An Electrochemical Approach toward the Metastable Type II Clathrate Germanium Allotrope

By using an anodic conversion process at 280 °C, the type II clathrates Na1.7(6)Ge136 and Na23.0(5)Ge136 were obtained from Na12Ge17 as the starting material. An alkali-metal iodide molten-salt electrolyte complied with the reaction conditions, allowing for the formation of microcrystalline products. Characterization by powder X-ray diffraction, scanning electron microscopy, and energy-dispersive X-ray spectroscopy also revealed Na4Ge13 as an intermediate and α-Ge and Cs8–xGe136 as byproducts, with the latter likely resulting from cation exchange between the starting material and electrolyte. Taking such minor side reactions and a small contribution of material without suitable electrical contact into account, anodic conversion of Na12Ge17 to Na1.7Ge136 proved to proceed without parasitic processes and to comprise the material bulk. The hitherto existing preparation method for Nax→0Ge136 by gas–solid oxidation of Na12Ge17 has thus been translated into a scalable high-temperature electrochemical approach with enhanced tools for reaction control, promising access to pure Ge(cF136) and Na24Ge136 after process optimization.

S emiconductor materials continue to be of major interest for, e.g., more efficient solar cells, battery electrode materials, or optoelectronic devices. This coincides with fundamental-science investigations on related intermetallic phases 1−3 and on metastable allotropes of semiconductor elements or their derivative compounds, 4−12 particularly also those with clathrate types of structure. 13−24 One prominent representative is the germanium allotrope Ge(cF136) 16,17 (Figure 1), which has been closely linked to the development of redox-preparation methods for metastable allotropes and intermetallic phases. First identified as an oxidation product of Na 12 Ge 17 25 after reaction with an ionic liquid, 16 it has become available in the laboratory scale by gas−solid redox approaches 26,27 and, more recently, by an elaborate thermal degradation technique. 19 However, more easily scalable preparation methods, at best precluding the formation of byproducts by an improved reaction control, are desirable if application as an electronic material 19,28 is further pursued.
Herein, a high-temperature electrochemical redox approach toward Ge(cF136) and the derivative Na 24−δ Ge 136 clathrate is introduced. Again, Na 12 Ge 17 was the starting material. In the course of bulk anodic conversion at 280°C, it depleted in Na and transformed into a microcrystalline type II clathrate material. Depending on gross conversion, Na 24−δ Ge 136 and Na x→0 Ge 136 were found to be the main crystalline products. With a sufficiently low melting point, a mixture of LiI, NaI, and CsI was applied as a molten-salt electrolyte. Aluminum metal served as the cathode material, acting as a selective acceptor for Li + ions ( Figure S1). The formation of elemental alkali metals at the cathode with possible adverse effects 29−31 was not observed. The overall cell reaction is given in eq 1. To control the potential of the Na 12 Ge 17 anode, diamond-type germanium (α-Ge, cF8) was used as a pseudoreference electrode. The potential of metastable Ge(cF136) as the target of an exhaustive oxidation of Na 12 Ge 17 was therefore assumed close to 0 V versus α-Ge. To keep the reference potential stable and unaffected by the cell reaction within the applied threeelectrode setup (Figure 2), an electrolyte melt saturated in CsI and NaI was chosen. The nominal LiI−NaI−CsI composition (mass ratio 3.62:1:5.35) consequently yielded three phases at 280°C: 32 the electrolyte melt and solid NaI and CsI. At constant temperature, liberation of Na + from the anode and consumption of Li + at the cathode thus result in the further precipitation of NaI only (eq 1). Molybdenum served as an inert contacting material applicable in halide melts. 33,34 More experimental details are provided in ref 35 and the Supporting Information.
Gradual anodic conversion of the Na 12 Ge 17 working-electrode material ( Figure S2) was controlled by a chronoamperometric approach: The working-electrode potential was increased stepwise toward the target potential of 0 V (vs α-Ge), starting from the initial open-circuit potential (−0.65 V), and gross conversion was monitored by the transferred charge ( Figure 3). Anodic conversion was allowed to proceed freely at the respective adjusted potential, until falling below a minimum current (conversion rate). Subsequently, the next potential was adjusted, if applicable (see the Supporting Information). In a remarkable analogy to chemical gas−solid oxidation of Na 12 Ge 17 to Ge(cF136) at 280°C, 26,27 anodic conversion proceeded for almost 10 days to completion ( Figure 3). After 20% gross conversion, the target potential of 0 V had to be adjusted. Afterward, a steady reaction was indicated by an almost constant current (slope of the transferred-charge curve) for the next 3 days. At about 80% gross conversion, the current (conversion rate) distinctly decreased, but a steady further reaction at a lower rate was observed for a further 6 days. Such behavior might result from the disappearance of a particular phase, the kinetically less hindered Na depletion of which had been rate-determining by then. Possibly, the lowered rate indicates exclusive Na depletion of the type II clathrate Na x Ge 136 . More detailed studies are needed here.
The phases obtained after complete anodic conversion of Na 12 Ge 17 to 0 V versus α-Ge were elucidated by investigating the working-electrode material after cooling and washing with alcohol and water. The final product was separated with a yield of about 67 mass %, as calculated for complete conversion to Ge(cF136). According to Rietveld refinement 36 of the powder X-ray diffraction (PXRD) data (Figure 4), it consisted mainly of a crystalline type II clathrate phase (about 70 mass %) at the composition Na 1.7(6) Ge 136 . Consistent with the low Na content assigned, the lattice parameter of 15.2272(3) Å determined versus the internal LaB 6 standard 37 was only slightly larger than that observed for Ge(cF136) from gas−solid oxidation 27 and close to that of Na x Ge 136 (x ≈ 4.5) obtained by thermal degradation 19 (Table S1). α-Ge (18 mass %) and a second type     38 However, Na + Cs mixed occupancy might likewise explain the observed scattering power in the hexakaidecahedral cages of that phase, and more elaborate studies are needed to elucidate that issue. Investigations by scanning electron microscopy (SEM) and electron-dispersive X-ray spectroscopy (EDXS) adequately confirmed the presence of Cs, particularly in individual particles (see the Supporting Information), and estimated a suitable overall content in the specimen of about 2 mass %. Likewise, in agreement with PXRD, Na was hardly detectable in the sample, while Ge was the main component (about 98 mass %). The presence of a Cs-containing clathrate actually indicates partial cation exchange of the working-electrode material with the electrolyte. However, former experiments have shown that Cs atoms do not refill the empty clathrate cages of Ge(cF136) at 280°C, although such a reaction is known for Na, K, and Rb. 27 Moreover, thermal degradation of Na 16 Cs 8 Ge 136 stops at Cs 8 Ge 136 , 38 and Cs atoms do not escape. Cation exchange with an already existing crystalline Na-containing clathrate seems thus unlikely. Rather, it should occur with the salt-like Na 12 Ge 17 25 starting material or an intermediate. To avoid Cs contamination, methodical optimization might comprise electrolytes such as NaAlCl 4 , provided that they are inert toward the redox-sensitive starting material. 39 Such an investigation might also disprove the possibility that the residual electron density assigned to 1.7(6) Na atoms in the main phase might actually originate from, e.g., 0.34(1) nonremovable Cs atoms.
An additional PXRD investigation of the clathrate product, by adding LaB 6 as an internal intensity standard, revealed practically complete crystallinity (see the Supporting Information). The product, therefore, does not contain noncrystalline contaminants in significant amounts. This investigation also confirmed that the unidentified crystalline phase (Figure 4), not being accounted for in Rietveld refinement, may be present with a low mass fraction only, in agreement with the weak reflections observed. In conclusion, the phase and sample compositions revealed by PXRD are in line with SEM/EDXS and with gross conversion of 85% based on the transferred charge ( Figure 3). Because the refinement results suggest nearly complete conversion to elemental Ge (Na x→0 Ge 136 and α-Ge), only about 10−15 mass % of the Na 12 Ge 17 starting material should not have reacted. Actually, the presence of unreacted material was indicated by gas formation upon washing with alcohol and water, and the aqueous solution was markedly basic (pH ≈ 11). The anodic conversion is, nevertheless, evidenced to comprise the bulk of the Na 12 Ge 17 starting material, while providing a high selectivity for Na x→0 Ge 136 .
To identify intermediates involved in the anodic conversion process of Na 12 Ge 17 , a reaction was stopped at about 50% gross conversion ( Figure 3). After washing, a type II clathrate of composition Na 23.0(5) Ge 136 was revealed as the main crystalline phase (56 mass %). Different from Na 1.7(6) Ge 136 , the clathrate cages are practically completely Na-filled in that case. The lattice parameter of a = 15.4412(7) Å agreed with the Na-rich Na x Ge 136 compositions reported earlier. 13,27 As minor products, Na x→0 Ge 136 (23 mass %) and the Cs-containing phase (5 mass %) were detected, again, with equal composition and lattice parameters (see the Supporting Information). This finding indicates that the potential of 0 V versus α-Ge determines the final products of anodic conversion. α-Ge (9 mass %) was revealed as a byproduct, again. However, different from the completed anodic conversion, also Na 4 Ge 13 40,41 was detected (8 mass %). Not being present after the completed reaction, Na 4 Ge 13 and Na 24−δ Ge 136 consequently occur as intermediates of anodic conversion of Na 12 Ge 17 toward elemental Ge, which is in line with reports on thermal degradation of Na 4 Ge 4 42 and chemical gas−solid oxidation of Na 12 Ge 17 . 43 However, so far Na 24−δ Ge 136 has never been conserved as the main product. Thermal degradation of epitaxially grown Na 4 Ge 4 at suitably low temperature has provided access only to thin films of that phase. 44 Higher temperatures and rapid conversion applied to prepare bulk products favored the formation of the more temperature-stable Na 4 Ge 13 , which, however, rapidly degraded to α-Ge rather than to Na 24−δ Ge 136 . 42 On the other hand, although gas−solid oxidation of Na 12 Ge 17 at 280°C may lead to small amounts of Na 24−δ Ge 136 after short reaction times, 43 the method does not provide sufficient control of the redox potential to hinder Na depletion of the clathrate during bulk conversion. Once emptied, refilling of the Ge(cF136) cages to Na 24−δ Ge 136 by using Na vapor was similarly hard to control. 27 Anodic conversion of Na 12 Ge 17 is, therefore, the first preparation method with a realistic prospect to access the completely Nafilled type-II germanium clathrate in bulk. By adjustment to a slightly lower anode potential presumed to target Na 24 Ge 136 , even the unintended formation of α-Ge, usually occurring with only few mass percent, but besides X-ray amorphous products in the gas−solid oxidation products, 26,27 might be suppressed. From hitherto experience in gas−solid oxidation, 26,27,43 α-Ge is indicative of reactions bypassing the formation of Na 24−δ Ge 136 , which is required to yield Na x→0 Ge 136 with pure Ge(cF136) as the end point of the subsequent Na depletion. Optimization of the yield and selectivity seems thus feasible.
Coming along with both potential control and in situ traceability for the redox reaction and leading to a wellcrystalline bulk material because of applicability at suitably high temperatures, the anodic conversion approach in inorganic salt melts provides a substantial further development of the redoxpreparation method. For preparative usage, it may overcome the restrictions of recent related room-temperature approaches for kinetically favorable topotactical reactions. 8,9 Anodic conversion complements electrochemical preparation methods for crystalline intermetallic compounds, alloys, or semiconductor materials, which so far have been successful mostly for cathodic processes. 2,45−48 It promises not only to promote the fabrication of Ge(cF136) and other crystalline metastable semiconductor allotropes such as Si(cF136) 14,15 or Ge(oP32) 4 but also to make a variety of intermetallic phases of the Na−Ge and comparable systems preparatively accessible.
Details on chemicals and the preparation of Na 12 Ge 17 , on the manufacturing of electrodes and cell assembly, on the chronoamperometric procedure, on PXRD investigations for phase analysis and crystallinity studies, on the structure refinement and refinement results, and on the