 250 µL Hamilton syringes by syringe pump infusion. All capillaries within the system were washed with dry DMF for 2 hours before and at least 10 minutes in between measurements to avoid decomposition reactions and consequent clogging.
The following ESI parameters were used: Spray Voltage

Mass spectra of the reaction solution yielding 1
ESI-MS measurements on single crystals of 1 failed, which we attribute to its larger molar mass in combination with a rather coordination-type bonding. Yet, we were able to identify important species in the reaction solution, which indicate the co-existence of (Ge2As2) 2− (in protonated, monoanionic form), (Ge3As) 3− , and also (GeAs3) − . The latter is not part of the cluster anion, yet must form alongside of (Ge3As) 3− , according to equation (1) provided in the main document. Hence, we can take the spectra as further indication of the correctness of our findings. The high resolution images of the said species are provided in Figures S13-S17. Note that ESI mass spectra of Zintl anions always show monoanions, even though the original charge was higher.   Figure S14. High-resolution ESI mass spectrum in negative ion mode recorded immediately upon injection of a fresh reaction solution of 1 in DMF, indicating the coexistence of (Ge3As) − and (GeAs3) -. Topmost: measured, below: simulated. Figure S15. High-resolution ESI mass spectrum in negative ion mode recorded immediately upon injection of a fresh reaction solution of 1 in DMF, indicating the coexistence of (Ge3As) − and (Ge2As2H) -. Topmost: measured, below: simulated.  Figure S16. High-resolution ESI mass spectrum in negative ion mode recorded immediately upon injection of a fresh reaction solution of 1 in DMF, indicating the coexistence of (Ge3As) − and (Ge2As2H) -, and (GeAs3) − . Topmost: measured, below: simulated.         Figure S22. High-resolution ESI mass spectrum in negative ion mode of (AuSn4Bi2) -, recorded immediately upon injection of a fresh solution of 2 in DMF. Top: measured, bottom: simulated.      Figure S26. High-resolution ESI mass spectrum in negative ion mode of (AuSn5Sb6) -, recorded immediately upon injection of a fresh solution of 3 in DMF. Top: measured, bottom: simulated.      Figure S30. High-resolution ESI mass spectrum in negative ion mode of (AuPb4Sb4) -, recorded immediately upon injection of a fresh solution of 3 in DMF. Top: measured, bottom: simulated.

Methods
All calculations were undertaken by employing density functional theory (DFT) methods as implemented in the program system TURBOMOLE V7.1.1. 6,7 We applied the TPSS functional. 8 The used basis sets were of the quality def2-TZVP (for the exchange reactions) and dhf-TZVP (for the study of the four dimeric anions [(AuPb5Bi3)2] 4− and 2-4). 9 We additionally used the corresponding auxiliary basis sets 10 and effective core potentials at Sn, Pb, Sb, Bi, and Au atoms. 11 The bonding situationin particular that of the Au atoms − was examined by means of population analyses based on occupation numbers (Paboon) 12 and upon inspection of localized molecular orbitals (LMOs). 13 The pictures of the LMOs were created with gOpenMol. 14  Figure S32 shows the calculated minimum structure of {[AuPb5Bi3]2} 4− and gives the corresponding structural data. The atom labelling corresponds to Figure 5 in the manuscript.

Bonding situation of the Au atoms
The results of the Paboon analysis are provided in Table S10. Additional aurophilic interactions between the two Au atoms may be present, but the fact that the two subunits hold together, obviously may be rationalized without them. Figure S33. Localized molecular orbital of the 3-center interaction between the atoms Pb1, Bi2, and Au1 (Pb: yellow, Pb: blue, Au: purple; contours drawn at ±0.048 a.u.). S25 Figure S34. Localized molecular orbital of the 3-center interaction between the atoms Pb5 i , Bi2 i , and Au1 i (Pb: yellow, Pb: blue, Au: purple; contours drawn at ±0.048 a.u.). Figure S35. Localized molecular orbital of the 3-center interaction between the atoms Pb1 i , Pb2 i , and Au1 (Pb: yellow, Pb: blue, Au: purple; contours drawn at ±0.048 a.u.).