Catalytic Upgrading of Ethanol to 1-Butanol Biofuel Additive Using Pd/MgO-Al₂O₃ and Bimetallic Pd-Cu/MgO-Al₂O₃ Mixed Oxide Catalysts

Abstract

Catalytic upgrading of bioethanol via a C–C coupling reaction is a sustainable method of producing 1-butanol, a high-performance biofuel. This reaction was studied using a flow-through microreactor system with Pd/MgO-Al₂O₃ and bimetallic Pd-Cu/MgO-Al₂O₃ mixed oxide-based catalysts in a H₂ carrier gas at a pressure of 21 bar and temperatures ranging from 200 to 350 °C. The effect of the metal promoter(s) on the hydrogen transfer reaction steps in the overall reaction was investigated. The palladium promoter significantly improved the activity and butanol selectivity across the entire temperature range. However, the yield of liquid products decreased significantly at temperatures higher than 250 °C, primarily because the decarbonylation side reaction of the acetaldehyde intermediate accelerated. The promoting effect of Pd was most beneficial below 250 °C because the decarbonylation reaction was inhibited by the reversible poisoning effect of CO on multiple Pd sites responsible for decarbonylation. Diluting the Pd phase with Cu increased liquid yields due to gradually decreasing decarbonylation activity. However, the dehydrogenation–hydrogenation activity decreased as well, as did the promoting effect on the corresponding reaction steps in the coupling reaction. Additionally, the product distribution changed dramatically, decreasing 1-butanol selectivity, because metallic Cu can catalyze the formation of ethyl acetate and ketone products.

1. Introduction

Biobutanol is considered a potential alternative to bioethanol presently used in large quantities as renewable blending component of gasoline. It has several advantages over bioethanol, including a higher energy density (29.2 MJ/L vs. 24.0 MJ/L), a non-corrosive nature due to its low miscibility with water, a lower vapor pressure, a higher allowable blend ratio, and research octane number close to that of gasoline [1,2,3]. It can be used as a pure fuel or blended with gasoline in high concentration without any modification of the conventional engines [1].

The Guerbet self-coupling reaction of bioethanol [4,5] provides a promising perspective for 1-butanol synthesis from bioethanol. The reaction requires basic catalysts or catalysts with both base and acid sites. The mechanism of the ethanol self-coupling reaction is complex and remains debated due to its sensitivity to catalyst type and reaction conditions. In the most commonly proposed pathway, C–C bond formation occurs through the aldol condensation of two acetaldehyde intermediate molecules via surface enolate species [4,5]. Because the reaction intermediates are a primary enolate and an aldehyde, the self-coupling reaction of ethanol produces 1-butanol. The aldol condensation pathway consists of a series of reaction steps involving four main reaction steps: (i) ethanol is dehydrogenated to produce acetaldehyde, (ii) acetaldehyde undergoes aldol self-condensation to form acetaldol, (iii) acetaldol is dehydrated to form crotonaldehyde, and (iv) crotonaldehyde is hydrogenated (in two steps) to produce 1-butanol. During the final hydrogenation steps, if molecular hydrogen and an active metal component are not present, hydrogen must be transferred from an alcohol molecule to the unsaturated aldol product. This most likely occurs via the Meerwein–Ponndorf–Verley (MPV) reduction mechanism [4,5,6,7]. During the MPV reduction process, the transfer of hydrogen from ethanol results in the formation of acetaldehyde, which can then participate in the chain reaction.

Other proposed mechanisms are the so called direct condensation and semi-direct condensation pathways [4,5,8,9,10,11]. In the direct coupling pathway, the β-hydrogen of the ethanol molecule becomes activated, causing it to participate in condensation reaction with another ethanol molecule. Note that the direct coupling pathway was suggested for strong base catalysts, such as MgO or alkaline metal zeolites, in the absence of a metal promoter at relatively high reaction temperatures (over 350 °C), where unwanted side reactions could have been significantly accelerated [4,9]. The semi-direct pathway begins with the dehydrogenation of an ethanol molecule, giving an acetaldehyde intermediate. This intermediate then reacts with an ethanol molecule to form either butanal or crotyl alcohol. During the final hydrogenation step, hydrogen is transferred from an alcohol molecule to the unsaturated aldol product most likely via the MPV mechanism as suggested for the aldol condensation pathway (vide supra). The contribution or dominance of the direct, semi-direct, or aldol condensation pathways in the ethanol coupling reaction to butanol depend on the catalyst and the reaction conditions, especially on the reaction temperature, and still is a matter of debate [10,11,12].

The catalyst is a key factor in the performance of the ethanol coupling reaction and the distribution of its products. Heterogeneous catalytic systems are preferred over homogeneous catalytic systems for industrial-scale ethanol coupling because there is no catalyst separation problem, and, ideally, the system can operate continuously for long periods without extensive regeneration, improving efficiency, activity, and sustainability. These catalytic systems include basic, alkaline metal form zeolites [8,13], magnesia [9], non-stoichiometric hydroxyapatite [6,7], and the most widely studied magnesia–alumina mixed oxides [14,15,16], which are also the main subject of this work.

The mixed oxide catalysts, formed by incorporating aluminum into the structure of basic magnesia, are among the most often used catalysts, as they have strong and moderately strong base sites and sites of medium and strong acid strength necessary for the aldol condensation and dehydration steps to take place [14,17]. They can be prepared by calcining the corresponding hydrotalcite precursor [14,18]. An advantage of this catalyst is that by varying the Mg/Al ratio in the hydrotalcite precursor and/or the calcination temperature, the strength of the base and acid centers can be varied to shift the selectivity towards the Guerbet alcohol product and minimize the undesired side reactions, e.g., formation of ether and ester and dehydration of alcohols to olefins [4,14,15,16,19]. In a recent work, we demonstrated in a combined experimental and kinetic modeling study that, for ethanol–ethanol coupling to 1-butanol, aldol condensation is the dominant pathway below 335 °C, while semi-direct coupling to crotyl alcohol prevails above 340 °C when using a MgO–Al₂O₃ mixed oxide (Mg/Al = 2:1) derived from a hydrotalcite precursor [20].

The ethanol self-coupling reaction does not require external H₂ source and/or dehydrogenating–hydrogenating metal sites. However, the initial and final hydrogen transfer steps of the aldol condensation and semi-direct condensation pathways can be promoted by introducing active metal catalyst component. Transition metal promoters have been widely used to promote ethanol conversion to butanol by facilitating the dehydrogenation of ethanol and/or the hydrogenation of the unsaturated reaction intermediates. Transition metals such as Pd, Ru, Ni, Co, and Cu were found to improve the activity of mixed oxides in the Guerbet coupling reaction by promoting the hydrogen transfer reaction steps [21,22,23,24]. The promoters can effectively lower the required temperature for the overall coupling reaction; however, a metal of high activity, such as Pd, can facilitate undesired hydrodeoxygenation (HDO) reactions, mainly decarbonylation reactions leading to significant amount of alkane byproducts and CO [23,25,26]. Conversely, a transition metal with low hydrogenation–dehydrogenation activity, such as Cu, is less effective in accelerating the hydrogen transfer steps, but exhibits low HDO and/or decarbonylation activity [25,26]. These results suggest that the hydrogenation–dehydrogenation and acid-base functions of the catalyst must be balanced to achieve a high reaction rate and selectivity toward the desired coupling product, butanol, although the effects of these functions on the reaction are still far from being fully understood.

The selectivity of noble metal-containing heterogeneous catalysts can be controlled by adding a second metal with lower hydrogenation–dehydrogenation activity. Optimal changes in the surface structure and/or electronic properties can inhibit side reactions and accelerate the desired reaction [27]. In recent years, bimetallic Pd-Cu catalysts have been used in several heterogeneous catalytic reactions such as C-C bond formation, selective hydrogenation of alkynes, and hydrogenation of CO₂ [25,26,28,29,30]. For example, Goulas et al. [25] observed increased selectivity of ethanol-acetone-butanol (ABE mixture) and 1-octanol coupling reactions using a Pd-Cu alloy catalyst (with a Pd/Cu atomic ratio of 3) supported on MgO-Al₂O₃/activated carbon support compared to a supported Pd catalyst. They attributed the significantly improved oxygenate coupling selectivity to the dramatic suppression of the decarbonylation side reaction.

In a previous work, we investigated the promoting effect of Pd, Pt, Ni and Ru metals on a MgO-Al₂O₃ mixed oxide catalyst in the Guerbet coupling reaction of ethanol [19]. The most favorable effect was obtained with the Pd promoter (1 wt.% Pd on MgO-Al₂O₃) under reducing atmosphere using H₂ carrier gas. However, the beneficial effect of Pd could be observed only within a narrow temperature range (200–250 °C), where the equilibrium allowed significant concentrations of both dehydrogenated and hydrogenated products formed in the initial dehydrogenation and closing hydrogenation steps, respectively. At higher reaction temperatures (≥275 °C) mainly the decarbonylation side reaction on the metallic active sites, producing C₁–C₃ alkanes and CO, was significantly accelerated.

In the present work, the strategy suggested by Goulas et al. [25,26] was applied to suppress the decarbonylation activity of Pd via modification with Cu. The Pd atoms in a 1 wt.% Pd/MgO-Al₂O₃ catalyst were partially replaced by Cu atoms to obtain Pd-Cu/MgO-Al₂O₃ catalysts with increasing molar fraction of Cu (0.25–0.75) to improve the butanol selectivity of the ethanol self-coupling reaction. Our results show that the introduction of Cu can effectively suppress the decarbonylation activity of the catalyst. However, an increasing molar fraction of Cu, greater than 0.25, weakens the promoting effect of Pd in the dehydrogenation–hydrogenation steps of the coupling reaction and reduces the selectivity of butanol due to the acceleration of side reactions that lead to C₄ oxygenates, such as ethyl acetate, as well as C₄ ketones and aldehydes. The underlying reasons of the changing catalytic behavior are discussed.

2. Materials and Methods

2.1. Catalyst Preparation

Details on the preparation of MgO-Al₂O₃ mixed-oxide catalysts can be found elsewhere [19]. In brief, the hydrotalcite precursor with a Mg/Al atomic ratio of 2 was prepared using the co-precipitation method. An aqueous solution was prepared using Mg(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O (Sigma-Aldrich, Merck KGaA, Darmstadt, Germany). The solution containing Mg²⁺ and Al³⁺ ions at concentrations of 1 M and 0.5 M, respectively, was slowly added to NaOH buffer solution at a pH of 11.5 with continuous stirring at room temperature. Simultaneously, 2 M NaOH solution was added dropwise to maintain the pH at about 10. Based on our previous study, the hydrotalcite precursor was calcined in air at 550 °C for three hours to produce a MgO-Al₂O₃ mixed oxide catalyst with favorable acid-base and catalytic properties [19]. This MgO-Al₂O₃ mixed oxide catalyst was also used as a support material to prepare the Pd and/or Cu containing MgO-Al₂O₃ mixed oxide catalyst samples. The metal promoters were introduced to the support via wet impregnation method using aqueous solutions containing calculated amount of Pd(NO₃)₂·4NH₃ (5% Pd containing solution, abcr, Karlsruhe, Germany) and/or Cu(NO₃)₂·3H₂O (Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) metal precursor compounds. Monometallic catalysts containing 1 wt.% Pd (93.96 μmol Pd/g_cat.) or 0.59 wt.% Cu (93.96 μmol Cu/g_cat.) on MgO-Al₂O₃ and three additional bimetallic catalyst samples were prepared, having 93.96 μmol metal/g_cat., wherein the catalyst contained Cu beside Pd atoms in molar fractions of 0.25, 0.50, and 0.75. After drying at 70 °C for 24 h, the samples were heated at 550 °C in air to decompose the precursor compounds. The catalyst samples are designated as Pd_1.0/MgO-Al₂O₃, Pd_0.75Cu_0.25/MgO-Al₂O₃, Pd_0.5Cu_0.5/MgO-Al₂O₃, Pd_0.25Cu_0.75/MgO-Al₂O₃ and Cu_1.0/MgO-Al₂O₃. The catalyst samples were reduced in situ in H₂ flow at 500 °C for 2 h before the catalytic runs.

2.2. Catalysts Characterization

2.2.1. X-Ray Powder Diffraction (XRPD)

The XRPD measurements were performed using a Philips PW 1810/3710 (Amsterdam, The Netherlands) diffractometer equipped with a graphite monochromator (CuKα radiation, λ = 1.5418 Å) and an Anton Paar (Graz, Austria) HT1200 high-temperature chamber. The XRPD patterns were recorded in steps of 0.02° 2θ with a scan duration of 5 s per step, while the X-ray tube operated at 40 kV and 35 mA.

2.2.2. Transmission Electron Microscopy (TEM)

TEM investigations were carried out using a JEOL JEM 1011 microscope (Tokyo, Japan) equipped with a side-mounted Morada 11-megapixel camera (Olympus Soft Imaging Solutions, Münster, Germany)., to study the morphology of the samples. The samples were suspended in ethanol and drop-casted onto a 300 mesh copper grid, coated by Formvar and Lacey carbon support film. The TEM images were obtained at an accelerating voltage of 80 kV.

2.2.3. The Specific Surface Area (SSA) of the Catalyst

The SSA of the catalysts was determined based on the N₂ adsorption isotherms. The samples were pretreated at 250 °C in vacuum of 10⁻⁶ mbar for 4 h, then the adsorption isotherms were measured at the temperature of the liquid nitrogen (−196 °C) using an automatic, volumetric adsorption analyzer (The “Surfer”, Thermo-Fisher Scientific, Waltham, MA, USA). The SSA of the samples were calculated using the Brunauer–Emmett–Teller (BET) method.

2.2.4. Temperature-Programmed H₂-Reduction (H₂-TPR)

The H₂-TPR measurements on the metal-containing catalyst samples were carried out in a home-made apparatus. About 200 mg sample from the 0.5–0.8 mm sieve fraction of the catalyst was calcined in air at 550 °C, placed in a U-shape quartz reactor tube (I.D. = 4 mm) and pretreated in a 30 cm³/min N₂-flow at 550 °C for 1 h, then allowed to cool to room temperature in the N₂ stream. The gas flow was switched to a 30 cm³/min flow of a 10 vol.% H₂/N₂ reducing gas mixture. The sample was then heated at a ramping rate of 10 °C/min to 600 °C and held at this temperature for 1 h. The effluent gas was passed through a liquid nitrogen cooled trap to remove the water from the effluent gas before reaching the thermal conductivity detector (TCD) used to monitor the change in the H₂ concentration during the reduction process. The hydrogen consumption was determined from the area under the H₂-TPR curve using a previously determined calibration value obtained by the reduction of a known amount of CuO under the same conditions.

2.2.5. CO Pulse Chemisorption

The concentration of surface metal sites was determined by CO pulse chemisorption experiments carried out using the same apparatus as for the H₂-TPR measurements. Approximately 100 mg of the catalyst sample (sieve fraction of 0.5–0.8 mm) was placed in the quartz reactor tube and reduced in situ in a 50 cm³/min H₂ flow at 500 °C for 1 h. The sample was flushed then with a 20 cm³/min He flow and allowed to cool to room temperature in a He flow. At room temperature, pulses of 10 µL CO were injected into the He flow that was passing through the catalyst sample. A thermal conductivity detector (TCD) monitored the CO concentration of the He. The injection of CO pulses was continued until the sample was saturated by CO. The amount of chemisorbed CO was calculated from the TCD peak areas of the CO pulses.

2.2.6. Fourier-Transform Infrared (FT-IR) Spectroscopy

The electronic state of the MgO-Al₂O₃ supported metal particles was characterized by analyzing the FT-IR spectra of adsorbed CO. The FT-IR spectrum of the surface species from the adsorption of CO were recorded in transmission mode using a Nicolet 6700 FT-IR (Thermo-Fisher Scientific, Waltham, MA, USA) equipped with a home-made in situ IR-cell equipped with CaF₂ windows. A self-supported wafer (with a thickness of 5–10 mg/cm²) was prepared from the catalyst sample and placed into the IR-cell. Prior to CO adsorption, the catalyst wafer was reduced in situ in H₂ flow at 450 °C for 1 h, then evacuated under high vacuum (10⁻⁶ mbar) at this temperature for 30 min and cooled to room temperature under evacuation. The spectrum of the activated catalyst sample was collected at room temperature, then the sample was exposed to 5 mbar CO at room temperature, followed by evacuation to remove the gas phase CO from the cell. The sample spectrum, containing the spectrum of the surface species from the adsorption of CO was collected and then a difference spectrum was generated by subtracting the spectrum of the wafer collected before CO adsorption from the spectrum of the CO-loaded wafer. All spectra were collected by averaging 128 scans with a resolution of 2 cm⁻¹. To enable quantitative comparisons, each spectrum was scaled to 5 mg/cm² wafer thickness.

It should be noted that the development of the spectra of the adsorbed species was monitored over time in the presence of 5 mbar of CO gas and during evacuation at 10⁻⁶ mbar to determine the steady-state CO coverage of the catalyst surface. Since the band intensities stabilized within 15 min in both cases, spectra collected after 20 min equilibration were used for further analysis.

2.2.7. Quasi-Operando Diffuse Reflectance Infrared Fourier-Transform (DRIFT) Spectroscopy

The formation of surface species from ethanol and their transformation on the catalyst surface were studied using a Nicolet iS10 spectrometer (Thermo Scientific) equipped with a DiffuseIR mirror system and a flow-through DRIFT reactor cell (PIKE Technologies, Fitchburg, WI, USA). The cell design allows the flow of a carrier gas or a gas phase reactant mixture to pass through the catalyst bed at a controlled flow rate. The sample holder of the reactor cell was filled with the finely powdered catalyst sample (about 20 mg). The metal-containing catalysts were in situ reduced in a 30 cm³/min H₂ flow at 450 °C for 1 h then was cooled to 30 °C in H₂ flow. The adsorption and reaction of ethanol (EtOH) was initiated by switching the H₂ flow to a saturator unit filled with ethanol. The H₂ gas flow, saturated with ethanol vapor at 0 °C (1.58 vol.% EtOH/H₂), was passed through the catalyst bed for 30 min at 30 °C. Then the reaction temperature was raised at a rate of 2 °C/min up to 350 °C under continuous flow of the EtOH/H₂ reactant mixture. Spectra were taken in every five minutes (10 °C increments) between 30 and 350 °C. The series of the sample spectra were corrected with the spectra of the catalyst in pure H₂ flow collected at about the same temperatures. The resulting difference spectra are related to both molecularly adsorbed ethanol and surface species formed in a surface reaction, as well to gas-phase ethanol above the catalyst. However, the contribution of the gas-phase spectrum was proved to be nearly negligible under the applied conditions. Therefore, the absorption bands appearing in the difference spectra can be attributed predominantly to surface species formed (positive bands) or consumed (negative bands) during the adsorption and/or reaction process on the catalyst surface. Note that the catalytic system was in a transient state because the reaction temperature was increased continuously throughout the experiment.

2.3. Catalytic Tests

The ethanol coupling reaction was investigated using a high-pressure, fixed-bed, flow-through catalytic reactor system. The reaction temperatures varied between 200 and 350 °C in a H₂ carrier gas, while the total pressure was 21 bar. The weight hourly space velocity (WHSV) of ethanol was 1 ${\mathrm{g}}_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}\cdot }{\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}^{-1}{\cdot \mathrm{h}}^{-1}$, whereas the H₂/ethanol molar ratio was set to 5. Two grams of the catalyst sample (sieve fraction of 0.5–0.8 mm) were loaded into a stainless-steel catalytic reactor tube (I.D. = 9 mm) and was in situ reduced in the reactor at 500 °C in a 50 cm³/min H₂ flow for two hours prior to the catalytic run. The ethanol feed was provided by a high-precision HPLC pump (Gilson, Madison, WI, USA), while the H₂ flow was controlled by a mass flow controller (Brooks, Hatfield, PA, USA). The effluent gas mixture from the reactor was passed through a separator unit to separate the liquid and gas phase products. The pressure in the catalytic system was set by an electronic back pressure regulator (Brooks, Hatfield, PA, USA) placed upstream of the reactor and the separator unit. The liquid product mixture was collected in every hour, when the steady state was reached in the catalytic system (usually after 1–3 h time-on-stream). The liquid products were analyzed by a gas chromatograph (Shimadzu 2010, Shimadzu Corporation, Kyoto, Japan) equipped with a Thermal Conductivity Detector (TCD) and a Flame Ionization Detector (FID) connected in series and a Zebron ZB-WAXplus capillary column (L 30.0 m × ID 0.32 mm × df 0.25 µm; Phenomenex, Torrance, CA, USA). Gaseous products, including CO₂, CO, and volatile C₁–C₄ hydrocarbons, were analyzed by GC-TCD-FID system (HP 5890 Series II online, Hewlett-Packard, Palo Alto, CA, USA) using a Carboxen^® 1006 PLOT capillary column (L 30 m × ID 0.32 mm; Supelco, Bellefonte, PA, USA).

The activity of the catalysts was evaluated based on ethanol conversion, yield of liquid products and product selectivity in the organic liquid phase product mixture, calculated by the following equations:

$$\mathrm{E}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n} (X_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}})={\displaystyle \frac{m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}^0-m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}}{m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}^0}}\times 100\%,$$

(1)

$$\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d} \mathrm{o}\mathrm{f} \mathrm{l}\mathrm{i}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{d} \mathrm{p}\mathrm{h}\mathrm{a}\mathrm{s}\mathrm{e} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{s}={\displaystyle \frac{m_{\mathrm{l}\mathrm{i}\mathrm{q}. }}{m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}^0}}\times X_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}},$$

(2)

$$\mathrm{S}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y} \mathrm{o}\mathrm{f} \mathrm{a}\mathrm{n} \mathrm{o}\mathrm{r}\mathrm{g}\mathrm{a}\mathrm{n}\mathrm{i}\mathrm{c} \mathrm{l}\mathrm{i}\mathrm{q}\mathrm{u}\mathrm{i}\mathrm{d} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}={\displaystyle \frac{c_i}{\Sigma c_i}}\times 100\%.$$

(3)

In Equations (1) and (2), $m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}^0$, and $m_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}$ mean the amount of EtOH in the feed and in the product mixture (g/h), respectively, whereas $m_{\mathrm{l}\mathrm{i}\mathrm{q}. }$ means the amount of the liquid product mixture (g/h). In Equation (3), $c_i$ and $\Sigma c_i$ mean the concentration of an organic product (wt.%) and the total concentration of all organic products (wt.%) in the liquid phase.

3. Results and Discussion

3.1. Catalyst Structure

3.1.1. The MgO-Al₂O₃ Mixed Oxide Used as Support Material

The characteristics of the MgO–Al₂O₃ mixed oxide catalyst used as a support material in the present study have been discussed in our previous work [19]. Briefly, calcining the hydrotalcite precursor at 550 °C produced a MgO–Al₂O₃ mixed oxide with a specific surface area of 218 m²/g. This mixed oxide contained slit-like pores of non-uniform size, which is typical of calcined hydrotalcites [31,32]. XRD measurements confirmed that the hydrotalcite precursor transformed into the corresponding mixed oxide during the calcination process, which exhibits only the characteristic reflections of the cubic periclase form of MgO with no evidence of an aluminum oxide phase. The absence of the aluminum oxide phase indicates that the aluminum is finely dispersed in the magnesium oxide (MgO) structure, possibly resulting in the formation of Al–O–Mg linkages. Characterization of the acid-base properties by temperature-programmed desorption of CO₂ and NH₃ (CO₂- and NH₃-TPD) revealed that the MgO–Al₂O₃ mixed oxide catalyst calcined at 550 °C contained medium-strong and strong basic sites at concentrations of 70 and 110 µmol/g_cat. It also contained medium-strong and strong Lewis acid sites at a concentration of 160 and 240 µmol/g_cat, respectively. These balanced acid-base properties were optimal for achieving the highest butanol selectivity in the ethanol coupling reaction compared to catalyst samples obtained at lower (450–500 °C) or higher (600 °C) calcination temperatures [19]. The literature agrees on the necessity of acid-base pair sites as active sites for the C-C coupling process [14,17]. However, the strength required for these pairs in each reaction step is still a matter of debate. Both the dehydrogenation of ethanol to acetaldehyde and the aldol condensation to n-butanol proceed via the initial formation of surface ethoxide species on a Lewis acid–strong base pair [14]. Acetaldehyde condensation to form 1-butanol is a bimolecular reaction between adjacent acetaldehyde species that requires strong basic centers. These centers are believed to be the most catalytically relevant function in the system because they control the formation of enolates, which leads to the C-C coupling step necessary for 1-butanol formation [17].

Note that neither the number nor the distribution of basic sites was altered in the presence of 1 wt.% Pd metal (see Figure S1 in the Supplementary Materials).

3.1.2. The Metal-Promoted MgO-Al₂O₃ Mixed Oxide Catalysts

The measured metal content of the catalyst samples, as determined by elemental analysis (see

Table 1. Characterization of the catalysts.

Catalyst ID a	Pd Cont. b, μmol/gcat.	Cu Cont. b, μmol/gcat.	SSA c, m2/g	Pore Vol., cm3/g	H2 Consumption d, μmol/gcat.	Ads. CO f, μmol/gcat.	CO/Pd g Molar Ratio
Pd1.0/ MgO-Al2O3	93.7	0	176	0.22	29.99 (0.64) e	14.97	0.16
Pd0.75Cu0.25/ MgO-Al2O3	70.1	23.1	134	0.24	36.79 (0.78) e	27.22	0.39
Pd0.5Cu0.5/ MgO-Al2O3	46.9	46.2	140	0.23	52.90 (1.13) e	12.60	0.27
Pd0.25Cu0.75/ MgO-Al2O3	23.3	70.3	134	0.23	70.55 (1.50) e	2.61	0.11
Cu1.0/ MgO-Al2O3	0	94.1	152	0.22	92.59 (1.97) e	0.28	-

^a The numbers in the catalyst ID indicate the nominal molar fractions of the metals. ^b Pd or Cu content of the catalysts determined by elemental (ICP-OES) analysis. ^c Specific surface area (SSA), determined by the Brunauer–Emmett–Teller (BET) method. ^d Hydrogen consumption calculated from the temperature-programmed H₂-reduction (H₂-TPR) curves. ^e The numbers in parenthesis indicate the H/(Pd + Cu) atomic ratio calculated from the H₂-TPR measurements. ^f Determined by CO pulse chemisorption method. ^g CO/Pd molar ratio calculated from the CO pulse chemisorption experiments, assuming that CO chemisorption on Cu metallic sites is negligible.

), is very close to the nominal value of 93.96 µmol (Pd + Cu)/g_cat., indicating that the calculated amounts of metal precursor compounds were nearly completely retained by the catalysts during wet impregnation.

The specific surface area of the metal containing catalyst samples ranged from 134 to 176 m²/g, and the pore volumes were similar (0.22–0.24 cm³/g). Note that during the wet impregnation process using the aqueous solution of the metal precursor salt, the MgO-Al₂O₃ mixed oxide support can partially reconstruct to the original hydrotalcite structure. After calcining the catalyst, containing the metal precursor, the metal oxide species become finely dispersed on the surface, potentially filling some of the mesopores [32]. This results in a lower surface area for the metal-containing samples compared to the MgO-Al₂O₃ mixed oxide support.

Figure 1 shows the XRPD patterns of the Pd-Cu/MgO-Al₂O₃ catalysts obtained after calcination and subsequent reduction of the catalyst precursors. The catalysts exhibit the same characteristic reflections of the cubic periclase form of magnesium oxide observed in the MgO-Al₂O₃ mixed oxide support [19], indicating that introducing the metal promoters did not alter the support’s structure. No characteristic reflections for metal phases were observed, suggesting that the size of the metal particles in the reduced catalysts is below the detection limit of XRPD (approximately 5 nm). These results are supported by the TEM images of the reduced catalyst samples, which show that the size of the metal particles in each sample is well below 5 nm (see Figure S2).

The XRPD patterns measured on the used catalyst samples (Figure S3) show no noticeable structural changes occurred during the catalytic reaction.

Figure 2 shows the H₂-TPR curves measured on the calcined metal containing catalyst precursors. The H₂-TPR curve measured on the Pd_1.0/MgO-Al₂O₃ sample presents a derivative-shape feature between about 70 and 150 °C consisting of an overlapping positive and negative peak (Figure 2, a). The positive peak can be assigned to the hydrogen consumption during the reduction of easily reducible Pd species, whereas the overlapping negative band is due to the release of hydrogen from β-PdH (beta palladium hydride) formed on the previously formed Pd⁰ particles [33,34,35]. The reduction of PdO-like species weakly interacting with the support, if present, would reduce at or below room temperature (this fraction remains undetected under the conditions applied during the H₂-TPR measurement). Therefore, the low-temperature reduction peak, starting around 70 °C, is probably due to the reduction of solvated Pd²⁺ weakly interacting with the mixed oxide support [34,36]. Meanwhile, the reduction peak that appears around 240 °C (Figure 2, a) is related to the relatively strong interaction between PdO and the support, whereas the peak at 370 °C is probably attributed to the combined effects of the strong interaction between PdO and the support, as well as the reduction of the support’s surface hydroxyl groups [35,37]. The H/Pd molar ratio of 0.64 calculated from the hydrogen consumption (Table 1) is about one third of the theoretical value of 2.0, which suggest that about two third of the Pd content was reduced at room temperature in the reducing gas flow before starting the temperature ramp during the H₂-TPR measurement.

The H₂-TPR curve obtained on the Cu_1.0/MgO-Al₂O₃ sample exhibits a relatively low intensity shoulder at about 220 °C and a broad reduction peak (probably consisting two overlapping component peaks) in the temperature range of 250–550 °C (Figure 2, e). The low-temperature peak can be assigned to the reduction of small CuO crystallites, whereas the peak(s) appearing over about 250 °C can be attributed to reduction of better stabilized Cu²⁺/Cu⁺ species probably forming solid solution within the oxide lattice [21,32]. The H/Cu atomic ratio, calculated from the H₂ consumption and the Cu content, was 1.97 for the Cu_1.0/MgO-Al₂O₃ sample (Table 1). This value matches well with the theoretical value of 2.0, suggesting that the reduction of Cu is complete in this sample during the H₂-TPR measurement.

The temperature range for the reduction of the bimetallic samples gradually shifted to lower temperatures at increasing Pd content (Figure 2, b–d) as the presence of Pd facilitates the reduction of Cu in the bimetallic particle [37,38,39]. This effect clearly suggests that close proximity between Cu and Pd species was achieved in the bimetallic samples [38]. The hydrogen consumption and the corresponding H/(Pd + Cu) atomic ratio for the bimetallic samples increases proportionally between 0.64 (Pd_1.0/MgO-Al₂O₃) and 1.97 (Cu_1.0/MgO-Al₂O₃) with the Cu content (Table 1). Note that the peak maxima of the reduction peaks, indicating the maximum rate of the reduction process, always appear below 500 °C (Figure 2), suggesting that both Pd and Cu in the catalyst samples are in the reduced state after reduction at 500 °C in H₂ flow for 2 h before the catalytic run.

3.1.3. Characteristics of the Metal Components of the Metal-Promoted MgO-Al₂O₃ Mixed Oxide Catalysts

The amount of chemisorbed CO (μmol CO/g_cat.) determined by the CO pulse chemisorption method (Table 1) reflects the concentration of the surface metal atoms available for adsorption in the catalyst samples. The CO/Pd molar ratio calculated from the amount of chemisorbed CO for the Pd_1.0/MgO-Al₂O₃ sample was 0.16 (Table 1). This value suggests that the ratio of the surface Pd atoms to the total number of Pd atoms (the Pd dispersion) is about 16%, assuming an adsorbed CO to surface Pd atom ratio of 1. Note, however, that due to the abundancy of bridge-bonded CO species on Pd (vide infra) a stoichiometry of 1/2 was suggested for Pd [40,41], which gives a Pd dispersion value of 32%. The average Pd metal particle size can be calculated from the metal dispersion (D_Me) using the following equation from ref. [41]:

$$d=6{\displaystyle \frac{v}{a{D}_{\mathrm{M}\mathrm{e}}}},$$

(4)

where v is the volume occupied by a single metal atom in the bulk of metal (0.0147 nm³) and $a$ is the average surface area occupied by one metal atom (0.0793 nm²). Based on the Pd dispersion value of 32%, the calculated average metal particle size in the Pd_1.0/MgO-Al₂O₃ sample is 3.5 nm, which agrees well with the TEM result (Figure S2).

The amount of chemisorbed CO is an order of magnitude lower for the Cu_1.0/MgO-Al₂O₃ catalyst sample than for the Pd_1.00/MgO-Al₂O₃ sample (Table 1) indicating that the CO chemisorption is nearly negligible on Cu sites relatively to Pd sites in the monometallic samples. Note that the number of surface Pd atoms involved in CO chemisorption is approximately twice as high due to the abundance of bridge-bonded CO (see below). This further confirms that the Pd sites are the primary bonding sites for CO. Assuming that CO chemisorption on Cu metallic sites in the bimetallic samples is also negligible, the CO/Pd molar ratios were calculated from the CO pulse chemisorption experiments (Table 1). Substitution of about 25% of the Pd atoms by Cu significantly increased the CO/Pd ratio (Table 1) due to the dilution of Pd with Cu, which resulted in a higher concentration of surface Pd atoms. At higher Cu contents, when 50 to 75% of the Pd atoms was replaced by Cu, the CO/Pd ratio gradually decreased, possibly due to the copper enrichment on the surface of the bimetallic particles [29,42].

Infrared spectroscopy of adsorbed CO was used to investigate the surface properties of the metallic components of the metal-promoted MgO-Al₂O₃ mixed oxide catalysts. The spectra of adsorbed species formed from CO adsorption on the monometallic and bimetallic catalyst samples are shown in Figure 3.

The carbonyl bands of CO adsorbed on highly dispersed Pd catalysts are attributed to linearly bound (Pd–CO) and nonlinearly bound carbonyl species at two-fold (Pd₂–CO) and three-fold (Pd₃–CO) coordination sites, appearing either above or below 2000 cm⁻¹ depending on the type of coordination [38,39,43,44]. The carbonyl bands due to linearly bound CO appear at 2050 and, as a shoulder, at around 2070 cm⁻¹ on the monometallic Pd_1.0/MgO-Al₂O₃ sample (Figure 3, a). In a recent study by De Castro et al. [39], based on combined infrared spectroscopic and density functional theory investigation, these bands were assigned to linear Pd–CO species formed on steps/edges (2050 cm⁻¹) and terraces (2070 cm⁻¹) of the Pd particles, respectively. Note that at high CO coverage, this latter band appeared with a significantly higher intensity and shifted to higher wavenumbers (2085 cm⁻¹) (Figure S4) in line with former studies [39,44]. The carbonyl bands appearing in the nonlinear region below 2000 cm⁻¹ (Figure 3, a) can be assigned to Pd₂–CO species at two-fold coordination sites on terraces (1950 cm⁻¹) or steps/edges (1900 cm⁻¹), and Pd₃–CO species at three-fold coordination sites on terraces (1845 cm⁻¹) of the Pd particles [38,39].

The characteristic carbonyl bands appear with a significantly lower intensities on the monometallic Cu_1.0/MgO-Al₂O₃ catalyst (Figure 3, e) than on the monometallic Pd catalyst due to the weaker adsorption of CO on the Cu surface than on the Pd surface [29,44], as was also confirmed by the CO pulse chemisorption experiments. The low intensity carbonyl band at 2110 cm⁻¹ and the most intense carbonyl band at 2008 cm⁻¹ (Figure 3, e) can be assigned to linearly bound Cu–CO species formed on the terraces and steps/edges of the Cu particles, respectively [39]. Additionally, the band at about 1980 cm⁻¹ (appearing as a shoulder) and the broad band around 1880 cm⁻¹ can be attributed to bridging Cu₂–CO species formed on two-fold coordination sites on terraces and steps/edges, respectively [39]. Note that at high CO coverage, additional pair of bands can be observed at about 2050/2100 cm⁻¹ due to the formation of gem-dicarbonyl species (Figure S4e) [39].

Substituting 25% of the Pd atoms with Cu atoms in the bimetallic samples reduced the intensity of the carbonyl bands assigned to the Pd₂–CO species (bands at 1950 and 1900 cm⁻¹) and, in particular, the carbonyl band attributed to the Pd₃–CO species (a broad band around 1845 cm⁻¹), while the intensity of the band assigned to linearly adsorbed CO species at 2050 cm⁻¹ increased (see Figure 3, spectra a and b). This results clearly indicate the alloying of the Pd surface with Cu, causing the surface dispersal of Pd ensembles resulting in increasing number of isolated Pd atoms and lowering the ability of the Pd surface to form strong multiple bonds with CO molecules [29,38,44]. It is important to note that new carbonyl bands appearing as a shoulder at 2025 and 1915 cm⁻¹ became visible on the Pd_0.75Cu_0.25/MgO-Al₂O₃ sample (Figure 3, b). The formation of similar new component bands have been substantiated on a Pd-Cu bimetallic catalyst at around 2045, 2020, and 1910 cm⁻¹ and assigned to CO adsorbed on isolated Pd on four types of copper surface facets with an embedded and adatom configuration [39]. Following these assignments, the bands appearing at 2025 cm⁻¹ and 1910 cm⁻¹ can be assigned to CO bound to Pd single atoms embedded in a Cu(111) facet and CO forming a bridge between Pd and Cu on a Pd atom embedded in a Cu(211) terrace site, respectively [39].

Substituting 50 to 75% of the Pd atoms with Cu atoms in the bimetallic samples further reduced the intensity of all the carbonyl bands (Figure 3 and Figure S4). The band intensities appear to be nearly proportional to the Pd content at high CO coverage (see Figure S4), but they significantly decrease at lower CO coverage, which is achieved after evacuation in high vacuum. This is especially true for samples containing highly copper-substituted Pd–Cu bimetallic particles (Figure 3). These results suggest that Pd sites are the primary binding sites for CO and that their ability to strongly bind CO decreases with increasing Cu content in the bimetallic particles. The number of available Pd sites for CO adsorption decreases by the dilution with Cu due to the ensemble effect that reduces the ability of Pd in the bimetallic Pd–Cu particles to form bridge-bonded Pd carbonyls, while the latter effect can be attributed to ligand effect of Cu induced by modification of electronic structure of Pd upon alloying with Cu in the bimetallic particles [44].

3.2. Catalytic Properties

The results of the catalytic experiments for the ethanol coupling reaction over the metal-promoted MgO-Al₂O₃ catalysts are presented in Figure 4. The ethanol conversion (solid lines), the yield of liquid products (dashed lines), and the composition of organic products in the liquid phase, based on Equation (3) (color coded bars), in the function of the reaction temperature are presented. The results of the catalytic measurements were comparable to those reported in the relevant scientific literature (see in Table S1). The carbon balance (moles of carbon in all products divided by moles of carbon in the reacted ethanol) was greater than 95% for each measurement (see in Table S2).

For the non-promoted MgO-Al₂O₃ mixed oxide catalyst, shown for comparison in Figure 4A, ethanol conversion primarily produced liquid products, with only a negligible amount of gaseous products formed (Figure S5A) even at the highest reaction temperature (less than 1% at 350 °C). These results indicate that decarbonylation side reactions leading to CO and hydrocarbon moieties (mainly methane) practically do not take place in the absence of active metal sites [19]. The ethanol conversion is below 5% in the low temperature range below about 250 °C, and then continuously increases up to about 35% at higher reaction temperatures (Figure 4A); however, the alcohol selectivity (including butanol and hexanols) drops significantly at elevated temperatures due to the formation of other oxygenates, mostly diethyl ether formed in undesired dehydration reactions. Note that the MgO-Al₂O₃ catalyst showed stable performance in the function of time-on-stream up to about 30 h after reaching steady state (Figure S6).

Introduction of Pd active metal significantly improved the activity of the MgO-Al₂O₃ mixed oxide catalyst in the whole temperature range (Figure 4B). The effect of the active metal on ethanol conversion and alcohol selectivity (mainly butanol and some hexanols) was most beneficial between 200 and 250 °C. In line with our former observations [19], the presence of Pd and H₂ could significantly improve the catalytic activity and alcohol selectivity in this temperature range without adversely affecting the yield of organic liquid products (Figure 4B). Ethanol conversion was approximately an order of magnitude higher with the Pd_1.0/MgO–Al₂O₃ catalyst than with the non-promoted MgO–Al₂O₃ catalyst, and alcohol selectivity increased substantially (see Figure 4A,B). For instance, at 250 °C, ethanol conversion increased from 1.6% to 14.4%, and alcohol selectivity increased from 82.5 wt.% to 95.6 wt.%, resulting in butanol yields of 1.3 wt.% and 13.8 wt.%, respectively. The beneficial effect of the Pd promoter can be explained by the promoting effect of the metal in the initial ethanol dehydrogenation step and the hydrogenation of the unsaturated reaction intermediates in the closing reaction step [4,5,6,11]. Note that using H₂ in the reaction system has the advantage that, when activated on the metal palladium, H₂ can participate in closing H-transfer reactions. In an inert atmosphere, these H-transfer reactions can only proceed with ethanol via the Meerwein-Ponndorf-Verley (MPV) mechanism, and promoting effect of Pd was found to be less effective [19]. As shown in Figure 4B, alcohol selectivity and liquid yield decreased considerably in the high temperature range above 250 °C, while gas phase product yield increased proportionally (Figure S5B). This was mainly due to the acceleration of undesirable decarbonylation side reactions, which led to the formation of ≥C4 alkanes in the liquid products and C1–C3 alkanes in the gas phase product mixture. The dominance of CH₄ and CO in the gas phase (Figure S5B) suggests that the accelerated decarbonylation of the acetaldehyde intermediate over metallic sites at high temperatures is primarily responsible for the decrease in liquid yields.

Dilution of Pd with increasing amount of Cu in the bimetallic particles resulted in a gradually decreasing ethanol conversion (see Figure 4B–E); however, the ratio of the liquid- and gas-phase products considerably increased (Table S3). It follows that the decarbonylation activity of the catalyst, responsible for the formation of gas phase products (mainly CO and CH₄), was suppressed in a larger extent than the dehydrogenation–hydrogenation activity of the catalyst. These results are in line with former observation for a similar supported bimetallic Pd-Cu catalysts [25,26,45]. It is important to note that at low Cu content the desired main product 1-butanol is formed with a high selectivity (>85%) below about 300 °C (Figure 4C), while extensive substitution of Pd with Cu (>25%) resulted in a radical change in the organic liquid composition (Figure 4D–E). The butanol selectivity significantly decreased mainly due to the formation ethyl acetate, ketones (mainly 2-butanone below 300 °C and increasing amounts of C3-C7 ketones over 300 °C), and aldehydes (mainly acetaldehyde and some butanal). These byproducts form due to the enrichment of the bimetallic particles’ surface with Cu, which catalyzes the aforementioned reaction routes. Similar conversion curves were obtained for the MgO-Al₂O₃ mixed oxide catalyst and the Cu_1.0/MgO-Al₂O₃ catalyst (see Figure 4A,F), which suggests that the presence of Cu alone does not have a significant promoting effect. However, a similar product distribution can be observed on the monometallic Cu_1.0/MgO·Al₂O₃ catalyst (Figure 4F) as on the bimetallic catalysts with high Cu substitution (Figure 4D–E), indicating that the Cu⁰ active sites are primarily responsible for the ester-forming reaction route, as shown in previous studies [21,46,47,48]. The formation of esters has been proposed to proceed via the reaction of alkoxide species with an aldehyde (Tishchenko-type reaction) or via a hemiacetal intermediate formed when alcohols react with an aldehyde intermediate [4,21,46,48]. The formation of ketones also became significant in the high temperature range over about 300 °C (Figure 4C–F). Their formation was found to predominantly occur at high conversions at the expense of alcohols via reaction pathways possibly involving intramolecular H-transfer in the 3-hydroxybutanal (enolate) intermediate formed in the aldol condensation step of the coupling reaction [4,48,49,50]. The dehydration-hydrogenation reactions of the resulting keto form (4-hydroxybutan-2-al) can lead to the formation of 2-butanone (methyl ethyl ketone) [4,49,50]. The appearance of 2-butanone as the main ketone product on the copper containing catalysts clearly supports this latter reaction pathway.

3.3. Formation and Transformation of Surface Species

The quasi-operando DRIFT spectroscopic investigations provide information about the formation of surface species from ethanol and their transformation on the catalyst surface under catalytic conditions. The results obtained in transient mode in a continuously flowing ethanol/hydrogen gas mixture at an elevated reaction temperature for each monometallic and bimetallic catalyst sample are presented in detail in Figures S7–S12. The results obtained on metal-free MgO-Al₂O₃ catalyst are shown in Figure S7 for comparison. The spectra measured at selected reaction temperatures between 100 and 350 °C on the monometallic and bimetallic MgO-Al₂O₃ catalysts are compared in Figure 5.

The assignments of the characteristic bands observed on the MgO-Al₂O₃ catalyst to surface species formed from the adsorption of ethanol were given in a previous study [19]. Similar bands developed on the metal promoted MgO-Al₂O₃ catalysts (Figures S7–S12), so these assignments can also be applied to these catalyst samples. The negative band at 3730 cm⁻¹ in the region of the ν_OH stretching vibrations indicate the involvement of the surface OH groups in the adsorption interaction. The appearance of a broad absorption band around 3200 cm⁻¹ (bottom spectra in Figures S7A–S12A) can be attributed to the characteristic frequency of surface OH groups being red-shifted from its original position (3730 cm⁻¹) due to H-bonding interaction of molecularly adsorbed EtOH with the surface OH groups [51,52,53]. This broad band disappears above about 150 °C as expected due to the decreasing coverage of the catalyst surface with molecularly adsorbed EtOH (Figures S7A–S12A). The absorption bands related to molecularly adsorbed EtOH in the range of C-H deformation vibrations can be identified at 1483 (β_s[CH₂]), 1460 (δ_as[CH₃]), 1400 (δ_s[CH₃]), 1280 (δ_OH), and 1105 cm⁻¹ (ν_C-O) (Figures S7B–S12B) [54,55,56]. These bands also disappear around 150 °C due the shift in the adsorption–desorption equilibrium. However, the negative ν_OH band at 3730 cm⁻¹ persists at higher temperatures (>150 °C), indicating a surface reaction consuming surface OH groups. The product of the surface reaction can be identified as monodentate ethoxy species giving characteristic bands at 1165, 1130, and 1070 cm⁻¹ (Figure 5) assigned to the ν_C-O/ν_C-C vibrations, whereas the corresponding C-H deformation vibrations are discernible at 1485 (β_s[CH₂]), 1450 (δ_as[CH₃]) and 1385 cm⁻¹ (δ_s[CH₃]) [51,56,57,58,59]. The formation of ethoxy species has been shown to be accompanied by the formation of water [51,52,56,60], which explains the consumption of the surface OH groups during the surface reaction.

The ethoxy species also exhibit characteristic bands in the region of the ν_CH stretching vibrations at 2965, 2930, and 2870 cm⁻¹ (Figures S7A–S12A), which are attributed to ν_as[CH₃], ν_as[CH₂], and overlapping bands of ν_s[CH₃]/ν_s[CH₂] vibrations, respectively [51,56,61]. As the temperature increases, the relative intensity of the ν_as[CH₂] band increases in comparison to the ν_as[CH₃] band. This suggests that the transformation of the ethoxy species resulted in surface species containing more -CH₂- groups, consistent with the formation of coupling products. The additional bands appearing at approximately 2820 cm⁻¹ (a shoulder) and 2725 cm⁻¹ (Figures S7A–S12A) are likely due to the ν[CH] stretching vibrations of aldehyde-like surface species formed in the dehydrogenation reaction, whereas the corresponding C-H deformation band appears at 1360 cm⁻¹ [17,61,62,63].

The absorption bands attributed to the monodentate ethoxy species (1165, 1130, and 1070 cm⁻¹) decrease simultaneously with the formation of coupling products on each catalyst sample at elevated temperatures, indicating that ethoxy species are the key surface intermediates for the coupling reaction. Consumption of the ethoxy species is accompanied by the appearance of a new pair of bands at 1585 and 1442 cm⁻¹ (Figure 5A–F). These bands can be attributed to the ν_as[O-C-O] and ν_s[O-C-O] vibrations of carboxyl species [59,61,64,65,66]. The separation of the asymmetric and symmetric components (Δν = 143 cm⁻¹) suggests that chelating bidentate carboxyl species are formed [64,67]. Their formation has been shown to proceed via an aldehyde intermediate from the surface ethoxy species [64,65,66]. As shown in Figures S7B–S12B, the concentration of these carboxyl species continuously increases on the catalyst surface at elevated temperatures, which is typical of thermally stable, low-reactivity species, such as spectator species [61,65].

The characteristic absorption bands and the formation of the corresponding surface species are similar for both the metal-free and the metal-promoted MgO-Al₂O₃ catalysts (see Figures S7–S12). However, new surface species with a characteristic absorption band at 1345 cm⁻¹ form only for the latter catalysts at temperatures above 200 °C (Figure 5A–F). Furthermore, these surface species form much more prominently on the monometallic Cu_1.0/MgO-Al₂O₃ and Cu-containing bimetallic catalysts than on the monometallic Pd_1.0/MgO-Al₂O₃ catalyst (Figure 5D–F). The intensity of the band at 1345 cm⁻¹ gradually increases up to the highest reaction temperature, which suggests the formation of thermally stable, low-reactivity species. A similar thermally stable absorption band is typically observed for the dissociative adsorption of acetic acid. This band is assigned to the δ_s[CH₃] vibration of the acetate group [68,69]. The formation of surface acetate is likely indirectly related to a reaction product formed in the surface reactions mainly catalyzed by metallic Cu sites, such as ethyl acetate. In a secondary reaction, this ester can hydrolyze to give acetic acid, which can form strongly adsorbed acetate species on the catalyst surface and/or can condensate to ketons with the liberation of CO₂ [4,24]. The appearance of some CO₂ among the gas phase products (Figure S5) suggests that this minor reaction route probably also contributed to the formation of ketons at higher reaction temperatures.

The catalytic results indicated that the formation of methane and CO becomes significant in the high-temperature range (>250 °C) on the Pd-containing catalysts due to accelerated decarbonylation of the acetaldehyde intermediate, which leads to decreased liquid yields and increased gas yields (see Figure 4 and Figure S5). The DRIFT spectroscopic results support this notion. Both methane (see Figures S8A–S11A) and CO (Figure 6) in the gas phase can be detected at temperatures above 250 °C in the spectra obtained on metal-promoted catalysts within the C-H and C-O stretching vibration ranges, respectively. It is important to note that the carbonyl bands assigned to nonlinearly bound CO at two-fold (Pd₂–CO) and three-fold (Pd₃–CO) coordination sites (see Figure 3) appear at significantly lower temperatures than gas-phase CO. Meanwhile, the carbonyl band attributed to linearly bound CO (Pd–CO) appears at the same time as gas-phase CO (see Figure 6A–D). These observations suggest that the decarbonylation reaction of the acetaldehyde intermediate is facile at relatively low temperatures. However, the reaction product, CO, strongly binds to the Pd sites at temperatures below about 250 °C, which is consistent with previous studies [25,45]. Therefore, the strongly bound CO can hinder the decarbonylation reaction effectively. These observations also substantiate that the decarbonylation reaction primarily occurs on multiple Pd sites (represented by two- and three-fold coordination sites for CO adsorption). The decarbonylation reaction speeds up at around 250 °C. At this temperature, the adsorption–desorption equilibrium allows CO to be released from the active Pd sites. CO chemisorption experiments proved that diluting Pd with Cu in the bimetallic particle decreased the concentration of multiple Pd sites (see Figure 3). Consistent with the catalytic results, this leads to a lower decarbonylation rate, as evidenced by the lower gas-phase CO concentration observed at the same reaction temperatures (see Figure 6A,D). The decarbonylation reaction, and thus the CO formation, barely occurs on the monometallic Cu_1.0/MgO-Al₂O₃ catalyst (Figure 6E) due to the significantly lower decarbonylation activity of Cu compared to Pd.

4. Conclusions

The self-coupling reaction of ethanol for producing a high-performance biofuel component 1-butanol was studied over Pd/MgO-Al₂O₃ and bimetallic Pd-Cu/MgO-Al₂O₃ mixed oxide-based catalysts. The promoting effect of the active metal(s) on the hydrogen transfer reaction steps in the overall reaction was investigated. The catalytic properties were investigated by means of a flow-through microreactor system in the temperature range between 200 and 350 °C at 21 bar total pressure, a WHSV of 1 ${\mathrm{g}}_{\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}}\cdot {\mathrm{g}}_{\mathrm{c}\mathrm{a}\mathrm{t}}^{-1}{\cdot \mathrm{h}}^{-1}$, and at a H₂/ethanol molar ratio of 5. The formation and transformation of surface species under catalytic conditions were studied using quasi-operando DRIFT spectroscopy to establish structure-activity relationships.

Results show that both metals are in the reduced state and the metal particles are highly dispersed (average particle size well below 5 nm) on the MgO-Al₂O₃ support both in the monometallic and bimetallic catalyst samples. Substitution of about 25% of Pd atoms with Cu significantly increased the concentration of surface Pd atoms in the bimetallic particles (detected by CO chemisorption), while higher Cu contents (50 to 75% substitution) gradually decreased their concentration, possibly due to surface enrichment of the particles with copper. The FTIR spectroscopic investigation of adsorbed CO revealed that low substitution (25%) of the Pd with Cu caused the surface dispersal of Pd ensembles and resulted in increasing number of isolated Pd atoms and decreasing number of Pd ensembles that able to form strong multiple bonds with CO molecules. Beside this ensemble effect, further dilution of Pd with Cu decreases the ability of Pd in the bimetallic particles to strongly bind CO due to ligand effects.

The catalytic results show that the presence of Pd promoter significantly improved the activity and butanol selectivity across the entire temperature range compared to the metal-free MgO-Al₂O₃ catalyst. However, the yield of liquid products decreased significantly at temperatures higher than 250 °C mainly due to the acceleration of the undesired decarbonylation side reaction of the acetaldehyde intermediate.

The results of the quasi-operando DRIFT spectroscopic investigations revealed that the facile decarbonylation side reaction of the acetaldehyde intermediate primarily occurs on multiple Pd sites (represented by two- and three-fold coordination sites for CO adsorption). However the reaction product, CO, can strongly bind to these Pd sites and cannot desorb at temperatures below about 250 °C. Thus, the promoting effect of Pd was beneficial only in the lower temperature range up to about 250 °C, where the decarbonylation reaction was inhibited due to the reversible poisoning effect of CO.

Diluting the Pd phase with an increasing amount of Cu resulted in higher liquid yields due to the gradually decreasing decarbonylation activity. However, a molar fraction of Cu, greater than 0.25, significantly weakens the promoting effect of Pd on the dehydrogenation–hydrogenation steps of the coupling reaction. Additionally, the distribution of products changed dramatically at the expense of 1-butanol selectivity because metallic Cu can catalyze side reactions that lead to C₄ oxygenates, such as ethyl acetate, as well as ketone products.