**1. Introduction**

Bioethanol is a renewable, biodegradable, non-toxic raw material nowadays produced in an ever-larger amount as the base for a possible green transition to energy production. Its global production is expected to grow up to 41.4 billion liters by 2025, with a progressive reduction in price [1]. It is well known that bioethanol can selectively be dehydrogenated to ethyl acetate under mild conditions, using copper-based catalysts [2–11].

$$\text{\textbullet 2C}\_2\text{H}\_5\text{OH} \rightleftharpoons \text{CH}\_3\text{COOC}\_2\text{H}\_5 + 2\text{H}\_2\tag{1}$$

Pure hydrogen exempt from CO and directly usable in fuel cells can be obtained from ethanol through this route. Moreover, the reaction is reversible and ethyl acetate can then be hydrogenated to ethanol closing the chemical cycle. On this basis, in the present work we suggest the possibility to use ethanol or derivatives as hydrogen vectors (LOHC, Liquid Organic Hydrogen Carrier) operating in a new original way. As known, hydrogen is a powerful vector of energy, but it is difficult to transport [12]. In fact, its volumetric energy density is very low due to its very scarce density under standard temperature and pressure conditions (0.0824 kg/m3 under ideal gas conditions). This low density causes a very low

**Citation:** Santacesaria, E.; Tesser, R.; Fulignati, S.; Raspolli Galletti, A.M. The Perspective of Using the System Ethanol-Ethyl Acetate in a Liquid Organic Hydrogen Carrier (LOHC) Cycle. *Processes* **2023**, *11*, 785. https://doi.org/10.3390/pr11030785

Academic Editor: Davide Dionisi

Received: 1 February 2023 Revised: 1 March 2023 Accepted: 3 March 2023 Published: 7 March 2023

**Copyright:** © 2023 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 (https:// creativecommons.org/licenses/by/ 4.0/).

volumetric energy content of hydrogen, i.e., 0.01 MJ/L H2 for the gas at ambient conditions and 8.5 MJ/L H2 for liquefied H2. One possible solution is to store hydrogen inside a liquid molecule that can easily be transported and that can release it when needed. In this regard, many different LOHCs have been proposed in the literature such as methylcyclohexane, decaline, dodecahydro-N-ethyl carbazole, and other molecules containing aromatic rings. A more detailed list of possible LOHC molecules is reported in some recently published reviews [13–15]. All these LOHCs have the advantage of a high hydrogen density but also many disadvantages such as: (*i*) aromatic molecules are harmful, not renewable, and not biodegradable; (*ii*) their dehydrogenation occurs at relatively high temperatures with high consumption of heat; (*iii*) catalysts promoting the reaction are normally based on precious noble metals (Pt, Pd) susceptible to deactivation or poisoning. Another proposed approach is the use of methanol, ethanol, and hydrocarbons to produce hydrogen by steam reforming but in this case, CO and CO2 are obtained as by-products. CO2 must be separated, and CO contaminates the produced hydrogen. Moreover, the process is irreversible and consumes the reactants causing CO2 emissions [16]. On the other hand, several researchers have described the ethanol dehydrogenation to ethyl acetate by using different catalysts [2–11] and some patents have also been published on this reaction describing the employment of different catalysts and related operative conditions [17–19]. Moreover, an industrial process has been developed and patented by Davy Process Technology, a Johnson and Matthey company [20–22]. More recently, Eurochem Engineering developed and patented in 2011 an industrial process for ethanol dehydrogenation to ethyl acetate carried out in a packed bed tubular reactor containing a more efficient copper-copper chromite commercial catalyst [23]. Under the best conditions, an ethanol conversion of 50–60 mol% and selectivity to ethyl acetate higher than 97% were obtained. The reaction was conducted in a temperature range of 473–493 K and 20–25 atm of pressure with an ethanol contact time of about 90–100 g·h/mol. In that process, 1 mole of pure hydrogen was obtained as a co-product for each mole of converted ethanol. The obtained hydrogen can easily be separated by condensing ethanol and ethyl acetate. All the aspects of this process have been studied in a detailed way and the results have been already reported in some different publications [24–27]. The kinetics of the reaction has been accurately studied by performing many experimental runs in different operative conditions, in the perspective of a suitable scale-up towards the industrial plant and a dedicated paper has been published [28]. The dehydrogenation reaction is reversible, and this is the reason for the limited ethanol conversion obtainable in a single step. The reverse reaction is therefore feasible and different works have studied the hydrogenation of ethyl acetate to ethanol by using different copper-based catalysts [29–33]. All these works, respectively devoted to ethanol dehydrogenation and ethyl acetate hydrogenation, demonstrate the feasibility of the use of ethanol as a LOHC because it is possible to close the chemical cycle of dehydrogenation of ethanol and hydrogenation of ethyl acetate. However, until now no-one has suggested the use of ethanol as a LOHC except for the proposal reported in a work recently published by Tran et al. [34] and in the more recent one by Mevawala et al. [35]. Tran et al. proposed the use of ethanol as a LOHC considering the cycle dehydrogenation-hydrogenation promoted by a ruthenium complex in homogeneous phase. Although the homogeneous catalytic approach is not convenient with respect to the heterogeneous one, the work is important because it is a further demonstration of the feasibility of the idea to use ethanol as a hydrogen carrier. On the other hand, the paper of Mevawala et al. [35] studied the thermodynamic and environmental aspects of the ethanol–ethyl acetate cycle for hydrogen storage applications, without an in-depth analysis of the catalytic systems. Their modeling of the overall cycle determined an energy efficiency up to 88%, a significantly high value with respect to those of other LOHCs, mainly addressed to the very low endothermicity of the reaction. Moreover, a preliminary eco-balance evidenced that the use of ethanol was more sustainable from a carbon emission perspective when compared with fossil LOHCs. Therefore, starting from the existing literature and considering, as a reference, the performance of the already wellknown copper-copper chromite catalyst in promoting ethanol dehydrogenation we would

like to define the best operative conditions for both ethanol dehydrogenation and ethyl acetate hydrogenation, also with the aim to find alternative catalysts exempt of chromium with similar or superior performances. In the perspective of the scale-up, the kinetics of both the dehydrogenation and hydrogenation reactions in conventional reactors will be defined for a preliminary determination of the techno-economic aspects of a structured process with the aim to optimize the characteristics of three different application areas: energy-storage, energy-transport, and mobility application.

#### **2. Methodology of Ethanol Dehydrogenation**

#### *2.1. Thermodynamics of the Occurring Reactions*

The conversion of ethanol to ethyl acetate occurs according to the following stoichiometry:

$$2\text{C}\_2\text{H}\_5\text{OH} \rightleftharpoons \text{CH}\_3\text{COOC}\_2\text{H}\_5 + 2\text{H}\_2\tag{2}$$

ΔH◦ 500 K = 33.64 kJ/mol; ΔG◦ 500 K = −6.44 kJ/mol

From experimental observation, this reaction is the addition of two consecutive reactions that are:

$$(1)\ \text{C}\_2\text{H}\_5\text{OH} \rightleftharpoons \text{CH}\_3\text{CHO} + \text{H}\_2\tag{3}$$

ΔH◦ 500 K = 71.30 kJ/mol; ΔG◦ 500 K = 10.59 kJ/mol

$$(2)\ \text{CH}\_3\text{CHO} + \text{CH}\_3\text{CH}\_2\text{OH} \rightleftharpoons \text{CH}\_3\text{COOCH}\_2\text{CH}\_3 + \text{H}\_2\tag{4}$$

ΔH◦ 500 K = −37.66 kJ/mol; ΔG◦ 500 K = −17.03 kJ/mol

It is opportune to observe that the overall reaction (2) is moderately endothermic, while reaction (3) is endothermic, and the successive reaction (4) is on the contrary exothermic. The equilibrium constants of the three reactions, at 500 K, are respectively:

$$\text{K}\_{\text{p2-500 K}} = 4.69; \text{ K}\_{\text{p3-500 K}} = 0.080; \text{ K}\_{\text{p4-500 K}} = 58.94$$

The equilibrium of reaction (3) is shifted to the left but as acetaldehyde is consumed by reaction (4) it proceeds to the right and the equilibrium of the overall reaction is moderately shifted to the right at 500 K. The overall reaction (2) is more favored at 600 K with the equilibrium constant being about 17. Thermodynamic calculations have been made on data reported by Stull et al. [36]. The choice of 500 K, as reference temperature, is a compromise between the thermodynamic and kinetic properties of the involved reactions, because higher temperatures favor the occurrence of undesired side reactions.

## *2.2. Catalysts Normally Employed for Promoting the Ethanol Dehydrogenation*

Many different catalysts have been used for promoting ethanol dehydrogenation [2–11,17–28]. It is possible to distinguish two different classes of copper catalysts respectively containing copper/copper chromite, and copper metal supported and/or promoted by different oxides such as Al2O3, Cr2O3, ZnO, ZrO2, and SiO2. The presence of oxide compounds has the scope to improve the dispersion of Cu thus, as consequence, slowing down the catalyst deactivation due to the metal sintering. Some other catalysts containing Ni or Pd have also been tested with lower performances. Interesting is the use of ruthenium complexes acting in a homogeneous phase, at very low temperatures, and promoting both ethanol dehydrogenation and ethyl acetate hydrogenation. Many papers and reviews have been published on the subject and nowadays some conclusions can be drawn [37–39]. According to Finger et al. [37], the dehydrogenation of ethanol to acetaldehyde occurs on the Cu surface, while the coupling of acetaldehyde with ethanol occurs mainly at the interface Cu-metal oxide (ZnO, ZrO2) where ethanol is adsorbed as alkoxide. According to Pang et al. [38], acetaldehyde is generated preferably on an unreduced Cu<sup>+</sup> site while H2 is liberated on Cu0. The ratio Cu+/Cu0 would be therefore important for the selectivity, and a high presence of Cu+ favors the formation of acetaldehyde, while the contrary favors the formation of ethyl acetate. As previously reported, the

Cu dispersion is of paramount importance, the activity being normally proportional to the specific surface area of Cu. In this regard, the support has a great influence on catalyst activity, selectivity and stability. For example, a high acidity of the support is detrimental, favoring the dehydration path instead of the dehydrogenation one; thus, the oxides, being weak acids or basic, favor dehydrogenation. Oxides trapping the Cu particles favor stability, hindering the deactivation due to sintering. It seems that stability can be induced also by alloying Cu with a moderate quantity of Ni. The operative conditions are also very important, considering that the formation of acetaldehyde is favored by low pressure (<1 atm) and high temperature (>573 K), whilst the selectivity to ethyl acetate increases at a relatively high pressure (10–30 atm) and low temperature (<523 K). The optimal operative conditions will be chosen through a compromise that allows the achievement of the maximum ethyl acetate yield. Phung [12] recently published a review on copper-based catalysts for ethanol dehydrogenation examining, on the basis of the previous literature, the respective roles of: Cu loading, Cu dispersion, the particle size, the Cu support and related acidity or basicity and contact time.

#### *2.3. A Reference Catalyst, Reaction Mechanism and Related Kinetic Model*

In this work, the behavior of a copper-copper chromite catalyst could be used as a reference starting point for comparing kinetic data obtained by employing other catalysts because the kinetics of that catalyst are well known from different sources [28,40]. Different kinetic models have been proposed in the literature. For example, Tu et al. [40] considered the main overall reaction to be of the pseudo-first order; thus, the model suggested by these authors was oversimplified. Moreover, the tested copper chromium catalysts were deactivated with time and a deactivation kinetic law was proposed by the same authors. More recently, a kinetic model based on reliable reaction mechanisms has been published by Carotenuto et al. [28]. The model proposed by these authors considered the following reaction scheme:

$$\text{CH}\_3\text{CH}\_2\text{OH} \rightleftharpoons \text{CH}\_3\text{CHO} + \text{H}\_2\tag{5}$$

$$\text{CH}\_3\text{CHO} + \text{CH}\_3\text{CH}\_2\text{OH} \rightleftharpoons \text{CH}\_3\text{COOCH}\_2\text{CH}\_3 + \text{H}\_2\tag{6}$$

$$\text{2CH}\_3\text{CHO} \to \text{Other products} \tag{7}$$

Probably occurring according to the following reaction mechanisms:

C2H5OH + σ<sup>0</sup> σ*EtOH* (8)

$$
\sigma\_{\rm EOH} + \sigma\_0 \to \sigma\_{\rm Ach} + \sigma\_{\rm H2} \text{ rate determining step} \tag{9}
$$

$$
\sigma\_{AcH} \rightleftharpoons \sigma\_0 + \text{CH}\_3\text{CHO} \tag{10}
$$

$$
\sigma\_{\rm H2} \to \sigma\_0 + \rm H\_2 \tag{11}
$$

σ<sup>0</sup> is the fraction of free active sites on the catalytic surface, while σ<sup>i</sup> is the fraction of active sites occupied by adsorbed "i" molecules. The "rate determining step" should be the surface reaction between chemisorbed ethanol and a catalyst void site to form adsorbed acetaldehyde. For the second reaction the following mechanism was suggested:

C2H5OH + σ<sup>0</sup> σ*EtOH* (12)

CH3CHO + σ<sup>0</sup> σ*AcH* (13)

$$
\sigma\_{EOH} + \sigma\_{AcH} \rightleftharpoons \sigma\_{EA} + \sigma\_{\text{H2}}\text{ rate determining step}\tag{14}
$$

$$
\sigma\_{EA} \rightleftharpoons \sigma\_0 + \text{CH}\_3\text{COOCH}\_2\text{CH}\_3 \tag{15}
$$

$$
\mathfrak{o}\_{\text{H2}} \rightleftharpoons \mathfrak{o}\_{\text{0}} + \text{H}\_{\text{2}} \tag{16}
$$

In this case, the rate determining step would be the reaction between adsorbed ethanol and adsorbed acetaldehyde. Based on the postulated reaction mechanisms, it is possible to write the following kinetic law expressions:

R

$$r\_1 = \frac{k\_1 P\_{EtOH} \left(1 - \frac{1}{Kc\_1} \frac{P\_{AcH} P\_{H\_2}}{P\_{EtOH}}\right)}{\left(1 + b\_{EOH} P\_{EtOH} + b\_{AcH} P\_{AcH} + b\_H P\_H + b\_{EA} P\_{EA}\right)^2} \tag{17}$$

R

R

R

R

$$r\_2 = \frac{k\_2 P\_{EtOH} P\_{AcH} \left(1 - \frac{1}{\mathcal{K}c\_2} \frac{P\_{EA} P\_{H\_2}}{P\_{EOH} P\_{AcH}}\right)}{\left(1 + b\_{EtOH} P\_{EtOH} + b\_{AcH} P\_{AcH} + b\_H P\_H + b\_{EA} P\_{EA}\right)^2} \tag{18}$$

$$r\_3 = k\_3 P^2\_{AcH} \tag{19}$$

The reactions of acetaldehyde to other undesired products are normally reactions of acetaldehyde condensation as deeply investigated by Inui et al. [7] and Colley et al. [8]; many different undesired products could be obtained in the worst operative conditions as in the following scheme (Scheme 1):

**Scheme 1.** Pathways of by-products formation from acetaldehyde condensation.

To avoid the formation of acetaldehyde condensation by-products it is clearly imperative to keep the acetaldehyde concentration in the system low and to achieve this result the catalyst selectivity becomes of paramount importance. In contrasty, no-one observed an acetaldehyde decomposition reaction to CO by using copper-copper chromite catalysts.

In the kinetic approach followed by Carotenuto et al. [28], the reaction of acetaldehyde to other undesired side products was simplified to a pseudo-second-order reaction because it was characterized by a very low conversion. The best fitting parameters obtained by interpreting all the experimental runs performed are summarized in Table 1. The kinetic parameters reported in Table 1 seem reliable because the activation energies and the adsorption constants are reasonable values compatible with the postulated reaction mechanism. The very low value of k3 and the negligible value of the corresponding activation energy is justified by: (*i*) the approximation introduced by arbitrarily assuming a pseudo-second order rate law; (*ii*) the fact that reaction (7) is not a single reaction but is an assembly of different reactions exclusively consuming acetaldehyde; (*iii*) the very low amount of by-products that determines high analytical errors.


**Table 1.** Optimized kinetic parameters obtained for a commercial copper-copper chromite catalyst, from [28], reported with the permission of Elsevier.

This reaction system is singular because, looking at the reaction scheme, we can observe that ethanol transformation to ethyl acetate passes through the formation of acetaldehyde, an endothermic reaction (ΔH ≈ 71 kJ/mol), while the successive reaction is moderately exothermic (ΔH ≈ −40 kJ/mol). This means that using a tubular reactor it is opportune to separate it in different stages (at least 2) differently heated if the system should be maintained approximately isotherm. In the first part it is necessary to furnish heat, whilst in the second one the cooling of the reactor is required. On the other hand, if a unique adiabatic reactor is adopted and a hot reagents stream (e.g., 508 K) is fed, it is possible to observe initially a decrease in the temperature followed by a progressive moderate increase. For an isothermal tubular reactor, the following system of differential equations must be simultaneously solved:

$$\frac{dF\_{\text{EtOH}}}{dZ} = -\mathcal{W}\_{\text{cat}}(r\_1 + r\_2) \tag{20}$$

$$\frac{dF\_{\rm{AcH}}}{dZ} = W\_{\rm{cat}}(r\_1 - r\_2 - 2r\_3) \tag{21}$$

$$\frac{dF\_{AcOEt}}{dZ} = \mathcal{W}\_{cat}r\_2\tag{22}$$

$$\frac{dF\_{H2}}{dZ} = W\_{cat}(r\_1 + r\_2) \tag{23}$$

where *Fi* is the molar flow rate of component i, *Z* = L/Lbed is the dimensionless reactor length, *Wcat* is the loaded catalyst weight and *rj* are the rates of the j-reactions referred to the unit of catalyst weight. Figure 1 shows the conversion of ethanol and selectivities to respectively ethyl acetate and acetaldehyde as a function of the contact time W/F (g·h/mol), at 493 K and 20 atm, calculated with the described kinetic model and related parameters.

A similar kinetic approach has been published by Men'shchikov et al. [41], who interpreted runs performed on a catalyst composed of copper-zinc-chromium supported on alumina (CuO-ZnO-Cr2O3-Al2O3) in the temperature range 503–563 K and pressure of 10–20 atm. The runs were carried out in a continuous fixed-bed reactor and interpreted with a Langmuir–Hinshelwood kinetic model. The conversion to ethyl acetate changed with the temperature from 10 to 47%, acetaldehyde was produced in a small amount from 0.7 to 3%, while increasing the temperature greatly increased the presence of other by-products from 0.3% to 7%. Optimal kinetic parameters of the model were determined and interpreted for all the kinetic runs showing an error of ±20%. Starting from similar results, it would be possible to do a scale-up and formulate the best arrangement of the plant equipped with a conventional reactor for producing hydrogen and ethyl acetate.

**Figure 1.** Kinetic behavior of copper-copper-chromite catalyst. Ethanol conversion and selectivities to respectively ethyl acetate and acetaldehyde, for different contact times at 493 K and 20 atm. Curves have been obtained by solving the system of differential Equations (10)–(13) with the parameters of Table 1.
