Alkali-Free Hydrothermally Reconstructed NiAl Layered Double Hydroxides for Catalytic Transesterification

Abstract

NiAl layered double hydroxides (LDHs) are promising bifunctional catalysts comprising tunable redox and Lewis acidic sites. However, most studies of NiAl LDH employ alkali hydroxide carbonate precipitants which may contaminate the final LDH catalyst and leach into reaction media. Here, we report an alkali-free route to prepare Ni_xAl LDHs with a composition range x = 1.7 to 4.1 using (NH₄)₂CO₃ and NH₄OH as precipitants. Activation of LDHs by calcination–rehydration protocols reveal Ni_xAl LDHs can be reconstructed under mild hydrothermal treatment (110 °C for 12 h), with the degree of reconstruction increasing with Ni content. Catalyst activity for tributyrin transesterification with methanol was found to increase with Ni content and corresponding base site loadings; TOFs also increased, suggesting that base sites in the reconstructed LDH are more effective for transesterification. Hydrothermally reconstructed Ni_4.1Al LDH was active for the transesterification of C₄–C₁₂ triglycerides with methanol and was stable towards leaching during transesterification.

1. Introduction

Layered double hydroxides (LDH) of general formula [M²⁺_(1−x)Al³⁺_x(OH)₂]^x+(Aⁿ⁻)_x/n·yH₂O are interesting materials for photo-, electro-, and thermocatalysis [1,2] and electronic applications [3,4,5]. They typically adopt three-dimensional, lamellar structures of mixed metal oxyhydroxide platelets, with the interlayer voids providing nanoporous chemical reactors [6]. The catalytic versatility of LDHs arises from their structural flexibility, wherein a range of divalent M²⁺ and trivalent M³⁺ cations can be introduced into the platelets and diverse charge-balancing anions Aⁿ⁻ (e.g., CO₃²⁻, SO₄²⁻, Cl⁻, OH⁻) within the interlayers [7,8]. Consequently, LDHs offer tunable redox and electronic properties and morphologies amenable to delamination into individual two-dimensional platelets [9] or assembly into complex architectures such as sand roses [10,11]. The morphology of LDHs is largely dictated by their synthesis; low temperature co-precipitation favors random lamellar arrangements in which precursor impurities are retained [9], whereas moderate temperature calcination to remove impurities necessitates subsequent reconstruction [12,13] of mixed metal oxide crystallites by hydrothermal treatment to recover lamellar structures (calcination–rehydration) [14,15,16,17,18].

LDHs featuring transition metals such as Co, Fe, or Ni as divalent cations have been the focus of electrocatalytic applications [19] due to their good conductivity [20], low cost, and ease of generating host structures with high proton mobility [19]. NiAl LDH materials also find widespread application as precursors to Ni/Al₂O₃ catalysts [21] for hydrogenation/hydrodeoxygenation [19,22], H₂ production by steam reforming [23], and lignin depolymerization/hydrogenolysis [24].

LDH materials are typically synthesized by the coprecipitation of soluble M²⁺/M³⁺ ionic solutions using NaOH or KOH, in the presence of the corresponding counter ion, typically CO₃²⁻ [8,25]. The use of ammonia or urea as a precipitating agent is preferable for applications where intercalated alkali cations from the residual precipitating agent could impair performance of the resulting LDH [26,27,28]. While routes such as hydrothermal urea hydrolysis [29] have been employed to prepare NiAl LDHs, the use of NH₄CO₃/NH₄OH precipitants has not been explored, to our knowledge, and may offer milder synthesis conditions.

Activation of LDH materials is often required to increase their catalytic utility [16]. This commonly involves calcination to decompose the precipitated LDH by removing charge compensating anions (accompanied by dehydroxylation) to form an amorphous mixed oxide. Reconstruction of these mixed oxides into lamellar structures is observed on subsequent hydration, termed the ‘memory effect’, which also introduces vacancies into the metal oxyhydroxide layers, due to the replacement of interlayer CO₃²⁻ with OH⁻ [30,31]. Calcined MgAl LDH materials can be readily reconstructed by vapor phase rehydration [12,32], but relies on moderate temperature (<550 °C) calcination to avoid significant formation of tetrahedral Al³⁺ and spinel structures [33]. Indeed, not all calcined LDHs can be fully reconstructed by such treatments, as the partial oxidation of the M²⁺ cation or the formation of thermodynamically stable mixed oxides are also possible [34]. A consequence is that calcined ZnAl LDH requires more forcing hydrothermal treatments to partially regenerate the initial lamella structure, due to the stability of Al(OH) and ZnO formed on calcination [28]. NiAl LDH materials are also challenging to reconstruct due to the formation of Al-doped NiO and spinel-like phases at elevated temperatures [21,35,36]. The influence of the Ni:Al ratio on the hydrothermal reconstruction of calcined NiAl LDH prepared by NH₄OH/NH₄CO₃ precipitation has not been previously reported. However, as alkali-free precipitants typically yield smaller LDH crystallites than their alkali analogues [26], these may be more susceptible toward reconstruction.

NiAl LDH materials are expected to exhibit mild basic properties (albeit weaker than for MgAl LDH, due to the lower electropositivity of the transition metal) [36,37], and hence may be suitable for base-catalyzed reactions. Here, we explore the hydrothermal reconstruction and catalytic activity of a family of NiAl LDHs for the transesterification of model (C4-C12) triglycerides to fatty acid methyl esters (FAME).

2. Results and Discussion

2.1. Catalyst Characterisation

The successful synthesis of alkali-free NiAl LDH materials was verified by XRD, SEM, and EDS. XRD reflections of the as-precipitated LDHs (Figure 1a) were in good agreement with reference patterns for the takovite structure (JSPDS 15-0087), while EDS confirmed that Ni:Al bulk atomic ratios were similar (albeit slightly higher) to their nominal values, varying between 1.7:1 and 4.1:1 across the LDH family (

Table 1. Physicochemical properties of as-prepared and reconstructed NiAl LDH materials.

As-Prepared					Reconstructed
Nominal Ni:Al Atomic Ratio	Bulk Ni:Al Atomic Ratio a	Surface Ni:Al Atomic Ratio b	NiAl LDH Formula c	BET Surface Area d /m2·g−1	BET Surface Area d /m2·g−1	Base Site Loading e /mmol·g−1	Base Site Loading /molecules·g−1
1.5:1	1.7:1	1.0	[Ni0.6Al0.4(OH)2].(CO3)0.180.56H2O	149	183	0.062	3.7 × 1019
3:1	2.7:1	2.6	[Ni0.7Al0.3(OH)2].(CO3)0.140.58H2O	236	207	0.072	4.4 × 1019
4:1	4.1:1	4.1	[Ni0.8Al0.2(OH)2].(CO3)0.100.72H2O	77	134	0.099	6.0 × 1019

^a from EDX (error ±0.13), ^b from XPS, ^c CO₃²⁻, and H₂O quantified by TGA, ^d Standard error ±30 m²·g⁻¹, ^e from CO₂ pulse chemisorption.

). Corresponding BET surface areas were also in accordance with literature values for NiAl LDHs, spanning 260–60 m²·g⁻¹ [8]. Surface areas of LDH materials are a complex convolution of crystallite size and density; however, for the same morphology, increased Ni content is expected to lead to a decrease in the specific surface area. NiAl LDH compositions were also estimated from TGA by quantifying weight losses attributed to the desorption of intercalated H₂O (~150 °C) and CO₃²⁻ decomposition (261–355 °C) (Table S1 and Figure S2) and are in good agreement with expectations [38,39]. Higher Ni loadings are expected to expand the interlayers, thereby increasing water intercalation.

A comparison of surface and bulk Ni compositions from XPS and EDS, respectively, confirm an expected linear relationship for a pure phase LDH (Figure S1a) and, hence, a negligible surface segregation of Ni or Al pure phases. Ni 2p XP spectra of the NiAl LDHs (Figure S1b) reveal a well-resolved spin–orbit split doublet with 2p_3/2 and 2p_1/2 peaks at 854 eV and 872 eV binding energies, respectively, and shake-up satellites at 860 eV and 879 eV consistent with Ni²⁺ [40]. Corresponding Al 2p XP spectra reveal a similar spin–orbit split doublet with a 2p_3/2 peak at 75.2 eV consistent with Al³⁺ in an LDH. Activation of NiAl LDHs by calcination and hydrothermal reconstruction was subsequently assessed. A 350 °C calcination temperature was selected as this achieved sufficient dehydroxylation/decarbonylation to destruct the ordered LDH phase in all materials (Figure S2) while also minimizing sintering. All calcined NiAl LDHs exhibited reflections at 37.4°, 43.9°, and 63.3° (Figure 1b), consistent with the formation of NiO (JSPDS 01–1239) [41]. The genesis of a crystalline NiAl₂O₄ spinel phase was not observed, and requires higher temperatures (~900 °C) [42]. Calcination at 350 °C thus generates NiO crystallites and amorphous alumina, similar to previous reports [38,43,44]. Higher temperature calcination of an alkali-precipitated Ni₂Al LDH reports the formation of an Al-containing NiO-like oxide at 450 °C alongside an Al-rich amorphous component. Previous Rietveld refinement also revealed a smaller lattice parameter for the former phase than observed for pure NiO (lattice parameter of 0.418 nm) [43], which increased with Ni:Al content of the parent LDH. A slight increase in the lattice parameter (determined from the d(200) reflection) of the NiO phase with Ni:Al ratio was also observed for our NiAl LDHs (from 0.413 to 0.416 nm); this indicates that trace Al must likewise be incorporated into alkali-free NiO crystallites obtained after 350 °C calcination.

The reconstruction of 350 °C calcined NiAl LDHs was subsequently attempted at 110 °C in water. Note that it was not possible to reconstruct calcined NiAl LDHs using water vapor (Figure S3) in accordance with previous reports [21,35,36]. The complete reconstruction of calcined LDHs requires rehydration of mixed metal oxides and dissolution of any segregated M²⁺ oxide, which for Ni²⁺ requires more forcing conditions owing to the formation of stable (NiAl)O mixed oxides on calcination [13].

Hydrothermal treatments have proven more effective than water vapor in reconstructing the lamellar structure of calcined MgAl [45] and ZnAl [28] LDHs. In the present work, recovery of an LDH phase was strongly dependent on the Ni:Al ratio in the parent material (Figure 1c), with Ni_1.7Al and Ni_2.7Al LDHs retaining significant NiO. In contrast, the Ni_4.1Al material exhibited negligible NiO and a majority reconstructed LDH phase, with the lattice parameter for the c axis determined to be 2.25 nm (calculated from the d(006) refection at 23.6°). This is almost identical to that for the parent material (2.26 nm), and corresponds to an interlayer spacing (d) of 0.75 nm (note lattice parameter c = 3d [46]). This observation is significant since it is generally regarded that reconstruction of calcined NiAl LDHs by rehydration alone is not possible and requires harsh conditions (e.g., 160 °C in NH₄OH) [36]. Successful hydrothermal reconstruction in this work may reflect enhanced Al incorporation into the NiO phase obtained by lower temperature (350 °C) calcination versus literature reports, and concomitantly less phase-separated tetrahedral Al³⁺ (difficult to reincorporate during LDH reconstruction [33]).

The morphology of the Ni_4.1Al LDH was studied by SEM following calcination and hydrothermal reconstruction (Figure 2). The precipitated NiAl LDH exhibited a characteristic sand-rose structure comprising assemblies of fused platelets. These structures evolve due to initially fast crystallisation of LDH seeds which preferentially grow along their (001) planes, disfavouring nanoparticle aggregation [47]. HRTEM (Figure S4) reveals the lattice spacing of LDH crystallites within precipitated platelets as 0.33 nm, corresponding to the d(110) plane, which is slightly larger than that determined from the corresponding XRD reflection but within the range expected for LDH lattice fringes. Calcination transforms the parent LDH into spherical agglomerates (Figure 2b), which are converted back to a fused lamellar structure by hydrothermal treatment (Figure 2c). Lattice fringes of the reconstructed Ni_4.1Al LDH were determined by XRD and HRTEM as 0.30 and 0.31 nm, respectively. Slight contractions in the interlayer spacing relative to the parent material are attributed to the replacement of interlayer CO₃²⁻ by OH⁻ anions. NiAl LDHs are expected to exhibit basic character, originating from undercoordinated oxygen sites or Me–OH surface groups formed to charge balanced Al³⁺ ions replaced by Ni²⁺ in LDH layers as the Ni:Al ratio increases [48]. The quantification of base site loadings by CO₂ pulse chemisorption confirmed that basicity increases with Ni:Al ratio (Table 1).

2.2. Transesterification Activity

The impact of Ni:Al ratio on the activity of reconstructed NiAl LDHs was subsequently evaluated for the base-catalyzed transesterification of tributyrin with methanol (Figure 3 and Figure S5).

TOFs increased from 60 to 180 h⁻¹ with rising Ni:Al ratio, mirroring the trends in conversion (which ranged from 15% to 30% over 6 h reaction) (see Figure S5), indicative of increased accessibility or base strength of active sites, the former being consistent with a greater extent of hydrothermal reconstruction observed by XRD (Figure 1c inset). These TOFs are in line with literature values for tributyrin transesterification over related ZnAl (180–290 h⁻¹) [28] and Mg₂Al (329 h⁻¹) [45,49] LDHs, considering their relative basicity (NiAl ≈ ZnAl < MgAl [16,50]), which are comparable to values for nanocrystalline MgO (220 h⁻¹) [51] and Mg-ZrO₂ (100 h⁻¹) [52]. Selectivity to methyl butyrate was >95–80%, with a slight decrease observed with Ni:Al ratio attributed to increased conversion across the series, and reduced transesterification activity of diglycerides (see discussion later) [53].

The most active Ni_4.1Al LDH catalyst was selected for the transesterification of longer-chain C₈-C₁₂ TAGs (Figure 4 and Figure S6). Ni_4.1Al LDH was active for all three TAGs; however, conversions and TOFs fell with increasing alkyl chain length, likely reflecting poorer base site accessibility for the bulkier TAGs [45]. FAME selectivities also fell with increasing chain length, reaching ~40% for trilaurin after 24 reaction, suggesting poor mass transport [45,54] of bulky diglyceride (DAG) and monoglyceride (MAG) reactively-formed intermediates, and that their competitive adsorption with TAG hinders subsequent conversion under batch conditions [55,56]. Slow DAG conversion has previously been accounted for, owing to activation barriers for transesterification being higher than for their corresponding triglycerides. The removal of the first fatty acid chain leads to increased electron density at the -CH₂-OH center of the glyceride backbone, which hinders subsequent attack by adsorbed MeO- at the remaining ester groups [53].

The stability of the Ni_4.1Al LDH catalyst was investigated by a recycling study for tributyrin transesterification, which confirmed the heterogenous nature of the catalysis (Figure S7). The same catalyst could be recycled, retaining ~60% of activity for tributyrin transesterification after three cycles. ICP-OES revealed a negligible decrease in Ni:Al (Table S2) after the third recycle, suggesting that loss of activity was attributed to deactivation by poisoning of base sites or pore blockage by reaction intermediates [45].

There are few previous studies of NiAl LDH or NiO-derived catalysts for TAG transesterification, which hinders performance benchmarking, and of these all employed alkali precipitation syntheses and, hence, homogenous contributions to catalytic activity cannot be discounted [57,58]. Calcined NiAl and ZnAl LDHs are reported as inactive for sunflower oil transesterification with methanol at 65 °C [50]. Our reconstructed Ni_4.1Al LDH, which affords 14% trilaurin conversion (albeit at a higher temperature), is promising, despite their essentially microporous nature arising from a lamellar structure [59] with layer spacing of 0.3 nm observed in HRTEM (Figure S4), which hinders access of bulky TAGs to interlayers and in-pore base sites. Hence, most TAG activation is expected to occur over the external surface of the 2D NiAl LDH lamellar sheets in the sand rose structure observed in Figure 2c. The introduction of meso- and macroporosity is expected to significantly enhance TAG accessibility; hence, alkali-free, hydrothermally activated NiAl LDHs show potential as heterogeneous solid base catalysts for transesterification.

3. Experimental

3.1. Catalyst Synthesis

All reagents were of analytical grade and were used without further purification. Ni_xAl LDHs were synthesized in a 500 mL Radley’s reactor-ready stirred vessel. Aqueous solutions of Ni(NO₃)₂·6H₂O (1.5 M) and Al(NO₃)₃·9H₂O (1.5 M) were mixed together in desired proportions to yield Ni:Al molar ratios spanning 1.5 to 4, with a final solution volume of 100 cm³. Ammonium carbonate (2 M, 100 cm³) was used as the precipitating agent and was added by a syringe pump (1 mL·min⁻¹) to the metal nitrate mixture while being stirred. Once the addition of metal nitrate and ammonium carbonate solutions was completed, ammonia solution was added dropwise to maintain a constant pH of 9.5. The mixture was aged while stirring at 65 °C for 18 h, after which the resulting precipitate was filtered and washed repeatedly with deionised water until the washings were pH 7. The resulting solid powder was then calcined at 350 °C (ramp rate of 1 °C·min⁻¹) for 5 h under flowing O₂ (20 mL·min⁻¹). The calcined solid was subjected to hydrothermal treatment in 50 mL of water, in a stirred Teflon-lined autoclave at 110 °C under autogenous pressure for 12 h. The treated solid was recovered by centrifugation and oven-dried at 80 °C before being stored in a vacuum desiccator prior to use.

3.2. Catalyst Characterisation

Powder X-ray diffraction was performed on a Bruker D8 ADVANCE diffractometer (Coventry, UK) using Cu K_α X-ray radiation (0.15418 nm). The diffraction patterns were recorded between 2θ = 10–80°. Crystallite sizes were determined by application of the Scherrer equation. The ratio of LDH:NiO following reconstruction was determined from the relative heights of the most intense reflections for NiAl LDH (13.8°) and NiO (43.6°). X-ray photoelectron spectroscopy was performed on a Kratos Axis HSi spectrometer (Manchester, UK) fitted with a charge neutralizer and magnetic focusing lens employing monochromated Al K_α radiation at 90 W. Spectral fitting was performed using CasaXPS version 2.3.16, with binding energies referenced to the C 1s Peak at 284.5 eV. N₂ porosimetry was undertaken on a Quantachrome Nova 4000 porosimeter (Hook, UK) on samples degassed at 120 °C for 3 h. Surface areas were calculated by the Brunauer–Emmett–Teller (BET) method from the desorption isotherm for P/P₀ < 0.2. Base site densities were determined by CO₂ pulse titration using a Quantachrome ChemBET 3000 chemisorption analyzer (Hook, UK) with a thermal conductivity detector. Then, 50 mg of the sample was placed in a quartz cell, outgassed for 1 h under flowing He at 120 °C, cooled to 40 °C, and then titrated with 50 µL CO₂ pulses at room temperature until uptake saturated. Thermogravimetric analysis with online mass spectrometry (TGA-MS) was performed on a Mettler Toledo TGA/DSC2 Star system (Leicester, UK) under flowing nitrogen during sample heating to 800 °C at 10 °C·min⁻¹. TGA was performed under N₂ to avoid any complications from oxidation of defects formed when a loss of H₂O/CO₂ would complicate the calculation of absolute mass loss during the TGA experiment. Scanning electron microscopy (SEM) was performed on a JEOL JSM-7000F microscope fitted with an energy dispersive X-ray spectroscopy detector (EDS) (Welwyn Garden City, UK) using an accelerating voltage of 20 kV. Transmission electron microscopy (TEM) analysis was carried out using a JEOL 2100F FEG STEM microscope operating at 200 keV. Samples were dispersed in methanol and drop cast onto copper grids coated with a holey carbon film (Agar Scientific Ltd., Essex, UK).

3.3. Catalytic Activity

Transesterification was performed using Radley’s Starfish reactor, equipped with ACE^TM pressure flasks (Vineland, NJ, USA) modified with a dip tube to enable aliquots to be periodically withdrawn. Reactions were performed at 110 °C, using 10 mmol of triglyceride (TAG) and 308 mmol (12.5 mL) methanol, with 0.0025 of mol (0.59 cm³) dihexyl ether as the internal standard; an additional 20 wt% 1-butanol was introduced for longer chain C₈-C₁₂ TAGs (tricaprylin and trilaurin, 99%, Alfa Aesar (Lancashire, UK) to improve their miscibility. Aliquots were periodically withdrawn during reaction. They were then filtered and diluted with dichloromethane, and analyzed by off-line gas chromatography (GC) using a Varian 450-GC (Crawley, UK) fitted with a Phenomenex (California, USA) ZB-5HT Inferno capillary column (15 m × 0.32 mm × 0.1 μm) for C₄–C₈ TAGs. Heavier TAGs were analyzed using a temperature-programmed, on-column injector and a Phenomenex ZB-1HT Inferno wide-bore capillary column (15 m × 0.53 mm × 0.1 µm). Turnover frequencies (TOFs) were calculated by normalizing initial rates of TAG conversion derived from the linear portion of reaction profiles (<20% conversion) to base site loadings obtained from CO₂ chemisorption. Triglyceride conversion and selectivity to fatty acid methyl ester (FAME) were calculated as follows:

$$\mathrm{Conversion}=\frac{\mathrm{Concentration} \mathrm{of} \mathrm{TAG} \mathrm{at} \mathrm{T}=0\left(\mathrm{mol}/\mathrm{L}\right)-\mathrm{Concentration} \mathrm{of} \mathrm{TAG} \mathrm{at} \mathrm{T}=\mathrm{t}\left(\mathrm{mol}/\mathrm{L}\right)}{\mathrm{Concentration} \mathrm{of} \mathrm{TAG} \mathrm{at} \mathrm{T}=0\left(\mathrm{mol}/\mathrm{L}\right)}\times 100$$

(1)

$$\mathrm{Selectivity}=\frac{\mathrm{Concentration} \mathrm{of} \mathrm{product} \mathrm{formed} \left(\mathrm{mol}/\mathrm{L}\right)}{\mathrm{Total} \mathrm{product} \mathrm{concentration} \left(\mathrm{mol}/\mathrm{L}\right)}\times 100$$

(2)

Leaching studies were conducted by hot filtration and recycle tests to establish the catalyst stability during tributyrin transesterification in methanol at 110 °C. The recovered catalyst was filtered between recycles, repeatedly washed with methanol to remove weakly bound residues, and then dried at 80 °C. After the third cycle, the catalyst was analyzed by ICP-OES to determine the final Ni and Al content.

4. Conclusions

NiAl LDHs with Ni:Al molar ratios spanning 1.5–4.1 were synthesized via an alkali-free route using NH₄OH and NH₄CO₃ as pH regulator and precipitant, respectively. The influence of activation protocol on 350 °C calcined Ni_xAl LDH was explored, with hydrothermal reconstruction at 110 °C proving to be the most effective for regenerating the parent LDH crystalline lamellar structure and maximizing base site loading. The extent of hydrothermal regeneration increased with Ni:Al ratio, attributed to suppressed NiO and AlO_x phase separation during calcination for higher Ni content. Base site loading increased with Ni content, attributed to undercoordinated oxygen sites as Al³⁺ is replaced by Ni²⁺, and/or increased surface hydroxyls on lamellae. The initial rates of tributyrin transesterification with methanol (and final conversion) increased with Ni loading and, hence, the base site density. However, TOFs also increased with Ni:Al ratio, indicative of a change in base site strength or accessibility across the family (the latter consistent with enhanced reconstruction). The most active Ni_4.1Al LDH catalyst was moderately stable towards site-blocking, and effective for the transesterification of C₄–C₁₂ triglycerides to FAMEs. Future work will explore the versatility of these catalysts for transesterification of real oil feedstocks and the development of macroporous architectures in order to improve the mass transport of bulky TAGs. The somewhat weaker basicity of NiAl versus MgAl LDH counterparts may be advantageous with respect to poisoning by fatty acid impurities or adventitious CO₂.