Ionic Liquid‐Driven Formation of and Cation Exchange in Layered Sulfido Stannates – a CH₂ Group Makes the Difference

Abstract

Two types of layered sulfido stannates or a molecular cluster compound are obtained upon ionothermal treatment of the simple sulfido stannate salt K₄[SnS₄] $·$  4H₂O that is based on binary tetrahedral [SnS₄]⁴⁻ anions. The formation of the respective products, novel compounds (C₄C₁C₁Im)₂[Sn₃S₇] (1 a), (C₄C₁C₂Im)₂[Sn₃S₇] (1 b), and (C₄C₁C₂Im)₂[Sn₄S₉] (2) with layered anionic substructures, or the recently reported compound (C₄C₁C₁Im)_4+x[Sn₁₀O₄S₁₆(SMe)₄][An]_x (A) comprising a molecular cluster anion, is controlled by both the choice of the ionic liquid cation and the reaction temperature. We report the scale‐up of the syntheses by a factor of 100 with regard to other reported ionothermal syntheses of related compounds, and a procedure of how to isolate them in phase‐pure form – both being rare observations in chalcogenido stannate chemistry in ionic liquids. Moreover, the synthesis of compound 1 a can be achieved by rapid cation exchange starting out from 1 b, which has not been reported for organic cations in any chalcogenido stannate salt to date.

Salts of layered sulfido stannate anions are obtained by ionothermal treatment of K₄[SnS₄] $·$  4H₂O comprising tetrahedral [SnS₄]⁴⁻ anions. The products form selectively upon choosing a specific pair of temperature and ionic liquid, and they can be obtained in large scale in phase pure form. Moreover, post‐synthetic treatment of one compound with the ionic liquid of the other one causes fast cation exchange under formation of the other network.

1. Introduction

The optoelectronic properties of chalcogenido (semi)metalate compounds can be fine‐tuned by different approaches. One approach is the introduction of further elemental components – other types of (semi)metal or chalcogen atoms – in order to change the charger carrier concentration,[ ¹ , ² , ³ , ⁴ , ⁵ , ⁶ ] such as done in compounds comprising the cluster ions [M_xE_yCh₁₈]^z− (M=Ga, In; E=Ge, Sn; x+y=10, Ch=S, Se), [M₁₀S₁₈]⁶⁻ (M=Ga, In), or [Zn₂₅In₃₁S₈₄]²⁵⁻,[ ⁷ , ⁸ , ⁹ ] for instance. Another approach is the change of the chalcogenido metalate network structure (with or without alteration of the network composition), which has an impact on the electronic band structure, as well.[ ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ ] An example for this is found in four related compounds that are based on anionic selenido stannate networks of different dimensionalities, 3D‐{Sn₂Se₅}, 1D‐{Sn₆Se₁₄}, 2D‐{Sn₂₄Se₅₆}, and 3D‐{Sn₁₈Se₄₀}. ^[13]

To vary the network architecture, one may choose specific synthetic strategies, which mainly differ in the absence or presence of liquid reaction media, the variation of the latter in solution‐ or flux‐based approaches, potentially used additives, and in the temperature during the synthesis. Another means of affecting the chalcogenido metalate substructures is the application of post‐synthetic treatments. For this, the application of ionic liquids has become very popular in recent times.[ ¹⁹ , ²⁰ , ²¹ , ²² ] Especially in the presence of amines as auxiliaries, one observes processes that involve partial or local dissolution and re‐assembly of the metalate units, usually under slightly elevated temperatures – so‐called ionothermal conditions. ^[23]

Herein, we report a series of chalcogenido stannate compounds that were obtained by ionothermal treatment of K₄[SnS₄] $·$  4H₂O ^[24] in the presence of 2,6‐dimethylmorpholine (DMMP). The reactions were carried out in different imidazolium‐based ionic liquids of the type (C₄C₁C_nIm)Br (with C₄, C₁, and C_n representing a butyl, a methyl and an alkyl chain with n carbon atoms as substituents of the cationic imidazolium ring Im). To scan the chosen reaction room, which address the heretofore relatively rarely explored sulfido stannates, we chose two different ionic liquids, (C₄C₁C₁Im)Br and (C₄C₁C₂Im)Br, and two different reaction temperatures, 120 °C and 180 °C. Both parameters led to changes of the product spectrum, indicating the subtle influence of the counterions, which differ by one CH₂ group only.

We present the syntheses and single‐crystal structures of the products, which were obtained with a remarkably high phase purity, and demonstrate the impact of the sulfido stannate architecture on the optical absorption properties and demonstrate the feasibility of a cation exchange of one ionic liquid cation for another. Band gap energies of the new compounds were determined to be 2.93 – 2.96 eV.

2. Results and Discussion

Syntheses and crystal structures. The ionothermal treatment of the starting compound K₄[SnS₄] $·$  4H₂O in the ionic liquids (C₄C₁C₁Im)X (X=Cl, Br) and (C₄C₁C₂Im)Br was carried out according to Scheme 1.

Scheme 1.

Survey of the preparation of compounds 1 a, 1 b, A, and 2 by ionothermal treatment of K₄[SnS₄] $·$  4H₂O in imidazolium‐based ionic liquids (C₄C₁C_nIm)X (with C₄, C₁, and C_n representing a butyl, a methyl and an alkyl chain with n carbon atoms as substituents of the cationic imidazolium ring Im; X=Cl, Br). Second products of the salt metathesis reaction, KCl and KBr are not indicated here.

Salt metathesis in (C₄C₁C₁Im)X (X=Cl or Br) at 120 °C, which affords KX as a second product, yields the sulfido stannate compound (C₄C₁C₁Im)₂[Sn₃S₇] (1 a). As the metathesis caused the cations in the system to be replaced by the ionic liquid cations, an impact of the latter on the anionic substructure seems to be very likely. To explore this in more detail, we used a slightly different ionic liquid, (C₄C₁C₂Im)Br, in which the CH₃ substituent in 3‐position of the imidazolium ring was replaced by a CH₂CH₃ substituent.[ ²⁵ , ²⁶ ] The product, (C₄C₁C₂Im)₂[Sn₃S₇] (1 b), possesses the same nominal composition of the anionic substructure, yet with slight structural variations (see below). To explore the additional effect of the reaction temperature in this chemical reaction space, we also carried out both reactions at 180 °C. Here, the presence of (C₄C₁C₁Im)⁺ cations caused the formation of (C₄C₁C₁Im)_4+x[Sn₁₀O₄S₁₆(SMe)₄][An]_x (A) ^[27] This was expected, as the chosen reaction conditions were those reported for the synthesis of this known compound. However with the other ionic liquid, the combined impact of templating cations and higher reaction temperature led to the formation of another compound, (C₄C₁C₂Im)₂[Sn₄S₉] (2). Notably, this compound is not accessible by ionothermal syntheses in (C₄C₁C₁Im)⁺‐based ionic liquids. All compounds were obtained as single crystals. The corresponding structures of the three novel compounds were determined by X‐ray diffraction. They are illustrated in Figure 1.

Figure 1.

View of the crystal structure of 1 a along the crystallographic c axis (a, b) and b axis (c, d) in different representations each. View of the crystal structure of 1 b along the crystallographic b axis (e, f) and c axis (g, h) in different representations each. Unit cell of 2 in two perpendicular views and different representations (i–l). {SnS₄} units are shown as yellow tetrahedra if they are part of defect‐heterocubane‐type units, or as red tetrahedra if they represent a separate building units. Organic cations are drawn as wires. Displacement ellipsoids are shown at a 50 % probability level.

Compound 1 a crystallizes as colorless plates in the trigonal crystal system (space group type P31c) with four formula units in the unit cell. The structure (Figure 1, a–d) is based on an anionic, two‐dimensional {[Sn₃S₇]²⁻}_n framework parallel to the ab plane. The latter resembles the famous honeycomb network of defect‐heterocubane‐type units {Sn₃S₄S_6/2}, connected via edge‐bridging by two μ‐sulfur atoms per contact. This substructure has been known from many other chalcogenido stannate salts, for instance (NMe₄)₂[Sn₃S₇] $·$  H₂O, ^[28] (trenH)₂[Sn₃S₇] (tren=tris‐2‐aminoethyl‐amine), ^[29] (AEPzH_x)_y[Sn₃S₇] (AEPz=1‐(2‐aminoethyl)‐piperazine), ^[30] and (Me₂NH₂)_1.33(Me₃NH)_0.67[Sn₃S₇] $·$  1.25H₂O. ^[10] Cations of the ionic liquid care for the charge balance, and furthermore separate the sulfido stannate layers from each other. In addition, each of the six‐membered rings (regarding each building unit as a member) accommodates one counterion molecule, which unfortunately could not be located on the Fourier map due to heavy positional disorder.

A very similar arrangement of cations and anions is found in compound 1 b (Figure 1, e–h), which crystallizes as colorless plates in the triclinic crystal system (space group type P1) with eight formula units in the unit cell. Beside the different ionic liquid counterions both compounds are topologically identical, but larger sterical demand of the counterions due to the additional CH₂ group of the organic cation causes a distortion of the heavy‐atom framework. This is illustrated in Figure 1 (a vs. e), by red highlight of the edges of the inner S⋅⋅⋅S edges of the defect‐heterocubane units of one of the “honeycombs”. The less symmetric cations furthermore cause a significantly lower crystal symmetry (only translation symmetry in 1 b versus a trigonal space group type in 1 a). Most probably, the size and shape of the (C₄C₁C₁Im)⁺ cation fits exactly to the structural demands of the honeycomb network, while the additional CH₂ group makes it just a bit too large, so the cations need to arrange in a different way, see Figure 1 (b vs. f). The slightly elongated shape of the (C₄C₁C₂Im)⁺ cation comes along with a slightly elongated version of the honeycomb rings, which can be taken as being correlated. We assume that the distorted network in 1b is less favorable, which allows for a cation exchange that will be discussed below.

Compound 2 also forms colorless plates, which crystallize in the monoclinic crystal system (space group type P2₁/m) with four formula units in the unit cell. Again, the structure consists of inorganic sulfido stannate layers and ionic liquid counterions that serve to separate the anionic layers from each other (Figure 1, i–l). However, the sulfido stannate network is constructed in a different, more complicated way than the honeycomb network on which compounds 1 a and 1 b are based. In 2, only two of three defect‐heterocubane units are directly connected, while the third is linked to the former via two {SnS₄} units (highlighted in Figure 1 as red tetrahedra). Six defect‐heterocubane‐type and four tetrahedral units form a macrocycle, yet unlike in 1 a and 1 b, these are further assembled into wave‐like layers that run parallel to {8 0 13}. The alkyl chains of the imidazolium cations extend into the voids of the anionic structure, which allows a relatively close approach of the imidazolium rings to the inorganic framework. A corresponding architecture of the anionic substructure has previously been reported for (C₄C₁C₁Im)₂[Ge_0.83Sn_3.17Se_9.06] ^[31a] and [N(C₃H₇)₄]₂[Sn₄S₉], ^[31b] which were obtained from K₄[Ge₄Se₁₀] $·$  3H₂O, SnCl₄  $·$  5H₂O and (C₄C₁Im)[BF₄] in presence of DMMP or SnS₂, elemental sulfur, (C₃H₇)₄NOH and H₂O, by ionothermal or hydrothermal reactions, respectively. Different linkage of the building units can lead to 3D structures, which were reported in compounds with the composition (C₄C₁C₁Im)₂[Sn₄S₉]. ^[32]

We succeeded in performing a scaled‐up reaction approach, and additionally managed to isolate the products 1 a, 1 b and 2 without ionic liquid residues and well‐separated from the binary side‐products. This is notable, as both scale‐up and purification of chalcogenido metalates obtained from ionic liquids is not trivial. To achieve higher quantities, the mixture of the starting materials was prepared in a Schlenk tube and treated (under stirring) in an oil bath instead of being kept un‐stirred in an oven. This methodology allowed to increase the amount of reacted material by a factor of 100. For obtaining pure products, we developed a washing process consisting of three steps: (1) Removal of the ionic liquid by treatment with acetonitrile that serves to dissolve the ionic liquid, yet not the sulfido stannate salt. (2) Removal of binary salts (KCl or KBr) by washing the raw‐product with an aqueous ionic liquid solution, 10 % (m/m) (1 a and 2) or an ionic liquid solution in methanol, 25 % (m/m) (1 b). (3) One more washing step with acetonitrile to eliminate ionic liquid residues that were imported by the previous step. The purity of the product was verified by powder X‐ray diffraction (PXRD, see Figure 2) and elemental analysis (see Tables S1 and S2, and Figure S1).

Figure 2.

PXRD patterns of 1 a (a), 1 b (b), and 2 (c), shown as black lines, with the simulated PXRD patterns (using the corresponding SC‐XRD data) shown as mirrored lines in red.

Upon treatment of 25 mg isolated, pure single‐crystals of 1 b with 1.0 g of a (C₄C₁C₁Im)X:MeOH mixture (X=[BF₄], Br; 1 : 3, m/m), a change of the structure is observed and confirmed by means of PXRD (Figure 3). The diffraction pattern recorded on the product of this procedure does not correlate to the PXRD pattern of compound 1 b anymore (blue line in Figure 3), but is in excellent agreement with the PXRD pattern of compound 1 a, which clearly demonstrates the clean structure conversion of 1 b to 1 a (beside a certain degree of amorphization, as recognized by a larger background in the measured diffractogram). We take this observation as a strong indication for a cation exchange. The driving force behind this procedure is probably the recovery of the undistorted anionic substrucure in 1 a. There are several evidences for ion exchange properties of this compound class in the literature, yet notably not for the replacement and incorporation of ionic liquid counterions or other (organic) cations of a comparable size.[ ¹⁰ , ¹⁶ , ³³ ]

Figure 3.

PXRD pattern recorded upon treatment of purified compound 1 b with a solution of (C₄C₁C₁Im)[BF₄] in MeOH solution (1 : 3, m/m) (black), and PXRD patterns of 1 a (red; measured) and 1 b (blue; simulated from the single‐crystal data) for comparision. A reflection of very low intensity observed at 8.1° (between the major reflections at 7.8° and 8.6°) is a minimal reminiscence of the diffraction pattern of 1 b, thus indicating its nearly complete absence in the product.

In order to estimate the band gap energy and the nature of the electronic transition to take place in these layered sulfido stannate salts, we performed UV‐visible spectroscopy. The spectra were recorded in diffuse reflectance mode, and the data were transferred into a Tauc plot using the Kubelka‐Munk function.[ ³⁴ , ³⁵ ] (F(R_∞)hν)^1/γ was plotted as a function of the photon energy, while γ as the power coefficient corresponds to the nature of the transition (γ=0.5: direct allowed; γ=2: indirect allowed). Additionally, the spectra measured in diffuse reflectance mode were transformed into absorbance spectra. ^[36] The absorbance data show a typical behavior and indicate band gaps close to 3.0 eV, with the “more open” structure of 2 causing a slighly larger band gap: 2.94 eV (1 a), 2.93 eV (1 b), 2.96 eV (2). The Tauc plots indicate that the transition is of indirect allowed nature. The results of the optical absorption measurements are illustrated in Figure 4.

Figure 4.

UV‐visible absorbance spectra (left) and Tauc plots (right), generated by application of the Kubelka‐Munk function (F(R_∞)hν)^1/γ with γ=0.5 of 1 a (a), 1 b (b), and 2 (c). The measurement was performed under inert conditions employing a Praying Mantis device.[ ³⁴ , ³⁵ , ³⁶ ]

3. Conclusion

We presented the synthesis, scale‐up and purification of three new sulfido stannate salts with imidazolium‐based organic counterions, (C₄C₁C₁Im)₂[Sn₃S₇] (1 a), (C₄C₁C₂Im)₂[Sn₃S₇] (1 b), and (C₄C₁C₂Im)₂[Sn₄S₉] (2). The compounds were obtained upon ionothermal treatment of K₄[SnS₄] $·$  4H₂O with ionic liquids that comprised the respective cations besides X⁻ (Cl⁻ or Br⁻) anions, which serve to form KX in a salt metathesis process. The two types of ionic liquid cations and the application of two different reaction temperatures (120 or 180 °C) control the product formation, with the fourth combination affording the known compound (C₄C₁C₁Im)_4+x[Sn₁₀O₄S₁₆(SMe)₄][An]_x (A) as a fourth product. While the sulfido stannate substructure of 2 differs significantly from that observed in compounds 1 a and 1 b, the latter two also show differences in structural details that we put down on the subtle differences of sizes and shapes between (C₄C₁C₁Im)⁺ and (C₄C₁C₂Im)⁺ as templating counterions. Post‐synthetic treatment of 1 b with a (C₄C₁C₁Im)[BF₄]/MeOH mixture affords a product that produces the diffraction pattern of 1 a in high phase‐purity. We attribute this observation to an exchange of the imidazolium cations, which has not been observed for chalcogenido stannate compounds comprising such organic cations. Measurement of the optical absorption behavior indicates band gap energies of 2.93–2.96 eV.

Experimental Section

General

All manipulations were performed under dry Ar atmosphere using standard Schlenk or glovebox techniques. Elements were used as received: K lumps (Acros Organics, 98 %), Sn powder (Sigma‐Aldrich, 99 %), and S powder (Alpha Aesar, 99,999 %). Ethane‐1,2‐diamine (en, 99 %, Acros Organics) and 2,6‐dimethylmorpholine (DMMP; 97 %; Acros Organics) was distilled from CaH₂ and stored over molecular sieve (3 Å). 1‐Butyl‐2,3‐methylimidazolium chloride/bromide, (C₄C₁C₁Im)Cl/Br, were treated under reduced pressure for several days. The literature‐known synthesis of 1‐butyl‐2‐methyl‐3‐ethylimidazolium bromide, (C₄C₁C₂Im)Br, ^[25] was modified by several steps: 1) The reactions were not performed under inert conditions. NaH was not reacted in pure form, but as a suspension in white mineral oil. Therefore, the oil had to be separated from the product by extraction with pentane. 2) Monitoring the reaction by ¹H NMR spectroscopy indicated that longer reaction times of 1 week for the first reaction step and 1–2 weeks for the second step were necessary. 3) The reaction batch was scaled‐up by a factor of 50. Single crystals of (C₄C₁C₂Im)Br (denoted here as B) were obtained that were of sufficiently high quality for an X‐ray diffraction study (see Supporting Information). The inorganic starting material K₄[SnS₄] $·$  4H₂O was prepared by solid state reaction of K₂S, Sn and S according to the literature. ^[24] The raw product was purified by extraction with degassed water. Note that the isolated products are stable under (humid) air and do not decompose over a period of several weeks according to powder X‐ray diffraction analysis.

Synthesis of (C₄C₁C₁Im)₂[Sn₃S₇] (1 a)

50 mg of K₄[SnS₄] $·$  4H₂O (0.11 mmol), 500 mg of (C₄C₁C₁Im)Cl, and 100 μL of en or DMMP were placed in a Duran glass ampoule that was sealed thereafter, heated to 120 °C and kept at this temperature for three days. The mixture was then allowed to cool down to room temperature. Colorless crystals of 1 a were obtained as colorless hexagonal plates. The raw product was washed three times with acetonitrile, three times with a (C₄C₁C₁)ImCl/H₂O (1 : 9) and three times with acetonitrile again. The purified product was obtained as colorless powder in approximately 65 % yield.

Synthesis of (C₄C₁C₂Im)₂[Sn₃S₇] (1 b)

50 mg of K₄[SnS₄] $·$  4H₂O (0.11 mmol), 500 mg of (C₄C₁C₂Im)Br, and 100 μL of DMMP were placed in a Duran glass ampoule that was sealed thereafter, heated to 120 °C and kept at this temperature for three days. The mixture was then allowed to cool down to room temperature. Colorless crystals of 1 b were obtained as colorless rhombic plates. The raw product was washed three times with acetonitrile, three times with a (C₄C₁C₂Im)Br/MeOH (1 : 3) and three times with acetonitrile again. The purified product was obtained as colorless powder in approximately 50 % yield.

Synthesis of (C₄C₁C₂Im)₂[Sn₄S₉] (2)

80 mg of K₄[SnS₄] $·$  4H₂O (0.17 mmol), 500 mg of (C₄C₁C₂Im)Br, and 160 μL of DMMP were placed in a Duran glass ampoule that was sealed thereafter, heated to 180 °C and kept at this temperature for three days. The mixture was then allowed to cool down to room temperature. Colorless crystals of 2 were obtained as colorless plates. The raw product was washed three times with acetonitrile, three times with a (C₄C₁C₂Im)Br/H₂O (1 : 9) and three times with acetonitrile again. The purified product was obtained as colorless powder in approximately 60 % yield.

Single Crystal X‐Ray Diffraction

X‐ray data were collected on a Stoe StadiVari diffractometer using Cu Kα radiation (λ=1.54186 Å; T=100 K) equipped with an Oxford Cryosystems module. Structure solution by dual space methods and full‐matrix least‐squares refinement against F ² were carried out using using SHELXT15, SHELXL15, and OLEX2 software. ^[37] The non‐hydrogen atoms were anisotropically refined. https://www.ccdc.cam.ac.uk/services/structures?id=doi:10.1002/open.202000287, deposition numbers 2032832 (1a), 2032833 (for 1 b), 2032834 (2), 2032835 (B), contain the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service https://www.ccdc.cam.ac.uk/structures. Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK [fax.: (internat.)+44 1223/336‐033; e‐mail: deposit@ccdc.cam.ac.uk].

Powder X‐ray diffraction

PXRD data were obtained on a Stoe StadiMP Cu Kα ₁ in transmission using a Mythen detector system. The data were examined by using WinXPOW.

UV‐visible spectroscopy

Optical absorption spectra were recorded on a Varian Cary 5000 UV/VIS/NIR spectrometer between 200 and 800 nm in the diffuse reflectance mode employing a Praying Mantis^TM accessory from Harrick. For ease of viewing, the raw data was transformed from %Reflectance (R) to Absorbance (A) according to the relation A=log(¹/_R). ^[36] Additionally, the measured diffuse reflectance spectra were converted in Tauc plots by application of the Kubelka‐Munk function (K‐M) in order to estimate the respective band gap energies of allowed (in‐)direct transitions:[ ³⁴ , ³⁵ ]

$${\displaystyle F\left(R\right)=\frac{k}{s}=\frac{{\left(1-R_{\infty }\right)}^2}{{2R}_{\infty }}}$$

where k is the K‐M absorption coefficient, R_∞ is the diffuse reflection, and s is the K‐M scattering coefficient.[ ³⁸ , ³⁹ ] Tauc plots were generated by plotting

$${\displaystyle {(F\left(R\right)·h\nu )}^{1/\gamma }}$$

as a function of the photon energy hν. The power coefficient can be γ=$1/2$ , $2/3$ , 2 or 3, depending on the nature of the transition, which corresponds to direct allowed, direct forbidden, indirect allowed or indirect forbidden transitions, respectively. E _g is estimated from the intercept with the x axis of the linear fit from the corresponding region. ^[40]

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

B. Peters, M. Möbs, N. Michel, F. Tambornino, S. Dehnen, ChemistryOpen 2021, 10, 227.