SynthesesThe used syntheses for the nitrosyl, silver and [NBu4]+ salt were already described in earlier publications of our group36,102,103. Nonetheless, it is worth mentioning again that the synthesis for the nitrosyl salt could be done in a at least 10 gram scale with regard to the lithium salt, using the commercial NO+[BF4]− salt and without lithium impurities in the product36,104. The synthesis of the solvent-free silver salt was a great progress to avoid potential harmful dichloromethane in further reactions25,29,102. Additionally, we could develop a new synthetic route for the known Fc+ salt using the nitrosyl, instead of the silver salt as oxidizing agent, to avoid colloidal silver which could be disturbing during further experiments. For a better allocation in the following NMR spectra, instead of the abbreviation [pf]− the detailed description [Al{OC(CF3)3}4]− will be used.Nitrosonium tetrakis(nonafluorotertbutanolato)aluminate(III) (NO+[pf]−)Caution! This procedure involves the work with liquified sulfur dioxide, which has a vapor pressure of ca. 4 bar at room temperature. Therefore, the synthesis requires trained personnel and proper equipment.NO+[Al{OC(CF3)3}4]− was synthesized based on the procedure published in ref. 36: Li+[pf]− (9.74 g, 10 mmol, 1.0.eq) and NO+[BF4]− (1.75 g, 15 mmol, 1.5 eq.) were filled in one side of a double-bulb Schlenk vessel equipped with a G4 frit plate. Sulfur dioxide (10 mL) was condensed onto the mixture of the reagents at −78 °C. The vessel was brought to room temperature and stirred for 7 days. Afterward, the solution was filtered and the solvent was removed by vacuum. NO+[pf]− was obtained as a colorless solid (8.86 g, 89%).Characterization1H-NMR (300.18 MHz, CD2Cl2, calibration to CHDCl2 = 5.32 ppm105, 298 K): No signals observed.7Li-NMR (116.66 MHz, CD2Cl2, 298 K): No signals observed.11B-NMR (96.31 MHz, CD2Cl2, 298 K): No signals observed.14N-NMR (21.69 MHz, CD2Cl2, 298 K): δ = 364.5 (s, NO+, 1 N) ppm.19F-NMR (282.45 MHz, CD2Cl2, 298 K): δ = −75.7 (s, [Al{OC(CF3)3}4]−, 36 F) ppm.27Al-NMR (78.22 MHz, CD2Cl2, 298 K): δ = 34.7 (s, [Al{OC(CF3)3}4]−, 1Al) ppm.FTIR (ZnSe, ATR):ν/cm−1 = 2342 (vw), 1354 (vw), 1301 (m), 1248 (vs), 1209 (vs), 968 (vs), 863 (vw), 830 (vw), 757 (vw), 726 (vs), 642 (vw), 568 (vw), 560 (vw).FT Raman (1000 scans, 250 m0W):ν/cm−1 = 2937 (vw), 2756 (vw), 2340 (m), 1355 (vw), 1315 (vw), 1275 (w), 1248 (vw), 1204 (vw), 1120 (vw), 975 (vw), 829 (vw), 815 (vw), 796 (vs), 746 (s), 571 (w), 562 (w), 537 (m), 366 (m), 325 (vs), 290 (m), 237 (m), 208 (w), 174 (w), 118 (m), 100 (m).Silver tetrakis(nonafluorotertbutanolato)aluminate(III) (Ag+[pf]−)Caution! This procedure involves the work with liquified sulfur dioxide, which has a vapor pressure of ca. 4 bar at room temperature. Therefore, the synthesis requires trained personnel and proper equipment.Ag+[Al{OC(CF3)3}4]− was synthesized based on the procedure published in ref. 102 The reaction was performed analogously to the one yielding NO+[pf]−. Instead of NO+[BF4]−, AgF (1.91 g, 15 mmol, 1.5 eq.) was used. Additionally, the reaction was performed in the absence of light. Ag+[pf]− was obtained as a colorless solid (9.03 g, 84%).Characterization1H-NMR (400.17 MHz, CD2Cl2/Et2O, calibration to CHDCl2 = 5.32 ppm105, 298 K): 3.50 (q, 3JHH = 7.0 Hz, O(CH2CH3)2, 2H) and 1.20 (t, 3JHH = 7.0 Hz, O(CH2CH3)2, 3H) ppm.19F-NMR (282.45 MHz, CD2Cl2/Et2O, 298 K): δ = −75.8 (s, [Al{OC(CF3)3}4]−, 36 F) ppm.27Al-NMR (104.27 MHz, CD2Cl2/Et2O, 298 K): δ = 34.6 (s, [Al{OC(CF3)3}4]−, 1Al) ppm.7Li-NMR (116.7 MHz, CD2Cl2/Et2O 2, 298 K): No signals observed.Tetrabutylammonium tetrakis(nonafluorotertbutanolato)aluminate(III) ([NBu4]+[pf]−)[NBu4]+[Al{OC(CF3)3}4]− was synthesized based on the procedure published in ref. 103: Li+[pf]− (19.5 g, 20 mmol, 1.0 eq.) and [NBu4]+Br− (6.45 g, 20 mmol, 1.0 eq.) were dissolved in a mixture of water and acetone (85:15 v/v, 150 mL) at room temperature. The solution was kept at a warm place/heated at 30 °C overnight, allowing the acetone in the solvent to evaporate, yielding a microcrystalline precipitate. The remaining solvent was removed by filtration and the residue was washed with water until all the bromide was removed (test with, e.g., silver nitrate). Afterward, the product was washed two times with hexane (2 × 100 mL). [NBu4]+[pf]− was obtained as a colorless powder (22.9 g, 94%).Characterization1H-NMR (300.18 MHz, CD2Cl2, calibration to CHDCl2 = 5.32 ppm105, 298 K): δ = 3.07 (m, [N(CH2CH2CH2CH3)4]+, 8H), 1.60 (m, [N(CH2CH2CH2CH3)4]+, 8H), 1.43 (m, [N(CH2CH2CH2CH3)4]+, 8H), 1.03 (t, 3JHH = 7.3 Hz [N(CH2CH2CH2CH3)4]+, 12H) ppm.19F-NMR (282.45 MHz, CD2Cl2, 298 K): δ = −75.7 (s, [Al{OC(CF3)3}4]−, 36 F) ppm.27Al-NMR (78.22 MHz, CD2Cl2, 298 K): δ = 34.6 (s, [Al{OC(CF3)3}4]−, 1Al) ppm.7Li-NMR (116.7 MHz, CD2Cl2, 298 K): No signals observed.14N-NMR (21.9 MHz, CD2Cl2, 298 K): No signals observed.Bis(η
5-cyclopentadienyl)iron(III) tetrakis(nonafluorotertbutanolato)aluminate(III) (Fc+[pf]−)NO+[pf]− (1.00 g, 1.01 mmol, 1.00 eq.) and Fc (0.23 g, 1.23 mmol, 1.22 eq.) were weighed, inside a glovebox, in one side of a double-Schlenk tube separated by a G3 or G4 frit and equipped with grease-free PTFE valves. Under reverse flow of Argon, 2FB (1,2-difluorobenzene, 10 mL) was added and led immediately to the formation of NO(g) and a dark blue solution. The solution was stirred at RT overnight and the solvent was removed under vacuo. To remove the excess of ferrocene, the residue was washed with n-hexane (5 mL). Therefore, the ferrocene solution in n-hexane was filtered through the frit and the solvent was condensed back to the side of the crude product, as many times, as the n-hexane solution was still colored yellowish before the filtration. Afterward, the crude product was dried under vacuo (10−3 mbar) to yield a blue powder of Fc+[Al{OC(CF3)3}4]− (0.99 g, 0.86 mmol, 85%).Characterization1H-NMR (300.18 MHz, 1,2-F2C6H4 (2FB), calibration to 1,2-F2C6H4 = 6.96 ppm against Si(CH3)4, 298 K): 33.87 (br. s., [Fe(C5H5)2]+, 10H) ppm.19F-NMR (282.45 MHz, 2FB, 298 K): δ = −75.7 (s, [Al{OC(CF3)3}4]−, 36 F), −139.6 (s, 1,2-F2C6H4, 2 F) ppm.27Al-NMR (78.22 MHz, 2FB, 298 K): δ = 34.7 (s, [Al{OC(CF3)3}4]−, 1Al) ppm.FTIR (ZnSe, ATR):ν/cm−1 = 3126 (vw), 1423 (vw), 1352 (vw), 1299 (w), 1273 (m), 1266 (m), 1253 (m), 1239 (m), 1213 (vs), 1163 (w), 1064 (vw), 1014 (vw), 972 (vs), 856 (w), 832 (vw), 792 (vw), 756 (vw), 728 (vs), 571 (vw).FT Raman (1000 scans, 200 mW):ν/cm−1 = 3133 (vw), 1425 (vw), 1363 (vw), 1304 (vw), 1273 (vw), 1113 (m), 1065 (vw), 851 (vw), 797 (vw), 746 (vw), 562 (vw), 538 (vw), 367 (vw), 321 (w), 299 (vs), 234 (vw), 170 (vw), 120 (vw), 82 (vw).Nitrosonium bis{tris(nonafluorotertbutanolato)aluminum(III)}-(μ
2)-fluoride (NO+[al-f-al]−)Caution! This procedure involves the work with liquified sulfur dioxide, which has a vapor pressure of ca. 4 bar at room temperature. Therefore, the synthesis requires trained personnel and proper equipment.NO+[F{Al(OC(CF3)3)3}2]− was synthesized based on the procedure published in ref. 26. NO+[PF6]− (560 mg, 3.20 mmol, 1.0 eq.) and (H3C)3Si–F–Al(OC(CF3)3)3 (5.04 g, 6.1 mmol, 2.0 eq.) were filled in a Schlenk vessel inside a glovebox. Sulfur dioxide (10 mL) was condensed onto the mixture of the reagents at −78 °C. The vessel was equipped with a bubbler, brought to −35 °C and the temperature was held for 1 h. Subsequently, the reaction solution was slowly warmed to room temperature and the sulfur dioxide evaporated. The white powder was dried at 10−3 mbar for 2 h. [NO]+[al-f-al]− was obtained as a colorless powder (4.36 g, 90%).Silver bis{tris(nonafluorotertbutanolato)aluminum(III)}-(μ
2)-fluoride (Ag+[al-f-al]−)Caution! This procedure involves the work with liquified sulfur dioxide, which has a vapor pressure of ca. 4 bar at room temperature. Therefore, the synthesis requires trained personnel and proper equipment.Ag+[F{Al(OC(CF3)3)3}2]− was synthesized based on the procedure published in ref. 26. The reaction was performed analogously to the one yielding NO+[al-f-al]−. Instead of NO+[PF6]−, Ag+[PF6]− (810 mg, 3.20 mmol, 1.0 eq.) was used. Additionally, the reaction was performed in the absence of light. Ag+[al-f-al]− was obtained as a colorless solid (4.68 g, 92%).Decafluoroanthracene (anthraceneF)AnthraceneF was synthesized based on the procedure published in ref. 106: 9,10-dichlorooctafluoroanthracene (2 g, 5.11 mmol, 1.0 eq.) and KF (0.98 g, 16.9 mmol, 3.3 eq.) were dissolved in a mixture of sulfolane (10 mL) and toluene (20 mL). The reaction mixture was heated for 2 h at 120 °C. Afterward, the toluene was removed from the reaction mixture. The reaction mixture was further heated for 4 h at 210 °C. At room temperature, water (50 mL) was added to the mixture. The mixture was filtrated and the brown residue was washed with water (3 × 20 mL). After drying at room temperature, the residue was taken up in dichloromethane (20 mL) and slowly cooled down to −40 °C. The solution was decanted and anthraceneF was obtained as a yellow to brown solid (0.63 g, 32%).Decafluoroanthracenium bis{tris(nonafluorotertbutanolato)aluminum(III)}-(μ
2)-fluoride ([anthraceneF]+[al-f-al]−)[NO]+[F{Al(OC(CF3)3)3}2]− (0.10 g, 1.0 eq, 66 μmol) and anthraceneF (28 mg, 1.2 eq., 78 μmol) were placed in a Schlenk flask and 1,2,3,4-tetrafluorobenzene (1 mL) was added. The solution instantaneously turned green-blue and a gas formation was observed. The solution was stirred 5 min at room temperature and was then layered with n-pentane (10 mL). Slow diffusion of the solvents over days led to the crystallization of [anthraceneF]+∙[F{Al(OC(CF3)3)3}2]− in blue plates, suitable for scXRD (90 mg, 49 μmol, 74%).Table 3 Properties relevant to apply the combination xFB with deelectronators de[WCA] and [NBu4]+[pf]− as supporting electrolyte saltCharacterizationFTIR (ZnSe, ATR):ν/cm−1 = 1600 (vw), 1517 (vw), 1491 (w), 1468 (vw), 1443 (vw), 1355 (w), 1300 (w), 1265 (s), 1243 (vs), 1211 (vs), 1177 (s), 1118 (w), 1038 (vw), 970 (vs), 957 (s), 863 (w), 810 (vw), 760 (vw), 726 (vs), 716 (m), 664 (w), 631 (w), 569 (w).FT-Raman (1000 scans, 50 mW):ν/cm−1 = 1570 (s), 1548 (s), 1436 (vs), 1415 (s), 1395 (vs), 1294 (s).Decafluorophenanthrenium bis{tris(nonafluorotertbutanolato)aluminum(III)}-(μ
2)-fluoride ([phenanthreneF]+[al-f-al]−)Ag+[F{Al(OC(CF3)3)3}2]− (84 mg, 1.0 eq., 53 μmol), I2 (6.7 mg, 0.5 eq., 27 μmol) and phenanthreneF (19 mg, 1.0 eq., 53 μmol) were placed in a Schlenk flask and 1,2,3,4-tetrafluorobenzene (0.8 mL) was added to the solution. The solution was filtered and brownish crystals suitable for scXRD formed out of the dark solution upon slow removal of the solvent. The product was obtained as dark brown crystalline blocks (78 mg, 42 μmol, 78%).CharacterizationFTIR (ZnSe, ATR):ν/cm−1 = 1673 (vw), 1662 (vw), 1645 (vw), 1584 (vw), 1522 (w), 1495 (m), 1468 (vw), 1433 (vw), 1379 (w), 1354 (vw), 1301 (w), 1265 (m), 1242 (vs), 1213 (vs), 1180 (m), 1105 (vw), 1093 (w), 1069 (vw), 1011 (vw), 972 (vs), 866 (vw), 847 (m), 799 (vw), 727 (vs), 707 (m), 636 (w), 569 (w).Dielectric spectroscopyComplex permittivity spectra were recorded using an Anritsu MS4647A vector network analyzer connected to an open-ended coaxial probe at frequencies ranging from 1 to 70 GHz107. The reflectometer was calibrated using air, conductive silver paint, and N,N-dimethylacetamide108. Samples were placed into a double-walled sample holder connected to a Julabo-E12 thermostat to control the temperature at 5, 10, 18, 25, 33, or 40 °C. All complex permittivity spectra were modeled using a Debye relaxation70 to obtain εs, the low-frequency limit of the permittivity, ε∞, the high-frequency limit of the permittivity, and the dielectric relaxation time τ. Recorded spectra, together with the details of the analysis, are given in Supplementary Note 5.Cyclovoltammetry, general proceduresAll cyclic voltammograms were recorded in an argon-filled glovebox. A three-electrode arrangement was used with a 1 mm Pt disc working electrode (WE), a Pt mesh or wire as counter electrode (CE) and a Pt wire as pseudo reference electrode (RE) for NO+ (Supplementary Notes 3.1), ferrocene (Fc) and N(4-BrC6H4)3 or Ag wire for the Ag+ and NO+ evaluation (triangular) measurements in the Supplementary Notes 3.5.3 / 3.7.3. For all measurements [NBu4]+[pf]− (c = 100 mm) was used as supporting electrolyte. For each measured electroactive sample, c(de+) = 10 mm was used. The RE (Pt|Fc+, Fc or Ag+|Ag reference) was added in a glass compartment with a frit to allow for a direct measurement and was filled with solutions of [NBu4]+[pf]− (100 mm), Fc (10 mm) and Fc+[pf]− (10 mm) or [NBu4]+[pf]− (100 mm) and Ag+[pf]− (10 mm). For each solvent S, a potential stability window (ECW) was recorded in a solution of [NBu4]+[pf]− (100 mm) in S to identify any impurities and to determine the stability range of the pure solvent against the RE used. Scan rates were varied from 20 mV s−1, 50 mV s−1, 100 mV s−1 up to 200 mV s−1, and if not stated otherwise, the half-wave potentials E1/2 did not change with the rate (full details in Supplementary Note 3). Since several of the published potentials in the Geiger/Connelly Review5 may have been afflicted by ion-pairing and other effects, we also measured the Ag+ and NO+ potentials in the like setup, but the solvents CH2Cl2, 1,2-Cl2C2H4, dimethylformamide (DMF), acetonitrile (AN). In addition, the ECWs of nitromethane (MeNO2), propylene carbonate (PC) and tetrahydrofurane (THF) were determined with [NBu4]+[pf]− (100 mm) as supporting electrolyte salt for comparison.CV-evaluation by triangular Born-Fajans-Haber-CyclesThis evaluation is exemplarily shown for the two solvents 5FB and 1,2-Cl2C2H4 in Fig. 7, all other triangular cycles are deposited in the Supplementary Notes 3.5.3 / 3.7.3.Fig. 7: CV-evaluation by triangular Born-Fajans-Haber-Cycles.Measured half-wave potentials of [Fc]+ and [NO]+ versus Ag+(10 mm)/Ag and [NO]+ versus [Fc]+/Fc in exemplarily selected 5FB and DCE = 1,2-Cl2C2H4 solution at a scan rate of 100 mV s−1. Knowing two out of the three values of the measurements, the third can be calculated in a Born-Fajans-Haber-Cycle approach. Hence, |ΔE| and |ΔG| errors can be calculated by using the relation ΔG = −zFΔE, with z = number of electrons, ΔE = potential difference, F = Faraday constant = 96,485 C mol−1.68 The mean errors of the calculated potential difference (ΔE) / the corresponding Gibbs energy difference (ΔG) are given in red in the center of the triangle. Note, E1/2 potentials were rounded to two decimal places and rounding errors can occur.Quantum chemical calculationsAn extended search of the potential energy surface was manually performed to find the lowest energy structures of all particles [de(S)n]+ (de = Ag, NO), {Ag[pf]}ip,solv and {(S)Ag[pf]}ip,solv at the dispersion109,110 corrected (RI-)BP86(D3BJ)/def2-TZVPP85 DFT86 level of theory. Corrections to statistical thermodynamics (ZPE, H° and S°) were taken from the frequency calculations111 at this level. The energies of the DFT structures were refined in a series of DLPNO-CCSD(T) single point calculations87,88,89 with Dunning’s112,113,114,115,116,117 basis sets cc-pVDZ, cc-pVTZ, cc-pVQZ and then extrapolated to the complete basis set limit (CBS). These accurate CCSD(T)/CBS values were used as electronic energies for the calculation of the thermodynamics of all particles, augmented by DFT-corrections to ZPE, H° and S°. Finally, contributions of solvation enthalpies and free energies in S were calculated with the COSMO-RS90,91,92 model at the BP86(D3)/def2-TZVPD//BP86(D3)/def-TZVP level, so that we overall derive the quantities ΔrH°(g) / ΔrG°(g) in the gas phase as well as ΔrH°(solv) / ΔrG°(solv) in solution in S at standard conditions (g: 298 K, 1 bar; solv.: 298 K, a = 1 mol L−1).
Pushing redox potentials to highly positive values using inert fluorobenzenes and weakly coordinating anions
