Synthetic Routes to Crystalline Complex Metal Alkyl Carbonates and Hydroxycarbonates via Sol–Gel Chemistry—Perspectives for Advanced Materials in Catalysis

Abstract

Metal alkoxides are easily available and versatile precursors for functional materials, such as solid catalysts. However, the poor solubility of metal alkoxides in organic solvents usually hinders their facile application in sol–gel processes and complicates access to complex carbonate or oxidic compounds after hydrolysis of the precursors. In our contribution we have therefore shown three different solubilization strategies for metal alkoxides, namely the derivatization, the hetero-metallization and CO₂ insertion. The latter strategy leads to a stoichiometric insertion of CO₂ into the metal–oxygen bond of the alkoxide and the subsequent formation of metal alkyl carbonates. These precursors can then be employed advantageously in sol–gel chemistry and, after controlled hydrolysis, result in chemically defined crystalline carbonates and hydroxycarbonates. Cu- and Zn-containing carbonates and hydroxycarbonates were used in an exemplary study for the synthesis of Cu/Zn-based bulk catalysts for methanol synthesis with a final comparable catalytic activity to commercial standard reference catalysts.

1. Introduction

Metal alkoxides are compounds of the form M(OR)_n (M = metal, R = alkyl rest) and can be regarded as versatile single-source precursors for inorganic functional materials [1,2,3,4,5,6,7,8,9] and pre-formed solid-state entities at the molecular scale [10,11]. Through alkoxides, access can be granted to a variety of amorphous or phase-pure oxides via metal organic chemical vapor phase epitaxy (MOVPE), thermal decomposition or sol–gel processes [12,13,14]. The resulting metal oxides have found very wide applications in, for example, solar cells [15], batteries [16], transistors [17], supercapacitors [18] and superconductors [19], among others.

The physical properties of metal alkoxides are greatly influenced at the molecular level, i.e., the choice of metal and the nature and number of the alkyl groups R present in the molecule. Due to the strong binding within the M-O-M motif, metal alkoxides, especially transition metal alkoxides, tend to form oligomeric structures (Figure 1), which significantly reduce the solubility of the alkoxides in most organic solvents. Whereas Group IV metal alkoxides exhibit an extremely high solubility in a range of organic solvents [12], transition metal alkoxides incorporating Fe, Ni or Cu are almost insoluble in any organic solvent. Obviously, this lack of solubility hinders precise chemical modification and poses hurdles to their suitability as soluble precursors for sol–gel processes. However, the quasi-polymeric character can be suppressed to a certain extent via steric and electronic tuning of the alkoxide OR rest [20,21,22,23,24,25]. Additionally, the solubility can be greatly influenced through the implementation of pendant donor groups (e.g.,OC₂H₄OMe or OC₂H₄NMe₂) [26] or via the formation of multimetallic alkoxides [20]. However, by using the mentioned modification approaches [27], the inherent stoichiometric limitations can hardly be overcome.

Depending on whether one or more types of metal cations are involved in the corresponding alkoxides, compounds are referred to as homometallic or heterometallic alkoxides, both of which [28] are highly reactive compounds due to the strongly polarized M^δ+-O^δ−-C bond, and, consequently, require handling under inert conditions. Whereas the oxygen atom is likely to react with electrophilic reagents, the metal center is prone to nucleophilic attack, for example, by water, whereby the reaction follows a complex sequence of hydrolysis and condensation reactions (Scheme 1) [29]. During the hydrolysis, the alkoxide groups are either replaced by hydroxo (OH⁻) or oxo (O²⁻) ligands. In the presence of sufficient amounts of water and x ≤ 2, other nucleophilic attacks of water take place. Subsequently, the M-O-M chain is prolonged, and the sol–gel process starts. In excess water, the simultaneous nucleophilic attack of water on all alkoxy groups takes place and leads to the formation of solvated metal hydroxides. The final product of the hydrolysis is the formation of complex oxidic or typically amorphous hydroxy-based solids.

As well as the reactivity of metal alkoxides towards water, the reaction of metal alkoxides with small molecules, such as CO₂, has been investigated in the past (Scheme 2). Computational studies suggest that the electron lone pair on the oxygen atom of the metal alkoxide acts as a nucleophile and attacks the carbon atom of CO₂, resulting in a zwitterionic intermediate. This intermediate then rearranges in a rate determine step, resulting in a new M-O bond [30]. Lastly, the cleavage of the corresponding metal alkoxide M-OR bond takes place to give the metal mono-alkyl carbonate (MMAC) product [31]. The formation of metal mono-alkyl carbonates is not only described for alkali metal-based alkoxides, but also known for Ti-, Zr-, Fe-, Nb-, Mo- and Cu-based alkoxides [32,33,34]. However, in the case of transition metal-based MMACs, the isolation of pure materials has rarely been reported and often only spectroscopic evidence is given for their formation. As well as the challenges involved in the isolation of pure metal alkyl carbonates, the potential of the CO₂ insertion into the M-O bond of metal alkoxides is utilized in a variety of catalytic transformations, such as in ring-opening polymerizations, ring-opening co-polymerizations and transesterification reactions. Hereby, it was found that metal alkoxides can be employed as catalysts for the preparation of polyoxygenates [35].

Spurred on by the possibility of adjusting the reactivity, solubility and composition of metal alkoxides at the molecular level, we investigated the utilization of metal alkoxides as precursors for advanced functional materials in catalysis. However, to exploit the full potential of metal alkoxides and to make them widely accessible for sol–gel chemistry, the solubility of many homometallic alkoxides must be tremendously enhanced. Therefore, we explored various solubilization strategies of metal alkoxide precursors, such as the derivatization via alcohol exchange reactions, the hetero-metallization and the CO₂ insertion to obtain metal alkyl carbonates. Although such carbonate species have been described in the context of polymerization catalysis before, we present them as alternative precursors for sol–gel synthesis for the first time. Whereas metal oxides or hydroxides are usually used as the source for sol–gel materials, [13] the utilization of metal alkyl carbonates can give rise to new materials, also for catalytic applications. Hereby, the amount of water for the hydrolysis reaction and the addition of a cosolvent can give control over the meso-structures of the resulting materials.

2. Results and Discussion

2.1. Solubilization Strategies of Metal Alkoxide Precursors

Based on the high solubility of organometallic precursors in various organic solvents and their controllable reactivity, organometallic compounds have widely been reported as single-source precursors for the synthesis of ceramic and oxidic materials [36]. However, the complex multi-step synthesis of such organometallic precursors makes them unsuitable precursors in the scope of this work. Moreover, when using organometallic compounds, safety aspects must be considered, especially with respect to upscaling for application in the chemical industry. Therefore, our devised strategy to synthesize solid oxidic catalysts is the widely reported sol–gel approach [37,38,39,40,41,42] in combination with tailor-made metal alkoxide systems, which are usually insoluble. As the solubility and reactivity of the starting materials have been demonstrated as being crucial to fine-tune the properties of the final materials when sol–gel processes are used, in the current work, we have employed the three following strategies to break the polymeric structure of the easily available and versatile metal alkoxide precursors (Figure 2):

(a)

Derivatization (alcohol exchange)

(b)

Hetero-metallization

(c)

CO₂ insertion

All these strategies have yielded soluble precursors in organic solvents that have allowed us to synthesize nanomaterials with controlled size, composition, and nano- and meso-structure. Noteworthy to mention here is that while derivatization and hetero-metallization are well known procedures reported in the literature, [43,44,45] the use of CO₂ insertion to solubilize metal alkoxides, is a fully new approach in the context of sol–gel chemistry. This elegant path enables the synthesis of soluble metal alkyl carbonates from insoluble metal precursors, which can then easily be utilized for the synthesis of sol–gel materials.

(a) 

Derivatization

The derivatization strategy gives the possibility to exchange the alkoxide group with alcohols that have a higher boiling point than the respective alcohol of the alkoxide rest present in the precursor. To verify the effectiveness of the alcohol exchange reaction, a set of experiments was performed and the future exchange alcohol (mROH_free) and the theoretically bound alcohol (mROH_bound) in the alkoxide were quantified (

Table 1. Derivatized alkoxides. EF > 0.95 in all alkoxides (OMEE: 2(2-methoxyethoxy)ethanol, OME: 2-methoxyethanol).

	Derivatized Alkoxide M(OR’)n	Starting Alkoxide M(OR)n	Exch. Alcohol R’OH
1	Al(OHex)3	Al(OiPr)3	1-Hexanol
2	Al(OMEE)3	Al(OiPr)3	OMEE
3	Mg(OHex)2	Mg(OEt)2	1-Hexanol
4	Mg(OMEE)2	Mg(OEt)2	OMEE
5	Zr(OMEE)4	Zr(OnPr)4	OMEE
6	Ti(OMEE)4	Ti(OnPr)4	OMEE
7	Cu(OMEE)2	Cu(OiPr)2	OMEE
8	Zn(OMEE)2	Zn(OiPr)2	OMEE
9	Zn(OME)2	Zn(OiPr)2	OME
10	Co(OMEE)2	Co(OiPr)2	OMEE
11	Mn(OMEE)2	Mn(OMe)2	OMEE

). Due to the identical molar masses, the exchange factor (EF) can be defined as EF = mROH_free/mRO_bound. It was shown that a complete replacement of the alkoxide group occurs for all metal alkoxides studied. Interestingly, the broad applicability of 2-(2-methoxyethoxy)ethanol for all metals tested indicates that this alkoxide rest allows an easy adjustment of the desired molar metal ratio of the final material at the molecular scale. Moreover, the resulting metal alkoxides were soluble in the same alcohol and several organic solvents.

(b) 

Hetero-metallization

As a second solubilization strategy, the formation of heterometallic alkoxides was exploited. However, this strategy is only applicable if the additional heteroatom functionality is needed in the final material. Additionally, the solubilization of the resulting alkoxide is strongly dependent on the oxidation state of the metal alkoxide that will be solubilized. In a series of experiments, various alkoxides were solubilized in the presence of 1-hexanol and Al(OHex)₃. After adding M(OR)₂ to the Al(OHex)₃-1-hexanol mixture and heating the samples to reflux temperature of 1-hexanol, a homogeneous slightly colored solution was formed with all heterometallic compositions shown in

Table 2. Heterometallic alkoxides, n[M(OR)₂] = 30 mmol, n[Al(OHex)₃] = 60 mmol and n[1-hexanol] = 300 mmol. All compositions are formed 5–10 min after reaching T = 353 K. Photographs are shown exemplarily.

	Starting Alkoxide M(OR)n	Heterometallic Alkoxide M[Al(OHex)4]z
1	Mg(OEt)2	Mg[Al(OHex)4]2
2	Cu(OiPr)2	Cu[Al(OHex)4]2
3	Zn(OiPr)2	Zn[Al(OHex)4]2
4	Co(OiPr)2	Co[Al(OHex)4]2
5	Mn(OMe)2	Mn[Al(OHex)4]2

.

(c) 

CO₂ insertion

As mentioned before, the insertion of CO₂ is known for many metal alkoxides. However, we utilize the metal alkyl carbonates for the first time as precursors for sol–gel chemistry [32,33,34].

Table 3. CO₂ inserted alkoxides, n[M(OR)₂] = 3 mmol, V[CO₂] = 0.25 L∙min⁻¹ and n[solvent] = 300 mmol, T = 313 K.

Alkoxide	Solvent	Time until Solubilized
Cu(OiPr)2	nHexane	Not solubilized
Cu(OiPr)2	Pyridine	0.2 h
Cu(OtBu)2	tButanol	1.3 h
Cu(OiPr)2	THF	1.8 h
Cu(OiPr)2	iPropanol	2.1 h
Cu(OMe)2	Methanol	3.5 h

shows the results of the solubilization of different Cu alkoxides via CO₂ insertion. For the insertion experiments, CO₂ was bubbled into a suspension of primary, secondary and tertiary Cu alkoxides in different solvents. Interestingly, the CO₂ insertion took place in different organic solvents and independently of the alkoxide rest. The only exceptions were non-polar and non-coordinating solvents, such as n-hexane, where the solubilization was not successful.

Apart from the fact that the copper alkoxides were solubilized through the reaction with CO₂, the success of the CO₂ insertion into various metal alkoxides was further confirmed by IR spectroscopic analyses for different solubilized alkoxides. Hereby, typical absorption bands of metal alkyl carbonates with signals at around 1625 cm⁻¹_, which are attributable to C=O bond stretching and several peaks at wavenumbers of 1372 cm⁻¹, 1100 cm⁻¹ and 1088 cm⁻¹, that are associated with C-O bond stretching, were observed (Figure 3).

In order to confirm the IR results further, NMR spectroscopic studies of Mg[Al(OⁱPr)₄]₂ before and after the insertion of ¹³C-enriched CO₂ were carried out in d₆-DMSO (Figure 4). After the insertion, two new peaks at 156.6 and 159.0 ppm become obvious in the ¹³C NMR spectrum, as well as a peak for free CO₂, which verifies the formation of alkyl carbonate groups (O-CO₂) and, therefore, metal alkyl carbonates [46]. The broadening and splitting of the peaks are interpreted to be due to the non-equivalent chemical environments of the O-CO₂ group within a heterometallic alkyl carbonate.

The amount of inserted CO₂ was quantified via a simple CO₂-adsorption experiment. The insoluble Cu(OⁱPr)₂ was used as a test case, which was transformed to Cu(O(O)COⁱPr)₂ after CO₂ was inserted. The fully solubilized Cu(O(O)COⁱPr)₂ solution was then treated with a flow of argon, which led to the release of the inserted CO₂ and subsequent precipitation of the starting material Cu(OⁱPr)₂. The released CO₂ was captured and added to a solution of Ba(OH)₂ leading to the formation of BaCO₃ and water. From the mass of dry isolated BaCO₃, it was confirmed that CO₂ inserts quantitatively. The same procedure was carried out for several alkoxides to calculate the so-called insertion factor (IF), determined with the given formula: IF = nCO₂/nOR (

Table 4. Insertion factor for different metal alkoxides, n[alkoxide] = 40 mmol, V[CO₂] = 2 L∙min⁻¹, n[solvent] = 6 mol, T = 313 K.

	Starting Alkoxide M(OR)n	Solvent	tCO2 (h)	IF	Formulae of Compound Based on CO2 Adsorption Experiment
1	Cu(OiPr)2	iPrOH	6	0.93	Cu(O(O)COiPr)1.86(OiPr)0.14
2	Zn(OiPr)2	iPrOH	6	0.89	Zn(O(O)COiPr)1.78(OiPr)0.22
3	Al(OiPr)3	iPrOH	6	0.65	Al(O(O)COiPr)1.95(OiPr)1.05
4	Zr(OPr)4	nPrOH	6	0.94	Zr(O(O)COPr)3.76(OPr)0.24
5	Cu(OMEE)2	OMEE	1	0.88	Cu(O(O)COMEE)1.76(OMEE)0.24
6	Zn(OMEE)2	OMEE	1	0.84	Zn(O(O)COMEE)1.68(OMEE)0.32
7	Zn(OME)2	OME	1	0.96	Zn(O(O)COME)1.92(OME)0.08
8	Al(OMEE)3	OMEE	1	0.82	Al(O(O)COMEE)2.46(OMEE)0.54
9	Zr(OMEE)4	OMEE	1	0.90	Zr(O(O)COMEE)3.6(OMEE)0.4

). Deviations from the ideal insertion factor (IF = 1) were seen in all metal alkoxides, which most likely resulted from the incomplete adsorption of CO₂ in the washing solution. Regarding experimentally determined Cu- and Zr-insertion factors, there is a high conformity of the values published in the literature [32,34]. However, Al(OⁱPr)₃ (see Table 4, entry 3) shows a clear difference from the ideal insertion factor, which points towards the fact that only two of the three existing alkoxy groups are accessible for an insertion reaction. Moreover, it was demonstrated that glycol ether alkoxides (see Table 4, entry 5–9) are converted to metal alkyl carbonates much faster than the solid alkoxides. Hereby, the insertion was completed after only 1 h vs. 6 h in the case of ⁱPrOH alkoxides. The reason for this is most likely the liquid nature of these alkoxides that facilitates the CO₂ insertion to form the metal carbonate precursor.

2.2. Hydrolysis of Metal Alkoxides

Metal hydroxides, carbonates and hydroxycarbonates are important precursors for metal oxides, which have found numerous applications [15,16,17,18,19]. In this context, layered double hydroxides (LDHs) have attracted considerable attention, with [47,48] LDHs consisting of positively charged brucite-like layers together with charge-balancing anions and water. LDHs are often referred to as hydrotalcite (HTC), which is a layered double hydroxide of the general formula Mg₆Al₂CO₃(OH)₁₆·4(H₂O), since this mineral was one of the first well-characterized examples of an LDH [49]. HTCs can be synthesized starting from a mixture of aluminum and magnesium alkoxides, which is subsequently hydrolyzed in a controlled manner to afford the solid material after filtration and washing.

To obtain an aluminum magnesium alkoxide mixture, we applied the derivatization strategy (Figure 5) and reacted magnesium ethoxide and aluminum iso-propoxide with 1-hexanol to afford Mg(OHex)₂ and Al(OHex)₃ (and probably also mixed metal alkoxides). Through distillative removal of the low boiling alcohols, ethanol and iso-propanol, full conversion could be achieved (see Figure 5). In order to obtain the corresponding HTC, two separate strategies have been developed (see Figure 5, Strategy A and B). Within strategy A, the carbonate source is added to the metal alkoxide solution by directly adding CO₂ to the molecular precursor (see Figure 5, Strategy A). In the second case, the metal alkoxide is hydrolyzed with (NH₄)₂CO₃, as a carbonate source, using a cosolvent strategy (see Figure 5, Strategy B).

The second strategy B (Figure 6) was not simply used to introduce carbonate, but also to control the particle formation into a homogeneous system instead of a 2-phase system prior to nucleation. Therefore, only strategy B was considered for the following HTC syntheses and the cosolvent study.

Due to the low solubility of water in 1-hexanol, a 2-phase system is formed during the hydrolysis reaction, which causes only a very small temporal oversaturation through which only low amounts of nucleation seeds are present. Subsequently the hydrolysis tends to result in the formation of large particles. To avoid this 2-phase system, a third solvent, such as methanol or ethanol, which is miscible with 1-hexanol and water, can be added. Through the addition of such a cosolvent, the formation of a homogeneous mixture in an appropriate stoichiometric area is guaranteed and the hydrolysis rate is not dependent on the mass transport between the organic and water phase anymore. Therefore, the influence of the variation in the cosolvent during the hydrolysis reaction of the Mg-Al(OHex) mixture on the specific surface area and the crystallite size was investigated as depicted in Figure 7 (see ESI for further information in Supplementary Materials). It becomes obvious that the addition of a cosolvent leads to an increase in specific surface area, whereby this effect is more pronounced in the case of methanol. Additionally, the application of a cosolvent affects the crystallite size and the corresponding intensity ratio. Hereby, the effect is more evident for the crystallite size along d(003), instead of the crystallite size along d(012). In the case of ethanol, the crystallite size along d(003) is reduced to 9 nm, whereas when methanol is added to the hydrolysis mixture a further decrease to 7.5 nm becomes obvious. Through the use of a cosolvent, the I(012)/I(003) ratio is decreased, whereby again this effect is more pronounced in the case of methanol. The different results obtained for methanol and ethanol point towards some kind of interaction between the cosolvent and the forming HTC.

In Figure 8, the effect of the solvent composition (methanol, 1-hexanol and water) on the HTC crystallite formation and the corresponding aspect ratios of the crystallites are depicted. In principle the ternary phase diagram of methanol, 1-hexanol and water can be divided into four areas (I–IV):

• Classic biphasic system. HTC crystallite appearance almost independent from solvent composition. Aspect ratio ≈ 1
• HTC crystallite formation is strongly dependent on the water content. With decreasing water content, an increased anisotropic character of the HTC crystallites can be observed. x[H₂O] = 0.45, aspect ratio ≈ 3; x[H₂O] = 0.25, aspect ratio ≈ 7
• Formation of film-like HTC morphologies with a high degree of anisotropy. x[H₂O] = 0.05, aspect ratio ≈ 25
• Formation of gel-like structures, which cannot be assigned as HTC structures according to the characteristic [0 0 3] and [0 0 6] reflexes.

Whereas in the classical 2-phase system, the crystallite size and shape are almost independent of the solvent composition, there is clearly an influence of the solvent composition for the monophasic regime. This finding is plausible since in the biphasic system, the hydrolysis reaction is mainly limited by the mass transport between the aqueous and the organic phase and therefore not greatly affected by the solvent composition. In the biphasic regime the formation of the biggest crystallites can be observed with an aspect ratio d_a,b/d_c of about 1.

In the monophasic region within area II, there is no mass transport limitation coming into play, which causes the HTC crystallite formation to be strongly dependent on the water content. Whereas with a water content of x[H₂O] = 0.45 the calculated aspect ratio is about 3, at lower water concentrations of x[H₂O] = 0.25 the aspect ratio is already increased to 7. Despite this strong influence of the water content, the 1-hexanol/methanol ratio does not have an impact on the aspect ratio of the crystallites in this case.

Within the areas III and IV in the ternary phase diagram, at very low water contents (x[H₂O] = 0.05), a limitation of the HTC crystallite growth, due to the lack of water, can be witnessed. Whereas in the case of area III, film-like HTC structures are formed with a high degree of anisotropy (aspect ratio ≈ 25), in area IV, the formation of a gel without crystallinity can be observed. The different HTC formation behaviors in the areas III and IV in the ternary phase diagram can probably be attributed to an alteration in polarity, which directly influences the kinetics of crystallization of the system. Whereas the high methanol content in the area III causes the mixture to be very polar, area IV is far more non-polar due to the enhanced 1-hexanol content.

2.3. Hydrolysis of Metal Alkyl Carbonates for Cu/ZnO/Al₂O₃ Catalyst Synthesis

Methanol, with an annual global consumption of 53 × 10⁶ t, is one of the most important base chemicals [50]. In industry, methanol is produced from H₂, CO and CO₂ over a Cu/ZnO/Al₂O₃ catalyst at 5–10 MPa and 493–503 K [51]. The reaction mechanism of industrial methanol synthesis on Cu/ZnO-based catalysts has been under-investigated and is still a controversial topic. In this context, a combined experimental and theoretical study showed that the active sites consist of Cu steps decorated with Zn atoms, which are stabilized by bulk defect [52]. However, it was also found that the Cu/ZnO/Al₂O₃ catalyst is a dynamic system with a continuously changing structure [53]. Mechanistically, it was also found that both CO and CO₂ hydrogenation pathways are active for methanol synthesis [54,55]. Copper-based catalysts for the methanol synthesis are usually obtained from co-precipitation reactions of metal nitrates [56,57,58]. Due to the exothermic reaction, large limits with regard to equilibrium yield and the resulting high demands due to low temperature activity, already small improvements of the catalyst activity and stability have a large impact on the process.

In the context of Cu/ZnO/Al₂O₃ catalysts for methanol synthesis, we utilized alkoxide precursors and solubilized them via CO₂ insertion to form metal alkyl carbonates (see Section 2.1). Subsequently, the metal alkyl carbonates were hydrolyzed with 600 equivalents of water to obtain the oxidic material, which was then calcined under synthetic air. This approach is beneficial since on the one hand the nucleation can be decoupled from growth processes, and on the other hand the absence of any foreign cations, such as Na⁺ and NO₃⁻ as used in conventional precipitation processes, opens the potential of a simplified synthesis scheme.

(a) 

Cu/Zn-based complex carbonates

For the synthesis of alkoxide-based precursors for the methanol synthesis, Cu(OⁱPr)₂ and Zn(OⁱPr)₂ were treated with 2-(2-methoxyethoxy) ethanol. Afterwards Al(OMEE)₃ was added and the mixture was reacted with gaseous CO₂ (see Section 2.1). The resulting solution was then hydrolyzed and washed with water. Hereby, a clear dependency of the phase formation of the water equivalents used during the hydrolysis became obvious (Figure 9). Whereas small amounts of water (h_equ = 2.5 or 20) led to the formation of amorphous materials, the addition of larger amounts of water caused crystalline hydrotalcite-like compounds and zinc malachite to form (h_equ = 200 or 600). Consequently, the material, hydrolyzed with 600 equivalents of water, was applied as a catalyst for the methanol synthesis.

After the hydrolysis, the resulting materials were calcined under synthetic air and a solid of the composition Cu_0.61Zn_0.21Al_0.18 was obtained. The appearance of CuO, ZnO and ZnAl₂O₄ only occurs at calcination temperatures above 873 K (Figure 10). Very similar observations with respect to the thermal treatment were made by Behrens et al., for CuZnAl hydrotalcite-like precursors with Al contents between 30 and 40% [59].

Further to this, the influence of the calcination time and temperature on the methanol space time yield (STY) was investigated. The calcination of the samples was carried out at a temperature in the range of 723–973 K (723, 823, 873, 898, 928, 973 K) for 2 to 24 h. This parameter corridor was selected based on the analysis of the specific surface area of the materials depending on the calcination temperature and time (see ESI for further information). As reference catalysts, a commercially relevant sample Cu/Zn/Al (NP-Refcat-Cu_0.5Zn_0.3Al_0.2) and a low temperature sample Cu/Zn/Al (HP-Refcat-Cu_0.61Zn_0.21Al_0.18) were utilized.

The methanol space time yield of the materials calcined at 873 K is shown in Figure 11, whereby a clear dependency on the calcination time becomes obvious. The STY can be improved when the calcination time is increased from 1 to 2 h. Afterwards a decrease in the STY with increasing holding period can be witnessed. With a holding period of 2 h, STYs comparable to the commercial MeOH catalyst (NP-Refcat) could be reached. When other calcination temperatures and holding times were selected, the MeOH STY of the reference catalysts could not be reached (see ESI for further information).

When the gas hourly space velocity is increased from 2400 h⁻¹ to 4000 h⁻¹, a similar trend is observed (Figure 12). Whereas the samples, calcined for 0 or 1 h, show an almost identical MeOH STY, a holding period of 2 h leads to a clear enhancement of the STY, which is only 15% under the performance of the commercial MeOH catalyst (NP-Refcat). As seen before, a further increase in the calcination time caused the MeOH STY to drop drastically.

To explore the dependency of the MeOH STY on the calcination time, the materials calcined at 873 K for 0 to 8 h were analyzed by PXRD (Figure 13). From the decrease in the half width of the CuO reflexes, it becomes obvious that an enhanced calcination time influences the grain growth of CuO. Simultaneously, the ZnO and ZnAl₂O₄ reflexes become more pronounced, pointing towards the decomposition of the ht-carbonate [59].

Based on the diffractograms shown in Figure 13, the crystallite copper oxide crystallite sizes were calculated (Figure 14). An almost linear increase in the CuO crystallite sizes with enhanced calcination time can be observed. Therefore, the calcination temperature and the time influence the thermally induced conversion processes within the catalyst.

To rationalize the influence of the calcination temperature and duration on the methanol space time yield, the CuO crystallite sizes before and after the catalytic reaction as well as the specific surface area after the catalytic testing were analyzed (Figure 15). Although there is a clear increase in the calculated copper crystallite sizes before and after the catalytic reaction with increasing calcination temperature and time, there is no clear correlation with the observed methanol space time yield. A similar conclusion can be drawn from the specific surface area. Again, the increase in calcination temperature has a direct influence on the specific surface area; however, no correlation with the MeOH STY can be made.

Although it was empirically found that the calcination temperature and time has a direct influence on the catalytic performance of the Cu/ZnO/Al₂O₃ catalysts, no clear correlations with other structural parameters, such as the copper oxide crystallite sizes or the specific surface areas, was observed. This points towards a complex interplay of different structural and electronical factors. The complexity of the structural parameters that influence the catalytic activity under industrially relevant conditions and the fact that only a small and varying fraction of the metallic Cu surface is catalytically active has already been described by Schlögl and co-workers [52]. These findings demonstrate the need for a very detailed parameter study, which can definitely be supported by high throughput technology, [60,61,62] to find and optimize suitable materials for heterogeneous catalysis.

(b) 

Zr-promoted Cu/ZnO/Al₂O₃ catalysts

The use of zirconium for the synthesis of MeOH catalysts is described extensively in the literature [63,64,65,66,67,68,69,70,71]. In copper-based methanol catalysts, zirconium is most likely to act as chemical and structural promoter. Certain authors argue that migration of zirconium cations into the metallic copper leads to an increased formation of Cu⁺ species [64,69] and therefore an enhanced catalytic activity. In terms of structural promotion, zirconium tends to lead to an enhanced specific surface area [66,69,72] and copper dispersion (and higher specific Cu surface area) [69,72,73] as well as smaller copper crystallites [69] after reduction.

Zirconium-promoted catalysts were obtained from the reaction of Cu(OⁱPr)₂ and Zn(OⁱPr)₂ in 2-(2-methoxyethoxy) ethanol. Subsequently, Al(OMEE)₃ and Zr(OMEE)₄ were added and the mixture was treated with CO₂. Afterwards the hydrolysis was carried out with 600 equivalents of water and a blue green solid was obtained, which was found to consist of Cu_0.61Zn_0.21Al_{0.18-(0.18∙xZr)}Zr_(0.18∙xZr) after washing with water and drying. Based on the results shown before for the unpromoted Cu/ZnO/Al₂O₃ materials, the samples were calcined at 873 K for 2 h (see ESI for further information) and the zirconium-promoted materials were tested as catalysts in the synthesis of methanol from H₂, CO and CO₂. From Figure 16, it becomes obvious that the zirconium amount has an enormous influence on the methanol space time yield at a GHSV of 2400 h⁻¹. When 40% of the aluminum in the materials is displaced by zirconium, higher space time yields than the commercial reference MeOH catalyst (NP-Refcat) can be achieved. However, a larger amount of zirconium leads only to a minor performance enhancement. At a reactor temperature of 493 K the MeOH STY is only 20% lower than the STY obtained with the commercial low temperature reference MeOH catalyst (HP Refcat).

To investigate the influence of the Zr promoter amount further, the CuO crystallite sizes were calculated from the CuO (111) reflexes before and after the catalytic testing (Figure 17). With an increasing degree of substitution, the CuO crystallite sizes before and after the testing decrease. After the catalytic testing, the CuO crystallites are about 50% smaller when going from X_Zr = 2.5% to X_Zr = 60%.

To explore the role of Zr as structural promoter in the Cu/ZnO/Al₂O₃ system, the specific surface areas of the materials were determined before and after the catalytic testing depending on the degree of Zr substitution (Figure 18). Hereby, it becomes obvious that an increase in the degree of substitution to 20% leads to an enhanced specific surface area. Above X_Zr = 20% leads to a reduction in the specific surface area. However, even at X_Zr = 60%, the specific surface area is larger than for the ones observed for very low degrees of substitution. This trend is observed for the materials before and after the catalytic tests. Overall, it can be said that Zr acts as a structural promoter and leads to an enhanced Cu crystallite size and an increased specific surface area.

3. Materials and Methods

3.1. Analytical Methods

• Powder X-ray diffraction

For the measurement of powder diffractograms, an AXS D8 Discover/GADDS (General Area Detector Diffraction System, Bruker Corporation, Billerica, MA, USA) with HI-STAR area detector and a Bruker D8 Advance diffractometer with Lynxeye XE Detector, both with Bragg–Brentano geometry, were utilized. Both diffractometers were equipped with a Cu Kα radiation source with a wavelength of 1.540598 nm. The determination of the crystallite sizes was carried out according to the Scherrer equation, based on the full half-widths $∆\left(2\theta \right)$ obtained from the diffractograms, and the Scherrer form factor K = 0.9.

$$∆\left(2\theta \right)=\frac{K·\lambda }{L·\cos \left({\theta }_0\right)}$$

(1)

• N₂ physisorption

The specific surface areas were determined via N₂ physisorption measurements using a Micromeritics TriStar II 3020. Before the actual measurements, the samples were preconditioned at 423 K for 60 min under vacuum. Afterwards the samples were cooled to 77 K and the adsorption isotherms were recorded. The analyses of the isotherms were carried out according to the Brunauer, Emmett und Teller BET-method [74].

• ATR-IR spectroscopy

ATR-IR spectroscopic measurements were carried out using an ALPHA-T FT-IR spectrometer with a diamond ATR unit from Bruker (Bruker Corporation, Billerica, MA, USA). To avoid any hydrolysis of metal alkoxides, the sample preparation was carried out in a glovebox under argon atmosphere.

3.2. Experimental Details

In the following section exemplary examples of the synthesis of different alkoxide precursors, HTCs and bulk catalysts are given. Granular lithium, Al(OⁱPr)₃, Ti(OⁿPr)₄ and Mg(OEt)₂ were purchased from Sigma Aldrich (Sigma Aldrich Corporation, St. Louis, MO, USA), whereas CuCl₂, ZnCl₂, CoCl₂, MnCl₂, Zr(OⁿPr)₄ were obtained from Alfa Aesar (Haverhill, MA, USA).

• Preparation of Cu(OⁱPr)₂

In total 3.82 g (0.55 mol) elemental lithium in granular form was reacted slowly with 750 g iso-propanol at 80 °C. After no elemental lithium was visible anymore, the mixture was added to a homogeneous dark green solution of 36.98 g (0.275 mol) anhydrous CuCl₂ in 830 g anhydrous iso-propanol, which was heated before at 70 °C for 1 h. Immediately, Cu(OⁱPr)₂ was formed as dark green solid and the viscosity of the mixture strongly increased. Subsequently, the mixture was stirred at 68 °C for 1 h and afterwards cooled to room temperature with an ice bath. The copper(II) alkoxide was isolated via centrifugation, washing with iso-propanol and drying. The yield typically amounts 85–93%.

• Preparation of Zn(OⁱPr)₂

In total, 5.30 g (0.76 mol) elemental lithium was reacted slowly with 940 g iso-propanol at 80 °C. As soon as the reaction was completed, the lithium iso-propoxide solution was reacted with a pale yellow Zn(OⁱPr)₂ solution, which was synthesized before from 52.04 g (0.38 mol) ZnCl₂ and anhydrous iso-propanol (780 g) at 50 °C for 1 h. Immediately after the addition, Zn(OⁱPr)₂ was formed as white solid, whereas an enhanced viscosity of the reaction mixture was observed. The mixture was stirred for another hour at 69 °C and subsequently cooled to room temperature with an ice bath. The products were obtained after centrifugation, washing with iso-propanol and drying under vacuum. The yield typically amounts 91–95%.

• Preparation of Co(OⁱPr)₃

In total, 17.95 g (0.14 mol) anhydrous CoCl₂ was reacted with anhydrous iso-propanol. After the mixture was heated to 50 °C for 1 h, a dark purple solution was obtained. This solution was added to a ⁱPrOH solution of 1.92 g (0.28 mol) lithium, which was stepwise reacted with 315 g iso-propanol at 80 °C. Immediately after the addition, Co(OⁱPr)₂ was formed as dark purple solid and, simultaneously, the viscosity of the mixture was increased. This mixture was stirred at 69 °C for 1 h and then cooled down with an ice bath. The product was isolated via centrifugation followed by washing and drying under vacuum. Typically, yields of 94% can be achieved.

• Preparation of Mn(OMe)₂

In total, 27.20 g (0.22 mol) anhydrous MnCl₂ was reacted with 280 g dry methanol. The mixture was heated to 50 °C for 1 h and a pale pink solution was obtained. This solution was then added to a before prepared methanolic lithium solution (obtained from 3.0 g (0.43 mol) lithium in methanol at 80 °C), whereby a pale pink solid was formed. Simultaneously the viscosity of the mixture was increased. The mixture was heated at 69 °C for 1 h and subsequently cooled down with an ice bath. The product was isolated by centrifugation, washing with methanol and drying under vacuum. Typical yields of 82% can be achieved.

• Preparation of Al(OMEE)₃

In total, 100 g (0.49 mol) Al(OⁱPr)₃ was suspended in 176.5 g (1.47 mol) of liquid 2-(2-methoxyethoxy) ethanol. After stepwise heating of the suspension to 180 °C, a homogenous solution was formed and, simultaneously, iso-propanol was generated, which was removed from the mixture via distillation. Yields of 98% can be achieved.

• Preparation of Zr(OMEE)₄

In total, 165 g of a 70 wt% solution of Zr(OⁿPr)₄ in n-propanol was mixed with 169.4 g (1.41 mol) of liquid 2-(2-methoxyethoxy) ethanol. The mixture was heated stepwise to 180 °C under formation of n-propanol, which was removed via distillation. Yields of 94% can be achieved.

• General preparation of hydrotalcites

To a mixture of 3.1 g (0.03 mol) Mg(OEt)₂ and 1.85 g (0.10 mol) Al(OⁱPr)₃, 9.2 g (0.09 mol) 1-hexanol was added. After the removal of ethanol and iso-propanol via distillation at 393 K, Mg_0.75Al_0.25-hexanolate was obtained. Depending on the composition of the hydrolysis mixture, further 1-hexanol was added. The hexanolate was added to a mixture of methanol, water and ammonium carbonate under stirring. The resulting precipitate was filtered off after 24 h, washed twice with 50 g acetone and three times with 150 g water. Subsequently, the precipitate was dried at 80 °C for 16 h.

• Preparation of Mg₆Al₂(OH)₁₈∙4 H₂O

In total, 12.43 g (0.11 mol) magnesium ethoxide and 7.40 g (0.04 mol) aluminum iso-propoxide were filled into a 500 mL reaction vessel under inert gas atmosphere. Subsequently, 66 mL of 1-hexanol was added. Ethanol was distilled off at 80 °C and the iso-propanol was distilled off at 84 °C. The resulting metal hexanolate solution was heated at 80 °C for 1 h under stirring. Then 65 mL methanol was added. To trigger the hydrolysis, 31 mL of an aqueous solution of 3.48 g ammonium carbonate was added to the metal hexanolate solution under vigorous stirring, while the temperature of the hexanolate solution was kept at 80 °C After the addition of the aqueous carbonate solution, the heating of the mixture (at 80 °C), as well as stirring of the mixture, was continued for 5 h to equilibrate and to age the precipitate. Afterwards the precipitate was filtered off and subjected to a washing process, which consisted of a first washing step with acetone and a second washing step with water. In the next step the precipitate was placed into a drying oven for 12 h at 80 °C. The yield of dried product corresponded to 98 wt% of the theoretical yield.

• Preparation of Mg₆Al₂(CO₃)(OH)₁₆∙4 H₂O

In total, 12.43 g (0.11 mol) magnesium ethoxide and 7.40 g (0.04 mol) aluminum iso-propoxide were filled into a 500 mL reaction vessel under inert gas atmosphere. Then, 18.4 mL dry 1-hexanol were added to the metal salts. The metal salts were converted into the metal hexanolates by exchanging the alcohol groups and removal of the low boiling alcohols via distillation, whereby ethanol was distilled off at 80 °C and iso-propanol distilled off at 84 °C. In a subsequent step the metal hexanolate was converted into the metal hexyl carbonate solution by treating the solution with gaseous CO₂. The CO₂ was delivered to the solution by bubbling gaseous CO₂ (with a tube equipped with a jet nuzzle) into the solution for one hour while the solution was heated at 80 °C. The flow rate of the CO₂ supply was 1000 mL per minute. The yield of the resulting solution containing metal hexyl carbonate was 98% of the theoretical yield. Then the metal carbonate hexanolate solution was heated at 80 °C for 1 h under agitation. In the next step, 65 mL of methanol was added to the solution. This step was followed by the hydrolysis step, which was performed by the addition of 31 mL H₂O under vigorous stirring. The stirring of the mixture was continued for a period of 5 h. The resulting precipitate was separated from the liquid by filtration and subjected to a washing process. Washing included washing treatment with acetone and washing treatment with water. The washed precipitate was dried for 12 h at 80 °C within a drying oven. The yield of dried product was 98 wt% of the theoretical yield.

• Preparation of Cu/ZnO/Al₂O₃ bulk catalysts

In total, 6.65 g (0.04 mol) Cu(OⁱPr)₂ and 2.31 g (0.01 mol) Zn(OⁱPr)₂ were treated with 160 g of 2-(2-methoxyethoxy) ethanol. After 30 min at 80 °C, a homogenous dark blue solution was formed. To this solution 4.15 g (0.01 mol) Al(OMEE)₃ was added and gaseous CO₂ (1 L∙min⁻¹) was bubbled through the mixture for 1 h. The alkoxide solution was then transferred to water (1413 g, 78.48 mol, 600 equivalents). A blue green solid was formed, which was washed with 100 g of water and dried at 80 °C for 16 h. After drying, a blue green solid of the composition Cu_0.61Zn_0.21Al_0.18 was obtained. The materials were calcined under synthetic air (further information about the temperature and the time are given in the main text).

• Preparation of zirconium-promoted Cu/ZnO/Al₂O₃ bulk catalysts

In total, 6.65 g (0.04 mol) Cu(OⁱPr)₂ and 2.31 g (0.01 mol) Zn(OⁱPr)₂ were treated with 160 g of 2-(2-methoxyethoxy) ethanol. The mixture was heated to 80 °C and resulted in a dark blue solution after 30 min. Subsequently, 1.66 g (0.004 mol) Al(OMEE)₃ and 3.68 g (0.006 mol) Zr(OMEE)₄ were added and gaseous CO₂ (1 L∙min⁻¹) was bubbled through the mixture for 1 h. For the hydrolysis, the mixture was added to 1483 g (82.37 mol, 600 equivalents) of water. A blue green solid was formed, which was washed with 100 g of water and dried at 80 °C for 16 h. After drying, a blue green solid of the composition Cu_0.61Zn_0.21Al_0.072Zr_0.108 was obtained. The materials were calcined under synthetic air (further information about the temperature and the time are given in the main text).

3.3. Catalytic Testing

For the catalytic testing of the materials in the gas phase, a high throughput reactor system with 16 reactors was utilized (Figure 19). The catalytic bulk materials were mixed with corundum in a 1:1 volumetric ratio to ensure an efficient heat transfer and avoid local overheating of the catalyst and filled into the reactor with a layer of Alodur Korund WSK F100 at the bottom and at the top of the reactor. The different feed gases were added via different mass flow controllers. Gaseous samples of the reaction were taken via a multiport valve at the end of the reactor and the products were analyzed using gas chromatography (Agilent 6890). For the analysis of the permanent gases and CO₂, a thermal conductivity detector with a molecular sieve 5 Å-column (50 m × 0.53 mm × 50 μm) was utilized. For the analysis of organic components, a flame ionization detector with a DB1-column (60 m × 0.32 mm × 3 μm) was used.

4. Conclusions

Metal alkoxides were solubilized using different strategies, such as derivatization, hetero-metallization or/and CO₂ insertion. Through the reaction of metal alkoxides with carbon dioxide, access to soluble metal alkyl carbonates could be granted via CO₂ insertion into the metal–oxygen bond of the alkoxide. This CO₂ insertion strategy is therefore not only an alternative solubilization approach but also provides access to phase-pure, well-characterized metal alkyl carbonates, which can be applied in controlled hydrolysis and condensation processes for sol–gel syntheses as a way to access relevant inorganic functional materials, such as LDHs and mixed metal oxides. Further kinetic studies of the CO₂ insertion into metal alkoxides are currently under investigation.

To prevent the precipitation in biphasic systems (aqueous/organic) and, inherently, not to compromise regarding mass transport limitations between the two phases during the hydrolysis of the metal alkoxide precursor, the application of a cosolvent was investigated. Overall, the application of a cosolvent influences the specific surface area, the size and the morphology of the final materials. It is worth mentioning that the solvents used in this work were antisolvents for the final product (LDHs). This in turn allowed us to synthesize the ultra-thinnest LDH materials.

The application potential of the metal alkyl carbonates as catalyst precursors was demonstrated in the synthesis of methanol from CO, CO₂ and H₂. Therefore, the CO₂-inserted alkoxide precursors were hydrolyzed to obtain Cu/ZnO/Al₂O₃ and Zr-promoted Cu/ZnO/Al₂O₃ bulk catalysts. The calcination time and duration of the precursor affects the catalytic performance of the resulting materials in the direct synthesis of methanol. The reached STY is only 15% below a commercial reference methanol catalyst, and by using zirconia as promoter, the MeOH STY could even be increased further with the Cu_0.61Zn_0.21Al_0.072Zr_0.108 system. This STY is about 10% higher than the one obtained with a commercial reference MeOH catalyst and only 15% below a commercial low temperature MeOH catalyst.

5. Patents

Schunk, S.A.; Lizandara-Pueyo, C.; Futter, C.; Emmert, T.P. Process for the Preparation of Oxide Materials, Layered Double Hydroxide Materials, Hydroxide Materials and Carbonate-Based Materials. WO 2016/096990 A1, 23 June 2016.

Schunk, S.A.; Lizandara-Pueyo, C.; Jevtovikj, J.; Emmert, T.; Lejkowski, M.; Müller, R. Process for the Preparation of Carbonate Containing Metal Oxides-Hydroxides by Employing Metal Alkoxide Solutions Containing Metal Alkyl Carbonate Groups. EP 3 192 772 A1, 15 January 2016.