**Layered Double Hydroxides as Bifunctional Catalysts for the Aryl Borylation under Ligand-Free Conditions**

#### **Lorenna C. L. L. F. Silva 1, Vinícius A. Neves 1, Vitor S. Ramos 2, Raphael S. F. Silva 3, José B. de Campos 2, Alexsandro A. da Silva 4, Luiz F. B. Malta 1,\* and Jaqueline D. Senra 1,4,\***


Received: 29 January 2019; Accepted: 18 March 2019; Published: 27 March 2019

**Abstract:** Organic derivatives of boron, such as boronic esters and acids, are important precursors for a wide range of environmental, energy, and health applications. Several catalytic methods for their synthesis have been reported, even though with the use of toxic and structurally complex ligands. Herein, we demonstrate preliminary studies envisaging the synthesis of boronic esters from an inexpensive catalytic system based on Cu/Al layered double hydroxides (LDH) in the presence of Na2PdCl4. The Cu/ Al LDHs were prepared according to coprecipitation method and characterized by X-ray diffraction (XRD) (with Rietveld refinement) to evaluate the contamination with malachite and other phases. Preliminary catalytic results suggest that pure Cu/Al LDH has potential for the borylation of aryl iodides/ bromides in the absence of base. Indeed, a synergic effect between copper and palladium is possibly related to the catalytic efficiency.

**Keywords:** boronic esters; borylation; Suzuki–Miyaura; layered double hydroxides; copper; palladium

#### **1. Introduction**

Organoboron compounds have received great attention in last years due to significant impact in analytical [1], technological [2], and medicinal fields [3]. Particularly, low-weight compounds containing boronic acid/ester moieties have played an important role in the synthesis of new hybrid materials, sensors and complex organic molecules, through synthetic protocols such as Suzuki–Miyaura reactions. However, mild and selective synthesis of boronic acids and esters still represent a challenge. In general, the classic reactions involving the generation of organometallics (e.g., arylmagnesium or aryllithium) followed by the reaction with a borate have been substituted by catalytic protocols [4]. Recently, new methods involving metal-free conditions have also been discovered [5,6]. However, most of them suffer from low reactivity/selectivity along with relatively high costs of the boron reagents. Some advantages of the catalytic protocols are the improved reaction selectivity and the possibility of the whole system reuse. Since the general steps for the catalytic synthesis of boronic acids and esters involve a sequence of oxidative addition and reductive elimination, several semihomogenous or heterogeneous systems are described based on the employment of noble metals (e.g., Pd and Ir) as catalysts [7].

In 1995, Miyaura first reported the Pd-catalyzed aryl borylation protocol by using a phosphine-based catalyst—Pd(dppf)Cl2—in the presence of KOAc for the activation of diboron reagent [8]. Since then, several efficient methodologies have been reported based on similar reaction conditions [9–11]. Recently, Ratniyom and coworkers [12] reported an efficient cooperative catalysis based on Pd(0)/Cu(I) in the presence of triphenylphosphine with good results towards aryl iodides and bis(pinacolate)diboron. However, the yields were very sensitive to the nature of the base and reduced drastically with the use of aryl bromides/chlorides.

Inorganic networks with basic properties represent one alternative to avoid the use of expensive ligands as well as strong bases in order to achieve the B–B bond activation. Mostly, bimetallic systems usually involve low loadings of a noble metal in combination with an early transition metal to allow a cost-effective, broader functional group tolerance, and scalable reaction condition. Layered double hydroxides (LDHs)—anionic clays also known as hydrotalcites—have a structure based on layers of bi- and trivalent metals hydroxides and interlayer spaces occupied by anions and neutral molecules [13], see Figure 1. They can be composed of a wide range of different metals and anions, which make them useful as anionic exchangers, flame retardants, catalysts, and supports for metallic nanoparticles [14], and have low-cost of synthesis and basic surface properties [15]. Silva and coworkers [16] previously reported a catalyst system based on Mg/Al LDHs, PdNPs, and cyclodextrins for the efficient Suzuki–Miyaura reaction between aryl bromides and arylboronic acids. Lately, Sreedhar [17] employed Cu/Al LDHs as catalysts in Ullmann reactions between aryl chlorides and amines. So far, cross-coupling with heterogeneous ligand-free Cu(II) based cocatalysts has not been reported to date in borylation reactions.

**Figure 1.** Representation of layered double hydroxide (LDH) hydrotalcite-like structure, with the Cu/Al hydroxide layers and the carbonate containing interlayer space. Cu → green, O → red, H → blue, C → grey, Al → light red.

In the present work, we disclosed Cu/Al LDHs as bifunctional cocatalysts in the presence of Na2PdCl4 for the borylation reactions between aryl halides and bis(pinacolato)diboron. The key points examined here are the influence of LDH purity and the catalytic conditions towards the reaction outcome.

#### **2. Results**

#### *2.1. Synthesis and Characterization of Cu/Al LDH*

A set of conditions were varied to accomplish the synthesis of the Cu/Al LDH. Figure 2 shows the X-ray diffraction (XRD) patterns of samples obtained using or not sodium carbonate as an additive. These diffractograms were refined using the Rietveld method in order to obtain the composition of phases. In addition, the refinement also afforded the unit cell parameters for LDH phase: these data are available in the supporting information file (Tables S1–S4). The LDH synthesized without sodium carbonate (Figure 2a) exhibited a XRD pattern composed of reflections from the mineral-like phases of gerhardtite (Cu(OH)NO3, JCPDS-ICDD 01-082-1991), hydrotalcite (Mg0.67Al0.33(OH)2)(CO3)0.165(H2O)0.48, JCPDS-ICDD 01-089-5434), and nitratine (NaNO3, JCPDS-ICDD 00-036-1474). The LDH phase was obtained as a low crystallinity, minority phase with wide and low intensity peaks at 9.7◦ and 19.9◦. The Cu(OH)NO3 phase (gerhardtite) was obtained as the majority phase with the most intense peaks at 12.7◦, 25.6◦, 33.9◦, 36.1◦, 40.4◦, and 42.9◦. Finally, NaNO3 (nitratine) was found as another minority phase, however showing a thin peak at 2θ = 22.8◦, besides 29.4◦, 31.8◦, and 38.9◦.

**Figure 2.** Rietveld-refined X-ray diffraction (XRD) patterns for Cu/Al LDH obtained (**a**) without and (**b**) with Na2CO3 dissolved in the precipitation agent solution.

In contrast, Figure 2b shows the XRD pattern for the sample synthesized using sodium carbonate. In this case, the majority phase was the layered double hydroxide with peaks at 2θ = 11.6◦, 23.5◦, 35.6◦, 39.5◦, and 47.0◦. Another significant phase was malachite (Cu2(OH)2CO3, JCPDS-ICDD 01-075-1163) related to the peaks 2θ = 11.9◦, 14.7◦, 17.5◦, 24.0◦ 29.7◦, 31.7◦, 32.6◦, and 35.4◦. As minority phases, it was found that spertiniite (Cu(OH)2 JCPDS-ICDD 01-080-0656) related to peaks at 2θ = 16.7◦, 23.8◦, 34.0◦, 35.9◦, and 39.8◦, and nitratine with peaks at 2θ = 22.8◦, 29.4◦, 31.8◦ and 38.9◦.

Regarding the LDH unit cell parameters (Table S1, supporting information), both XRD were refined as having a rhombohedral crystal system and belonging to the R-3m spatial group. However, the sample synthesized without carbonate presented the biggest unit cell dimensions (a = 3.039 Å and c = 26.921 Å against a = 2.979 Å and c = 22.519 Å for the carbonate Cu/Al LDH), which means that an expanded structural network was obtained.

The addition of sodium carbonate to the reaction medium allowed obtaining LDH as the main phase, but did not avoid forming malachite, a by-product. Still, it was decided to use the carbonate salt dissolved in the precipitation agent solution in the fore coming tests.

In the next set of experiments the LDH precipitation pH and postsynthesis work-up were evaluated.

The precipitation pH was varied between two values, 8 and 10, which are the most used in LDH synthesis. The XRD patterns for both samples are presented in Figure 3. Figure 3a evidences the refined XRD pattern for the LDH precipitated at pH = 8, which, compared to that of Figure 3b, related to LDH precipitated at pH = 10; thus pH is not a significant parameter in the synthesis. This is also supported by strong similarities of unit cell dimensions between these two samples (Table S2, supporting information). Therefore, a pH of 8 was chosen as the working pH in the next tests.

**Figure 3.** Rietveld-refined XRD patterns for Cu/Al LDH obtained at (**a**) pH = 8 and (**b**) pH = 10.

In all experiments the LDH postsynthesis work-up proceeded using centrifugation, but one sample was isolated using filtration. Their refined XRD patterns are shown in Figure 4. Concerning the unit cell dimensions, no significant differences between these two samples can be perceived (Table S3, supporting information) As evidenced when nitratine phase percentages are compared the centrifuged sample (Figure 4a) it exhibits 5-fold more NaNO3 than the filtrated solid (Figure 4b), which implies that the postsynthesis work up is an important parameter to be aware of.

To reinforce that postsynthesis steps are very important one last sample was synthesized following the conditions established in the present study; however, this LDH was submitted to washing, using organic solvents (EtOH/Acetone). Its refined XRD measurements are presented in the Figure 5. The calculated unit cell parameters show no significant differences from those previously shown (Table S4, supporting information). In the phase distribution it is clearly observed that the malachite % dropped from 31.93% (Figure 4b) to 10.23% (Figure 5), increasing LDH % from 47.92% to 65.70%.

Besides, the XRD peaks shown in Figure 5 appear wider and less intense than those in Figure 4b, for instance. According to the Full Width Half the Maximum (FWHM) criterium used in the Scherrer formula, the mean crystal size evolved from 21 nm (for samples of Figures 3 and 4) to 14 nm, signaling a decrease in crystallinity upon postsynthesis treatment, such as washing with organic solvents.

To evaluate the effect on the borylation reaction, catalytic tests were carried out by using malachite and LDH as catalysts.

**Figure 4.** Rietveld-refined XRD patterns for Cu/Al LDH obtained after (**a**) centrifugation and (**b**) filtration.

**Figure 5.** Rietveld-refined XRD pattern for Cu/Al LDH obtained after washing with ethanol and acetone.

#### *2.2. Synthesis of Aryl Boronic Esters Employing Cu/Al LDH Catalyst*

The reaction between bis(pinacolato)diboron, (B(pin)2) and 1-iodo-4-nitrobenzene was taken as a model. Initially, purified Cu/Al LDH and malachite were evaluated as catalysts with acetonitrile as solvent. From these data, it is possible to confirm that malachite did not show catalytic activity towards the borylation reaction. To evaluate the copper loading, we used the LDH composition according to the method previously described by our group [18]. It was observed that both copper-based catalysts were not able to catalyze the reaction (Table 1, entries 1 and 2). Similarly, the use of Na2PdCl4 as the sole catalyst was ineffective (Table 1, entry 3). Indeed, addition of CuSO4 in combination with Na2PdCl4 did not react under this condition (Table 1, entry 4). The comparison of the Pd precursor led us to evaluate a semihomogeneous system composed of palladium nanoparticles (PdNPs) stabilized by cyclodextrins (Table 1, entries 5 and 6). Our research group have already described this catalytic system to the carbon–carbon cross-coupling reactions [19]. However, it was not observed significant conversion even when using acetonitrile. Remarkably, the addition of 2 mol% Na2PdCl4 in the presence of LDH (30 mol% Cu) allowed a yield of 98% of the expected product (Table 1, entry 7). Since the positive effect could be related to the LDH structural and compositional properties, we also tested the most common Mg/Al LDH in the presence of CuSO4 and Na2PdCl4, which rendered a good yield (Table 1, entry 8). Having in mind that the Cu/Al LDH is an anionic exchanger, it has been tested a [PdCl4] <sup>2</sup><sup>−</sup> exchanged Cu/Al LDH, and surprisingly, the yield obtained was significantly lower, 35% (Table 1, entry 8).

#### **Table 1.** Survey of reaction condition.

2


<sup>1</sup> Determined by gas chromatography–mass spectrometry (GC-MS); <sup>2</sup> Use of Mg/Al LDH, CuSO4 and Na2PdCl4; <sup>3</sup> LDH intercalated with Na2PdCl4; <sup>4</sup> solvent = THF; <sup>5</sup> solvent = dioxane; <sup>6</sup> Addition of Cs2CO3.

In order to see the effect of copper in the reaction, we have tested different copper loadings (Table 1, entries 10–12). According to the results, a relatively high loading of copper has shown to be the most adequate for a high conversion (Table 1, entry 7). Similarly, to verify the influence of the Pd loading it was varied from 2 to 0.05% under the same conditions (Table 1, entries 13–15). In this case, 2 mol% Pd was necessary to keep an acceptable turnover. Analogously to Cu%, it was observed a direct relationship between the conversion rate and the Pd%. With the aim of evaluating the effect of noncoordinating solvents, THF and dioxane (Table 1, entries 16 and 17) were also tested under the conditions described in Table 1, entry 3. Surprisingly, no appreciable yield of the product was observed in both cases. Additionally, the use of base hampered the reaction (Table 1, entry 18).

Since the above results demonstrated that the catalytic system based on Cu/Al LDH and Pd(II) can be efficiently used for the borylation of an aryl iodide under ligand-free conditions, we carried out preliminary reactions with some aryl halides in order to examine its applicability. In this case, we have tested the influence of electron-withdrawing/donating groups substituted in the aromatic core. By considering their comparatively low-cost and availability, we have mainly focused on the evaluation of the reactivity of aryl bromides.

In general, it was possible to note an influence of the electron-withdrawing/ donating capabilities of the substitutional groups. Remarkably, it was evident that the strongly electron-donating amine group did not favor the reaction in both cases (Table 2, entries 1 and 5). In general, the absence of a substituent led to small yields.


**Table 2.** Preliminary scope for the borylation reaction catalyzed by Cu/Al LDH and Na2PdCl4 1.

<sup>1</sup> Determined by GC-MS.

Encouraged by the preliminary results, we next investigated whether this catalyst could have recycling potential. In such case, it could shed a light on the behavior of Cu/Al LDH as a reservoir or as real catalytically active species. However, attempts to recover the material after reaction work up failed.

#### **3. Discussion**

Malachite is a common second phase present in Cu/Al LDH synthetic samples [20–26], and it is known to affect the general properties of the LDH, such as its catalytic properties [23–25]. In most cases, it arises from the need to use sodium carbonate as a precipitation co-agent, otherwise LDH is not formed as the majority phase, as observed in Figure 2a. The way the unit cell dimensions obtained for this phase (Table S1, supporting material) evidenced an expanded network in comparison to the carbonate containing LDH phases. This can be understood in terms of attraction electrostatic forces acting between layer and interlayer parts of the material that permit to pack more efficiently when CO3 <sup>2</sup><sup>−</sup> instead of NO3 − is used as the intercalated ion.

Some authors in the literature have tried to eliminate the malachite impurity by a number of different approaches.

Muñoz et al. [23] sought for pure Cu/Al LDH in order to obtain its calcined-oxide derivative to catalyze the reduction of NO and CO gases. Through coprecipitation method synthesis they were able to verify that precursors solutions with lower concentrations prevented the formation of malachite, and that the oxide derived from malachite-free LDH showed better catalytic results than the derived from impure LDH. Gao et al. [24] synthetized Cu/Zn/Al/Zr LDH in order to obtain its calcined oxide for the catalysis of CO2 hydrogenation, and also found that catalysts with malachite prior to the calcination presented worse catalytic activity than the ones without the contamination. Both Muñoz and Gao agree that higher ratios of Cu/Al induce the formation of malachite. Ichikawa et al. [25] tried a different way to obtain noncontaminated Cu/Al LDH: they used a coprecipitation method followed by an aging process at 90 ◦C under air bubbling for 1 h. They stated that this process helped to smooth the crystallization process removing the excess of CO3 <sup>2</sup><sup>−</sup> ions. They also found that the calcined oxide derived from the pure Cu/Al LDH was a better catalyst to the conversion of acrylonitrile to acrylamide than the contaminated one. Recently, Qu et al. [26] reported a new mechanochemical synthesis method for Cu/Al LDH, in which they dry milled Cu2(OH)2CO3 and Al(OH)3 at a planetary ball mill for 2 h at 600 rpm, producing an amorphous solid mixture that was treated in aqueous medium at room temperature under magnetic stirring for 4 h, producing a pure Cu/Al LDH.

Another important feature is the presence of nitratine (NaNO3) in the phase mixture composition after LDH synthesis. Firstly, this phase arises from the combination of sodium of NaOH solution and nitrate from salt precursors of Cu2+ and Al3+. However, considering sodium nitrate is soluble in aqueous medium, it is not clear how this phase coprecipitates with LDH. Figure 4 gives a hint: the centrifugation process leads to 5-fold increase of nitratine% in the phase mixture composition. In addition, Figure 5 points to a decrease by half of this phase percentage guaranteed by washing with EtOH/Acetone. Therefore, the combined use of filtration and washing in the postsynthesis work-up of LDH is beneficial for this catalyst.

The catalytic conditions studied for the borylation model reaction were initiated with an activated aryl iodide since electron-withdrawing groups are known for the accelerated effect on the reaction rate [7]. Under the basic conditions of the classical Pd cycle, it is assumed that the higher the Ar-Pd(II)-X electrophilicity, the faster the transmetallation with bisorganodiboron. In fact, the results pointed to a high catalytic conversion by using 4-iodonitrobenzene as an electrophile. To evidence the clean and selective conditions for the borylation reaction, the crude 1H NMR spectrum of **3** is presented in the Supporting Information. A relatively high copper loading was, however, necessary to increase the substrate conversion. As pointed out, this fact has already been observed by Ratnyom [12] and other groups [13] but involving a typically homogeneous catalytic system based on Cu or Pd/Cu combined with phosphines. Regarding the usual organometallic mechanism, it is interesting to note that the LDH surface is possibly responsible for the presence of basic species involved in the preactivation step. The addition of base, however, was detrimental to the reaction, suggesting a partial decomposition of the LDH structure.

Even though the cooperative catalysis of Pd and Cu was already reported in some recent works, the ligand-free condition has been scarcely explored mostly because of the difficult electronic tuning towards the transmetallation step even though the soft coordination of acetonitrile cannot be ruled out. However, the surprising fact here is related to the superior activity of a heterogeneous layered Cu matrix, which allowed high substrate conversion and product selectivity in the reaction model. In detriment, common copper salts, such as the basic copper carbonate (malachite), were ineffective under the same conditions, indicating that the chemical environmental around copper is decisive to the reaction outcome even in the presence of palladium. This fact could be reinforced when testing Mg/Al LDH in the presence of CuSO4 and Na2PdCl4 that resulted in an inferior product yield compared to the same catalytic conditions by using the Cu/Al matrix.

In addition, the preliminary scope of the method showed an unclear electronic effect of substituents: apparently, the reaction is favored by both electron-donating and -withdrawing groups but, intriguingly, it was not the case for the amine substituent. It is conceivable that a hybrid mechanism is operative with the moderate to low yields of α-naphtyl, 4-bromobenzyl and phenyl bromides arising from a predominant oxidative addition modulation. However, the low reactivity of bromobenzene was intriguing. We presume the occurrence of possible alternative pathways (e.g., Ullmann-type reaction and hydrodehalogenation) in minor extension but the solubility factor cannot be ruled out. Indeed, the role of LDH matrix as a catalytic reservoir of active Cu species can be considered based on experiments with coordinating solvents [27]. According to the frustrated catalyst recycling attempts, it is also possible that acetonitrile act as a moderate LDH exfoliation agent. On the other hand, the effect of PdCl4 <sup>2</sup><sup>−</sup> intercalation in LDH seems to slow down the possible Pd(II) reduction. Lastly, (Bpin)2 seems to be involved in the reduction of Pd(II) species, as suggested by the solution darkening according to visual inspection.

Taken together, these results suggest that a synergic effect between the Cu(II) and Pd(0) species can arise from a combination of factors related to the electronic effects on substrate along with the chemical environment of both metals.

A schematic view containing some of these effects is shown in Scheme 1.

**Scheme 1.** Proposed mechanism for the Pd and Cu-catalyzed aryl borylation in the presence of LDH.

#### **4. Materials and Methods**

All the reactants and solvents in this work were commercially acquired and used without previous purification.

The powder X-ray diffractograms were recorded by a diffractometer Panalytical X´Pert PRO MPD (Malvern Panalytical Ltd., Royston, United Kingdom), with generator of Cu X-ray (λ = 1.5418 Ǻ), line focus (1.8 kW), universal Theta-2Theta goniometer with 240 mm radius, divergence slit fixed <sup>1</sup> 2 ◦, flat-diffracted beam monochromator, and proportional detector Xe with 40 kV of tension and 40 mA of current in the X-ray tube. The samples were analyzed self-supported in aluminum sample holder and used under the Bragg–Brentano geometry; it used ranges between 5◦ < 2θ < 70◦, with 0.05◦ step and 2.5 s per step.

The calculation based on the Rietveld Refinement applied in this work, through the program TOPAS academic V5.0 (Coelho Software, Brisbane, Australia), was based on the Fundamental Parameters Approach, based on the instrumental parameters with background correction. If necessary, the following parameters were refined; unit cell dimensions, sample height displacement, zero-shift, weight fraction (scaling), preferred orientation, atomic species/substitutions, atomic coordinates, site occupancies, thermal displacement parameters, crystallite size, and lattice strain.

All the organic products were characterized by 1H and 13C NMR spectroscopy (Bruker Analytics, Berlin, Germany).

#### *4.1. Synthesis of the Cu/Al Layered Double Hydroxides*

The Cu/Al layered double hydroxides (Cu/Al LDH) and Cu4Al2(OH)12CO3·4H2O, were synthetized using a constant pH method: one aqueous solution of Cu(NO3)2·3H2O (0.225 mol·L−1) and Al(NO3)3·9H2O (0.075 mol·L−1) and other of NaOH (0.50 mol·L−1) and, alternatively, Na2CO3 (0.15 mol·L−1), were added (40 mL each) simultaneously by continuous dropping to 40 mL of a NaOH solution (pH = 8 or 10), under magnetic stirring at room temperature. The pH was maintained during the process. A blue slurry was formed, which was then filtrated or centrifuged and dried at room temperature. The filtrated material was alternatively washed with ethanol:acetone 1:1 volume mixture.

#### *4.2. Synthesis of Boronic Esters Employing Palladium Nanoparticles*

The synthesis of PdNPs followed method described by Senra et al. [19], Scheme 2. At first, 1 mL of a Na2PdCl4 0.005 mol·L−<sup>1</sup> aqueous solution was mixed with 69 mg of <sup>β</sup>-cyclodextrin in a 10 mL glass flask and the mixture was kept at 70 ◦C for 1 h under magnetic stirring. Then, it was added to the flask 0.25 mmol of 1-bromo-4-methoxybenzene, 0.25 mmol de bis(pinacolato)diboron, 0.375 mmol of K2CO3, and 2 mL of solvent (ethanol/H2O 50% or acetonitrile). The system was kept at 70 ◦C for 24 h, and then the reaction mixture was extracted with dichloromethane and NaCl aqueous saturated solution (1:1). After the extraction, the mixture was dried with anhydrous sodium sulfate and evaporated under reduced pressure.

#### *4.3. Synthesis of Boronic Esters Employing Na2PdCl4/Cu/Al LDH*

Into a 5-mL screw cap flask was added 0.9 mmol de bis(pinacolato)diboron, 0.6 mmol de 1-iodo-4-nitrobenzene, 2 mL of solvent, and 0.15 mmol (Cu) of LDH/Pd or 0.15 mmol (Cu) of Cu/Al LDH and 2 mL of Na2PdCl4 0.005 mol·L−<sup>1</sup> aqueous solution, Scheme 3. In the latter, the percentage of palladium varied from 2% to 0.5%. The system was kept at 70 ◦C for 24 h under magnetic stirring, and its extraction was done as described in Supporting Information.

**Scheme 3.** Model reaction.

#### **5. Conclusions**

Cu/Al LDH had some synthesis parameters evaluated in order to produce the least quantity of by-products. For such material, the list of contaminants includes the mineral-like phases malachite, spertiniite, and nitratine. The catalytic system based on the least-contaminated LDH heterogeneous matrix and aqueous Na2PdCl4, under ligand-free conditions, showed efficient catalytic properties in the borylation reactions by using bis(pinacolato)diboron as boron precursor. This catalyst system can be a promising low-cost alternative for use in classical homogeneous phosphine-based systems. Additional experiments to investigate the possible reaction intermediates and their correlation with the LDH structure are under investigation and will be reported in due course.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4344/9/4/302/s1: Figure S1: 1H NMR of **3**, Figure S2: 13C NMR of **3**, Figure S3: 1H NMR of **4**, Figure S4: 13C NMR of **4**, Figure S5: 1H NMR of **6**, Figure S6: 13C NMR of **6**, Figure S7: 1H NMR of **7**, Figure S8: 13C NMR of **7**, Figure S9: 1H NMR of **8**, Figure S10: 13C NMR of **8**, Figure S11: 1H NMR of **9**, Figure S12: 13C NMR of **9**).

**Author Contributions:** L.C.L.L.F.S. conducted all the borylation reactions. V.A.N. prepared the catalysts. V.S.R. and J.B.d.C. conducted the XRD analysis and the Rietveld refinements. R.S.F.S. and A.A.d.S. conducted the product characterization. L.F.B.M. and J.D.S. supervised the work and wrote the paper.

**Funding:** This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES), Finance Code 001. All authors acknowledge the financial support by CNPq and FAPERJ. Specifically, L.F.B.M. would like to thank the funding from CNPq, Universal project code 425613/2016-0, and from FAPERJ, Jovem Cientista do Nosso Estado project code E-26/203.212/2017.

**Acknowledgments:** The authors thank CBPF (Centro Brasileiro de Pesquisas Físicas), UFRJ and UERJ for the analytical support.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Aldol Condensation of Furfural with Acetone Over Mg/Al Mixed Oxides. Influence of Water and Synthesis Method**

#### **Almudena Parejas, Daniel Cosano, Jesús Hidalgo-Carrillo \*, José Rafael Ruiz, Alberto Marinas , César Jiménez-Sanchidrián and Francisco J. Urbano**

Departamento de Química Orgánica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUIQFN, Universidad de Córdoba, Campus de Rabanales, Edificio Marie Curie, E-14071 Córdoba, Spain; q12pabaa@uco.es (A.P.); q92cohid@uco.es (D.C.); qo1ruarj@uco.es (J.R.R.); alberto.marinas@uco.es (A.M.); qo1jisac@uco.es (C.J.-S.); fj.urbano@uco.es (F.J.U.)

**\*** Correspondence: jesus.hidalgo@uco.es; Tel.:+34-957-218-638

Received: 23 January 2019; Accepted: 20 February 2019; Published: 23 February 2019

**Abstract:** Aldol condensation of furfural and acetone (an important initial step to obtain diesel from biomass) was studied over MgAl mixed oxides. The influence of the utilization of microwaves and/or a surfactant (Pluronic 123) during the synthesis as well as the use of water (either pre-hydrating the solids before catalytic studies or in water/toluene mixtures as the reaction medium) is discussed. The combined use of Pluronic 123 and microwaves led to solids with bigger pore sizes, exhibiting lower basicity and higher acidity than the conventional synthetic method, thus resulting in an increase in the yield of the desired product of condensation, comprising two molecules of furfural and one of acetone (F2Ac). As for the influence of water, re-hydration of the mixed oxides was detrimental to activity, probably as a result of the partial blocking (solvation) of active sites. On the contrary, the increase in water percentage in the reaction medium resulted in higher conversions, though selectivity to F2Ac decreased. The weakening of the C=O bond of furfural in the presence of water as well as the higher solubility of the first condensation product (FAc) in toluene, as compared to water, could account for that. A 44.5% yield of F2Ac (66% conversion) after 16 h was obtained with the most active solid, which maintained the activity for three consecutive reactions.

**Keywords:** aldol condensation; biomass valorization; Mg/Al mixed oxides; surfactant; microwaves; influence of water

#### **1. Introduction**

Fossil fuel depletion and environmental concern have boosted the search for renewable energies, one of the possible sources being biomass [1,2]. Furfural is a so-called platform molecule from biomass obtained through xylose dehydration [3,4] and can be transformed into a wide range of chemicals via hydrogenation, oxidation, decarbonylation, nitration, or condensation processes, just to cite some of them [5]. For instance, aldol condensation and subsequent hydrogenation and hydrodeoxygenation can lead to liquid hydrocarbons for use as diesel [6–8].

Aldol condensation is a well-known C–C bond formation process which can occur in acidic or basic sites, the latter being more frequently reported in the literature [6–15]. It requires the existence of a reactive hydrogen in alpha position, with respect to a carbonyl compound able to form an enol, which reacts with another carbonyl compound, and after dehydration, yields a conjugated enone. Focusing on aldol condensation between furfural and acetone (Figure 1), it can initially lead to 4-(2-furanyl)-3-buten-2-one (FAc), a subsequent aldol condensation with another furfural molecule, forming 1,5-bis-(2-furanyl)-1,4-pentadien-3-one (F2Ac) (Figure 1a) [16,17]. Some side reactions include

acetone self-condensation to form diacetone-alcohol and mesityl oxide (Figure 1b), condensation between FAc and acetone (Figure 1c), and multiple aldol condensations between different carbonyl compounds, thus forming polymers [18,19] (Figure 1d).

Aldol condensations have been traditionally performed in organic media, using base catalysts such as sodium or calcium hydroxides. Nevertheless, the existence of corrosion problems and the difficult reutilization have led to the use of some other base heterogeneous catalysts, such as hydrotalcites and hydrotalcite-derived mixed oxides [20,21], amorphous aluminophosphate [15], and diamine-functionalized MCM-41 [22], just to cite some examples.

In the present work, different AlMg mixed oxides were obtained through calcination of layered double hydroxides (LDHs) and tested for aldol condensation with acetone to form F2Ac. The influence on the catalytic results of two synthetic variables (conventional or microwave heating with the presence or absence of Pluronic 123 as the surfactant) was explored. Furthermore, the effect of water (either pre-hydrating the solids before catalytic studies or in water/toluene mixtures as the reaction medium) is discussed.

**Figure 1.** Reaction scheme for aldol condensation of furfural and acetone. Some side reactions have also been included. (**A**) aldol condensation between furfural and acetone, (**B**) acetone self-condensation, (**C**) condensation between FAc and acetone and (**D**) multiple aldol condensations between different carbonyl compounds.

#### **2. Results and Discussion**

#### *2.1. Textural, Structural and Acid–Base Characterization of the Solids*

X-ray diffractograms of uncalcined and calcined hydrotalcites are shown in Figure 2. As can be seen, uncalcined solids exhibit a typical hydrotalcite crystallinity profile (JCPDS 22-700), with symmetric reflections at 2θ = 11◦, 22◦, 36◦, 37◦, 45◦, 60◦, and 62◦. Therefore, sharper peaks corresponded to (003), (006), (010), and (013) reflections, whereas broader signals were obtained for (009), (015), and (016) reflections, all of them representative of layered materials. As for calcined solids (Figure 2b), diffraction patterns are very similar to each other, exhibiting (111), (200), and (220) reflections ascribed to periclase. In a previous paper, Aramendia et al. [23], using 27Al NMR-MAS, demonstrated that the coordination of Al3+ changed from octahedral to tetrahedral upon calcination of hydrotalcites, Al3+ ions thus isomorphically substituting Mg2+ ions, forming MgAlOx periclase.

**Figure 2.** X-ray diffractograms of the different solids synthesized in the present work. (**a**) Uncalcined solids. (**b**) Solids calcined at 450 ◦C.

Thermal stability of hydrotalcites was determined by TG-DTA (Figure 3). In all cases, weight loss percentage is in the 42–45% range (Figure 3a). HTCON and HTMW thermogravimetric profiles are consistent with those reported in the literature for hydrotalcites [24,25]. Therefore, two main weight losses are observed. The first one at 100–200 ◦C is ascribed to the loss of intercalated water molecules, whereas nitrate coming from both the precursor and hydroxyl groups can account for the second loss at higher temperatures (250–500 ◦C). For the solids synthesized using the surfactant (HTCONP and HTMWP), the second weight loss seems to be produced quicker (i.e., at lower temperatures), thus suggesting that for those systems, re-structuration to form periclase is somehow favored by Pluronic 123. Heat flow profiles (Figure 3b) seem to confirm this hypothesis, the exothermal peak centered at 450 ◦C in HTCON and HTMW being shifted to lower temperatures (300–350 ◦C) for HTCONP and HTMWP.

**Figure 3.** TG analyses (**a**) and heat flow (**b**) of hydrotalcites.

Raman spectra of uncalcined solids (i.e., hydrotalcites) are represented in Figure 4. The band appearing at 557 cm−<sup>1</sup> can be assigned to vibrations of brucite-like octahedral layers, Al-O-Mg, which are present in all Mg-Al hydrotalcites [26]. Moreover, the spectra also exhibit bands at 710 and 1055 cm−1, corresponding to nitrate vibrations [27] and bands at ca. 3500 cm−1, due to surface hydroxyl groups. In the case of HTCONP and HTMWP solids, there are also some intense bands of C-H stretching Pluronic 123 at 2986, 2941, and 2933 cm−<sup>1</sup> [28].

**Figure 4.** Raman spectra of the uncalcined solids (hydrotalcites).

N2 adsorption–desorption isotherms of calcined solids are shown in Figure 5. In all cases, type IV isotherms corresponding to mesoporous materials were obtained. BET surface areas, pore volume and average pore diameter values are given in Table 1. With regards to the BET areas, they are in the 160–210 m2·g−<sup>1</sup> range, the highest value corresponding to HTCONP-450. Modification of conventional synthesis by using microwave irradiation and/or in the presence of the surfactant (Pluronic 123) led in all cases to an increase in BET area. Solids that aged under microwave irradiation exhibit smaller pore diameters than their conventionally-heated counterparts (compare HTMW-450 vs. HTCON-450 or HTMWP-450 vs. HTCONP-450). Systems synthesized in the presence of the surfactant present bigger pores (compare HTCONP-450 vs. HTCON-450 and HTMWP-450 vs. HTMW-450). Therefore, the effect of microwaves and the presence of a surfactant on pore volume is the opposite. However, if both variables are changed simultaneously, the influence of the surfactant is more important, thus resulting in the pore diameter increasing (compare HTCON-450 and HTMWP-450 with pore diameters of 6.8 and 8.4 nm, respectively).

**Figure 5.** Nitrogen adsorption–desorption isotherms corresponding to the mixed oxides.


**Table 1.** Summary of the main features of the mixed oxides synthesized in this work.

X-Ray fluorescence results (Table 1) evidenced a good incorporation of Mg and Al to the solids, with Mg/Al ratios very similar to the nominal value (Mg/Al = 2).

Base characterization of the solids was performed using thermal programmed desorption of pre-adsorbed CO2 (CO2-TPD) and the results are given in Tables 1 and 2, and in Figure 6. In all cases, signals were deconvoluted in peaks, which, depending on the desorption temperature, were ascribed to weak (80–200 ◦C), medium (200–300 ◦C), or strong (>300 ◦C) basic sites, respectively. Taking HTCON-450 as the reference, the use of microwave irradiation and/or the presence of the surfactant in the synthesis results in a drop in total basicity. Interestingly, as far as the base site distribution is concerned, the effect of microwave irradiation, the presence of the surfactant, or both variables simultaneously considered is different. Therefore, in the absence of Pluronic 123, microwave irradiation does not vary base site distribution. On the contrary, the presence of Pluronic 123 results in an increase in the strong base sites' percentage, to the detriment of weak ones. Finally, simultaneous use of microwaves and Pluronic 123 leads to an increase in the percentage of weak base sites.

**Figure 6.** Temperature-programmed desorption profiles of CO2 for the mixed oxides synthesized in this work.

**Table 2.** Base site distribution of the solids expressed as μmol CO2/g. Values in brackets represent the percentage of the total basicity.


Acid characterization of the solids was performed by thermal programmed desorption of pre-adsorbed pyridine (Py-TPD), results being given in Tables 1 and 3, and in Figure 7. Taking HTCON-450 as the reference, contrary to basicity, total acidity increases when the solids are synthesized utilizing a microwave and/or in the presence of Pluronic 123. Furthermore, with regards to acid site distribution, only microwave irradiation has some effect (acid strength decreases), whereas the presence of Pluronic 123, either under conventional heating or microwave irradiation, does not vary acid site distribution.

**Figure 7.** Temperature-programmed desorption profiles of pyridine for the solids synthesized in this work.

**Table 3.** Acid site distribution of the solids expressed as μmol Py/g. Values in brackets represent the percentage of the total basicity.


All in all, microwave irradiation and/or the use of Pluronic 123 as the surfactant result in an increase in total acidity and a decrease in total basicity together with an increase in BET areas.

#### *2.2. Catalytic Activity*

The solid synthesized under conventional heating and in the absence of the surfactant (HTCON-450) was used as the reference material in order to study the influence of reaction temperature and the presence of water in the reaction medium on catalytic activity.

#### 2.2.1. Influence of Reaction Temperature

Table 4 summarizes catalytic results obtained for t = 3 h at different temperatures. From that table, it is evident that the increase in temperature results in the increase in conversion, whereas selectivity to F2Ac, the desired product, hardly changes. Therefore, from then on, 100 ◦C was selected as the reaction temperature.

**Table 4.** Aldol condensation of furfural and acetone on HTCON-450: Influence of temperature on catalytic activity. Reaction conditions: Reactor pressurized to 5 bar with N2; 10 mmol of furfural, 20 mmol acetone, 20 mL toluene, and 400 mg catalyst, t = 3 h.


#### 2.2.2. Influence of Water

Two different approaches were made to study the influence of water on catalytic activity of HTCON-450. On the one hand, the solid was re-hydrated using a N2 flow saturated in water. On the other hand, reactions were performed in water/toluene mixtures (0%, 5%, 10%, 50%). Regarding the former approach (Table 5), the pretreatment of HTCON-450 with a flow of nitrogen saturated in water results in a drop in conversion (35.0% and 23.7% for HTCON-450 and HTCON-450-rehydrated, respectively, at t = 3 h), whereas selectivity values to F2Ac are quite similar. This suggests that rehydration results in the elimination of active sites to a certain extent, which could be ascribed to solvation. Results obtained for the uncalcined HTCON solid (which exhibits very low catalytic activity) are also given for the sake of comparison.

**Table 5.** Aldol condensation of furfural and acetone on HTCON-450: Influence of calcination and rehydration of the solid. Reaction conditions: Reactor pressurized to 5 bar with N2; 10 mmol of furfural, 20 mmol acetone, 20 mL toluene, and 400 mg catalyst, t = 3 h, 100 ◦C.


Table 6 summarizes the results obtained for the study conducted using different water/toluene ratios as the reaction medium. As can be seen, the higher the percentage of water, the higher the conversion, but in general, the lower the selectivity to the desired product, F2Ac. In a previous paper [29] on chemoselective hydrogenation of alpha,beta-unsaturated carbonyl compounds, our research group found evidence by Raman spectroscopy that water interacted with the carbonyl group and made the double bond weaker and thus more reactive (carbonyl band shifted to lower wavelength values). The same could occur in the C=O group in furfural and account for its higher conversion in the presence of water. With regards to the change in selectivity, one should consider that we are working in a biphasic media and thus the catalyst hydrophilic character, as well as the relative solubility of reactants and products both in toluene and water, is important. Active sites in the catalyst will probably interact better with water than with toluene. Furfural is partially soluble in water (50–100 mg·mL<sup>−</sup>1). Its condensation with one molecule of acetone will produce FAc, whose solubility in water is much lower (1–10 mg·mL<sup>−</sup>1). Therefore, once formed, FAc will pass to the organic phase (toluene) and will not be able to undergo subsequent condensation with another acetone molecule to produce F2Ac. This results in the increase in selectivity to FAc and probably in conversion, since FAc is retired of the aqueous phase as the reaction proceeds. All in all, the highest yield to F2Ac, the desired product, is

achieved with pure toluene. Therefore, this reaction medium was selected for subsequent studies on other catalysts.

**Table 6.** Aldol condensation of furfural and acetone on HTCON-450: Influence of the presence of water in water/toluene mixtures. Reaction conditions: Reactor pressurized to 5 bar with N2; 10 mmol of furfural, 20 mmol acetone, 20 mL (toluene + water) and 400 mg catalyst, t = 3 h, 100 ◦C.


#### *2.3. Catalytic Activity of the Other Mixed Oxides*

Once pure toluene and 100 ◦C had been selected as the reaction conditions for aldol condensation of furfural and acetone in order to obtain F2Ac, the study was extended to the other mixed oxides. Reactions were conducted at 3 h and 16 h, the main catalytic results being summarized in Table 7.

A first conclusion from that table is that the lowest conversion values correspond to HTMW-450. It is important to note that this solid was the one exhibiting the lowest pore diameter (4.2 nm, Table 1), which could account for that. Focusing on the other solids, the highest conversion value at 3 h is achieved for HTCON-450, whereas as the reaction proceeds, the rate is higher for the systems synthesized using Pluronic 123, which together with their higher selectivity to F2Ac results in F2Ac yields of 24.6%, 28.3%, and 44.5% for HTCON-450, HTCONP-450, and HTMWP-450, respectively, at t = 16 h. As evidenced by thermal-programmed desorption of pyridine and CO2, the use of microwave irradiation and/or the presence of Pluronic 123 in the reaction medium during the synthesis resulted in an increase in total acidity and a decrease in total basicity. In a previous paper, Climent et al. [15] described the cooperative effect of weak acid and base sites of an amorphous aluminophosphate in aldol condensation, thus resulting in higher selectivities than those presented by stronger acid or base catalysts. This effect together with the increase in pore size could explain the higher yields obtained for the solids synthesized using the surfactant.

**Table 7.** Results obtained for aldol condensation of furfural and acetone on the different solids. Reaction conditions: Reactor pressurized to 5 bar with N2; 10 mmol of furfural, 20 mmol acetone, 20 mL toluene and 400 mg catalyst, 100 ◦C.


#### *2.4. Reutilization of HTCON-450 and HTMWP-450*

Finally, some reutilization studies were conducted on HTCON-450 and HTMWP-450 solids, results being summarized in Table 8. In all cases, the Mg/Al ratio of the solids was quite similar to the nominal value (Mg/Al = 2). Moreover, after the reactions, the reaction medium was analyzed by inductively coupled plasma mass spectrometry (ICP-MS). No Mg or Al was detected which is evidence of the stability of the solids, which do not undergo leaching.


**Table 8.** Results obtained for reutilization studies. Reaction conditions: Reactor pressurized to 5 bar with N2; 10 mmol of furfural, 20 mmol acetone, 20 mL toluene and 400 mg catalyst, 100 ◦C.

As far as the catalytic activity is concerned, neither HTCON-450 nor HTMWP-450 exhibited any remarkable deactivation keeping F2Ac yield in the ca. 20% order after three hours. In the case of the most active solid at long reaction times (HTMWP-450), its activity and selectivity only decreased slightly (from 67.1 to 58.5%) after three consecutive uses.

#### **3. Materials and Methods**

Hydrotalcites were synthesized by a co-precipitation method, starting from two solutions containing 0.2 mol Mg(NO3)2·6H2O and 0.1 mol Al(NO3)3·9H2O in 25 mL deionized water, respectively (Mg/Al = 2). The mixture was slowly added to a pH 10 aqueous solution under continuous stirring and an inert atmosphere (N2), with temperature maintained at 60 ◦C. During precipitation, the pH value was maintained, adding NaOH 1M. The suspension was divided into four portions for further treatment. One part was kept under conventional heating at 80 ◦C for 24 h, followed by filtration and washing with deionized water, thus obtaining the solid called HTCON. A second portion was aged under microwave heating at 80 ◦C for 1 h, thus leading, after filtration and washing, to the solid termed as HTMW. A flexiWave platform for microwave synthesis (22 V, 50 Hz) with an IR temperature sensor (p/n IRT0500) was used. The other two portions were submitted to the same conventional or microwave heating while performing the synthesis in the presence of surfactant Pluronic 123 (2% by weight), thus leading to the solids named HTCONP and HTMWP, respectively. Finally, all four solids were calcined at 450 ◦C in the air for 8 h (1 ◦C·min−<sup>1</sup> ramp). Nomenclature of these solids include the suffix 450, referring to calcination temperature (HTCON-450, HTMW-450, HTCONP-450, and HTMWP-450). Subsequent treatment of HTCON-450 for 2 h at 450 ◦C in the presence of a N2 flow (50 mL·min<sup>−</sup>1) saturated in water at 20 ◦C led to a solid called HTCON-450-rehydrated.

A Setaram SetSys 12 instrument (SETARAM Instrumentation, Caluire, France) was used for thermogravimetric analyses (TGA). Experiments were performed on 20 mg samples placed in an alumina crucible and heated in the 30–600 ◦C range (10 ◦C·min<sup>−</sup>1, 50 mL·min−<sup>1</sup> air stream).

Textural properties (BET surface area, cumulative pore volume, and average pore diameter) were measured in a Micromeritics ASAP-2010 instrument (Micromeritics, Norcross, GA, USA.). Samples were heated at 120 ◦C and degassed to 0.1 Pa before measurement.

The measure of magnesium or aluminium leaching (presence in filtered reaction medium) was performed by inductively coupled plasma mass spectrometry (ICP-MS) on a Perkin–Elmer ELAN DRC-e instrument.

The Mg/Al ratio of solids was measured by X-ray fluorescence (XRF) spectroscopy (Rigaku ZSK PrimusIV wavelength X-ray spectrometer (Rigaku, The Woodlands, TX, USA). Further details are given elsewhere [30].

Raman spectra were recorded on a Renishaw spectrometer (InVia Raman Microscope, Renishaw, Gloucestershire, UK), equipped with a Leica microscope with various lenses, monochromators, filters, and a CCD detector. Spectra were recorded over the 150–4000 cm−<sup>1</sup> range, using green laser light excitation (532 nm) and gathering 32 scans.

X-ray diffraction (XRD) analysis was performed on a Siemens D-5000 diffractometer (Bruker Corporation, Billerica, MA, USA) using CuKα radiation over the range 5–80◦.

Surface acidity of samples was measured by thermal programmed desorption of pre-adsorbed pyridine (Py-TPD) using TC detection. Samples (30 mg) were cleaned by heating to 450 ◦C (10 ◦C·min−<sup>1</sup> ramp) under He flow (75 mL·min<sup>−</sup>1) and then cooled down to 50 ◦C. The catalysts were subsequently saturated with pyridine for 30 min, cleaned for 60 min with He and TPD monitored from 50 to 450 ◦C (10 ◦C·min<sup>−</sup>1), the final temperature being held for 45 min.

Surface basicity of the catalysts was determined on a Micromeritics Autochem II instrument by thermal programmed desorption of pre-absorbed CO2 (CO2-TPD) with TCD detection. Samples (100 mg) were cleaned in an Air stream (20 mL·min−<sup>1</sup> Ar, heating at 450 ◦C at a rate of 10 ◦C·min−<sup>1</sup> for 60 min and then cooled down to 40 ◦C). Then, solids were saturated with CO2 (5% CO2/Ar flow at 20 mL·min−<sup>1</sup> for 60 min), physisorbed CO2 removed with Ar flow (20 mL·min−<sup>1</sup> for 30min) and TPD monitored from 50 to 450 ◦C (5 ◦C·min<sup>−</sup>1), the final temperature being held for 60 min.

The solids were tested for aldol condensation of furfural using a Berghof HR-100 stainless steel high-pressure autoclave equipped with a 75 mL PTFE insert vessel. Under standard conditions, 10 mmol of furfural, 20 mmol acetone, 20 mL toluene, and a 400 mg catalyst were introduced in the vessel. Reactor was purged with nitrogen and pressurized to 5 bar of N2. The reaction temperature was set to 100 ◦C and started by switching on the stirring at 750 rpm. To stop the reaction, the vessel was submerged in an ice bath. The choice of toluene as the organic medium was motivated by a previous paper [3] on xylose dehydration to furfural where toluene was found to give the highest yield to furfural. The final strategy would be to make the one-pot transformation of xylose to furfural and then F2Ac.

Experiments to evaluate the influence of the presence of water in the reaction medium were conducted varying the water/toluene ratio (0%, 5%, 10%, and 50% volume) while keeping the total solvent volume constant (20 mL).

Once the reactions were finished, the products were analyzed by gas chromatography (Agilent 7890) with a flame ionization detector (GC-FID), using a Supelco NukolTM capillary column. In the case of using biphasic media (toluene/water mixtures), products were extracted from the aqueous phase with dichloromethane before GC-FID analysis. Quantification of furfural and condensation products was performed using the appropriate calibration curves. In all cases, mass balance considering unreacted furfural, FAc, and F2Ac was over 95%.

For reutilization experiments, after the reaction, the solids were filtered, washed with ethanol, and dried at 100 ◦C, followed by calcination at 450 ◦C under the same conditions as described in the synthesis. Nomenclature of reused catalysts include the suffix R (one reuse) or R2 (two reuses).

Furfural conversion and FAc and F2Ac selectivity were defined by Equations (1)–(3):

$$\text{Furfural conversion } (\%) \text{: } \frac{\text{initial furfural concentration } - \text{final furfural concentration}}{\text{initial furfural concentration}} \times 100 \quad (1)$$

$$\text{FAc velocity } (\%) \text{:} \frac{\text{FAc concentration}}{\text{FAc concentration} + 2 \cdot \text{F2Ac concentration}} \times 100 \tag{2}$$

$$\text{F2Ac velocity } (\%) \text{: } \frac{\text{2-F2Ac concentration}}{\text{FAc concentration + 2-F2Ac concentration}} \times 100\tag{3}$$

#### **4. Conclusions**

The synthesis of hydrotalcites in the presence of Pluronic 123 led, after calcination, to MgAl mixed oxides with bigger pore sizes than untreated solids. On the other hand, microwave irradiation led to smaller pore sizes as compared to conventional thermal treatment. As far as acid–base characteristics are concerned, the use of both microwave irradiation and Pluronic 123 during the synthesis resulted in a decrease of total basicity and an increase in total acidity.

Rehydration of mixed oxides by treating them with a nitrogen flow saturated with water led to solids exhibiting lower catalytic activity in aldol condensation of furfural, probably as a result of the partial blocking (solvation) of active sites. By contrast, the increase in the percentage of water in water/toluene biphasic media resulted in an increase in conversion values, though selectivity to FAc also increased at the expense of the desired product F2Ac. A plausible explanation is that

water weakens the C=O bond in furfural, thus favoring its transformation. Moreover, once FAc is produced, its higher solubility in toluene, as compared to water, favors its transfer to the organic medium, thus avoiding its subsequent reaction with another furfural molecule to yield F2Ac. The fact that the produced FAc is retired to the organic phase could also account for the observed increase in conversion.

A comparison of catalytic activity of the reference material (HTCON-450) with that of the other solids allows us to conclude that the use of Pluronic 123 during synthesis (especially in combination with microwave irradiation) resulted in solids exhibiting higher F2Ac yields at long reaction times. This could be the result of the combination of two factors: The above-mentioned larger pore size achieved with the surfactant and the increase in total acidity which could favor aldol condensation.

HTMWP-450 exhibited a good stability without any significant loss of activity after three uses.

**Author Contributions:** Conceptualization, C.J.-S. and F.J.U.; methodology, A.M. and J.R.R.; validation, F.J.U., J.R.R. and J.H.-C.; formal analysis, A.M. and J.R.R.; investigation, A.P. and D.C.; data curation, A.M., J.H.-C. and A.P.; writing—original draft preparation, A.P. and D.C.; writing—review and editing, J.H.-C and A.M.; supervision, C.J.-S. and F.J.U.

**Funding:** This research was funded by Ramón Areces Foundation.

**Acknowledgments:** The scientific support from the Central Service for Research Support (SCAI) at the University of Cordoba is acknowledged.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Catalysts* Editorial Office E-mail: catalysts@mdpi.com www.mdpi.com/journal/catalysts

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18