Recent Advances in Homogeneous/Heterogeneous Catalytic Hydrogenation and Dehydrogenation for Potential Liquid Organic Hydrogen Carrier (LOHC) Systems

Abstract

Here, we review liquid organic hydrogen carriers (LOHCs) as a potential solution to the global warming problem due to the increased use of fossil fuels. Recently, hydrogen molecules have attracted attention as a sustainable energy carrier from renewable energy-rich regions to energy-deficient regions. The LOHC system is one a particularly promising hydrogen storage system in the “hydrogen economy”, and efficient hydrogen mass production that generates only benign byproducts can be applied in the industry. Therefore, this article presents hydrogenation and dehydrogenation, using homogeneous or heterogeneous catalysts, for several types of LOHCs, including formic acid/formaldehyde/ammonia, homocyclic compounds, nitrogen- and oxygen-containing compounds. In addition, it introduces LOHC system reactor types.

1. Introduction

Since the rise in CO₂ concentrations caused by the increased use of fossil fuels following the Industrial Revolution, global warming has been a critical environmental issue. For decades, environmentally friendly renewable energy sources such as solar, wind, geothermal, tide, and biomass have piqued the interest of those seeking to replace fossil fuels. The most difficult issue with the use of renewable energy is the unstable supply characteristics caused by fluctuating weather conditions. Efficient energy storage and transportation systems are required to redistribute renewable energy from abundant areas to energy-scarce areas. In this regard, hydrogen with a high gravimetric storage density (33.0 kWh/kg) can be a clean and sustainable energy carrier because it can be produced from the water via electrolysis using renewable energy in sunny or windy regions and then transported to energy-deficient regions for electrical energy generation. [1,2,3]. However, gaseous hydrogen molecules are not efficient for the storage and transportation of sufficient energy due to a low-volume energy density (0.003 kWh/L) [4]. To overcome this problem, compressed gaseous hydrogen (CGH2, 1.3 kWh/L at 700 bar) or liquefied cryogenic hydrogen (LH2, 2.4 kWh/L) can be used, but these still have a lower volumetric energy storage density than fossil fuels such as diesel (9.8 kWh/L), heavy oil (10.7 kWh/L) and natural gas (2.4 kWh/L at 200 bar) [4]. Furthermore, expensive high-pressure tanks (350–700 bar) or cryogenic systems are necessary to store gaseous or liquefied hydrogen for transportation, and it is also problematic regarding safety issues because of its explosive nature [5,6,7,8,9]. Thus, various hydrogen storage methods (liquid hydrogen, metal hydride, inorganic chemical hydrides, and metal–organic frameworks) have been investigated for the “hydrogen economy” in terms of efficient hydrogen storage and easy transport from energy-rich locations to high-value hydrogen applications.

The liquid organic hydrogen carrier (LOHC) is a promising candidate for the absorbance of hydrogen where renewable energy is abundant and release of hydrogen to generate electric energy where energy is scarce due to its high safety levels and low cost. To develop these potential LOHC systems, several criteria should be considered:

(1)

Physical characteristic: Various physical properties, as given below, should be investigated to discover promising LOHC systems.

• Melting point: Liquids and low-melting solid organic compounds are considered LOHCs, but hydrogen-rich LOHCs, hydrogen-lean LOHCs, or a mixture of hydrogen-rich/hydrogen-lean LOHCs should exist in a liquid state for easy storage and transportation in an existing gasoline infrastructure at ambient temperature.
• Boiling point: LOHCs with a high boiling point are more effective in easily separating pure hydrogen gas from hydrogen-lean LOHCs after completion of dehydrogenation.
• Thermal stability: Decomposed LOHCs, through hydrogenation and dehydrogenation at high temperatures, could affect the ability of catalysts. Thus, the high thermal stability of LOHCs is an important factor for sustainable catalytic cycles.
• Viscosity: From a transportation standpoint, the lower viscosity of LOHCs makes them useful for pumping into existing tanks and pipelines.

(2)

Reversibility: Most LOHC systems can be recycled because hydrogen-rich and hydrogen-lean LOHCs are reversible in the presence of suitable catalysts via hydrogenation (exothermic reaction) and dehydrogenation (endothermic reaction). Irreversible circular hydrogen carriers, such as formic acid and ammonia, can also be used in LOHC systems, but re-use is inefficient as gaseous carbon dioxide or nitrogen is released into the atmosphere during dehydrogenation and must be injected into a new batch for hydrogenation with each cycle.

(3)

Hydrogen storage capacity: In 2020, the hydrogen and fuel cell technologies office (HFTO) in the U.S. Department of Energy suggested research and development goals for advanced hydrogen storage system, shown below. Thus, promising LOHC systems need to meet these conditions

• Hydrogen gravimetric storage density—6.0 wt% hydrogen (1.5 kWh/kg system).
• Hydrogen volumetric storage density—0.030 kg hydrogen/L (1.0 kWh/L).
• Cost—$333/kg stored hydrogen capacity ($10/kWh).

(4)

Eco-friendly characteristic: To minimize environmental damage, it is necessary to develop LOHC systems that produce only benign byproducts, such as water, during hydrogenation and dehydrogenation. Moreover, the toxicity of LOHCs should be investigated to prevent its effects on human health.

This review will cover a variety of potential liquid organic hydrogen carriers (LOHCs) that have several advantages regarding the above criteria, including both irreversible hydrogen carriers (formic acid/formaldehyde and ammonia) and reversible LOHCs (homocyclic compounds, nitrogen-containing compounds, and oxygen-containing compounds) (Figure 1, bottom).

Along with the discovery of potential LOHCs, the development of efficient catalytic systems is required. A pair of hydrogen-rich and hydrogen-lean LOHC compounds can be reversibly converted, using thermally activated bond-breaking and -formation [10]. The endothermic dehydrogenation of benzene (300 °C–350 °C) and heterocycles (50 °C–200 °C) requires extremely high temperatures [11]. To accelerate the hydrogenation and dehydrogenation process, the development of novel catalysis enables mild reaction conditions by lowering the activation energy and controlling the kinetics [12,13]. There are two types of catalyst systems: homogeneous and heterogeneous catalysts. Each have advantages and disadvantages.

• Homogeneous catalysts: The performance of homogeneous catalysis at lower temperatures is generally better than heterogeneous catalysis because it is more useful to modify the catalyst with various ligands to enhance the catalyst’s ability [14]. As a result, homogeneous catalysis is also applicable to LOHCs. However, recyclability is not good, since it is difficult to recover the catalyst after the reaction.
• Heterogeneous catalysts: Heterogeneous catalysts are typically effective for application in potential LOHC systems in terms of thermal stability and recyclability because it is easier to separate the catalyst from LOHCs by filtering. Furthermore, heterogeneous catalysts can be used for continuous flow processes to scale up for hydrogenation and dehydrogenation to take and release hydrogen in industry.

Several reviews of LOHC systems, especially for comprehensive concepts/research [7,14,15,16,17], formic acid-based LOHCs [18], nitrogen-containing LOHCs [13], and storage/transportation [19,20,21,22,23], have been published in the last 10 years. In this review, we aim to focus on advances that occurred in recent decades in representative hydrogenation and dehydrogenation using homogeneous and heterogeneous catalysts for application in a variety of potential LOHC systems (Figure 1, top). In addition, we will also introduce reactor types for catalytic hydrogenation and dehydrogenation in the academy and industry.

2. Circular Hydrogen Carriers

In contrast to reversible LOHCs, the process of absorbing and releasing hydrogen in circular hydrogen carriers is irreversible due to the release of gaseous carbon dioxide or nitrogen through dehydrogenation. As a result, every hydrogenation requires new material (CO₂ or nitrogen) from the atmosphere along with hydrogen to form formic acid or ammonia. These prototype circular hydrogen carriers allow hydrogen to be stored and transported to energy-poor areas, then produce hydrogen via dehydrogenation while simultaneously releasing CO₂ or nitrogen into the atmosphere [14,24,25].

2.1. Formic Acid

In industry, three main processes have been used to produce formic acid: (1) acidolysis of formate salts; (2) liquid-phase hydrocarbon oxidation; (3) carbonylation of methanol to methyl formate, followed by direct ester hydrolysis [26]. The catalytic dehydrogenation of formic acid and hydrogenation of carbon dioxide to generate formic acid have recently been investigated for use in LOHC systems. Formic acid can be produced by reducing carbon dioxide, and catalytic dehydrogenation of formic acid can produce carbon dioxide and molecular hydrogen. In this regard, formic acid (4.4 wt%) can be an attractive LOHC. In general, the most difficult part of formic acid dehydrogenation is suppressing undesired CO generation through dehydration because carbon oxide deactivates the catalyst. Thus, developing efficient catalytic hydrogenation and dehydrogenation using homogenous (Fe, Ru, Co, Mn, and Ir) and heterogeneous (Pd and Co) catalysts is important to produce extremely pure formic acid and hydrogen.

In 2011, the Beller group reported an efficient method for producing hydrogen from formic acid through dehydrogenation using a highly active iron catalyst [27]. The active iron catalyst Fe-1 system can be generated from a mixture of a commercially available cationic iron(II) precursors, Fe(BF₄)₂·6H₂O, and a tetradentate phosphine ligand, [P(CH₂CH₂PPh₂)₃, PP₃], in propylene carbonate (PC) as an environmentally friendly solvent. For dehydrogenation, only 0.005 mol % of Fe-1 was used to produce hydrogen with a TOF of 9425 h⁻¹ at 80 °C in the absence of additives (Table 1, Entry A). Furthermore, the catalyst system is very stable against water, air and high temperatures but sensitive to trace amounts of chloride impurity. As a result, continuous hydrogen production was successfully achieved with a total TON of 92,417 due to the high robustness of Fe-1. However, deactivation of the catalyst system occurred because of chloride accumulation over time.

In 2014, Pidko and colleagues reported the formation of formic acid using a homogeneous ruthenium pyridyl-phosphine (PNP)-pincer catalyst via carbon dioxide hydrogenation. [28]. The Ru-1 system with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) for the hydrogenation of carbon dioxide to formic acid was identified as an optimal catalyst and base. The best performance in the dehydrogenation of formic acid was achieved using Ru-1 with triethylamine as a base, delivering H₂ and CO₂ (TOF = 257,000 h⁻¹) at 90 °C (Table 1, Entry B). The proposed catalytic cycles for the hydrogenation of CO₂ and the dehydrogenation of formic acid are illustrated in Scheme 1. For the dehydrogenation of formic acid, the Ru complex 4 undergoes base-assisted H₂ formation via recombination of the hydride ligand and a proton from BH⁺ to form Ru complex 5, which is converted to Ru cationic complex 6 upon the release of H₂. Ru complex 7 is produced by the coordination between the formate anion and complex 6. Finally, the initial complex 4 is regenerated through rearrangement of 7 into a transient species 8, followed by CO₂ emission.

To develop consecutive charge (hydrogenation) and discharge (dehydrogenation) cycles, Olah and colleagues studied an efficient reversible hydrogen storage system in the presence of ruthenium pincer complex Ru-2 without amine ligands in 2015 [29]. Conceptually, the formate salts can be selectively converted to hydrogen and CO₂, which can be recycled back to the formate salts using a suitable catalyst system. As shown in Scheme 2, carbon dioxide/hydroxide and bicarbonate can be interconverted at equilibrium; then, bicarbonate undergoes hydrogenation to form formate and water in the presence of the Ru-2 catalyst at higher pressure (75 bar). In situ-generated formate can also be converted to bicarbonate via dehydrogenation in the same batch at a lower pressure (atmospheric pressure). To prove the concept of reversibility, six consecutive carbon dioxide hydrogenation and sodium formate dehydrogenation were successfully performed at 70 °C by pressure swing without the addition of external carbon dioxide. As a result, a pure mixture of hydrogen and carbon dioxide was obtained with a total TOF of >11,500 and, no CO impurity was detected due to the high stability of Ru-2 over six cycles (Table 1, Entry C).

In 2016, Cao and coworkers developed heterogeneous palladium catalysts, supported by pyridinic nitrogen-doped carbon, for the dehydrogenation of the formic acid [30]. Palladium is an active component in formic acid decomposition under mild conditions; however, the stability of palladium can be reduced due to low H₂ selectivity and poor CO tolerance. To solve these problems, the surface electronic structure of the supported palladium catalyst was modified with flexible materials, such as carbon-based supports, and a series of CN_x carbon hybrids was synthesized (CN_x: x denotes the N/C molar ratio). Thus, Pd is more active in the N-doped carbon support than other metals, such as Pt, Ru, Rh, and Ir nanoparticles. CN_0.25 exhibited the highest activity compared with conventional metal oxide and reduced graphene oxide (rGO). Finally, the N-doped carbon-supported Pd catalyst, Pd/CN_0.25, was found to be active for formic acid dehydrogenation, delivering ultraclean H₂ gas without the formation of carbon monoxide byproduct, with a TOF of 5530 h⁻¹ per surface Pd site at 25 °C. In addition, the heterogeneous catalyst, which has the turnover number (TON) 50,040 due to its excellent recyclability and ease of handling upon completion of the reaction, can be reused six times. Pd/CN_0.25 can also proceed with a reverse hydrogenation reaction from carbon dioxide to formic acid; therefore, it could serve as a new and effective way to increase the efficiency of formic acid dehydrogenation and carbon dioxide hydrogenation (Table 1, Entry D).

In 2018, the Laurenczy group reported the carbon dioxide hydrogenation reaction to produce formic acid using an iron(II) catalyst without additives in an aqueous medium at room temperature [31]. The active Fe(II)-PP₃TS complex (Fe-2) can be generated in situ from inexpensive catalyst precursor, Fe(BF₄)_2, and meta-trisulfonated-tris[2-(diphenyl-phosphino)-ethyl]phosphine (PP₃TS) ligand in aqueous solution. Carbon dioxide hydrogenation requires high pressure to increase the solubility of gaseous H₂/CO₂ in aqueous solution. In addition, since carbon dioxide reduction is an exothermic reaction, a mild temperature is also an important factor to accelerate the hydrogenation rate. After the optimized reaction conditions, the highest formic aid concentration (0.5 M) was obtained under optimal pressure (100 bar H₂ and 50 bar CO₂) at 25 °C for 60 h in the presence of Fe-2 catalyst (Table 1, Entry E). Recycling experiments for three cycles also show the stability of the Fe-2 catalyst in air and water without deactivation.

In 2019, Singh et al. reported the catalytic dehydrogenation of formic acid using a water-soluble Ru-arene complex containing N-H moiety [32]. The use of water-soluble catalysts in aqueous solutions allows for environmentally friendly catalysis by avoiding toxic organic solvents. Previous studies demonstrated that ethylenediamine (en)- and 8-aminoquinoline (AmQ)-based chelating ligands affect the activation of formic acid, which is used to facilitate transfer-hydrogenation. Thus, Ru-arene complexes containing various nitrogen donor-chelating ligands were investigated at 90 °C in the absence of a base for formic acid dehydrogenation. Among the various catalyst complexes, [(η⁶-C₆H₆)Ru(κ₂-N_pyNHMe-8-AmQ)Cl]⁺, (Ru-3), containing 8-(N-methylamino)quinoline, was the most efficient catalyst, with a TON of 2248. To understand the mechanism, pH-dependent studies were conducted, and the highest TOF (940) was obtained at mild acidic conditions (pH = 4.0, [HCOONa]:[HCOOH] = 2:1) (Table 1, Entry F). This result indicates that H₃O⁺ and HCOO⁻ formation is important in the dehydrogenation reaction. In addition, the formation of diruthenium intermediate was observed during the catalytic dehydrogenation, based on mass spectrometry and single-crystal X-ray crystallography.

In 2019, Beller and coworkers developed a homogeneous cobalt–pincer complex for formic acid dehydrogenation in an aqueous medium under mild conditions [33]. Since cobalt is a relatively abundant 3D metal, dehydrogenation using the cobalt catalyst is more economical because Co is cheaper than other metals such as Ru, Pd, and Ir. Several Co¹ complexes were synthesized and characterized using NMR, combustion analysis, and X-ray diffraction to investigate efficient catalysts. Formic acid was almost fully dehydrogenated and decomposed into carbon dioxide at 80 °C within 90 min, reaching a turnover number (TON) of 2260, using a cobalt PNP–pincer complex. However, it is air-sensitive. To address this problem, the precursor complex (Co-1) was used in situ activation by sodium triethylborohydride (NaBEt₃H) for convenient handling (Table 1, Entry G). Based on the DFT computations, a mechanism for the cobalt-catalyzed dehydrogenation of formic acid was proposed (Scheme 3).

In 2019, Tondreau et al. reported a robust and rapid manganese-mediated formic acid dehydrogenation [34]. The manganese catalyst complex, [(^tBuPNNOP)Mn(CO)₂][Br] (Mn-1), was synthesized using MnBr(CO)₅ with ^tBuPNNOP (=2,6-(di-tert-phosphinito)(di-tert-butylphosphinamine)pyridine) as a hybrid backbone chelate ligand (Table 1, Entry H). In the presence of the Mn-1 catalyst, formic acid dehydrogenation was achieved (TOF = 8500 h⁻¹) using a triethylamine base at 80 °C. This study demonstrated the recyclability of the manganese complex, Mn-1, for catalytic dehydrogenation; however, the reactivity slowly decreased after three cycles due to the loss of TEA during the pressure venting.

In 2020, the Beller group reported a phosphine-free manganese complex for formic acid dehydrogenation [35]. These ligand-free systems are more efficient than those in the previous study using [(^tBuPNNOP)Mn(CO)₂][Br] (Mn-1) in terms of cost and atom-economy. The most active and stable complex was Mn(pyridine-imidazoline)(CO)₃Br (Mn-2), yielding more than 14 L of H₂/CO₂ mixture for more than 87 h at 92.5 °C. The pH effect was tested to optimize the reaction conditions, and a pH lower than 7 was appropriate for the dehydrogenation. However, the reaction rate of the catalyst, Mn-2, decreased over time due to increased pH since the formic acid was consumed during the reaction. To solve this problem, a continuous injection system of formic acid was designed to efficiently produce H₂/CO₂ products (a TON of 2637 in 87 h) (Table 1, Entry I).

Bao and coworkers developed a series of biomimetic Ir complex, containing pendant N moieties to promote the catalytic activity (by proton transfer) for the dehydrogenation of formic acid [36]. The pendant amine groups in [Fe-Fe]-hydrogenase enzyme can act as proton relays for hydrogen oxidation and production. To mimic the active site of the Fe hydrogenase, a water-soluble and air-stable Ir-1 complex was synthesized (Table 1, Entry J). The TOF value of Ir-1 complex was 15 times higher than that of the complex without pendant pyridine. The variations in temperature, pH, and formic acid concentration were explored, and optimal reaction conditions (60 °C, pH = 1.7, 6 M formic acid aqueous solution) were found, reaching a high TOF (8250). Based on the mechanistic studies, the formic acid dehydrogenation was improved by the synergistic effect of using pendant N moiety as pendant bases in the outer coordination sphere. The NH moieties in pyridinium cation and the 1H-pyrazole unit can stabilize the transition state particularly well, and promote proton transfer in the RDS using hydrogen bonding with formic acid.

Beller et al. reported a stable single-atom cobalt catalyst (SAC) for the selective heterogeneous catalytic dehydrogenation of formic acid [37]. A metal–organic-framework-derived Co-N-C catalyst, with isolated single cobalt atoms, was synthesized. The dispersed Co-SAC, is more efficient, selective, stable, and acid-resistant than cobalt nanoparticles (NPs).

2.2. Formaldehyde

Formaldehyde (bp −19 °C, 400 g/L of solubility in water) can easily be produced by biodegradation from organic compounds and has a higher gravimetric hydrogen density (8.4 wt% in a gaseous state or 5.0 wt% in 37 wt% aqueous solutions) than formic acid (4.4 wt%). In addition, the hydrogen production from formaldehyde in an aqueous solution via dehydrogenation is thermodynamically favorable (ΔHr = −35.8 kJ/mol) [38,39,40,41,42,43]. Aqueous formaldehyde is in equilibrium with methanediol, which can be converted to formic acid by dehydrogenation, followed by another dehydrogenation to produce hydrogen and carbon dioxide. Therefore, formaldehyde could also be a potential circular hydrogen carrier.

In 2016, Prechtl and coworkers reported a microwave-assisted hydrogen-release system from aqueous formaldehyde using a homogeneous ruthenium arene complex catalyst, [RuX₂(arene)]₂ [44]. Contrary to the current methods of synthesizing various [RuX₂(arene)]₂, the application of microwave irradiation was used to avoid long reaction times and high temperatures. The best low-cost ruthenium arene complexes, [RuCl₂(p-cymene)]₂ (Ru-4) showed the highest TON (700) and the greatest TOF (3142 h⁻¹) under optimal conditions without additives (Table 1, Entry K). A high conversion was obtained during the first three cycles using a sample recharge reaction and a two-phase separation method.

Table 1. The pathway of formic acid/formaldehyde dehydrogenation using various catalysts. Data obtained from [27,28,29,30,31,32,33,34,35,36,44,45].

Entry	Starting Material	Catalyst	Dehydrogenation Condition	TON	TOF	Ref.
A	3	Fe-1	Propylene carbonate, 80 °C	92,417	9425	[27]
B	3	Ru-1	DMF, NEt3, 90 °C	-	257,000	[28]
C	3	Ru-2	70 °C	-	>11,500	[29]
D	3	Pd/CN0.25	H2O, 25 °C	50,040	5530	[30]
E	3	Fe-2	H2O, 25 °C	-	-	[31]
F	3	Ru-3	H2O, HCOONa, 90 °C	2248	940	[32]
G	3	Co-1	H2O, HCOOK, NaBEt3H 80 °C	2260	-	[33]
H	3	Mn-1	Chlorobenzene, NEt3, 80 °C	-	8500	[34]
I	3	Mn-2	H2O, Triglyme, HCOOK, 92.5 °C	2637	-	[35]
J	3	Ir-1	-	-	8250	[36]
K	1	Ru-4	-	700	3142	[44]
L	1	Ru-5	-	515	850	[45]
M	1	Ru-5	-	12,905	-	[45]

Table 1. The pathway of formic acid/formaldehyde dehydrogenation using various catalysts. Data obtained from [27,28,29,30,31,32,33,34,35,36,44,45].

Entry	Starting Material	Catalyst	Dehydrogenation Condition	TON	TOF	Ref.
A	3	Fe-1	Propylene carbonate, 80 °C	92,417	9425	[27]
B	3	Ru-1	DMF, NEt3, 90 °C	-	257,000	[28]
C	3	Ru-2	70 °C	-	>11,500	[29]
D	3	Pd/CN0.25	H2O, 25 °C	50,040	5530	[30]
E	3	Fe-2	H2O, 25 °C	-	-	[31]
F	3	Ru-3	H2O, HCOONa, 90 °C	2248	940	[32]
G	3	Co-1	H2O, HCOOK, NaBEt3H 80 °C	2260	-	[33]
H	3	Mn-1	Chlorobenzene, NEt3, 80 °C	-	8500	[34]
I	3	Mn-2	H2O, Triglyme, HCOOK, 92.5 °C	2637	-	[35]
J	3	Ir-1	-	-	8250	[36]
K	1	Ru-4	-	700	3142	[44]
L	1	Ru-5	-	515	850	[45]
M	1	Ru-5	-	12,905	-	[45]

Recently, Singh and colleagues reported the ruthenium-catalyzed dehydrogenation of aqueous formaldehyde and paraformaldehyde to produce hydrogen [45]. For efficient hydrogen production from formaldehyde, in situ-generated ruthenium catalyst was prepared using [(η⁶-C₁₀H₁₄)RuCl₂]₂ Ru-4 with various monodentate nitrogen-based ligands at 95 °C in aqueous solution. Based on the experiments, using these ligands, Ru-4 with imidazole was the best catalyst system (initial TOF of 850 h⁻¹ and TON of 515 after 3 h) (Table 1, Entry L), producing 438 mL of hydrogen gas in 7 h, although Ru-4 without any ligands (initial TOF of 560 h⁻¹ and TON of 380 in 3 h) was more effective than substituted imidazole ligands. The purified mono-imidazole ruthenium complexes, [(η⁶-C₁₀H₁₄) Ru(imidazole)Cl₂], Ru-5, were synthesized and surpassed in situ-generated Ru-4/imidazole catalyst, having hydrogen gas evolution rate of more than 400 mL in 6 h (300 mL in 3 h), therefore, it is assumed that the ruthenium–imidazole interaction promotes proton exchange to release hydrogen during the dehydrogenation process. To demonstrate the stability of the Ru-5 catalyst, eight consecutive cycles for the dehydrogenation of formaldehyde were conducted with an excellent conversion (>80%), and bulk hydrogen production was performed with a TON of 12,905 for more than 33 h (Table 1, Entry M).

2.3. Ammonia

Ammonia (bp −33.3 °C, 31% w/w at 25 °C of solubility in water) is typically produced by the Haber–Bosch process and has a high gravimetric hydrogen density (17.5 wt% in a gaseous state). Thus, it can also be a promising circular hydrogen carrier. However, it is difficult to handle because of its critical safety issues, including toxicity, flammability, and corrosiveness. To solve the safety issues of ammonia, various ammonia-modified hydrogen carriers (amide, urea, carbamate, imide, and amine borane) have also been developed. However, these LOHCs have a lower gravimetric hydrogen density than ammonia. In this section, two representative examples using ammonia will be introduced. Furthermore, ammonia-modified LOHCs will be discussed later in this review.

In 2010, Au and coworkers reported core-shell-structured iron nanoparticles for CO_x-free hydrogen generation using ammonia decomposition [46]. The iron nanoparticles were encapsulated inside a porous stable silica gel, resulting in Fe@microporous SiO₂ and Fe@mesoporous SiO₂ through the control of the shell porosity; thus, the core-shell structured catalysts were more active than the naked iron nanoparticles. The naked iron nanoparticles did not achieve the full conversion, even at 700 °C, whereas the ammonia was fully decomposed to produce CO_x-free hydrogen using nano-Fe@meso-SiO₂ at 670 °C and Fe@micro-SiO₂ at 650 °C. Based on the experiment, the silica shell could prevent the core iron nanoparticles from aggregation. The core-shell structure enhanced the catalyst stability of Fe@meso-SiO₂ more than the naked iron nanoparticles (73 h vs. 40 h).

In 2015, Eguchi et al. reported rare-earth elements for the catalytic decomposition of ammonia using alumina-supported nickel catalysts [47]. Several catalysts, such as metals, alloys, carbides, and nitrides using Ru, Ir, Ni, Rh, Pt, Pd, Fe, Mo, and V, were developed for the ammonia decomposition [48,49,50,51,52]. Among these catalysts, nickel catalysts attracted much interest for ammonia decomposition as they had the highest activity in nonnoble metal catalysts. To improve the catalytic activity for ammonia decomposition, Ni/Al₂O₃ catalyst was modified by various rare-earth elements. For example, La-modified Ni/Al₂O₃ exhibited 92% ammonia conversion at 550 °C because of the enhancement of nickel dispersion and suppression of hydrogen inhibition phenomenon caused by the addition of rare-earth elements.

3. Reversible Liquid Organic Hydrogen Carriers

In the presence of efficient homogeneous or heterogeneous catalysts, typical liquid organic hydrogen carriers are generally reversible between hydrogen-rich and hydrogen-poor LOHCs via hydrogenation and dehydrogenation cycles. As these LOHCs have low melting points and high boiling points, they exist mostly in liquid form, making them suitable for storage and transportation using existing infrastructure such as tanks and trucks. Furthermore, for practical application in industrial fields, toxicity and eco-toxicity should be considered. In this section, we will talk about homocyclic compounds, N-containing compounds (N-heterocyclic compounds, Amide, Imide, Urea/Carbamate, Nitrile, and Amine–borane), and O-containing compounds (Silane/Alcohol pair, Ketone, and Ester/Alcohol pair).

3.1. Homocyclic Compounds

Early LOHC studies focused on polycyclic hydrocarbons due to their several advantages, such as their high gravimetric hydrogen contents (5 wt%–8 wt%), high volumetric hydrogen storage capacities (>60 g H₂/L), and reversibility between hydrogenation and dehydrogenation. Simple aromatic compounds such as benzene [53,54] and toluene [55,56] were intensively investigated in the 1980s as early LOHCs, but these LOHCs will not be addressed in this review.

In 2014, Wasserscheid and coworkers reported the various physical characteristics of dibenzyltoluene (DBT) 12 and benzyltoluene (BT) compared to N-ethyl carbazole (NEC) [57]. Based on the experiments, both DBT (m.p. −39 to −34 °C, b.p. 390 °C and 6.2 wt%) and BT (m.p. −30 °C, b.p. 280 °C and 6.2 wt%) have low melting points, high boiling points, high hydrogen gravimetric storage density, low toxicity, and eco-toxicity. However, NEC (m.p. 68 °C, b.p. 270 °C, and 5.8 wt%) is solid at room temperature, so it has transportation limitations in existing infrastructures. To investigate thermal stability, all three LOHCs were also conducted in the presence of Pt/Al₂O₃ (0.2 mol%) for 72 h at 270 °C. As a result, DBT (<0.01% decomposition) and BT (<0.01% decomposition) has higher thermal stability than NEC (<0.2% decomposition). For dehydrogenation, DBT was explored using several heterogeneous catalysts (0.15 mol%) at 270 °C for 3.5 h, Pt/C (1 wt%) was the best, with a 71% degree of dehydrogenation (Scheme 4a), compared to Pt/Al₂O₃ (0.5 wt%, 51%), Pt/Al₂O₃ (5 wt%, 40%), Pt/C (5 wt%, 55%), Pt/SiO₂ (1 wt%, 10%), Pd/C (5 wt%, 16%) and Pd/Al₂O₃ (5 wt%, 8%). After further optimizations, excellent dehydrogenation (97%) was achieved using 0.1 mol% of Pt/C (1 wt%) at 310 °C. Therefore, DBT could be a promising LOHC for practical industrial application in terms of physicochemical properties.

In 2018, the Wasserscheid group also reported the heterogeneous transition metal-catalyzed hydrogenation of dibenzyltoluene (H0-DBT) 12 to perhydro-dibenzyltoluene (H18-DBT) 13 using H₂/CO₂ gas mixture (70%–75% hydrogen, 20%–25% carbon dioxide, a small amount of methane, and carbon monoxide) (Scheme 4a) [58]. For the hydrogenation, four alumina-supported heterogeneous catalysts were explored. The results show that rhodium and ruthenium catalysts were effective below 180 °C, whereas the hydrogenation of H0-DBT 12 using platinum and palladium requires a higher temperature, above 200 °C. Additionally, platinum is almost inactive in the presence of CO₂ for all temperature ranges, requiring pure hydrogen gas conditions. However, rhodium on alumina showed the highest activity between 120 and 150 °C, even in the hydrogen-rich gas mixture. Based on the results, it is revealed that rhodium catalyst can be activated by the exchange of gas mixture due to the hydrogen production plateau. In addition, gas separation by hydrogenation, as well as by the conversion of hydrogen gas into the storable and transportable LOHC, is achievable in a single-step semi-continuous system, which operated at 150 °C, 45 bar, and a H₂/CO₂ ratio of 4:1.

Kalenchuk et al. reported reversible carbon-supported platinum-catalyzed hydrogenation–dehydrogenation of o-terphenyl, which is among the most promising materials for LOHC [59]. For the hydrogenation of o-terphenyl, Pt/C is an effective catalyst, delivering perhydro-o-terphenyl in full conversion with a high selectivity of 97%–99%. In the dehydrogenation of perhydro-o-terphenyl (cis isomer: trans isomer = 1:3), in the flow process, the mixture of o-terphenyl, m-terphenyl, and a condensation product, triphenylene, was formed by the isomerization of o-terphenyl, followed by dehydrogenation. Thus, the overall rate and volume of the generated hydrogen were reduced.

In 2019, the Yoon group developed the eutectic mixture of biphenyl 14 (m.p. 68 °C–70 °C, b.p. 255 °C and 6.9 wt%) and diphenylmethane 16 (m.p. 22 °C–25 °C, b.p. 264 °C–266 °C and 6.9 wt%), as an efficient and reversible LOHC [60]. Regarding transportation, the mixture of biphenyl 14 (35 wt%) and diphenylmethane 16 (65 wt%) is required because biphenyl 14 exists in a solid-state at room temperature. In their group, the mixture, as a novel hydrogen carrier, has demonstrated an ability to store hydrogen by catalytic hydrogenation over Ru/Al₂O₃ at 120 °C, under hydrogen (50 bar) (Scheme 4b). Based on the experiment, almost full conversion (99%) was achieved without any byproducts. The eutectic mixture (1 kg) of 15 and 17 was also converted to the mixture of 14 and 16 in 94% conversion via dehydrogenation over Pd/C pellet catalysts at 340 °C, and no decomposition products were detected. This means that these LOHCs have high thermal stability. Hydrogen purity, one of the key factors affecting LOHC, was excellent (99.9%). In addition, reversibility tests between hydrogenation and dehydrogenation were performed nine consecutive times. Finally, this study showed application potential in a polymer electrolyte membrane fuel cell.

Scheme 4. Reversible hydrogenation and dehydrogenation of (a) dibenzyltoluene 12 (DBT) and (b) the biphenyl-diphenylmethane 14–16 eutectic mixture. Data obtained from [57,58,60].

Scheme 4. Reversible hydrogenation and dehydrogenation of (a) dibenzyltoluene 12 (DBT) and (b) the biphenyl-diphenylmethane 14–16 eutectic mixture. Data obtained from [57,58,60].

3.2. Nitrogen-Containing Compounds

Nitrogen-containing compounds are among the most promising LOHCs due to their high gravimetric hydrogen densities and favorable hydrogen-release kinetics. The presence of N atoms in carbocyclic compounds (5.3–7.3 wt%) lowers the endothermicity of dehydrogenation compared to the corresponding homocyclic compounds because of the bond strength of C-H adjacent to N atom and N-H is weaker than the bond strength of C-H adjacent to C atom and C-H [14]. Thus, hydrogen production can be achieved by dehydrogenation at a lower temperature [9,61,62]. In addition, the development of amide- or imide-based systems (5.3–6.6 wt%) can form potential LOHCs using C–N or C–O bond cleavage from available and inexpensive amines and alcohols [63,64]. In the hydrogenation of carbonic acid derivatives, carbamates (5.26 wt%) and ureas are most challenging in the hydrogenation due to their polarizability [64]. However, these LOHC systems are fascinating in their practical synthetic route, under mild conditions and facile preparation using the reaction between CO₂ and a corresponding amine/amine or amine/alcohol pair. In addition, nitrile (13.3 wt%) and amine-borane (4.7 wt%) can be alternative LOHCs with their high gravimetric hydrogen capacity. This section introduces recent representative hydrogenation and dehydrogenation methods using homogeneous and heterogeneous catalysts for N-containing LOHCs with more detail.

3.2.1. N-Heterocyclic Compounds

In 2014, Yamaguchi and coworkers reported an efficient methodology for homogeneous perhydrogenation and perdehydrogenation of fused bicyclic N-heterocycles using Cp*Ir complexes bearing functional bipyridonate ligands as a single precatalyst [65]. The substrate, 2,6-dimethyldecahydro-1,5-naphthyridine 19, contains the ring-fused structure of two N-heterocyclic parts, and two nitrogen atoms are distanced from each other to prevent bidentate coordination to the iridium center of the complex. To prevent the rotation around the pyridyl–pyridyl bond, a Cp*Ir complex with a rigid functional 1,10-phenanthroline-2,9-dione ligand (Ir-2) was selected as an optimal catalyst. Hydrogenation of 2,6-dimethyl-1,5-naphthyridine 18 and dehydrogenation of 2,6-dimethyldecahydro-1,5-naphthyridine 19 via 5 molecules of hydrogen transfer were conducted in quantitative and 92% yield, respectively (Scheme 5a).

In 2016, Kempe et al. reported an efficient means of high-wt% hydrogen storage in an N-heterocyclic compound and ammonia using a reusable bimetallic catalyst [66]. A novel heterogeneous PdRu@SiCN (Pd:Ru = 2:1) catalyst was synthesized by mixing commercially available Ru and Pd complex with the polysilazane HTT1800, followed by cross-linking and pyrolysis. Using the novel bimetallic catalyst, the hydrogenation of N-ethyl carbazole (NEC) and dehydrogenation of dodecahydro-N-ethylcarbazole (12H-NEC) were achieved in agreement with the calculated hydrogen storage/release wt% for three cycles, with a minor decrease in hydrogen uptake and release. To explore the expanded substrates, octahydrophenazine was prepared using ammonia and two equivalent cyclohexane-1,2-diols, synthesized from lignin hydrogenolysis products by dehydrogenation and condensation in the presence of an Ir@SiCN catalyst. Finally, the hydrogen uptake and release between tetradecahydrophenazine 21 and phenazine 20 was demonstrated for seven catalytic cycles using Pd₂Ru@SiCN catalyst (Scheme 5b).

In 2017, Fujita and colleagues developed a new method for the reversible interconversion between 2,5-dimethylpyrazine 22 and 2,5-dimethylpiperazine 23, via iridium catalyzed hydrogenation and dehydrogenation [67]. In the hydrogenation of 2,5-dimethylpyrazine 22 using a co-solvent of p-xylene and water, the H/S ratio (stored hydrogen in mmol)/(amount of solvent in mL) reached 60. In addition, a 2,5-dimethylpiperazine 23, having hydrogen gravimetric capacity (5.3 wt%), was used for quantitative dehydrogenation, under solvent-free conditions with a small amount of catalyst Ir-2 (0.5 mol%). Under these optimal systems, reversible and repetitive interconversion between 2,5-dimethylpyrazine 22 and 2,5-dimethylpiperazine 23 was achieved for four cycles (Scheme 5c).

In 2018, Albrecht and coworkers developed an efficient method for the hydrogenation and dehydrogenation of quinoline derivatives using a triazolylidene-based iridium complex without the use of additives in water under mild conditions [68]. For the hydrogenation of quinoline 24a to form 1,2,3,4-tetrahydroquinoline 25a, optimal conditions were reached using 0.5 mol% Ir-3 in aqueous solution at 90 °C, under 5 bar of H₂ gas. The Ir-4 complex was effective in IPA solvent with full conversion under the same conditions. Several quinoline derivatives, such as 6-Me 24b, 3-Me 24c, 2-Me 24d, and 6-Cl 24e substituents, were tolerated using both Ir-3 in water and Ir-4 in IPA, whereas 6-COOH substituent 24f was conducted using only Ir-4 in IPA (Scheme 5d). The best result for the dehydrogenation of 1,2,3,4-tetrahydroquinoline 25a was obtained using Ir-3 catalyst in water at reflux temperature, delivering quinoline 24a in 90% yield within 20 h. Based on these results, complex Ir-3 provides a catalytic system for hydrogenation and dehydrogenation in the same solvent, and these reactions could proceed depending on the partial H₂ pressure of the reaction medium.

Wasserscheid et al. (2019) reported a molten salt–LOHC liquid–liquid biphasic reaction for the dehydrogenation of 2-methylindoline to 2-methylindole 28d using a cationic iridium catalyst [69]. The homogeneous hydrogenation and dehydrogenation of N-heterocycles was demonstrated using Rh, Ru, Ir, and Os complexes with excellent catalytic activities, however, catalyst separation is often difficult and has low catalyst stability. To resolve these problems, a novel strategy was designed, combining the advantages of mild homogeneous catalysis for hydrogen-release with a molten salt-based catalyst immobilization. [Ir(cod)(py)(PCy₃)]PF₆ (Crabtree’s catalyst), the best catalyst for dehydrogenation reaction, and the molten salt [PPh₄][NTf₂] (m.p. 134 °C) were used as the catalyst immobilization medium. To increase the stability of the Crabtree’s catalyst through strong coordination, triphenylphosphine (PPh₃) was added and, as a result, 2-methylindoline 29d was dehydrogenated to 2-methylindole 28d.

In 2019, Milstein and coworkers reported a heterogeneous palladium-catalyzed LOHC system based on 2-methylpiperidine and 2,6-dimethylpiperidine [72]. Piperidine derivatives are ideal candidates for LOHCs because of their abundance, cost-effectiveness, low melting point (−70 °C–2.4 °C), wide liquid range, and high hydrogen storage capacity (5.3–6.1 wt%). In the studies, in situ-generated Pd/C catalyst, formed from Pd(OAc)₂ and dried activated carbon, is more efficient than the commercial Pd/C in the dehydrogenation of 2-methylpiperidine. In addition, the reversible interconversion of 2-methylpiperidine and 2-picoline via hydrogenation and dehydrogenation could be successfully achieved because the electron-donating methyl groups at 2-positions prevent deactivation of the catalyst and facilitate the product dissociation from the catalyst. As a result, 2-methylpiperidine was obtained via dehydrogenation in 91% yield in the presence of 0.15 mol% of Pd(OAc)₂ at 170 °C for 94 h. The hydrogenation reaction immediately after the dehydrogenation was carried out by directly pressurizing the H₂ (2–7 bar) produced via dehydrogenation, delivering 2-picoline in 100% conversion at 150 °C for 36 h.

In 2020, Rueping and coworkers reported a single-manganese-complex-catalyzed hydrogenation and dehydrogenation of indole 26, indoline 27 and their derivatives [70]. Indoline 27 was formed in high yield using 2 mol% of the Mn-PNP catalyst Mn-3 and 5 mol% of KO^tBu, as the base, for 36 h at 100 °C under 50 bar of hydrogen pressure. In the event, the application of the optimal conditions to various substituted indoles, including C-5, C-6, and C-7 simple methyl-substituted 26a-c and halogen-containing indoles 26d-f delivered the corresponding indoline derivatives by selective hydrogenation with excellent H₂ yield (Scheme 5e). Other N-containing heterocyclic compounds, such as quinoxaline and benzoxazine, were also hydrogenated under similar conditions. Furthermore, the dehydrogenation of indoline 27 was performed using 1 mol% of the same catalyst Mn-3 at 120 °C for 24 h, delivering indole 26 in 89%.

Recently, the Li group reported an efficient photocatalytic acceptor-free dehydrogenation of hydrogen-rich N-heterocycles using Ni(OH)₂@CdSe/CdS quantum dots (QDs). The catalytic system has several advantages, such as low-cost, earth abundance, and optical/electronic properties [71]. To investigate the catalytic activity of CdSe and CdSe/CdS QDs for dehydrogenation, 1,2,3,4-tetrahydroquinoline 25a (THQ) was explored to produce quinoline 24a (QL) by injection of NiCl₂ into the reaction mixtures under 420-nm light (intensity = 50 mW cm⁻²). As a result, the amount of hydrogen produced depended on the irradiation times. To broaden the scope of N-heterocycles’ dehydrogenation, 2- MeTHQ 25d, 4-MeTHQ 29a, indoline 29c, and 2-methylindoline 29d were conducted, delivering the corresponding 2-MeQL 24d, 4-MeQL 28a, indole 28c, and 2-methylindole 28d in good yields (82%–93%); however, iso-THQ was ineffective due to the nonconjugation of the secondary amine with the phenyl group, which interfered with the oxidation process (Scheme 5f). Based on the XPS experiment, the lower negative potential of Ni2⁺/Ni than the Ni(OH)₂ of CdS is an active site for proton reduction. This improves the photocatalytic activity by electron transfer from the CdSe/CdS QDs to in situ-formed Ni(OH)₂.

3.2.2. Amide

In 2015, Milstein and coworkers reported a novel LOHC system of 2-aminoethanol 30/cyclic dipeptide 31 using a ruthenium pincer catalyst [73]. The PNN ruthenium pincer complex is an efficient catalyst for several C–O and C–N bond-forming dehydrogenative reactions with the formation of hydrogen gas. In this study, inexpensive and abundant amino alcohols were used as an LOHC candidate to form cyclic dipeptide 31 and linear oligopeptide 32 by dehydrogenative coupling. To enhance the hydrogen storage capacity, the formation of cyclic dipeptide 31 as a major product was required, instead of linear oligopeptide 32. Ru-6 was used as a catalyst, activated by KO^tBu, for the dehydrogenation of 2-aminoethanol 30, providing cyclic peptide 31 in 78% conversion and 48% yield. In addition, cyclic peptides 31 were synthesized more efficiently with the use of Ru-6, bearing the N(H)(^tBu) amine group instead of the NEt₂ group, with 85% conversion and 60% yield. The hydrogenation reaction of the cyclic peptide 31 was successfully conducted using catalysts Ru-6 and Ru-7 in the quantitative yield. Finally, repetitive reversal reactions of dehydrogenation–hydrogenation were performed with high conversion without isolation of the crude reaction mixture for 3 cycles (Figure 2).

Prakash et al. (2017) reported a reversible hydrogen storage system based on a methanol/amine pair [74]. The secondary N,N’-dimethylethylenediamine 34 was selected as a potential LOHC due to its low carbon content (for efficient H₂ storage) and low volatility (for easy handling). The reaction between the secondary diamine 34 and methanol proceeded in the presence of the Ru-2 catalyst; the product, formamide 37, was obtained with the release of hydrogen and CO (Scheme 6a). The pure H₂ was produced using the catalyst Ru-2 without the generation of CO gas, compared to Ru-8, Ru-PNP^ipr catalysts. The proposed pathway for the amide 37 and CO formation is depicted in Scheme 6b. The formaldehyde 1 formed by the initial dehydrogenation of methanol 35 undergoes a nucleophilic addition of an amine 34 to form an α-amino alcohol 36, followed by the second dehydrogenation, to provide a formamide 37. In addition, the formamide 37 could be converted to secondary diamine 34 at 60 bar H₂ pressure using a Ru-2 catalyst.

In 2020, Liu and coworkers reported a reversible interconversion of methanol–diamine and diamide based on non-noble metal–manganese-catalyzed hydrogenation and dehydrogenation [75]. The full conversion of N,N’-dimethylethylenediamine 34 was achieved with the use of PhPNP-Mn complex (Mn-4) by dehydrogenation; however, a small amount of CO was detected due to the decarbonylation of the formaldehyde intermediate. To suppress the formation of a byproduct, the catalyst was added in two equal portions, before and after gas release. After the optimization, it revealed that Mn-4 was the best catalyst for the dehydrogenation of diamine–methanol, while Mn-4 and NNP-pincer (Mn-5) catalysts were effective for the hydrogenation of diamide 38 (Scheme 7). Hydrogenation using Mn-4 requires harsher conditions (higher temperature, 180 °C, and hydrogen pressure, 80 bar) than the use of Mn-5 (110 °C, and hydrogen pressure, 60 bar).

In 2020, Milstein et al. reported the ruthenium-catalyzed synthesis of oxalamides 39 by dehydrogenative coupling of ethylene glycol 40 and amines and the reverse hydrogenation [76]. It is an atom-economic and environmentally benign strategy for the direct synthesis of oxalamides 39. The dehydrogenation reaction occurs during the coupling of ethylene glycol 40 using hexan-1-amine in the presence of a PNNH-based ruthenium pincer complex Ru-9 to form an α-hydroxy amide, followed by a similar catalytic pathway with other molecules of amine to give the desired oxalamide 39 in 86%. Under the optimized reaction conditions (1 mol% of Ru-9, 2 mol% of KO^tBu at 135 °C in toluene), various aliphatic amines, diamine, carbamate, and secondary amines were tolerated. In addition, the hydrogenation of oxalamides 39, which forms ethylene glycol 40 and corresponding amines, was achieved in the presence of 1 mol% of Ru-9 with 4 mol% of KO^tBu under 40 bar of H₂ gas at 135 °C within 24 h (76 %–97 % yield) (Figure 3).

3.2.3. Imide

In 2018, Milstein and colleagues reported an LOHC system through the reversible hydrogenation and dehydrogenation between cyclic imides and diol/amine using ruthenium catalyst [77]. For the hydrogenation of N-benzylphthalimide 41, Ru-7 (1 mol%) with KO^tBu (3 mol%) was explored as a catalyst in THF under 20 bar of H₂ at 110 °C, and the 1,2-benezenedimethanol 42 and benzylamine 43 were formed in quantitative yield (Scheme 8a). Based on the results, Ru-7 exhibits good catalytic activity due to its dual modes of metal–ligand cooperation by H-M/N-H and aromatization/dearomatization of the lutidine backbone.

The proposed mechanism for the hydrogenation of cyclic imides to diols and amines is depicted in Scheme 9. The hydroxyamide 44 was formed in the presence of the ruthenium catalyst by hydrogenation of one of the carbonyl groups from the cyclic imide 41, followed by C–N hydrogenolysis. The amine could be eliminated, whereas lactone 47 or phthalide could be generated after intramolecular cyclization. Finally, the lactone 47 undergoes Ru-catalyzed hydrogenation to form the desired diol product 42. For application in the LOHC system, the dehydrogenative coupling of 1,4-butanediol and ethylenediamine 51 with bis-cyclic imide was studied. The shows that the use of catalyst Ru-10 provides the corresponding cyclic imide with 70% yield, with suppression of the side pathway-forming lactone and linear oligoamide (14%).

In 2020, the Milstein group also reported the selective hydrogenation of cyclic imide to diol and amine using inexpensive and earth-abundant metal, manganese pincer complex catalyst [78]. For hydrogenation, 99% conversion from N-benzylphthalimide 41 to 1,2-benzenedimethanol 42 and benzylamine 43 was obtained using complex Mn-6 (2 mol%) with KO^tBu (3 mol%) under 30 bar of H₂ at 130 °C for 48 h (Scheme 8b). It is considered that the Mn-6 catalyst, bearing a PPh₂ group, is a highly reactive catalyst due to the lower steric hindrance and higher Lewis acidity of the Mn center for a better binding effect of the carbonyl compounds.

3.2.4. Urea/Carbamate

In 2019, Milstein and coworkers reported a hydrogen storage system based on the hydrogenation of ethylene urea 50 to methanol/ethylenediamine 51, and the dehydrogenative coupling of methanol/ethylenediamine 51 using ruthenium catalyst [79]. The promising LOHC candidate, methanol (6.52 wt%), can be converted to ethylene urea 50; it has a higher hydrogen storage capacity than amide formation (5.26 wt%). In this study, ethylene urea 50 was hydrogenated to ethylenediamine 51 using Ru-7 (1 mol%) and KO^tBu (4 mol%) in 1,4-dioxane under 60 bar of H₂ at 170 °C in 100% yield and conversion for 1.5 d. In addition, diformamide 48 and ethylene urea 50 were generated, using the Ru-7 catalyst, via the dehydrogenative coupling of ethylenediamine 51 and methanol, whereas no product was observed using Ru-10 (Scheme 10a). The pathway for dehydrogenative coupling of ethylenediamine 51 and methanol were proposed in Scheme 11. The methanol and ethylenediamine 51 undergo ruthenium-catalyzed dehydrogenation to formamide 49 and diformamide 48 with the release of hydrogen gas. Ethylene urea 50 was formed from the formamide 49 via dehydrogenation or from diformamide 48 in the presence of ethylenediamine 51.

The Milstein group reported manganese-catalyzed hydrogenation of carbamate derivatives, which are the most challenging carbonyl compounds to hydrogenate [80]. The quantitative conversion of the linear carbamates, bearing an aliphatic or aromatic substituent, to methanol was obtained by hydrogenation in the presence of the manganese complex Mn-6 (2 mol%) and KO^tBu (3 mol%) under low hydrogen pressure (20 bar) at 130 °C for 48 h (Scheme 10b).

3.2.5. Nitrile

In 2010, Sabo-Etienne et al. reported the hydrogenation of nitriles to amines using a ruthenium complex, under mild conditions [81]. The amine/nitrile pairs, with a long hydrocarbon chain, could be a potential LOHC because two hydrogens are available from each CH₂NH₂ fragment in the compound (13.3 wt% of hydrogen per CH₂NH₂). In this study, the selective hydrogenation of benzonitrile 53 was achieved using ruthenium complex, RuH₂(H₂)₂(PCyp₃)₂ Ru-12, to provide primary amine 43 with the suppression of side pathways, forming imines and secondary amines (Scheme 12a). Finally, excellent conversion (98% benzylamine 43:dibenzylamine = 96:4) was obtained via ruthenium-catalyzed hydrogenation, using the catalyst Ru-12 (0.2 mol%) under 3 bar of H₂ in THF at ambient temperature for 24 h.

In 2013, Szymczak and coworkers developed oxidant-free selective dehydrogenation of primary amines to nitriles using a homogeneous amide-derived NNN–Ru hydride complex Ru-13 [82]. The conversion from benzylic amines 55 to corresponding benzonitriles 54a was achieved without any additives, such as an oxidant or hydrogen acceptor; thus, potentially oxidizable functional groups were tolerated in this system. Under optimized reaction conditions, the use of various aliphatic, aromatic, and benzyl was explored, delivering corresponding nitriles 54 in moderate yields (Scheme 12b).

In 2016, Mata et al. reported the feasibility of the LOHC system for the conversion of primary amines to nitriles via catalytic hydrogen production using an immobilized ruthenium catalyst [83]. The ruthenium complex was synthesized using [Ru(p-cym)Cl₂]₂ and imidazolium salt to form N-heterocyclic carbene (NHC) ligand ruthenium complexes. The dehydrogenation of benzylamine 43 to benzonitrile 54a was performed using Ru-14 catalyst in toluene for 8 h without an oxidizing reagent, forming the products in full conversion with the selectivity of nitrile:imine = 82:18 ratio. In the case of dehydrogenation of an alkylamine, the formation of the imine side-products was inhibited by increasing the number of carbon chains, resulting in high selectivity of the desired nitriles (Scheme 12c). To increase the stability and recyclability of the catalyst, an NHC-Ru-rGO heterogeneous catalyst was synthesized by immobilizing complex Ru-14 on graphene. In the end, the hybrid metal NHC-Ru-rGO was reused up to 10 times for the dehydrogenation of benzylamine 43.

3.2.6. Amine-Borane

In 2011, Liu and colleagues reported a single-component liquid-phase hydrogen storage system using boron–nitrogen heterocycles [84]. In this study, the cyclic amine–borane BN-methylcyclopentane 57 (4.7 wt%) was developed for the hydrogen storage system with the synthetic access of cyclic amine. The BN-methylcyclopentane 57 could release two hydrogen equivalents per molecule with the formation of the trimer 56 when thermally activated. Therefore, the catalytic dehydrogenation of the amine–borane was investigated using first-row transition metal halide. NiBr₂ and CuBr were the most active catalysts in the formation of the trimer 56 in 76% conversion at 80 °C in 5 min. The desired trimer 56 was also obtained with the release of six hydrogen equivalents in the presence of a FeCl₂ catalyst with 95% yield (Scheme 13). In addition, amine–borane trimer 56 could be converted back to the amine–borane heterocycle monomer in 92% yield through methanolysis using LiAlH₄.

3.3. Oxygen-Containing Compounds

Methanol is a useful C1 feedstock in industry and is produced in a cost-effective manner using natural gas, coal, and biomass [85,86,87,88]. Thus, it is considered an attractive LOHC due to its abundance and high hydrogen density (12.5 wt%) [89,90,91]. Methanol has been used for classical hydrogen production via an endothermic dehydrogenation (63.7 kJ/mol). However, a high concentration of CO was formed as a byproduct, and an extremely high temperature is required. Therefore, various catalytic systems have been developed for hydrogenation and dehydrogenation using silane/alcohol, ketone, or ester/alcohol pairs to produce efficiently produce hydrogen under mild conditions.

3.3.1. Silane /Alcohol Pair

In 2017, Mata and coworkers reported a silane/alcohol pair as a LOHC for the catalytic dehydrogenative coupling of hydrosilanes with alcohols [92]. The dehydrogenation of silane and alcohol is a spontaneous thermodynamic process, and the enthalpy of bonds and entropy are favored by the formation of strong Si-O bonds and the release of hydrogen gas, respectively. Therefore, the optimal conditions were found at low temperatures with the use of 1 mol% of the air-stable ruthenium complex, [Ru(p-cym)(Cl)₂(NHC)] (Ru-14) [44,80], under aerobic conditions. The dehydrogenative reactions of MeOH, EtOH, nPrOH, and nBuOH were fast, with completion in 30 s using 1 mol% of the catalyst (Scheme 14a). To broaden the substrate scope, various primary, secondary, and tertiary silanes were explored, and hydrogen was fully converted. The regeneration of hydrosilanes from silyl ethers could be achieved by a reduction using LiAlH₄ or other reducing agents because the dehydrogenative coupling of silanes and alcohols produces a single product, as opposed to other hydrogen carriers, which form a mixture or polymer. Finally, the Ru-14 complex was also immobilized on the surface of graphene to enhance the stability and recyclability of the catalyst and was reused for ten cycles without any decrease in activity.

In 2018, the Mata group also reported iridium-catalyzed high hydrogen production using silicon/alcohol pair [93]. The dehydrogenative coupling of silanes and alcohols was fast, and various silanes and alcohols were tolerated using the iridium complex, [IrCp*(Cl)₂(NHC)] Ir-5, as a catalyst. Among the substrate scopes, 1,4-disilabutane 61 and methanol are appropriate pairs for the LOHC system, which has a hydrogen storage capacity of 4.3 wt% (Scheme 14b). In addition, the proposed mechanism for the dehydrogenative coupling of silane/alcohol pair was shown in Scheme 15. In a methanol solution, the equilibrium between the Ir-5 complex and methanol coordinated cationic species 62 was formed. The Ir-η¹-H-SiR₃ 63 with the electrophilic character of silicon is attracted by methanol nucleophile to generate neutral iridium hydride 66 and the cation, R₃Si(H)OMe⁺. After the protonation of the iridium hydride 66 by the cation, R₃Si(H)OMe⁺, the Ir-dihydrogen complex 67 was formed, and finally released hydrogen gas.

3.3.2. Ketone

In 2020, Sundararaju et al. reported the iridium-catalyzed transfer-hydrogenation of ketones using methanol for application to the LOHC system [94]. Recently, reductions in carbonyl compounds into the corresponding alcohols have become one of the most important transformations; efficient catalytic systems have been developed without the use of a base under mild conditions. Using [Cp*IrCl₂]₂ catalysts, various ketones 69, with electron-donating–electron-withdrawing substituents, and halides on the aryl ring, delivered the transfer-hydrogenation product 70 in moderate to good yield (Figure 4). The method is an economical and environmentally friendly alternative to harsh reaction conditions and hazardous reducing agents.

3.3.3. Ester/Alcohol Pair

In 2019, Milstein and coworkers reported the ethylene glycol/ester pair as an efficient and reversible LOHC with a hydrogen storage capacity of 6.5 wt% [95]. The inexpensive and accessible ethylene glycol 40 might undergo dehydrogenative coupling to form liquid oligoesters 71, which could be hydrogenated back to ethylene glycol 40 (Scheme 16). To find the appropriate catalyst for the esterification, several ruthenium complexes containing phosphine–pyridyl-secondary amine (PNNH), phosphine–bipyridyl (PNN), phosphine–pyridyl–phosphine (PNP), and acridine-based PNP ruthenium complex were investigated. Among the complexes, the complex Ru-15 has the best catalytic ability, in the absence of a base, in toluene/DME co-solvents. The ethylene glycol 40 was recovered using 1 mol% of the same catalyst, Ru-15 in toluene/DME under 40 bar of H₂ with a 96% yield within 60 h. For dehydrogenation, 0.5 mol% of Ru-15 was used under reduced pressure (95 mbar) for 7 days, and 94% conversion was obtained.

In 2020, the Milstein group developed a new LOHC system of ethylene glycol/ester, exceeding 5.0 wt%, using the ruthenium pincer complex [96]. The dehydrogenative coupling of the ethylene glycol 40 using ethanol was performed in the presence of the acridine-based PNP ruthenium complex, Ru-15, with KO^tBu in toluene/DME at 150 °C for 72 h. In addition, the catalytic system using the stable dearomatized acridine PNP ruthenium catalyst Ru-15 worked well under base-free conditions, and the ester mixture was hydrogenated back to ethylene glycol 40 and ethanol (Scheme 17).

4. Reactor Type for LOHC System

For practical mass hydrogen production in industrial fields, efficient reactors should be developed in terms of scalability, cost-effectiveness, stability, and safety [97,98,99,100,101]. This section will introduce four types of reactor: heterogeneous catalytic reactor, membrane reactor, hot-pressure-swing reactor, and reactive distillation system.

4.1. Heterogeneous Catalytic Reactor

In 2018, Laurenczy and coworkers reported a heterogeneous catalytic reactor for continuous hydrogen production using formic acid [102]. In this study, the homogeneous ruthenium-meta-trisulfonated triphenylphosphine catalyst (Ru-mTPPTS) demonstrated high activity and stability. The heterogeneous catalytic system was developed based on the immobilized homogeneous Ru-mTPPTS catalyst. The catalytic support composed of cross-linked polystyrene beads was suitable in a continuous flow reactor [103,104,105]. The overall scheme of the experimental setup of the formic acid reformer unit is depicted in Figure 5a. The pure formic acid was continuously injected by an HPLC pump, and the reactor termperature was maintained using the insulator, oil circulation bath, heater, and thermocouple. The gas mixture was monitored by in-line IR analyzers and a gas chromatograph (GC). The catalytic activity of the lab-scale reactor for the formic acid dehydrogenation was measured at 270 h⁻¹ (TOF, at 105 °C), and the reactor was stable for 120 h under the operating conditions. The feasibility of the developed reactor was demonstrated to be combined with commercial fuel cells.

In 2020, Jung et al. reported a heterogeneous ruthenium-catalyzed hydrogenation of CO₂ to formic acid using a fixed-bed reactor in continuous flow process (Figure 5b) [106]. The heterogeneous catalytic system is very attractive due to the ease of catalyst separation. Thus, the immobilization of a homogeneous catalyst on catalytic support can combine the advantages of homogeneous and heterogeneous systems. The rationally designed catalyst support can improve porosity, and the presence of various heteroatom ligand sites enables the diversification of heterogeneous catalytic systems. Based on previous studies, the Ru/bpyTN-30-CTF was synthesized and employed in the continuous CO₂ hydrogenation to formic acid. A structural representation of Ru/bpyTN-30-CTF synthesis and a schematic of structured CO₂ hydrogenation in a trickle-bed reactor is depicted in Figure 5b. Continuous hydrogenation of CO₂ was performed in a homemade stainless-steel tubular reactor. The ruthenium catalyst powder was charged from the middle of the vertically positioned reactor and covered by a glass wool. To prevent the loss of the catalyst, the remaining space was filled with glass beads. The reactor was pressurized to the desired pressure and heated to the desired temperature. The flow rate of H₂ gas was controlled using a mass flow controller. Finally, the liquid product was collected and characterized using HPLC and NMR, and the gaseous product was analyzed using GC. Finally, the heterogeneous ruthenium-catalyzed hydrogenation, in a continuous process using the trickle-bed type fixed-bed reactor, showed a significant catalytic performance and superior long-term stability without noticeable deactivation (a total TON of 524,000 for 30 days).

4.2. Membrane Reactor

In 2015, Tsuru and colleagues reported the dehydrogenation of methylcyclohexane 75 to toluene 76 for hydrogen production, using a bimodal catalytic membrane reactor [107]. A catalytic membrane reactor is a combination of permselectivity membranes and heterogeneous catalysts. The membrane removes one of its components, shifting the equilibrium toward the product, resulting in higher reaction conversions even at lower temperatures. These types of membrane must provide sufficient permeability, permselectivity, and long-term stability. For the dehydrogenation of methylcyclohexane 75, a bimodal catalytic membrane reactor consisting of a Pt/γ-Al₂O₃/α-Al₂O₃ catalytic support and a bis(triethoxysilyl), ethane (BTESE)-derived silica layer was prepared (Figure 5c). Methylcyclohexane 75 was fed inside the membrane by a syringe pump and the hydrogen production proceeded. H₂ gas with a purity greater than 99.8% was obtained in the permeate stream, since the BTESE-derived membrane was almost impermeable to methylcyclohexane 75 and toluene 76. The H₂ recovery ratio reached 99%, which improved the efficiency of the H₂ production system. In addition, a system combining a fixed-bed pre-reactor and a catalytic membrane reactor was proposed for methylcyclohexane 75 dehydrogenation. Thus, the system enhanced H₂ extraction to improve the conversion.

In 2018, Pfeifer and coworkers reported a micro-structured multi-stage membrane reactor concept using hydrogen intermediate separation for LOHC-dehydrogenation [108]. The use of a micro-structured reactor allows for an almost isothermal reaction environment, preventing cold spots by limited heat input. Pd-based membranes are useful in a micro-structured environment to separate hydrogen from a gas mixture. The schematic of the micro-structured, multi-stage approach is shown in Figure 5d. The gas and liquid-phase can be separated at the reactor outlet when the gas phase enters the membrane separation module. The remaining liquid product stream is directly fed to the next stage of the reactor. Membrane module PdAg is more suitable for LOHC applications than pure palladium due to its low operating temperature and high hydrogen permeation flux.

4.3. Hot Pressure-Swing Reactor

Wasserscheid et al. (2017) reported that the hydrogenation of dibenzyltoluene (DBT) 12 and dehydrogenation of perhydro-dibenzyltoluene (H18-DBT) 13 could be conducted using a hot pressure-swing reactor [109]. Conventional approaches to LOHC hydrogen storage systems apply separate hydrogenation and dehydrogenation reactors because hydrogenation reactors can operate at low temperatures and pressures, whereas reactors for the dehydrogenation require high temperatures and low pressures to maximize thermodynamic driving forces. The novel hydrogen storage concept conducts hydrogenation and dehydrogenation reactions at the same temperature and catalyst (pressure is the only variable for the shift of reaction equilibrium). According to Le Chatelier’s principle, high pressure due to hydrogen production inhibits dehydrogenation and favors hydrogenation. Thus, the hydrogenation and dehydrogenation reaction is dependent on the pressure of equilibrium conversion. A hot pressure-swing hydrogen storage unit that operates at 290 °C can operate at ambient pressure and a pressure of 4 bar for almost complete dehydrogenation. This system simplifies the hydrogen storage system into a single reactor, maintaining an elevated reaction temperature without a cooling system, and reduces investment and operating costs.

4.4. Reactive Distillation System

In 2020, Wasserscheid and coworkers reported the dehydrogenation of perhydro-benzyltoluene using a reactive distillation system [110]. This strategy allows for the dehydrogenation of pure hydrocarbon LOHC compounds at temperatures below 200 °C. The reactive distillation reactor enables simultaneous Hx-LOHC dehydrogenation and separation of Hx-LOHC from H0-LOHC. The reactive distillation setup consisted of four zones: an evaporator, a separation zone, a reaction zone, and a condenser (Figure 5e). Perhydro-benzyltoluene has a low boiling point and is abundant in the upper part of the separation zone, whereas benzyltoluene is located in the lower separation zone. Therefore, hydrogen-charged LOHC compounds are at the top of the column with the catalyst, whereas hydrogen-lean LOHCs are at the bottom without catalyst. Therefore, the reaction zone was placed above the distillation zone to take advantage of the higher concentration of perhydro-benzyl toluene. Boiling perhydrobenzyl toluene can increase the thermodynamic driving force for dehydrogenation reactions by reducing the partial pressure of hydrogen in the gas phase. In addition, reactive distillation can increase the efficiency and rate of the reaction by separating hydrogen-rich and hydrogen-lean compounds into different zones.

5. Conclusions

Our review concentrated on recent advancements in hydrogenation and dehydrogenation in liquid organic hydrogen carrier (LOHC) systems. Since hydrogen has gained popularity as a sustainable energy carrier, various hydrogen storage systems have been developed over the years. LOHCs are a particularly promising candidate for storing and transporting renewable energy from energy-rich to energy-scarce regions using existing infrastructure such as tanks and pipes. To find promising LOHCs, several criteria should be considered, including physical properties (melting point, boiling point, thermal stability, and viscosity), reversibility, hydrogen storage capacity, toxicity, and eco-toxicity. In this regard, this review discussed recently developed efficient hydrogenation and dehydrogenation strategies using homogenous and heterogeneous catalysts for potential LOHCs, including a circular hydrogen carrier (formic acid/formaldehyde and ammonia), homocyclic compounds, N-containing compounds (N-heterocyclic compounds, Amide, imide, Urea/Carbamate, Nitrile, and Amine-borane) and O-containing compounds (Silane/Alcohol pair, Ketone and Ester/Alcohol pair). Furthermore, four reactors (heterogeneous catalytic, membrane, hot pressure-swing, and reactive distillation systems) were introduced for practical industrial application.