Potassium 15-crown-5 phenanthrenide and potassium 18-crown-6 phenanthrenide were synthesized and characterized by powder X-ray diffraction and 39K solid state NMR spectroscopy. While the radical carbanion is very reactive in solution, the crystals are stable and storable under inert atmosphere. For 15-crown-5, a sandwich-like complex of potassium is formed with two molecules of crown ether per potassium resulting in a coordination number of 10. For the larger 18-crown-6 ligand, a 1:1 complex is obtained and a coordination number of 6 for the potassium cation. In neither crystal structure solvent molecules are incorporated. The 15-crown-5 compound crystallizes faster and is less soluble in THF as compared to the 18-crown-6 compound. Both compounds form solid phenanthrenide that is easy to handle and can be applied for reduction reactions.
Aromatic radical anions are obtained upon reduction of aromatic molecules with alkali metals. This was first published by Berthelot  and further explored by Schlenk et al. . Since the late fifties of the twentieth century, aromatic radical anions have been used as initiators [3, 4] in macromolecular chemistry and as reducing agent in organometallic , ,  and organic , ,  chemistry. The radical anions are kinetically unstable and highly reactive against moisture and oxygen . Consequently, storage times need to be short and its preparation and usage demand anhydrous and anaerobic conditions. Therefore, isolation and characterization have only been published for a few examples, like K2(naphthalene)2(tetrahydrofuran), or [(K2tetrahydrofuran3)(anthracene)2] [12, 13]. Moreover, crystalline alkali salts of radical carbanions generally contain a high amount of solvent and tend to lose this over time , , .
Alternatively to solvent molecules, complexing agents like amines or crown ethers may be used to produce more stable crystalline compounds , , . Castillo et al.  synthesized salts with a variety of different aromatic radical anions using 18-crown-6 (18c6) as a complexing agent. They crystallized lithium, sodium, and potassium complex cations having one molecule of crown ether as counter ions to the aromatic radical anions. In most cases, the crystals additionally incorporated solvent molecules .
Crown ethers were introduced by C. Pedersen in 1967 and were rewarded with the Nobel Prize in 1987 [18, 19]. Depending on the amount of ether units included in the crown ether, the number of coordination sites and the size of the cavity are different, and thus the selectivity to the different cations varies. Consequently, the stoichiometry of cation to ligand is known to vary with the ratio of the ring size to the cation diameter. If the cation diameter is less than or equal to the cavity size of the crown ether, 1:1 complexes are formed. However, if the cation diameter is larger, 1:2 and 3:2 complexes were found. The potassium cation fits perfectly into the 18-crown-6 but is too big for the 15-crown-5 (15c5) . Therefore, it is known that 1:2 potassium-15c5 complexes are formed in gas, solution, and the solid state , , , .
We synthesized phenanthrenide salts of 15-crown-5 and 18-crown-6 and determined their crystal structures by applying powder X-ray diffraction (PXRD) and 39K solid state NMR spectroscopy.
2 Results and discussion
The reaction of potassium metal and phenanthrene (Phen) in tetrahydrofuran (THF) leads to a dark green solution as reported before for the phenanthrenide anion . After adding two equivalents of 15-crown-5, crystals precipitated very quickly (see Scheme 1). The crystals were filtered and washed with THF. The salt does not degrade when stored in an inert atmosphere. Analogously, two equivalents of 18-crown-6 were added to the potassium phenanthrenide solution and crystals were formed, but only after 1 h. The crystals were filtered, washed with cold THF, and dried under vacuum. The compound is more soluble than the 15-crown-5 compound .
For both compounds, the potassium to phenanthrenide stoichiometry was determined by titration with HCl of the KOH that forms upon addition of water (Table S1). In both cases, a 1:1 ratio was confirmed. The potassium to crown ether ratio was determined by CHN analysis. For the 15-crown-5 salt, a 2:1 ratio of crown ether to potassium was observed, while the 18-crown-6 formed a 1:1 complex with potassium (Table S2).
The diameter of the crown ether cavity is 1.7–2.2 and 2.6–3.2 Å for the 15-crown-5 and the 18-crown-6, respectively . The potassium cation has a diameter of about 2.76 Å , therefore it fits perfectly into the cavity of 18-crown-6 but is too big for the cavity of 15-crown-5. For stabilization, a second 15-crown-5 molecule has to shield the remaining coordination sites, creating a sandwich structure.
Because of the low stability of potassium phenanthrenide solutions, we failed to obtain single crystals of the compounds. Therefore, the structures were solved by using PXRD data. Indexing of the powder diffractogram yielded a monoclinic unit cell for the 15-crown-5 and an orthorhombic unit cell for the 18-crown-6 salt. The relevant crystallographic parameters for both structures are given in Table 1.
The CIF files are deposited in the Cambridge Structural Database with the deposition numbers 2126102 and 2126103 for K(15c5)2Phen and K(18c6)Phen, respectively.
For structure solutions applying a simulated annealing algorithm, the molecular ions were included as rigid bodies. The geometries of the phenanthrenide anion and of the 18-crown-6 potassium complex were taken from Castillo et al. . A model of the 15-crown-5 potassium 2:1 complex was taken from Pedersen et al. . The structure models obtained by simulated annealing were then geometrically optimized with density functional theory (DFT) methods prior to the Rietveld refinement.
Refinement confirmed a centrosymmetric structure in space group P21/c for the 15-crown-5 salt. In the last refinement step, the position and orientation of the two crown ether rings, the potassium cation, and the phenanthrenide anion were all refined individually. Within each moiety, a common displacement factor was assigned to the same atom types. Aromatic and aliphatic hydrogen atoms were, however, assigned a common atomic displacement factor.
A preferred orientation using spherical harmonics of the 4th order was applied to the model . The Rietveld refinement gave a good agreement between the measured and calculated diffractogram (Figure 1) as reflected by an R Bragg of 2.3% and an R wp of 4.2%.
For the 18-crown-6 complex, refinement confirmed a centrosymmetric structure in space group Pnma. Both molecular ions reside on special positions. The K+ cation is located at the inversion center in the origin of the unit cell, while the phenanthrenide anion lies on the mirror plane on ¼ b. In the last refinement step, the position and orientation of the crown ether rings, the potassium cation, and the phenanthrenide anion were all refined individually within the frame of the symmetry restrictions.
The same atomic displacement factor was assigned to all atoms of the phenanthrenide anion and another one to all crown ether atoms. The displacement parameter of the potassium was refined anisotropically. A preferred orientation using spherical harmonics of the 4th order was applied to the model . These refinements (Figure 2) led to an R Bragg of 2.8% and an R wp of 5.5%.
The crystal structure of the 15-crown-5 salt consists of alternating layers of phenanthrenide anions and 1:2 potassium-crown-ether complexes along the a-axis. Within these layers, the molecules are tilted with respect to the a axis. The potassium cation has a coordination number of 10 with K–O distances between 2.84 and 3.24 Å (Figure 1 inset).
The crystal structure of the 18-crown-6 salt also consists of anion and cation layers along the b-axis. The phenanthrenide anions reside on the mirror plane at ¼ b. The potassium crown ether complexes are slightly tilted with respect to the b-axis. The orientation of the tilt is altered in every other row. The coordination number of the potassium cation is six with K–O distances between 2.81 and 2.85 Å (Figure 2 inset). The phenanthrenide to K+ distance is rather short (3.45 Å for the closest carbon atom of the anion), indicating a weak interaction (see NMR discussion) assisting in the stabilization of the structure.
To validate the coordination environment, we performed a 39K (spin I = 3/2) solid state NMR spectroscopy study. Since the 39K quadrupolar interaction is characteristic of the charge distribution around the nucleus, its parameters (quadrupolar coupling constant C Q and asymmetry parameter η Q) are a perfect local probe of the surrounding of the nucleus observed .
The 39K NMR spectrum of the 18-crown-6 complex (Figure 3) exhibits a broad signal with a shape typical of the second order quadrupolar interaction with C Q = 3.5 MHz and η Q = 0.11. These parameters fit quite well with the expected values calculated using CASTEP (C Q = 4.4 MHz, η Q = 0.19), considering that ab-initio calculations very often over-estimate the C Q values since they do not take any possible motion into account. For comparison, we also calculated the C Q and η Q values for an isolated 18-crown-6 potassium complex. In this case, η Q is very similar (0.19), however, C Q is significantly larger (6.4 MHz). This result corroborates a weak phenanthrenide to potassium interaction.
The 39K NMR spectrum of the 15-crown-5 complex, however, only exhibits a narrow (350 Hz) Lorentzo-Gaussian shaped signal. This is in complete contradiction to the CASTEP calculated parameter predicting a relatively large quadrupolar interaction (C Q = 5.1 MHz and η Q = 0.88). However, the cavity of that structure, created by the two 15-crown-5 molecules, allows for some degree of freedom for the potassium. At room temperature, this additional motion can thus partially or, in this case, totally average out the quadrupolar interaction, resulting in the observed narrow Lorentzo-Gaussian shape.
To probe for potential alternative potassium crown ether stoichiometries, the molar ratios of the reactants applied during synthesis were varied. For 15-crown-5, even when reducing the amount of crown ether to a 1:1 ratio, only the 1:2 complex crystallizes. The excess potassium crystallizes in a side phase (Figure S1). The excess crown ether in the 18-crown-6 salt can be washed out with THF (Figure S2).
Phenanthrenide radical anions may be stabilized as potassium crown ether salts of 1:2 (15-crown-5) and 1:1 (18-crown-6) stoichiometry. The salts can be handled conveniently and stored under inert atmosphere. Hence, this potent reductant can be put on the shelf for extended periods.
4 Experimental section
General: All experiments were performed under inert atmosphere in a glovebox or using conventional Schlenk techniques. Potassium, 15-crown-5, 18-crown-6, and phenanthrene were purchased from Sigma-Aldrich. The THF solvent was dried over molecular sieve (4 Å) and distilled. The pre-dried THF was stored over potassium and freshly distilled prior to use. 15-crown-5 were dried over molecular sieves (4 Å).
Synthesis of K(15c5) 2 Phen and K(18c6)Phen: Freshly cut potassium (14.4 mmol) and phenanthrene (14.3 mmol) were dissolved in THF overnight. The solution turned to a dark green color. Dried 15-crown-5 or 18-crown-6 (28.8 mmol) was added to the reaction mixture under stirring. The dark potassium crown ether phenanthrenide salt was formed within an hour. After washing the salt with (cold) THF, it was dried under vacuum (10−3 mbar) for at least 12 h.
Titration: For determination of the total potassium content, 40–50 mg of the salt were quenched in water and titrated with HCl (0.05 m). The concentration of phenanthrenide was determined by reacting the salt (40–50 mg) with dibromoethane (1 mL) under argon for 15 min, followed by quenching with water (1 mL) and titration with HCl (0.05 m). Phenolphthalein was used as indicator in both cases.
CHN analysis: An Elementar Unicode instrument equipped with a combustion tube filled with tungsten(VI) oxide granules was used at a combustion temperature of 1050 °C.
X-ray structure determination: In a glovebox, the dried powder sample was mortarted, filled in glass capillaries with 1 mm diameter (Hilgenberg) and sealed. Measurement of XRD patterns was performed in transmission mode on a STOE STADI P powder diffractometer (CuK α1 radiation, GE monochromator, linear position-sensitive detector). Indexing, simulated annealing and Rietveld refinement were done with the software package TOPAS . Models were built with jMol .
DFT calculations: The geometry optimizations [30, 31] on DFT level were carried out with the software package CASTEP  using the PBE functional and the Tkatchenko-Scheffler dispersion correction scheme . An electronic cut off energy of 900 eV and a Monkhorst k point grid spacing of 0.07 Å–1 were used . Calculations of the NMR parameters were also carried out with CASTEP [35, 36].
Solid state NMR spectroscopy: NMR spectroscopic experiments were performed on a Bruker Avance III 600 NMR Spectrometer (600.15 MHz). 39K spin echo experiments were performed at 28.0 MHz in a 3.2 mm ZrO2 rotor at 20 kHz MAS frequency with a nutation frequency of 64 kHz. During acquisition 1H nuclei were decoupled using the Spinal64  decoupling sequence with a nutation frequency of 65 kHz.
5 Supporting information
Results for the titration and elemental analysis of the phenanthrenide salts. Comparison of the powder diffraction patterns of the salts after adding 1 eq and 2 eq crown ether.
Dedicated to Professor Christian Näther on the occasion of his 60th birthday.
We thank Anna-Maria Dietel for CHN analysis of the samples. The authors thank Renee Timmins for English proofreading. N.E. and K.Z. thank the Elite Network of Bavaria (ENB) for a research fellowship and the elite study program “Macromolecular Science” for a fellowship.
Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.
Research funding: None declared.
Conflict of interest statement: The authors declare no conflicts of interest regarding this article.
23. Valentyn, H. G., Rudzevich, L., Miqueu, K., Sotiropoulos, J.-M., Pfister-Guillouzo, G., Romanenko, V. D., Bertrand, G. Angew. Chem. Int. Ed. 2002, 41, 3.Search in Google Scholar
27. Fernandez, C., Pruski, M. Probing Quadrupolar Nuclei by Solid-State NMR Spectroscopy: Recent Advances; Springer: Berlin, Heidelberg, 306, 2011; pp. 119–188.10.1007/128_2011_141Search in Google Scholar PubMed
32. Clark, S. J., Segall, M. D., Pickard, C. J., Hasnip, P. J., Probert, M. I. J., Refson, K., Payne, M. C. Z. Kristallogr. 2005, 220, 567–570; https://doi.org/10.1524/zkri.220.5.567.65075.Search in Google Scholar
35. Bonhomme, C., Gervais, C., Babonneau, F., Coelho, C., Pourpoint, F., Azaïs, T., Ashbrook, S. E., Griffin, J. M., Yates, J. R., Mauri, F., Pickard, C. J. Chem. Rev. 2012, 112, 5733–5779; https://doi.org/10.1021/cr300108a.Search in Google Scholar PubMed
The online version of this article offers supplementary material (https://doi.org/10.1515/znb-2021-0177).
© 2021 Walter de Gruyter GmbH, Berlin/Boston