Refined, white beeswax pearls and retinyl palmitate (vitamin A) of 1.7 MIU g^–1 (MIU = milli-international units) were purchased from Bulk Apothecary (Aurora, OH, USA) and Fisher Scientific USA (Pittsburg, PA, USA), respectively. Ethanol was obtained from Decon Laboratories, Inc (King of Prussia, PA, USA). Compression fabric (warp knit: 77% nylon and 23% spandex) was obtained from the Marena Group (Lawrenceville, GA, USA). Pampers^® Aqua Pure^TM nonwoven wipes were also used as a carrier to transfer microparticles from the substrate to skin.

We microencapsulated retinyl palmitate by melt dispersion technique and investigated the effect of four process variables on the produced microcapsules, such as different theoretical loading capacity (10%, 15%, 25%), types of wax (beeswax, carnauba wax, paraffin wax), emulsifier concentrations (0%, 1%, 2%) and stirring speeds (180 rpm, 230 rpm, 280 rpm) in our previous study (12). The statistical analysis showed that theoretical loading capacity and surfactant (%) were the most significant factors and we were able to determine that the highest theoretical loading (25%) and highest surfactant (2%) selected in that study can provide us high actual loading with the small size of the particles. There was no significant difference found among the effects of type of wax on loading capacity, encapsulation efficiency, antioxidant activity or mean size of particles. Hence we decided to conduct further study selecting beeswax as the shell material because of its natural skincare benefits as well as operational convenience due to low melting point (65°C). We selected 280 rpm stirring speed to facilitate dispersion of the oil-in-water emulsion and formation of small size particles.

Thermal analysis of the beeswax, retinyl palmitate and retinyl palmitate-loaded beeswax microcapsules was carried out by using Mettler-Toledo GmbH DSC821e (Greifensee, Switzerland) instrument, where a standard empty aluminium pan was used as the reference. The weight of the samples was within 2–9 mg, and the samples were scanned from 25°C to 100°C under nitrogen atmosphere with a heating rate of 10°C min^–1.

After preparing the microcapsules with 25% theoretical loading, we looked into the shelf life of microcapsules by measuring the actual loading percentage, i.e. the content of retinyl palmitate in a fixed amount of capsules over a period of time, both in powder and dispersion forms. We evaluated the shelf life of the beeswax microcapsules (approximately 71% encapsulation efficiency) in powder form, where they were filtered and dried before storing in an enclosed petri dish under room temperature; and also in dispersion form (approximately 75% encapsulation efficiency), where the particles were kept dispersed within the emulsion during preparation, stored inside dark vials in refrigerator and a portion was filtered on each day of measurement (Day 1, Day 4, Day 8, Day 15 and Day 31).

An extraction from 0.1 g of microcapsules was performed, by heating the capsule in 20 ml of ethanol solution to release the vitamin content and then filtering the wax residue. The concentration of supernatant aliquots was measured at 327 nm by a Shimadzu Corporation UV-2401PC spectrophotometer (Kyoto, Japan). The amount of retinyl palmitate was determined from a standard curve of known concentrations.

We conducted an in vitro kinetic release study similar to prior literature (27, 43) with some modification based on particle content, solvent type and machine parameters. The retinyl palmitate release profile from 3 g of suspended particles (approximately 77% encapsulation efficiency) was examined in 600 ml of pure ethanol. The study was performed in a New Brunswick Scientific C24 (Eppendorf, Germany) incubator shaker with a speed of 100 rpm and temperature set at 37±2°C. Supernatant aliquots of 2 ml were withdrawn and replaced by the fresh medium at appropriate time intervals (1 min, 5 min, 15 min, 30 min, 1 h, 2 h, 4 h). The supernatants containing dissolved retinyl palmitate were diluted and analysed by ultraviolet-visible (UV–vis) spectroscopy at 327 nm. The results were compared with a standard to calculate the vitamin A concentration and to evaluate the release ratio.

We used a robotic transfer replicator (Figure 1) to simulate the transfer of microparticles from a nonwoven wipe to the skin and evaluate the transfer percentage, by means of a similar method as described by Yu et al. (44). 1 g of microparticles was spread as evenly as possible by a spatula over a commercial nonwoven wipe containing 99% water that acted as a donor surface with a diameter of 133 mm. The receptor material was a compression fabric, i.e. a warp knit with a composition of 77% nylon/23% spandex (fabric weight 276 g cm^–2). This fabric was chosen because the study by Yu et al. (44) regarding transfer of particulates from carpet surface to human skin-like receptors revealed that this fabric replicated the human skin, particularly finger pads best as a receptor material. The receptor fabric was attached to an aluminium nose piece with the help of O-ring made of rubber. After the activation of the replicator, the nose piece descended the receptor fabric onto the donor surface and rubbed the receptor against the donor by executing a certain number of motions (imitating an hourglass pattern) under a constant pressure maintained by the programmed hydraulic system. After performing this rub cycle, the nose piece was raised and delivered onto a glass jar containing 20 ml of ethanol. The fabric was released into ethanol and shaken vigorously, followed by sonication for 2 h so that all the particle content is released into ethanol. Then aliquots were removed for assay in an UV–vis spectrophotometer to measure the content of retinyl palmitate. Finally, the amount of transfer of retinyl palmitate was calculated in percentage.

Fig. 1.

All the measurements for shelf life study were performed in triplicates, whereas the measurements of kinetic release study and simulated transfer study were performed in duplicates. The results have been reported as the mean values and their corresponding standard deviations.

Figure 2 shows differential scanning calorimetry (DSC) scans of beeswax, retinyl palmitate and beeswax microcapsules with 25% theoretical loading capacity. In the thermogram of retinyl palmitate, a sharp endothermic peak is observed at 34.33°C, which corresponds to its melting point. However, it is observed that the microcapsules show no endotherms corresponding to the melting point of retinyl palmitate. This implies that retinyl palmitate dissolved within the matrix of beeswax when the temperature reached its melting point (45). This observation was consistent with the result found by Milanovic et al. (46), where encapsulated ethyl vanillin dissolved in the carnauba wax matrix. Untreated beeswax and retinyl palmitate-loaded beeswax microcapsules show their melting peaks at 65.67°C and 63°C, respectively. The slight decrease in the melting point of the particle should be due to the mixing of retinyl palmitate and wax because of plasticisation. A second peak is observed for microcapsules at higher temperature (slightly higher than melting temperature), which could be because of fraction of large crystallites formed after encapsulation process that showed higher melting.

Fig. 2.

Figure 3 shows the shelf life study of the beeswax microcapsules in: (a) powder form stored under room temperature; (b) dispersion form stored in a refrigerator. When the particles were evaluated in powder form under room temperature, the microcapsules lost their active content within 8 days (Figure 3(a)). This phenomenon can be attributed to the diffusion of retinyl palmitate through the wax shell. The high compatibility between lipophilic, low molecular weight active ingredients with wax is the major cause of diffusion (47). Diffusion can be accelerated in small-sized particles due to the availability of larger contact areas as well as due to pores existing in the shell matrix (48).

Fig. 3.

Djordjević et al. (31) described the internal structure of particles produced by melt dispersion with the wax shell to be nonhomogeneous with matrix or hollow-shell morphology. Therefore in the prepared microcapsules, retinyl palmitate is distributed within the wax shell matrix. With the course of time, the core content comes up to the surface and diffuse through the shell. From Figure 4(a), the gradual change in the colour of beeswax microcapsules supports the phenomenon of diffusion as a plausible explanation. The particles stored as powder form appear to be bright yellow after the retinyl palmitate diffuses to the surface and they turn white (beeswax) when almost all of the core content leaches out.

Fig. 4.

On the other hand, when the retinyl palmitate-beeswax particles were stored in the dispersed aqueous emulsion in a refrigerator, they retained the core material and showed no significant decrease in retinyl palmitate content until the 15th day (Figure 3(b)). The variability in size distribution of different batches of filtered particles may account for the slight increase observed in actual loading capacity (Figure 3(b)). After 30 days, a decrease in loading was observed, which can be explained by ester hydrolysis of the beeswax while stored in aqueous emulsion resulting in the release of the content (49). retinyl palmitate-beeswax particles stored in the dispersed aqueous emulsion in the refrigerator do not show a significant visual difference in colour when filtered (Figure 4(b)).

The release profile (Figure 5) of retinyl palmitate-beeswax microcapsule showed an initial burst followed by a slower release of the vitamin entrapped inside the beeswax matrix. Due to the initial burst effect, 7% of the retinyl palmitate released at the first minute, leading to around 55% release in the first half an hour. After the initial rapid release, the release profile showed a sustained release over time. Within 4 h, approximately 98% of retinyl palmitate was released. A similar pattern of release was found by Kheradmandnia et al. (49) from ketoprofen-loaded solid lipid nanoparticles incorporated in the matrix of beeswax-carnauba wax mixture. Zigoneanu et al. (50) described the phenomenon of such initial burst as the result of the cumulative effect of diffusion of the core through the matrix, penetration of dissolution medium into the particle, and degradation of the shell matrix. As retinyl palmitate is soluble in ethanol, this explanation is agreeable to our result. Permeation of ethanol through the pores of the shell matrix and simultaneous diffusion of retinyl palmitate through the matrix facilitated the fast dissolution of the vitamin into ethanol. Duclairoir et al. has reported similar release profile for α-tocopherol from wheat gliadin nanoparticles, where mathematical models were demonstrated for the bistep release, i.e. the burst effect and the slower diffusion process (51). While the initial burst could not be described by their model, the time-dependent slow release showed a good fit (R² = 0.90) for the model in Equation (i):

(i)

Fig. 5.

Here, M₀ is the amount of active content incorporated, M_t is the amount of release core at time t, D is the diffusion coefficient and R is the radius of the particle. Thus the sustained release was related to the diffusivity of the active core inside the matrix system, the surface area of the particle and the loaded content.

From this result, we can understand that alcohol-based cosmetic formulations will not be stable over time as the core content would be released in the carrier substrate during the storage period, making retinyl palmitate susceptible to oxidation and degradation. On the contrary, as we already observed in the shelf life study, an aqueous medium prevents the active content from releasing from the capsule because of having no affinity to the lipophilic content. As a result, a water-based formulation would be suitable to contain the particles for cosmetic applications.

From the transfer study, we found that 21.7±0.02% of retinyl palmitate was transferred to the receptor material from the donor surface of wet nonwoven wipe after the preprogrammed rubbing cycle. The percentage falls within the range reported by Yu et al. in their study of transfer of particulates from carpet surfaces to human skin. Although this amount may vary depending on encapsulation efficiency, method of particle incorporation, and the amount of particle incorporated, this study demonstrates the potential of using such microparticles into facial wipes to impart skincare properties. Knaggs, in his skin-ageing handbook, mentioned that 0.05–0.1% tretinoin (retinoic acid) was effective to reduce signs of ageing in Asians (52). Oliveira et al. demonstrated in their study that topical application 1% retinyl palmitate has promising results for the treatment of skin ageing (53). According to Gangurde et al., the recommended concentration for topical semisolid formulation of vitamin A palmitate is 0.05%–0.3% (27). Thus, considering the approved dosage of retinoids, absorption and conversion rate of retinyl palmitate to retinoic acid within the skin, a proper formulation has to be developed in further study.

Although benzene, which is the hydrogen-lean form of the benzene/cyclohexane LOHC system (Figure 1), is a known carcinogen (the most significant drawback of this specific system), several studies have been conducted into the properties of effective hydrogenation/dehydrogenation catalysts for this model feedstock.

Fig. 1.

Hydrogenation is generally accepted to occur via associative adsorption of benzene onto the catalytic surface. During the hydrogenation step, it is believed the carrier molecules are arranged as ï-complexes and are hydrogenated in a stepwise fashion, through a series of cycloolefins (Figure 2) (44). This is supported with the detection of cyclohexene (44).

Fig. 2.

Early research concluded that monometallic iron-catalysts with body-centred structures were essentially inactive for the hydrogenation of benzene, while the activity of a mixed iron‐cobalt catalyst decreased with increasing iron content (45). Since cobalt and nickel, both with face-centred cubic structures, were found to be active for the reaction, it was assumed the body-centred structure of iron was responsible for its inactivity (45). However, a monometallic, face-centred copper catalyst was deemed to be completely inactive unless in the presence of a suitable promoter, such as nickel (45). The inactivity of the copper catalyst has since been attributed to difficulties in obtaining a suitably high dispersion on the catalyst support, while the mechanism of the promotional effect observed with the addition of nickel remains a topic for debate (35). Several hypotheses for this have arisen, including a changing in copper crystal structure, or an alteration in surface concentration as a consequence of nickel collection (45). Interestingly, even in the very early studies, it was clear that palladium has a high activity for the reaction, but only on the condition that the surface area was sufficiently high, while the conversion is also reported to be catalysed by rhodium(I) complexes (45, 46).

Nickel-based catalysts are frequently considered for ring-hydrogenation reactions, with several detailed studies revealing key information for future catalyst development. For example, the hydrogenation of benzene using nickel catalysts has been suggested to be a structure insensitive reaction, while an unexpectedly low activity at low nickel loadings (less than 5 wt%) is the result of a nickel aluminate spinel formation, when an alumina support is used (47). More commonly however, this spinel structure is formed at high reaction temperatures. Although expected, these findings are important, signifying not only the choice of metal, but also the properties of the support (such as acidity, interaction with the metal) and reaction conditions play a role in determining catalytic performance. The move from pgm-based catalysts to base metal catalysts could also lower the cost associated with the LOHC technology as well as reliance on highly fluctuating costs of pgms (48, 49). This is provided that comparable or even improved catalytic performance can be achieved.

Nickel is also widely referenced for its activity in catalysing the dehydrogenation reaction of cyclohexane. Given the low boiling point of cyclohexane (81°C), this reaction is often performed in the gaseous phase. The studies have shown that when using a cyclohexane feedstock, a bimetallic nickel/platinum combination (20 wt% nickel and 0.5 wt% platinum) can improve catalytic dehydrogenation activity 60-fold when compared to using a monometallic platinum catalyst (0.5 wt% platinum on activated carbon cloth), or 1.5-fold when compared to a monometallic nickel catalyst (20 wt% nickel on activated carbon cloth) (34). In addition, selectivity in the dehydrogenation reaction was improved using various monometallic nickel catalysts with the addition of platinum (50). For these cyclohexane dehydrogenation studies, a spray-pulsed reactor was used at a temperature of 300°C (34).

The concept of improved catalytic performance through the use of bimetallic combinations has also been reported for silver-based catalysts, employed for the benzene hydrogenation reaction (34). Here, the same spray-pulse reactor and reaction temperature (300°C) as discussed for the platinum catalyst above was used. For the most significant catalytic enhancement, it was concluded that the secondary promoter must be classified as a pgm (34). As an example, the hydrogen production rate was doubled when using a platinum-promoted silver catalyst on activated carbon cloth (1 wt% platinum + 10 wt% silver) compared to the non‐promoted 10 wt% silver equivalent (34). This was suggested to be the result of a synergistic effect between the two metals for breaking C–H bonds, high hydrogen reverse-spillover, or hydrogen recombination abilities of the catalyst (51). The improvement in catalytic activity of the silver catalysts promoted with 1 wt% of a pgm were found to follow the trend: platinum > rhodium > palladium, although catalyst stability improved in the order platinum > palladium > rhodium (51). The study did not comment on expected lifetime of the catalysts, highlighting the need for further research regarding key performance indicators (activity, selectivity, lifetime).

To catalyse the hydrogenation of toluene to methylcyclohexane, palladium-based catalysts have been reported to be effective. For example, in 1 h a 10 wt% palladium/carbon catalyst can achieve a 90% toluene conversion at 80°C and 15 bar hydrogen pressure, when the hydrogenation reaction is performed in a batch reactor (52). Interestingly, when this palladium on carbon catalyst was coated with a liquid coordination complex (LCC), the conversion of toluene increased to 99.9%, under otherwise identical reaction conditions (52). LCCs are described as ‘ionic liquid-like’ Lewis acid species, with their equilibrium composition containing cations, anions and neutral components (52). In this experiment, the LCC was synthesised from aluminium chloride and urea (52). It was concluded that only a thin film of the liquid (ionic liquid coating at 13 wt%) was required to improve the catalytic performance; when larger quantities were applied to the catalyst a decrease in activity was observed as a result of pore blocking and mass diffusion limitations (52).

As for benzene hydrogenation, platinum-based catalysts have also been found to be successful in catalysing the hydrogenation of toluene to methylcyclohexane. For example, a 0.3 wt% platinum catalyst supported on zeolite CBV‐780 (silicon:aluminium ratio of 40) is capable of achieving full hydrogenation at just 120°C, with a hydrogen pressure of 30 bar (53). The activity of the catalyst notably decreased when alloyed with palladium. A trimetallic combination of nickel, cobalt and molybdenum on a zeolite support was also found to effectively catalyse this hydrogenation reaction (54). In a batch reactor, with a reaction temperature of 200°C and a hydrogen pressure of 20 bar, it was found that a HY support (silicon:aluminium ratio of 5.1) allowed for superior catalytic activity after 30 min on-stream, when compared to using mordenite, HY (silicon:aluminium ratio of 80) or ZSM-5 alternatives. In explanation, the authors propose the larger pore volume and pore diameter of the HY support (silicon:aluminium ratio of 5.1) reduces pore diffusion limitations (54).

Moreover, monometallic nickel-based catalysts are effective for toluene hydrogenation (55). For example, one study highlighted the increase in catalytic activity observed for a 20 wt% nickel catalyst, compared to a 5 wt% nickel catalyst, both supported on gamma-alumina (55). Since it is generally accepted that an increase in the number of metallic active sites increases catalytic activity, this result is unsurprising (55). However, it was also found that employing supports of different alumina phases (γ-Al₂O₃ or κ-Al₂O₃) has different effects on catalytic activity when different nickel loadings are used. For instance, a significantly higher catalytic activity was reported for a 5 wt% Ni/κ-Al₂O₃ catalyst than for a 5 wt% Ni/γ-Al₂O₃ (toluene conversions of 98% and 28%, respectively), under the same reaction conditions (55). Yet at higher nickel loadings an opposite trend was observed: a 20 wt% Ni/κ-Al₂O₃ catalyst resulted in a lower toluene conversion than that of a 20 wt% Ni/γ-Al₂O₃ catalyst, although the difference in activities between the catalysts is smaller at this higher nickel loading (55). In explanation, it was concluded that changes in reducibility of the surface nickel sites has a larger effect on the observed activity than simply the number of active sites present (55). These reactions were performed in a plug-flow reactor, using 0.1 g of catalyst (55).

Nickel on kaolinite catalysts have also been considered for the hydrogenation of toluene (56). Using a flow-type system at ambient pressures, it was found that modification of the catalyst with a small amount (2 wt%) of either potassium or zinc (using KNO₃ or Zn(NO₃)₂) increased catalytic activity, which was attributed to a decrease in interaction strength between the nickel and kaolinite. Yet, a larger modification (3–7 wt%) resulted in a significant decrease in activity (56). This was proposed to be a result of potassium covering the nickel active sites, or an unfavourable zinc-nickel interaction (56).

The dehydrogenation reaction (methylcyclohexane to toluene, Figure 3) typically employs a platinum or nickel-based catalyst, supported on alumina (1). Reaction temperatures for such catalysts often vary between 350°C and 450°C (with a general increase in temperature increasing the conversion of methylcyclohexane to toluene before equilibrium is reached) (57, 58). Another study highlighted the influence of process conditions on increasing conversion of methylcyclohexane and lowering operating temperatures by using efficient palladium or palladium-alloy membranes in catalytic membrane reactors. Such reactors combine the catalytic dehydrogenation reaction with the extraction of hydrogen in a single unit (58).

Fig. 3.

Interestingly, a potassium-platinum on alumina catalyst was reported to achieve a hydrogen yield of 95%, when employed at 320°C in a fixed-bed reactor, which is slightly above the typical yield range given above (1). Moreover, an excellent selectivity (>99.9%) was reported for this reaction (1). A Raney-nickel catalyst has also been reported to achieve a 65% yield after 30 min on-stream, under multiphase reaction conditions (59). Although this reaction was performed at the lower reaction temperature of 250°C, the Raney-nickel catalyst is not a suitable candidate to catalyse the reactions of the LOHC technology (1). This is in view of the accompanying isomerisation and disproportionation reactions which would ultimately necessitate the need for more frequent LOHC replacement (1). It should be noted that from a thermodynamic viewpoint, the dehydrogenation of methylcyclohexane, like all LOHC candidates, should be performed at the lowest possible pressure as this facilitates the use of the lowest possible temperature for full conversion to toluene (34). However, in practice, it has been found that such conditions may not be compatible with the catalyst, potentially resulting in side reactions and catalyst deactivation (34).

As has already been stated, the catalyst support materials can be modified to increase catalytic performance (34). For the toluene/methylcyclohexane system, it was concluded that the hybrid composite support alumina-titania was responsible for a huge increase in catalytic dehydrogenation activity, when compared to alumina (99% and 16.5% methylcyclohexane conversion, respectively, where nickel is the active site) (34). In addition, properties of the perovskite La_0.7Y_0.3NiO₃, and metal oxides including La₂O₃, CeO₂ and MnO₂ have been evaluated in terms of their suitability as catalyst supports for the dehydrogenation of methylcyclohexane (34, 60). Most notably, it was found the hydrogen production rate more than doubled when comparing the Pt/La₂O₃ and Pt/La_0.7Y_0.3NiO₃ catalysts (21.1 mmol g_met^–1 min^–1 and 45 mmol g_met^–1 min^–1, respectively) (60). These reactions were conducted in a spray-pulsed reactor.

This particular LOHC system has also been studied alongside a palladium membrane reactor, which allows for purification of the hydrogen released from the LOHC (34). Contaminants of the hydrogen can include carbon monoxide, CO₂, methane and cyclic hydrocarbons and have been reported to be present in quantities between 100 ppm and 1000 ppm, depending upon the operating conditions of the LOHC system (61). The exact origins of the impurities remain unknown, but it is believed a combination of atmospheric oxygen and residual moisture in the LOHC is responsible for the formation of carbon monoxide and CO₂, while partial decomposition of the LOHC at high reaction temperatures, or contaminants from the production of the LOHC, can explain the hydrocarbon presence (61). The quality of the hydrogen released depends upon the intended application: for example, several fuel cells have specific regulations on compatible hydrogen purity (61). Despite this, only a few examples of hydrogen purification processes exist in the literature, most of which employ palladium membranes. This is not surprising, considering palladium has a high hydrogen solubility, permitting effective separation of hydrogen and contaminants at high temperatures (typically above 300°C) (61).

A palladium membrane reactor has been reported to have a dual functionality, acting both as a dehydrogenation reactor and a hydrogen purification system (34). The dehydrogenation of methylcyclohexane using such a reactor and a 1 wt% platinum on alumina catalyst, permitted the use of a 20°C decrease in temperature than would otherwise be required to achieve the same conversion (70%) without employing the membrane (225°C and 245°C, respectively) (34). In addition, the platinum catalyst demonstrated good stability and selectivity within the temperature range 150–325°C, with no significant deactivation for approximately 600 h time on-stream (62). A 5 μm palladium-silver membrane, coupled with a microstructured system, has also been reported to effectively purify the hydrogen released (61).

Boasting a high theoretical hydrogen storage capacity of up to 7.4 wt%, a naphthalene/decalin LOHC system has been considered (Figure 4). However, in practice, dilution with a solvent such as toluene is necessary to keep the LOHC cycle within the liquid phase (melting point of naphthalene: 80°C). This lowers the overall hydrogen capacity to 3.8 wt% (1). For this reason, the naphthalene/decalin system is less widely studied than other possible carriers, although important learnings have been gained, which can be transferred to other LOHC systems. It is worth noting that decalin exists as structural isomers, and thus an approximately equimolar mixture of the cis- and trans- isomers are produced upon hydrogenation of naphthalene, despite the thermodynamic favourability of the trans- isomer (1). The exact cis :trans ratio is naturally dependent on reaction conditions, in addition to the catalyst itself (1).

Fig. 4.

Much like other LOHC systems, platinum-based catalysts have also been the focus of most studies on the hydrogenation of naphthalene (1, 63). Yet, in contrast to other systems discussed, an aluminium Mobil Composition of Matter No. 41 (Al-MCM-41) support was investigated (63). Harsher reaction conditions (temperatures of 300°C and pressures of 69 bar) than employed for alternative systems were required to achieve full hydrogenation to decalin in 150 min, in a batch reactor (1). It was found that the reaction temperature could be lowered to 200°C under the same reaction pressure, but this imposed the requirement of a drastically increased reaction time (480 min) (1). Generally, conclusions from this study are similar to those of other systems: hydrogenation activity and selectivity are both strongly dependent on support properties (63).

Although the hydrogenation process presents its relative challenges, it is the dehydrogenation step which confirms the incompatibility of this system with the LOHC technology. For instance, this step is likely to produce intermediates including tetralin (Figure 5). Although tetralin itself can be dehydrogenated and become part of the LOHC process, it is unlikely that no tetralin molecules will remain following the dehydrogenation reaction. As tetralin is still partially hydrogenated, the hydrogen storage capacity of the LOHC is lowered in subsequent cycles. Thus, regular replacement of the carrier material would be required to ensure a constant hydrogen storage capacity (34). This would increase the operating expenditure of the technology (more LOHC required) in the case of naphthalene. In light of these challenges, a pilot scale demonstration is yet to be achieved (1).

Fig. 5.

Research into the dehydrogenation reaction of naphthalene has, however, produced some interesting results. Typically, when employing a platinum catalyst supported on carbon at conditions slightly milder than those used for the hydrogenation reaction (280°C for 150 min at atmospheric pressure), almost full conversion of decalin to naphthalene can be achieved in a batch reactor system. Furthermore, adding rhenium to the catalyst (platinum-rhenium on carbon) can decrease the reaction time from 150 min to 120 min under otherwise identical conditions (1, 64).

A 3 wt% platinum on carbon catalyst was also used to investigate the effect of varying the catalyst preparation method on catalytic performance for the dehydrogenation of decalin (34). It was found that advanced methods of preparation, including ion-exchange and polyol-assisted synthesis (in which ethylene glycol was used as a solvent and reducing agent), resulted in a greater dispersion of platinum (19.6% and 14%, respectively) than the more conventional precipitation and impregnation methods (10% and 5.4% platinum dispersion, respectively) (34, 65). Increasing the platinum dispersion increases the surface area of the metal over which the reaction can occur, consequently resulting in a greater rate of hydrogen release from decalin (34, 65). Again, optimisation of the support must also be considered. Several carbon-based possibilities including nanofibers, carbon black, carbon xerogel and ordered mesoporous carbon have been evaluated, with the high surface area of ordered mesoporous carbon believed to be responsible for demonstrating the highest activity for the dehydrogenation of decalin (34). Yet, over longer operation times catalyst deactivation was observed as a result of pore blockages with the bulky LOHC feedstock (34). Thus, it can be argued that as carbon black has wider pores than the ordered mesoporous carbon, this would be the most suitable support, despite a lower activity. These experiments were performed at 260°C, in a batch reactor (34).

Catalytic activity, in addition to selectivity, has also been reported to be improved with the addition of tin to platinum on activated carbon (66). The effect is two-fold: (a) the electronic modification of platinum by tin prevents the cleavage of C–C bonds on the catalyst surface, facilitating the adsorption and desorption of reactant and product molecules; (b) the addition of tin as a catalytic promoter improves catalyst stability by preventing the sintering and agglomeration of platinum at high temperatures (34).

The most significant advantage of the NEC/perhydro-NEC LOHC system (Figure 6) is the relatively low energy required for dehydrogenation (53.2 kJ mol^–1 hydrogen, Table I) (1). This lower reaction enthalpy facilitates a lower temperature for the dehydrogenation reaction and consequently improves the overall efficiency of the system. As such, interest into the potential applications of NEC has significantly grown in recent years, culminating in its use as a feedstock for the LOHC technology. However, like other LOHC candidates discussed thus far, NEC also possesses unfavourable properties for its employment within LOHC technology. Arguably the most significant of these is the solid nature of NEC at ambient temperatures. To ensure the key liquid property of the LOHC technology is not lost, employing NEC would thus require NEC dilution (1). This reduces the efficiency of the LOHC technology (1). Moreover, NEC is significantly more expensive than alternative LOHCs, such as toluene (€40.00 kg^–1 and €0.30 kg^–1, respectively) (1).

Fig. 6.

The dehydrogenation reaction is again most commonly performed with pgm-based catalysts, particularly palladium on alumina or platinum on alumina at metal loadings of around 5 wt%. These catalysts have been reported to have the highest catalytic activities, when compared to ruthenium and rhodium equivalents (67). Specifically, catalyst dehydrogenation activity follows the trend palladium > platinum > ruthenium > rhodium under atmospheric pressure and at 180°C (67). Full dehydrogenation to NEC was observed for palladium and platinum catalysts, after reaction times of 240 min and 300 min, respectively, while employment of a ruthenium catalyst over the same time frame produced a mixture of both the fully dehydrogenated NEC (71.28%) and the partially hydrogenated species, 4H-NEC (28.54%) (67). Using the catalyst with the lowest dehydrogenation activity (rhodium on alumina), even less of the fully dehydrogenated NEC is obtained (10.64%) (67). Important to note, all reactions are carried out below 270°C. Beyond this temperature, NEC becomes susceptible to dealkylation reactions, producing the byproduct carbazole (1, 68).

Like the dehydrogenation reaction, the hydrogenation of NEC is frequently catalysed by a pgm. However, given the comparative ease of the hydrogenation reaction (in comparison to the dehydrogenation reaction), a palladium or ruthenium-based catalyst is more frequently employed than for the dehydrogenation reaction. For example, full hydrogen loading onto NEC has been achieved in 3 h at 150°C and 50 bar hydrogen pressure using a 5 wt% ruthenium on alumina catalyst in a batch reactor (68). The molar ratio of LOHC to ruthenium used in these experiments was 400:1 (68). The catalyst Pd₂Ru on silicon carbonitride has also been reported to successfully catalyse the hydrogenation of NEC (69). In comparison to the 5 wt% ruthenium on alumina catalyst discussed previously, a milder reaction temperature (110°C) and hydrogen pressure (20 bar) can be employed (69). However, a significantly longer reaction time (36 h) is required for full hydrogenation of NEC to perhydro-NEC, and a greater amount of active metal is also required, compared to the ruthenium on alumina catalyst described above (0.52 mol% and 0.25 mol%, respectively) (68, 69). The same Pd₂Ru@SiCN catalyst has also been reported to successfully catalyse the dehydrogenation of perhydro-NEC at 180°C in a reaction time of 7 h, in a batch reactor (69).

More recently, full hydrogenation has also been attained using ruthenium supported on a rare earth hydride catalyst, Ru/YH₃ (70). Reporting both the mildest conditions (100°C, 10 bar) and the highest catalytic activity for the hydrogenation of NEC to date, it is clear that such rare earth supported catalysts have potential for LOHC applications (70). This was highlighted by concluding the high stereoselectivity of the Ru/YH₃ catalyst for the all-cis product would be advantageous for any subsequent dehydrogenation reactions, since the cis product is more easily dehydrogenated than the trans equivalent. In explanation, the authors propose the cis product is less sterically hindered on the catalyst surface (70). However, despite an excellent catalytic performance and satisfactory stability, a high hydrogen pressure is required to achieve high selectivity (70). Manufacturing costs, handling and scale-up considerations were not discussed and may present challenges for such a material.

The Ru/YH₃ catalyst has also been reported to effectively catalyse the hydrogenation of another N-heterocycle: 2-methylindole, which has been suggested as a suitable LOHC candidate (70, 71). However, much like NEC, 2-methylindole is a solid at ambient temperatures and possesses a slightly lower hydrogen storage capacity than NEC (5.7 wt% and 5.8 wt%, respectively) (71).

To achieve an effective and efficient hydrogen storage and transportation system, the choice of LOHC is of critical importance. For commercial implementation of the LOHC technology, DBT can be seen as the most suitable candidate, considering its key advantages: relatively low-cost, low toxicity and high hydrogen storage capacity (6.2 wt%, Table I). In addition, DBT is already mass produced as it is used as a heat transfer agent. As a result, there is a vast exploration of DBT as a potential carrier molecule in the literature.

Several catalysts have been evaluated for the hydrogenation and dehydrogenation reactions of the DBT/perhydro-DBT system (Figure 7). Much like NEC, DBT hydrogenation is also typically catalysed with an alumina-supported pgm, namely platinum or ruthenium (1). As an example, in a batch reaction, a ruthenium on alumina catalyst can achieve full hydrogen loading of DBT in 4 h at 150°C with a 50 bar hydrogen pressure, which is longer than that of NEC (3 h) under the same conditions (68).

Fig. 7.

The more in-depth studies on this feedstock have also revealed that the catalytic hydrogen-loading does not have to be performed with pure hydrogen. As DBT hydrogenation is selective, and provided the catalyst is not negatively affected by the presence of other components, a mixed gas stream (for example, including methane, CO₂) can be used (34). From an industrial perspective this is very attractive, enabling otherwise low-value hydrogen present in waste gas streams, from processes including reforming and gasification reactions, to be stored and transported. The first example of such a process was reported in 2017 by Dürr et al., in which a mixture of hydrogen and methane from the decomposition of methane (obtained from offshore drilling) was fed directly to the hydrogenation unit containing the DBT feedstock (72). Separation of the loaded (hydrogenated) LOHC and gaseous methane can then be facilitated. It has also been reported that the presence of methane does not negatively affect the hydrogenation or dehydrogenation of DBT/perhydro-DBT (30). In contrast, methane was remarkably found to slightly improve the hydrogenation rate, postulated to be a result of a lower DBT viscosity and thus improved hydrogen mass transport (34).

In a similar vein, hydrogenation of DBT has been investigated using a hydrogen/CO₂ stream (up to 30% CO₂) (73). However, with such a hydrogenating mixture, both methanation (CO₂ to methane) and reverse water gas shift (CO₂ to carbon monoxide) processes were found to occur, with the extent of such side-reactions being strongly dependent on the catalyst employed (34). Even with the most promising palladium on alumina and rhodium on alumina catalysts, the degree of hydrogenation in a batch reactor reached only 0.8 and required elevated temperatures (210°C and 270°C for the rhodium and palladium catalysts, respectively) when compared to using a pure hydrogen stream (34). For the rhodium catalyst, moderate methane formation (methane:CO₂ ratio of less than 0.1) can be achieved at temperatures between 120°C and 150°C (rhodium on alumina), while the palladium catalyst showed a lower selectivity for methane formation, with a methane:CO₂ ratio of less than 0.1 between 120°C and 270°C (palladium on alumina) (34, 73).

Despite the frequent employment of platinum and ruthenium catalysts with pure hydrogen streams, these catalysts were found to be ineffective for the selective hydrogenation of DBT using a hydrogen and CO₂ stream (34). Their respective unsuitability was concluded to be a result of platinum catalyst poisoning from the reduction product carbon monoxide formed via reverse water gas shift of CO₂ and facilitation of CO₂ methanation by ruthenium.

Palladium-based catalysts have low activities for the dehydrogenation reaction of perhydro-DBT, and have thus been identified as catalysts to avoid (34). Under the same reaction conditions (270°C, 3.5 h, batch reactor), a stark difference in the degree of dehydrogenation between a palladium on carbon (5 wt%) catalyst and its platinum equivalent has been observed (16% and 55%, respectively) (68). In a different study, a lower metal loading has proven to be advantageous (lowering the metal content of the platinum on carbon catalyst to 1 wt% from 5 wt% increased the degree of dehydrogenation from 55% to 71%), while the support preference for this particular dehydrogenation reaction was found to follow carbon > alumina > silica (34, 68). In these experiments, the catalysts were used in batch reactions in quantities of 0.15 mol% with respect to perhydro-DBT (68).

Interestingly, the wider research around the use of DBT as a LOHC has included a study in which the same platinum on alumina (0.3 wt%) catalyst has been used for both hydrogenation and dehydrogenation reactions (34). Under the same reaction temperature (291°C), Jorschick et al. reported that the two reactions could be interchanged by varying the pressure between 1.05 bar (dehydrogenation) and 30 bar (hydrogenation) in a hot pressure swing reactor (30). However, the use of a non‐optimised catalyst for the dehydrogenation process resulted in a lengthy reaction time (20 h), when compared to the examples discussed previously. In addition, it must be noted that the productivity of the dehydrogenation (and thus catalyst activity) decreased significantly over the first cycle, before stabilising for the following three cycles. This was concluded to be a consequence of bulky, high boiling point byproduct formation via thermal cracking of DBT. Such molecules, including diphenylmethane and 2,6-dimethyldiphenylmethane, were found to increase per cycle and can be expected to block active sites of the catalyst (34, 74). It has also been suggested that some of the possible byproducts of the dehydrogenation reaction are unstable and thus undergo cracking during a subsequent hydrogenation reaction. Such an example would be the conversion of perhydromethylfluorene to perhydrobenzyltoluene (74).

DBT exists as structural isomers and consequently, more than 24 stable intermediates can be found in a partially hydrogenated solution. These can be categorised into four main groups: DBT (H0-DBT), hexahydro‐DBT (H6‐DBT), dodecahydro‐DBT (H12‐DBT) and octadecahydro‐DBT (H18-DBT) (75). Laboratory studies regarding selectivity would require advanced analytical detection methods. DBT conversion can be studied using ultraviolet-visible spectroscopy (i.e. the degree of ring hydrogenation, providing information about conversion levels only), but detailed studies probing catalyst selectivity would be challenging, requiring either a complex high-performance liquid chromatography (HPLC) method or ¹H nuclear magnetic resonance (NMR) analysis (29, 76, 77).

As such, alternative model carriers are often sought for fundamental academic studies aimed at deepening the chemical understanding of how to improve the performance of catalytic materials for the LOHC technology. In addition to being easily characterisable, the ideal carrier would have high boiling points and low melting points of both hydrogen-rich and hydrogen-lean molecules. This enables the whole cyclic process to be carried out in the liquid phase. A high melting point would result in the formation of a solid at ambient temperatures, with a dissolving process or incomplete hydrogen-unloading significantly lowering storage efficiency. Whereas a low boiling point would require extra economic expense in gas condensation equipment. Frequently, toluene is employed as such a model compound. Given its structural similarities to the central motif of DBT, this is a logical choice. Yet, the environmental health and safety hazards associated with its usage are quite significant, being noted for serious concern regarding both human and aquatic toxicity (23). Thus, to lower the hazards associated with the feedstock, other aromatic compounds similar in structure to toluene might be chosen.

Energy-intensive industries as well as the transportation sector contribute significantly to global greenhouse gas (GHG) emissions (1). To mitigate climate change and achieve the goals of the Paris Agreement (2), it is necessary that each sector develops pathways towards GHG emission reductions and accelerate the transition towards deep decarbonisation. The production and use of a low-carbon hydrogen is seen as a ground-breaking aspect of a low carbon future, especially for hard-to-decarbonise sectors (3–14). A significant amount of renewable electricity will be required to enable energy-intensive industries and the transportation sector to reduce their emissions and meet decarbonisation goals when deploying green hydrogen produced via water electrolysis using renewable energy sources (11). Given that capacities of renewables and electricity costs for the production of green hydrogen are extremely heterogeneous (15–18), it is expected that the production of green hydrogen in all required locations at adequate costs will be challenging in the future. Similar to fossil fuels, which are imported and exported across the world, geographical locations with high renewable potential and low costs of electricity are expected to be focal points for the production of green hydrogen. Robust hydrogen storage and transportation systems are among the key components in the successful transition from fossil fuel-based energy systems towards hydrogen-based alternatives (19–23). Storage and transportation of hydrogen at scale are yet to be addressed. An innovative method for long-distance transport and long-term high density hydrogen storage is to use LOHCs. This process is a two-step cycle, which is based on loading of hydrogen via catalytic hydrogenation into LOHCs, such as unsaturated organic compounds, followed by unloading of hydrogen via catalytic dehydrogenation after transport and storage (24–27).

In our previous work we provided an overview and perspectives on the LOHC technology among different hydrogen storage and transportation technologies (27). This study describes the potential deployment and integration of LOHCs within different industries. These include: the transportation sector (automobiles, ships, trains); steel and cement industries; the use of stored hydrogen to produce fuels and chemicals from flue gases; a system integration of fuel cells and LOHCs for energy storage.

Renewable sources, typically wind or solar, provide the energy required for the electrolysis of water to produce hydrogen (26). For the electrolysis stage, polymer electrolyte membrane (PEM) electrolysis is the preferred method, due to the ability of the system to respond to the characteristic fluctuations in renewable energy power supplies (25). The hydrogen produced can then be stored in LOHCs (through the hydrogenation step, Figure 1).

Fig. 1.

Interestingly, it has been reported that when dibenzyltoluene is employed as the LOHC, the hydrogen produced via electrolysis does not need to be dried before the catalytic hydrogenation reaction (28). When a ruthenium-based catalyst was used in a pellet form rather than as a powder, the activity of the catalyst was found to be virtually unaffected by water, while only a small decrease in hydrogenation activity was recorded when employing platinum catalysts with wet hydrogen (28). Costs associated with energy-intensive hydrogen drying processes can therefore be avoided. To better understand this observation, the tolerance of the catalyst should be investigated with other LOHC candidates.

When required, the stored hydrogen can be released from the LOHC (through the dehydrogenation step) and converted back into electricity, again using fuel cells. In this step, either PEM fuel cells (PEMFCs) or solid oxide fuel cells (SOFCs) can be utilised. PEMFCs are better designed to produce variable quantities of energy and thus meet changes in energy demand (25). Thus, the SOFC technology is only advantageous when there is a constant electricity demand (25). Nonetheless, using the SOFC technology to convert green hydrogen back into electricity is widely believed to become a future conventional method for stationary, green electricity production (29).

Furthermore, the waste heat from SOFCs (operated at 600–1000°C) is of the correct level to allow its use in the dehydrogenation step of the LOHC process (25). The coupling of LOHC technology with SOFCs can therefore improve the overall efficiency of the LOHC technology. To emphasise this, scale-up calculations for a LOHC–SOFC integrated system predict that 1 kg h^–1 of hydrogen is capable of producing around 18.04 kW of power, corresponding to a SOFC efficiency of 54.1% (29). Another study suggests that an overall electrical efficiency of 45% is achievable with a 10-year SOFC lifespan (30). Such a long lifespan can be maintained if LOHC vapour does not damage the SOFC (31).

In contrast, a combination of LOHC systems with PEMFCs would require heat from another source, such as burning a portion of the hydrogen produced, to facilitate the dehydrogenation reaction (25). This is because the waste heat of PEMFCs (below 180°C) is much lower than the heat required for the endothermic dehydrogenation process needed for hydrogen release. Hence an integrated LOHC–PEMFC system would not be as efficient as a LOHC–SOFC equivalent (31).

Interestingly, the SOFC technology has also shown promise when combined with a mixed LOHC feedstock (29). In a temperature cascade dehydrogenation process, a eutectic mixture of N-ethylcarbazole (NEC), N-phenylcarbazole, ammonia and biphenyl-diphenylmethane increases the energy generated per unit mass of the LOHC (kilowatt hour per kilogram LOHC) by 1.3–2 times, compared to an individual LOHC (29). The system integration of the LOHC technology, fuel cell technologies and green hydrogen produced via electrolysis using electricity from intermittent renewable sources, can thus enable the local storage of excess energy or renewable electricity. Such a process is entirely sustainable as no harmful emissions are produced.

The concept of system integration can be used as a safety device for the storage of electricity on a large-scale, such as the National Grid, UK. There are currently several reports detailing the problem of electricity fluctuation in the National Grid, due to the lowered electricity demand during the COVID-19 pandemic (32). Despite a rise in the percentage of people working from home, resulting in an increased demand for domestic electricity, an overall decrease in electricity demand was observed (32). This can be attributed to the closure of non-essential offices, schools, hospitality and leisure venues. As a consequence, measures such as temporarily shutting down flexible windfarms are expected to be taken, in order to lower the excess of energy in the National Grid (32). Too much electricity in the National Grid is equally as concerning as too little electricity, as the rise in frequency increases the potential to damage infrastructure (33). As the proportion of renewable energy in the National Grid will predictably increase in the future, the fluctuations in electricity are also expected to increase. An energy storage technology, such as LOHC combined with fuel cells, could therefore be extremely beneficial. Similar would apply to ammonia and methanol, where such hydrogen carriers can be deployed in combination with fuel cells.

The above-mentioned energy storage ability of the LOHC technology has been reported to have implementation potential in residential and commercial buildings (localised energy storage) (Figure 2) (34, 35).

Fig. 2.

For such an application, heterocyclic aromatic compounds, such as NEC, have been considered. Although the fully dehydrogenated form is a solid at room temperature, which limits dehydrogenation to 90% and reduces the efficiency of the overall system, NEC is a safer feedstock than other possibilities (for example a toxic toluene/methylcyclohexane system). Thus, the requirements for domestic implementation are better accommodated (34).

In contrast to an application in which the loaded-LOHC is not stored in the vicinity of intended use, a ‘decentralised energy storage’ system is proposed. This technology also uses fuel cells to meet the local demands for electrical energy but has the added economic advantage that any unused energy can be sold back to the National Grid, for example. Furthermore, provided the building has such a resource as solar panels fitted onto the roof, the potential for a completely self-sufficient system exists, while any waste heat generated from the fuel cells, electrolyser and exothermic hydrogenation process may be used to heat the building (34). Yet, with the requirement of fuel cells, electrolysers, hydrogenation and dehydrogenation units, in addition to LOHC storage tanks, the physical space requirement and initial economic investment are high (34). It has, however, been suggested that houses already possessing a crude-oil storage tank may avoid the requirement for a subsequent tank (34).

The transformation of green hydrogen back into electricity is just one example which demonstrates how the LOHC technology can facilitate energy storage. Green hydrogen, stored and transported in LOHCs, can also be used as a ‘green’ feedstock for the synthesis of fuels and chemicals. Moreover, green hydrogen production and storage is a vital part of most carbon capture and utilisation (CCU) technologies which are focused on capturing CO₂ from either air or industrial flue gases and converting it into chemicals or fuels. For the moment, the hydrogen required for CCU technologies is considered to be generated from sustainable energy resources, however detailed integrated processes have often not yet been developed. The availability of green hydrogen for CCU processes is limited by competing demands such as hydrogen used in fuel cells for the transportation sector, and hydrogen used as domestic and industrial fuel supply.

Green hydrogen production is a key accelerator of CCU for production of chemicals and fuels at a commercial scale. Therefore, the production of hydrogen via electrolysis and its storage and transportation using LOHCs (or other hydrogen carriers such as ammonia or methanol) can be viewed as an integral part of sustainable chemicals and fuels manufacturing. One of such examples is methanol production, with a global production capacity of around 85 million metric tonnes in 2016, which is expected to rise in the coming years (36). Conventionally, methanol is synthesised using synthesis gas (syngas) produced from fossil fuels. However, a move towards the use of renewable hydrogen for sustainable methanol synthesis, using CO₂ captured either from the air or from industrial flue gases, would enable a reduction of global GHG emissions.

Thus, a scenario under which low-cost green hydrogen (i.e. produced at locations with an abundant energy supply and so cheaper electricity) is transported using LOHCs to the industrial sites (for example, cement, steel, refinery industries) with high CO₂ emissions to produce sustainable chemicals and fuels, might become viable in the future. Nevertheless, technoeconomic assessments and market penetration studies are required in order to understand under which circumstances this scenario can be realised.

In the coming years, the proportion of electricity to be obtained from renewable sources is expected to increase. By 2024, it is predicted as much as 30% of the global electricity demand will be met with renewable energy sources: an increase of 4% in four years (37). The expected 15–30% decrease in the cost of solar power within the same time frame is also expected to accelerate the growth of renewable energy generation sites (37). However, the fluctuations in renewable electricity supply make it unreliable for direct industrial use. Integration of the LOHC technology within a cement plant has thus been studied as method of energy storage, which can be utilised to equalise the plant’s energy output (Figure 3) (38).

Fig. 3.

The electricity for a cement plant can be supplied from renewable sources. During favourable conditions, the excess electricity is converted into hydrogen via electrolysis, followed by loading the hydrogen onto the LOHC and storing it. The dehydrogenation reaction is then performed at times of insufficient power supply from the renewable sources. The hydrogen released can be converted into electricity using fuel cell technology, a combustion engine or turbine (38). If an adequate power supply still cannot be achieved with the added electricity from the storage, the plant can be connected to the National Grid.

The higher temperatures required for the endothermic dehydrogenation reaction (i.e. hydrogen unloading) is a bottleneck of the LOHC technology, with the requirement of an external heating source resulting in a lowered system efficiency. In order to improve the efficiency of the LOHC system, the heat required for the dehydrogenation reaction can be coupled to the waste heat of a cement plant (Figure 3) (25). Arguably, this is a more suitable solution than the coupling of the dehydrogenation reaction to the waste heat from the exothermic hydrogenation reaction for heat recovery, as this is typically not of the same temperature level as is required for the dehydrogenation reaction. The waste heat from a cement plant, however, is more suitable (temperature level of up to 600°C) for the dehydrogenation process. Attractively, this avoids the need to reduce the efficiency of the process by burning a portion of the released hydrogen and there are no extra costs associated with employing an external heating source (25, 38).

Integrating a LOHC system with a cement plant allows an overall reduction in the working electricity cost of a cement plant. For a plant with an average power demand of 12.5 MW, it has been found that converting the hydrogen released in the dehydrogenation reaction to electricity and using this to power the plant would reduce electricity costs by around €1 million per annum (38). Although this saving has not been expressed as a percentage of the total electricity cost, it has been suggested that successful coupling of wind energy and the LOHC technology to a cement plant would achieve amortisation in fewer than 10 years. This is on the condition that investment costs are kept to a limit of €3.5 million. In addition, it should also be noted that the savings in electricity costs are calculated on the assumption that a thermal energy storage system is also installed, which ensures hydrogen can be released from the LOHC even in times of lower exhaust heat from the cement plant (38). In the absence of such a thermal storage system, fluctuations in the temperature of the waste heat must be provided for with the addition of a hydrogen burner (38).

The iron and steel industry is responsible for an annual output of approximately 2.5–3.0 Gt_CO2 year^–1, with up to 10% originating from within the European Union (EU). This represents 6% of total global CO₂ emissions, and 16% of total industrial CO₂ emissions. To reach the EU climate targets, the iron and steel industries must decrease their CO₂ emissions by up to 90% by 2050. Several processes are being explored to reduce CO₂ emissions from the steel industry. They can be broadly divided into two categories: carbon-based (coal- or natural gas-based) and hydrogen-based steel production (39). In carbon-based steel production, the residual gas emissions from the iron and steel industry can be transformed into valuable products, such as fuels or chemicals, or captured and stored or both. Hydrogen-based technologies, which use hydrogen as the reducing agent instead of carbon, avoid carbon emissions altogether, provided that hydrogen used in these processes is carbon-free hydrogen, produced by electrolysis of water using renewable electricity.

Numerous steel manufacturers have started to explore hydrogen-based technologies. As an example, voestalpine AG, Austria, has set up a goal of direct avoidance of CO₂ emissions in their steel manufacturing over the coming years by moving towards the use of green hydrogen for steel production (i.e. direct reduction of iron (DRI)). To this end, voestalpine together with their project partners have commenced the production of green hydrogen at the voestalpine premises in Linz, Austria within the framework of the EU-funded project called H2FUTURE (40, 41). In this project, the proton exchange membrane electrolysis technology is demonstrated on an industrial scale (6 MW), simulating rapid load changes in electricity generated from renewable energy sources and from electric arc furnace steelmaking (grid balancing). Thyssenkrupp Steel Europe AG, Germany’s biggest steelmaker, is also looking into using hydrogen for steel manufacturing. RWE AG, a German multinational energy company, and Thyssenkrupp Steel Europe AG have agreed to collaborate towards a longer-term hydrogen partnership to supply green hydrogen for steel manufacturing (42, 43). RWE plans to build a 100 MW electrolyser which can produce 1.7 tonnes of hydrogen per hour for Thyssenkrupp Steel Europe AG. This could potentially cover 70% of the quantity required by the Thyssenkrupp steelmaker’s blast furnace in Duisburg, Germany.

Another example is a joint venture between SSAB, LKAB and Vattenfall, all in Sweden, within the framework of the Hydrogen Breakthrough Ironmaking Technology (HYBRIT) project. The aim is again to reduce CO₂ emissions and decarbonise the steel industry by replacing coal with hydrogen in the steelmaking process to produce fossil-free steel at Sweden’s pioneering fossil-free steel production plant (Figure 4) (44–46).

Fig. 4.

Interestingly, the HYBRIT project faces two major challenges: (a) to develop an effective process to use 100% hydrogen on an industrial scale; (b) to produce hydrogen in an energy efficient way that is economically justifiable and commercially viable (46). To this end, the HYBRIT project has recently invested SEK 200 million (£17.5 million) in a pilot plant for storage of green hydrogen in Lulea, Sweden near the SSAB steel manufacturing site (47). The implementation study for the HYBRIT initiative has recognised the need for large-scale storage of hydrogen. To ensure an even flow of green hydrogen produced from renewable energy for steel manufacturing at the SSAB site, a large-scale hydrogen gas storage facility is required to balance the electricity system with an increasing proportion of weather-dependent power generation. It is expected that the integration of large-scale storage of green hydrogen to a fossil-free value chain for steel manufacturing will allow the fossil-free steel to be price competitive.

In the HYBRIT project, the 100 m³ subterranean (25–35 m underground) pilot hydrogen gas storage facility is built in a bedrock cavern with a steel lining as a sealing layer to store pressurised hydrogen (47). In contrast to SSAB, other steel manufacturing companies might not have the facilities for the underground storage of green hydrogen. Furthermore, geographically some of the steel manufacturing sites might not have access to low-cost surplus renewable electricity. In addition, if they are located in energy demanding industrial districts, numerous sectors will compete for green hydrogen or renewable electricity to reduce their CO₂ emissions in the near future. To ensure a competitive production cost for the fossil-free steel (i.e. DRI technology), an effective hydrogen storage and transportation technology, such as the LOHC technology, might thus be required to allow a reliable and on-demand supply of green hydrogen.

As large-scale storage of hydrogen is an important part for a fossil-free value chain for steel manufacturing, it should become an integral part of the DRI technology. For the deployment of the LOHC technology in the steel industry, two scenarios can be foreseen. One is an onsite hydrogen production and storage in LOHCs. Such a scenario can be realised when a steel manufacturing site has access to surplus renewable electricity from intermittent resources to produce hydrogen that should be stored to balance the electricity. Another one is using LOHCs to transport hydrogen, which is produced at geographical locations with high availability of renewable electricity, to a steel manufacturing site with low availability of renewable electricity but high demand for green hydrogen.

BMW AG, Germany, has been developing automobiles that employ hydrogen technology for around 40 years (48). The first model built using such a technology was named the BMW Hydrogen 7, and comprised the storage of liquid hydrogen in a cryogenic tank. However, during its early development serious technological challenges were realised (48). For instance, sufficient hydrogen storage space must be provided to enable longer-distance travel, but size and weight limitations for a practical motor vehicle must also be considered. In addition, the safety risks associated with burning hydrogen in the internal combustion engine must also be adequately minimised. As only 100 vehicles of the Hydrogen 7 model were ever released, it can be deduced that the disadvantages of the technology outweighed the advantages (48). More recently, BMW has announced a partnership with Toyota Motor Corporation, Japan, to develop a fuel cell-based system suitable for integration within its motor vehicles (49). It is predicted that the new model will be commercially available by 2022 (49).

As an alternative for onboard liquid hydrogen storage in mobility applications, the LOHC technology has been suggested (Figure 5) (50). Advantageously, onboard storage of hydrogen in the LOHC would resemble that of gasoline and diesel (liquid state of the LOHC at ambient pressures), which is widely understood, and the safety hazards of storing high-pressured hydrogen are removed (48). Moreover, a range of 500 km is reportedly achievable using 100 l of NEC loaded with hydrogen (equivalent to 5 kg of hydrogen) as the LOHC (48).

Fig. 5.

For motor vehicles, the loaded LOHC would be transported to the refuelling station, before a subsequent release of hydrogen from the carrier in a catalytic dehydrogenation reaction occurring onboard the vehicle (50). In contrast to diesel and gasoline fuels, the dehydrogenated LOHC would not be consumed, but stored within the vehicle until replaced with new hydrogen-loaded material at a designated station (48). The unloaded LOHC can then be transported back to a hydrogenation site and reloaded. Therefore, either two tanks, or a tank capable of separating the loaded and unloaded forms of the LOHC, is required. The hydrogen produced could then be used in an internal combustion engine or combined with fuel cell technologies. Despite the burning of a portion of hydrogen to meet the working temperature of a typical fuel cell, it has been predicted that the overall efficiency would still be higher than that of a combustion engine (48).

Highlighting the attractiveness of implementing the LOHC technology within the mobility sector, Hyundai Motor Company, South Korea, has recently announced plans to develop an onboard LOHC storage technology (51). This encompasses a partnership with Hydrogenious LOHC Technologies GmbH, Germany, who will supply dibenzyltoluene to be used as the LOHC. Initially, the technology will be introduced in South Korea, before being extended to the European market (51). It is expected that the development of LOHC compatible automobiles will raise the profile of the technology as an important tool in the transition to the hydrogen economy.

The use of hydrogen in maritime applications is also an active research area aimed at reducing the pollution from the maritime industry. One of the major challenges is the storage of hydrogen on board of ships (52). The LOHC technology is seen as one of the potential hydrogen storage solutions (52–54). For instance, the use of a double chamber tank system has been proposed, which is capable of separating the loaded and unloaded carriers during the fuelling process into different sections of the tank (53). Nevertheless, providing a suitable level of heat to facilitate the dehydrogenation reaction remains problematic.

Similarly, the use of LOHCs in trains has also been considered (55). Since a significant portion of trains are currently operated using diesel, or a combination of electricity and diesel, future environmental-focused objectives are likely to concentrate heavily on finding alternatives to these fuels (55). Although trains can be powered electrically with renewable electricity to meet zero-emission transportation goals, building the infrastructure of overhead wiring is relatively expensive. Currently, 42% of the UK railway routes are electrified and can become zero-carbon when using renewable electricity (56). The remaining 58% still rely on diesel (56). An alternative approach to electrification is the use of hydrogen fuel cells to generate electricity onboard to power trains (57). Hydrogen powered trains have a potential to revolutionise railway operations in Europe (56).

Using LOHCs for onboard hydrogen storage, coupled with hydrogen fuel cells for electricity production to power trains would avoid the production of hazardous emissions (i.e. CO₂, soot, nitrogen oxides), while still permitting long-distance travel, an essential criterion for trains. In a recent study, the LOHC technology (with dibenzyltoluene as the LOHC) was found to be a very promising option for hydrogen storage, transport and release and can be combined with electricity generation by hydrogen fuel cells to power trains (55). The choice of dibenzyltoluene as a LOHC was influenced by its favourable properties, such as low flammability, low toxicity and liquid-state within the range of hydrogenation and dehydrogenation temperatures, in addition to its commercial availability as a heat-transfer oil (55). Notably, this technology has been supposed to be favourable over alternatives, such as batteries, which are typically characterised by low energy densities (55). Furthermore, the hydrogen fuel cell technologies required for the integration with the LOHC technology for onboard hydrogen storage are expected to become a lower cost alternative to battery and diesel options in the second half of the decade.

Even though the studies into hydrogen-powered aviation are somewhat immature in comparison to trains and cars, it is anticipated that sustainable aviation will quickly become a central research focus in the coming years. Like other mobility sectors, hydrogen is again expected to be named as a primary energy source for propulsion in the aviation industry. This can be realised either through powering aircrafts with electricity generated through hydrogen fuel cells, direct burning of hydrogen in gas turbines or using hydrogen as a building block for production of sustainable synthetic aviation fuels (58). Notably, Airbus has already revealed three hydrogen-powered planes to lower aircraft emissions, which comprise the use of hydrogen gas-turbine engines (59, 60). In addition, Airbus SE, The Netherlands, has postulated the combined use of hydrogen fuel cells with a gas-turbine engine to create a ‘highly efficient hybrid-electric propulsion system’ (61). The urgency of this sustainable transition has also been emphasised by the Swedish government. A mandatory reduction in GHG emissions, originating from aviation fuels, will be introduced for fuel to be sold in the country from 2021 (62). This will start at a 0.8% reduction in 2021 and reach 27% by 2030, in preparation for reaching their national fossil-free target in 2045 (62).

In summary, the LOHC technology is an attractive solution for the storage and transportation of hydrogen to allow a reliable and on-demand hydrogen supply, enabling industrial decarbonisation. The potential deployment and integration of the LOHC technology within different industries, such as the transportation sector, steel and cement industries, the use of stored hydrogen to produce sustainable fuels and chemicals from flue gases, and a system integration of fuel cells and LOHCs for energy storage, is depicted in Figure 6.

Fig. 6.

The possibility of deployment and integration of LOHC systems within different industries is reviewed in this study. These include: the transportation sector, steel and cement industries, the use of stored hydrogen to produce fuels and chemicals from flue gases and system integration of fuel cells and LOHCs for storing renewable electricity. An effective system integration of the LOHC technology with different industries might help with the cost reduction of the LOHC technology, when for example, waste heat is used for dehydrogenation of LOHCs. Importantly, the deployment of the LOHCs for storage and transportation of hydrogen to allow a reliable and on-demand hydrogen supply might enable energy-intensive industries to reduce their emissions and meet decarbonisation goals.

Numerous possibilities for the deployment and integration of LOHCs within different industries might necessitate the use of different LOHC carriers in each instance. While a carrier choice offers a large amount of flexibility in the LOHC technology, the myriad of possible carriers and catalysts combined with reactor technologies might be considered as one of the factors impeding the integration and commercial deployment of the LOHC technology across different industries. Customer-tailored solutions and offerings might need to be developed to accommodate specific requirements. A review of the most prominent LOHC systems, focusing on properties of LOHCs and catalytic materials used for hydrogenation and dehydrogenation of LOHCs, is presented in our following work dedicated to the analysis of LOHC systems (63).

In recent years there has been great interest in reducing fossil fuel reliance. This comes in an attempt for countries to deliver on pledges outlined in the Paris Agreement of 2016, which was compiled to tackle climate change by lowering greenhouse gas emissions (GHG) (1, 2). Individual nationally determined contributions (NDCs) are central to this agreement, detailing the collective efforts required to achieve longer-term global aims. Within these NDCs, several countries made specific references to the increased use of renewable energy resources, given their widely recognised environmental advantages over more traditional fossil fuel equivalents. Now, atmospheric CO₂ levels are higher than any recorded in the previous 800,000 years, with a rise from 300 ppm to 400 ppm being recorded over the last 70 years (1950–2020) (3). These continually rising levels are commonly attributed to the increased consumption of fossil fuels, which are conveniently employed to satisfy ever-changing energy demands. Yet the consequences are serious: during the 20th century a 1°C increase in average global temperature accompanied the growth of CO₂ emissions, causing drastic changes in weather patterns and rising sea-levels (4, 5).

Fossil fuels are extremely widely used because of convenience, availability and the economic advantage of a lower initial capital expenditure. Developing a ‘greener’ system with a comparable energy density and transport efficiency is a challenge (6). Despite several years of research into alternatives, there is still a significant way to go before achieving the goals proposed in the Paris Agreement by 2050. In 2016 just 19.5% of the global energy demand was fulfilled by renewable sources (7). Germany, however, has pledged that by 2050, 80% of its energy will be produced from renewable sources, whereas other countries have implemented strategies to increase collaborative research to achieve net-zero emissions (7, 8). As an example, Portugal and The Netherlands signed a memorandum of understanding to develop a strategic export-import value chain to ensure production and transport of green hydrogen from Portugal to The Netherlands and its hinterland via the ports of Sines and Rotterdam (9).

Hydrogen is a well-studied alternative to fossil fuels. With its combustion producing only water as a byproduct, the environmental advantages of employing hydrogen as an energy carrier are obvious. However, at present, around 96% of the annual global hydrogen production is generated from fossil fuels (grey or brown hydrogen) (Table I). This comprises steam reforming of methane (48%), reforming of oil/naphtha (30%) and coal gasification (18%) (11). Thus, in order to completely remove reliance on the non-renewable, finite, fossil fuel resources, an alternative method of hydrogen production is required. This can be achieved with the electrolysis of water. Provided the energy for this process is obtained from renewable sources, sustainable, greenhouse gas emission-free hydrogen production is possible. Hence, hydrogen produced in this manner is termed green hydrogen.

The European Union (EU) has pledged billions of Euros to develop the so-called European Green Deal in the coming years. Here, hydrogen is considered ‘a key’ to fulfilling the ambitious target of halving carbon emissions by 2030 (8). To achieve this, the European Green Deal acknowledges the need to increase green hydrogen production capacity, setting targets for around 10 million tonnes of hydrogen to be produced annually by 2030 (8). Currently, the hydrogen supplied from electrolysers stands at just 4% of demand (8).

A third colour-coded category of hydrogen refers to blue hydrogen (Table I). Blue hydrogen is often considered a vital tool in transitioning from grey to green hydrogen production and, like grey hydrogen, uses a fossil fuel feedstock. However, carbon capture and storage (CCS) technologies are also employed to capture CO₂, reducing the greenhouse gas emissions (12). Given the increase in carbon-tax expected over the coming decades, blue hydrogen is an important improvement upon grey hydrogen, despite the currently higher initial capital expenditure (12–14). Brown, grey, blue and green hydrogen (Table I) are the most discussed colour-coded categories of hydrogen within the energy industry. Nevertheless, countless other hydrogen colours (i.e. the hydrogen colour spectrum), such as yellow, pink and turquoise hydrogen, also exist with each colour code describing the different type of source or process used to produce hydrogen (15, 16). For instance: pink hydrogen is generated by electrolysis of water using electricity from a nuclear power plant; yellow hydrogen is produced via electrolysis using solar power; turquoise hydrogen is generated via methane pyrolysis through direct splitting of methane into hydrogen and solid carbon (15, 16).

Some countries are well-positioned to the transition to blue hydrogen production, with a potential access to large, offshore CCS facilities (17). In contrast, many landlocked countries in Europe have a greater focus on the transition to green, rather than blue, hydrogen due to their limited access to offshore facilities (18). In addition, with blue hydrogen still requiring a finite resource, the demand of one country may eventually be pushed onto another, where resources are more available (12, 19). For example, an exhaustion of natural gas supplies in one location would require a country to source this elsewhere, resulting in the production of hydrogen and its utilisation at different locations. This dependence (expected to be governed by maintaining amicable international relations) is another factor causing several countries to seriously consider the favourability of blue over green hydrogen, or vice versa (12).

Producing hydrogen from renewable sources and development of technologies for this purpose have huge environmental benefits and supports the implementation of the Paris Agreement and the United Nations Sustainable Development Goals (SDG) (20). It contributes to multiple SDGs, such as SDG 7 (Affordable and Clean Energy), SDG 9 (Industry, Innovation and Infrastructure), SDG 11 (Sustainable Cities and Communities) and SDG 13 (Climate Action) (20). According to SDG 7: Affordable and Clean Energy, a substantial increase in the share of renewable energy in the global energy mix is required by 2030 (i.e. Target 7.2 of SDG 7) (21). However, the energy output from renewable sources is intermittent and dependent on geographic, seasonal and temporal factors. Furthermore, sites of highest energy potential for renewable hydrogen production (such as a desert, offshore wind and tidal farms) are rarely located in areas of highest energy demand, such as densely populated cities in central and southern Europe (1, 22). When moving towards a global low carbon hydrogen economy with the aim of meeting net-zero climate goals, a reliable storage and transportation of hydrogen at scale is a challenge which needs to be tackled to achieve a widespread usage of renewable hydrogen. The possibility to store and transport hydrogen is essential for the integration of high shares of renewable energy source with positive effects on SDGs (23).

Numerous technologies and options are currently being explored for effective hydrogen storage and transportation to facilitate a smooth transition to the hydrogen economy (24). LOHC is one such technology which has gained considerable attention in recent years (1, 6, 7, 25–27). In our work, which consists of the present paper and two accompanying papers (28, 29), we provide an overview and new perspectives on the LOHC technology among different hydrogen storage and transportation technologies. We analyse the advantages and disadvantages of the LOHC technology and future considerations for its optimisation which might accelerate its commercial deployment. Furthermore, in our following second paper we describe the potential deployment and integration of LOHCs within different industries: the transportation sector (automobiles, ships, trains); steel and cement industries; the use of stored hydrogen to produce fuels and chemicals from flue gases; and system integration of fuel cells and LOHCs for energy storage (28). Due to numerous possibilities for the commercial deployment and integration of LOHCs within different industries, the use of different LOHC systems might be considered to accommodate specific requirements. A review of the most prominent LOHC systems, focusing on properties of LOHCs and catalytic materials used for hydrogenation and dehydrogenation of LOHCs, is presented in our third paper dedicated to the analysis of LOHC systems (29).

Despite the attractively high gravimetric energy density of hydrogen (120 MJ kg^–1), the low volumetric energy density at ambient conditions necessitates the use of pressurised or liquified hydrogen to ensure economic viability (30). On a large-scale, the use of stored, highly pressurised hydrogen in transport systems presents serious safety concerns, such as an explosion. The technology is also costly, much like that of liquified hydrogen storage, which requires low temperatures (–252°C). In addition, such liquified hydrogen technologies can result in losses of between 0.3% and 3% of the hydrogen due to boil-off from the storage system (1).

Compressed hydrogen is typically transported through pipelines. As a result, long-distance transport is significantly more challenging when compared to the infrastructure, such as ships or railroads, currently available to transport liquid fossil fuels. Although attempts have been made to inject hydrogen into the natural gas grid and numerous countries are exploring blending hydrogen with natural gas, it can be argued that the intercontinental transport of hydrogen would require innovative solutions (31–33). These solutions are predicted to resemble current methods, relying on tanker vessels, trucks and railroads.

To overcome the challenges raised with the transportation and storage of hydrogen as an energy vector, more sophisticated concepts have been developed (Figure 1) (24). One such concept involves the use of low- and high-temperature metal hydride systems, where reversible adsorption and desorption of hydrogen is permitted through formation and breakage of chemical bonds with the storage material (1, 11, 34). Here, the terms low- and high-temperature systems refer to the dehydrogenation reaction temperature. These methods of solid-state hydrogen storage, therefore, do not require the demanding pressures associated with storing compressed hydrogen.

Fig. 1.

Low temperature metal hydrides are capable of releasing hydrogen at ambient temperatures and pressures but are typically limited to a maximum hydrogen capacity of 2.15 wt% (for example, LaNi₅H₆) (35). On the contrary, high temperature hydrides are attractive for hydrogen storage because of their high hydrogen storage capacities (6.5 hydrogen atoms per cm³ (7.6 wt%) for the metal hydride MgH₂, compared to a hydrogen storage density of 0.99 hydrogen atoms per cm³ for hydrogen gas and 4.2 hydrogen atoms per cm³ for liquified hydrogen). However, disadvantages of high temperature metal hydride systems include slow reaction kinetics of the adsorption-desorption process. Moreover, the high desorption enthalpies of these systems require high temperatures (300°C for MgH₂) for hydrogen release at atmospheric pressure, raising questions around the economic viability of the technology (34–36). The reversibility of the process is also limited by the decomposition of the metal hydride, making a regular replacement of this storage material necessary.

A different concept for solid-state hydrogen storage is the use of metal-organic frameworks (MOFs). Here, hydrogen is stored in the pores of the MOF framework, where the rate of hydrogen adsorption is dependent on its diffusivity in the particular MOF chosen (37). MOFs can be tailored to specific applications, including supercapacitors, fuel storage and batteries and have been noted for their short refuelling times (37). However, this technology typically requires low temperatures (ca. –196°C) and high pressures (100 bar) to achieve a reasonable energy density (7.2 MJ l^–1, 4.5 wt%). To illustrate this, the MOF displaying the highest hydrogen storage capacity to date (MOF‐210: total hydrogen gravimetric uptake of 17.6 wt%) must be operated at 80 bar and –196°C (38). It is therefore possible that a low system efficiency in terms of energy consumption, and thus cost, will be observed. It has been predicted that the hydrogen storage capacity of MOFs can be improved by increasing MOF surface area, but there are experimental limitations in synthesising such structures (38). Moreover, as with using compressed hydrogen, the requirement of a higher-pressure system implicates explosion risks (37).

Borohydrides have also been considered for hydrogen storage applications, given their attractively high theoretical storage capacities. For example, LiBH₄ has a gravimetric hydrogen storage capacity of up to 18.5 wt% (24). However, in practice this is rarely achievable as stable hydrides, such as LiH, can form during the hydrogen unloading process, for which high temperatures (typically greater than 300°C) are also required. Direct employment of borohydrides for hydrogen storage applications is therefore unfeasible. Yet, a technology combining both borohydrides and metal hydrides is now being considered. Such a system reduces the endothermicity of the dehydrogenation process, facilitating hydrogen release at lower temperatures, but the kinetics of both the hydrogenation and dehydrogenation reactions are slow and require suitable catalytic additives (24). This increases the complexity of the technology.

Chemical hydrides, both organic and inorganic, are another method available to store hydrogen, and rely on chemical reactions. Chemical hydrogen storage is used to describe storage technologies in which hydrogen is both released and restored through a chemical reaction. Like metal hydrides, chemical hydrides also form chemical bonds between the storage material and hydrogen. Yet, these two hydride classes have very different properties. Arguably the most significant of these is that chemical hydrides (such as methanol) are generally liquids at ambient conditions. This simplifies transportation and storage issues associated with either gaseous hydrogen or solid-state hydrides. Provided ammonia is in its liquid form, which can easily be achieved by applying a pressure of 10 bar at ambient temperature, ammonia can also be considered a chemical hydride, with a very attractive storage density (17.7 wt%) (24). In addition, given both methanol and ammonia are already synthesised on a large scale, there is the possibility of using existing infrastructure and production plants (24). Both chemicals have also been reported to have potential as hydrogen-alternative energy vectors (i.e. direct use as a fuel) rather than as hydrogen storage materials (24).

Ammonia is currently of industrial interest for hydrogen storage and transportation, given its high gravimetric hydrogen density (17.7 wt%) (24, 39–42). Additionally, the high ignition temperature of ammonia increases the safety of the technology and is considered advantageous (41, 42). When green hydrogen is used for the production of ammonia, such ammonia is classified as green ammonia (43–46). Green ammonia, when liquified, facilitates the transport and storage of hydrogen, allowing existing infrastructure to be used (42). With the potential for worldwide transport of the green ammonia, for example via ships, the distribution of green hydrogen to growing zero-emission markets will be facilitated (42). Once transported to the site of use, ammonia can be reconverted into carbon-free hydrogen which can be used at hydrogen refuelling stations, for example (42, 47).

Within the framework of the NEOM project it was recently announced that 650 tons of carbon-free hydrogen per day will be produced using 4 GW of renewable power from solar and wind in Saudi Arabia (48). To facilitate hydrogen storage and transportation, the hydrogen will be converted into 3500 tons of green ammonia per day (or 1.2 million tons per year) which will be transported around the world and then converted back into carbon-free hydrogen at hydrogen refuelling stations (48). By supplying the fuel cells currently used within the transport sector (specifically in buses and trucks) with this hydrogen, it is predicted that over 3 million tons per year of CO₂ emission can be prevented: equivalent to all emissions from 700,000 cars (48).

Although ammonia is less flammable than hydrogen, concerns around its toxicity to both humans and the environment have been raised (49). A spillage of this feedstock (for example during transport) could therefore have serious consequences, and questions around its suitability have been raised. However, it is important to remember that suitable controls have been employed to mitigate the risks associated with fossil fuels (in the forms of gasoline and diesel), which are also very harmful substances. Moreover, as ammonia is already synthesised on an industrial scale (Haber-Bosch process) for example, in the fertiliser industry, it can be argued that the safety concerns are known and can be managed effectively. If the ammonia were not to be dehydrogenated at its destination, but instead used as a fuel itself, the ammonia would be classified as a renewable fuel (50). As an example, using ammonia as a transportation marine fuel is currently being explored within the shipping sector to cut fossil fuel use in ocean-going vessels (51–55).

Liquid hydrocarbons and formic acid can also fall into the category of hydrogen carriers. Provided such carriers are produced using green hydrogen and atmospheric CO₂, or CO₂ from waste streams, the cycle can be labelled as carbon-neutral (11). However, as the carriers are used as liquid fuels, regeneration of the carrier material is not possible and thus new material must be purchased for every cycle, much like employing fossil fuels as energy vectors (11).

As previously discussed, renewable methanol can also be considered as a chemical hydride. Although the gravimetric energy density of methanol is lower than that of ammonia (12.5 wt% and 17.7 wt%, respectively), it is significantly higher than a typical metal hydride, such as MgH₂ (7.6 wt%, Table II) (24, 37). Most commonly, methanol is synthesised via the hydrogenation of CO₂ and carbon monoxide, whereas the release of hydrogen from methanol is done in the methanol steam reforming process, which involves the reaction of methanol with water (24). This process is generally preferred over methanol decomposition, permitting the release of three moles of hydrogen in comparison to the two moles from methanol decomposition, where the extra mole of hydrogen is provided by the water (24). Yet for the steam reforming reaction, temperatures of between 220°C and 330°C are often required to meet the thermodynamic demands of the endothermic reaction (24).

Although renewable methanol (produced via the hydrogenation of CO₂ waste streams) is produced on a much smaller scale than its non-renewable equivalent (in which natural gas is used to produce a mixture of carbon monoxide, CO₂ and hydrogen), similarities in the two processes exist (24). With an overlap in the technology, progress in renewable methanol production is better facilitated than other hydrogen storage technologies and as a result the first renewable methanol production plant was constructed in Iceland in 2011 by Carbon Recycling International (24). Interestingly, it has been reported that the separation of methanol and water (via distillation) is not required when using methanol as a hydrogen storage medium: the hydrogen can simply be released from the mixture in a steam-reforming reaction (24). This simplifies the process and eliminates costs associated with the energy intensive distillation step. However, CO₂ is also stored within the methanol-water mixture and hence is also released upon steam reforming. If a pure hydrogen output stream is required, it has been predicted that CO₂ could be separated from the hydrogen relatively easily, but this process would require additional separation technologies (24). Alternatively, the gaseous hydrogen and CO₂ mixture could be used directly within proton exchange membrane (PEM) fuel cells (56).

LOHCs, also categorised under chemical hydrides, are another option for the storage and transport of hydrogen. The first studies into this technology were completed in the 1980s by Japanese researchers, studying a benzene/cyclohexane system (1). The LOHC process comprises a two-step process, which is based on the loading of hydrogen onto the chosen LOHC in a catalytic hydrogenation reaction, followed by the unloading of hydrogen in a catalytic dehydrogenation reaction. This second step thus produces a stream of gaseous hydrogen alongside the unloaded form of the LOHC, which can then be reused in subsequent cycles (Figure 2). Between these two steps, the hydrogen-rich form of the LOHC can be easily stored and transported at ambient pressures, given its liquid state. The LOHC technology eliminates the expense associated with repeatedly purchasing a feedstock (i.e. the LOHC) and hence may be considered advantageous. Yet one must also consider the expense of returning the unloaded LOHC to the hydrogenation plant. In some cases, finding an alternative use for the unloaded LOHC may be most cost-effective.

Fig. 2.

Aromatic molecules are typically used as LOHCs due to their high hydrogen loading capacities (57). In addition, the cyclic compounds hydrogenated form of aromatic compounds have relatively good thermodynamic properties for the more challenging, endothermic dehydrogenation reaction. This can be explained by the stability gained on formation of the aromatic system (58). In contrast, dehydrogenation of bonds outside of such an aromatic system is difficult (even if the double bond formed can become conjugated with the system). For instance, dehydrogenation of ethylcyclohexane would form ethylbenzene not styrene as a result of thermodynamic limitations (58).

Importantly, the high hydrogen loading capacities of LOHCs enable high hydrogen storage and transport efficiencies, increasing the economic viability of the technology. This is highlighted by previous research that identified several potential carriers which meet the objectives for storage capacity and volumetric energy density, set by the United States Department of Energy (US DoE), as 6.5 wt% and 1.7 kWh l^–1, respectively (7). However, it is important to consider that the respective hydrogenation and dehydrogenation reactions may not always go to 100% completion.

As noted above, catalysts are required to facilitate hydrogen loading and unloading via hydrogenation and dehydrogenation reactions. Despite some reports detailing advantages of homogeneous catalysts (including lower operating temperatures and improved dehydrogenation product specificity), heterogeneous catalysts are considered preferable for both reactions in large-scale applications (26, 59).

A great majority of current studies examine the hydrogenation and dehydrogenation steps individually, often employing different catalysts to complete the two transformations. Most commonly, platinum group metal (pgm) catalysts, namely platinum, palladium or ruthenium-based, are employed (7, 30, 60, 61).

The LOHC technology is also an attractive option for stationary on-site energy storage (on-grid and off-grid), enabling long-term storage of large amounts of energy, such as seasonal storage and energy production buffering (26). In this case, a catalyst which allows both hydrogenation and dehydrogenation to be carried out in the same reactor through altering process conditions (such as temperature and pressure) can be deployed. The development of such a bifunctional catalyst would be an attractive research objective and facilitate a lower capital expenditure since a set-up in which only a single reactor and its associated pipework is necessary. From a process safety perspective, the ambient hydrogen storage pressures facilitated with the use of LOHC technology are an improvement on using a pressurised hydrogen storage tank.

Since the catalytic hydrogenation and dehydrogenation processes are exothermic and endothermic reactions, respectively, the hydrogenation step is typically performed at lower temperatures (100–240°C) and higher pressures (10–50 bar) than the dehydrogenation step (150–400°C, atmospheric pressure) (62). Thus, a higher level of heat is needed for the dehydrogenation reaction, which is often provided by an external heating source (62). To improve efficiency, system integration and intensification can be achieved, with the heat produced in the hydrogenation step being used to drive the dehydrogenation reaction when using LOHCs for on-site hydrogen storage (63). This would not be possible when the hydrogen is produced in a different location to where it is needed.

As discussed, a timely migration away from fossil fuel reliance is of the utmost environmental importance, and is achievable with a combination of renewable energy, production of green hydrogen via electrolysis and hydrogen storage and transportation technologies. However, implementation of a LOHC hydrogen storage and transportation system also has advantages over simply using compressed hydrogen, or direct employment of renewable electricity. On the contrary to storage and transportation of compressed hydrogen, the small quantity of gas present in a LOHC system minimises the risk of explosion and facilitates the safe handling of hydrogen. Importantly, this allows for the long-term storage of large amounts of loaded LOHCs at ambient conditions, without the loss of hydrogen or negative impact upon carrier storage density, which has been reported with the use of compressed hydrogen and alternative storage technologies such as metal hydrides (64).

The elimination of pressure-related hazards also makes LOHCs suitable for long-distance transport. The liquid state of the carriers enables existing infrastructure, originally built for crude-oil transportation such as pipeline networks, to be used (1). However, unlike liquified or compressed hydrogen, the term ‘infrastructure’ also includes the use of ships and trucks, which enables worldwide transport of hydrogen using LOHCs. This too is often considered as one of the main advantages of the LOHC technology, meaning the dehydrogenation plant required to release hydrogen from LOHCs does not need to be located within proximity of the hydrogen production site. Thus, the LOHC technology allows green hydrogen release (and indirect renewable energy use) in locations which are not best suited for renewable energy production. Moreover, as LOHCs are in the liquid state, they resemble current fossil fuel-based energy vectors, such as diesel and gasoline. Given societal familiarity with these systems, it is predicted that public acceptance of the LOHC technology will be increased in comparison to concepts such as metal hydrides, where public understanding is limited (7). This potential for greater acceptance of LOHC systems is considered advantageous (7).

In theory, the reversibility of the hydrogen loading and unloading processes allows LOHC to be used continually without replacement. Practically, however, this is unlikely due to LOHC material losses from side-reactions and incomplete unloading reactions (1). The choice of LOHC is also critical: for maximum efficiency, both the dehydrogenated and hydrogenated form must remain in their liquid state throughout the cycle. For instance, if the dehydrogenated form of the carrier material is a solid at ambient conditions, transport complications arise as it cannot be pumped through a pipe or into a truck or ship. To maintain the liquid state, incomplete hydrogen unloading or dilution of the LOHC would be required, reducing storage and transport efficiencies (7).

Moreover, the toxicity of the LOHCs themselves must be considered and evaluated in terms of projected applications; some LOHCs, such as benzene and toluene, have a toxicity so great that their use in practical applications is unfeasible, despite attractively high hydrogen storage capacities (65).

The choice of carrier can also influence the overall cost of the technology in other ways. Aside from the obvious cost associated with purchasing the LOHC feedstock, the released hydrogen from some carriers, like 1,2-dihydro-1,2-diazaborine, requires further hydrogen purification steps, while others (for example, dibenzyltoluene) have higher dehydrogenation heating demands (1). Both factors increase energy consumption. Moreover, the cost of transporting the unloaded LOHC via ships back to the site of hydrogenation (i.e. for hydrogen loading) should be taken into account. Mainly long-distance transportation of renewable hydrogen in LOHCs would be more economically viable than transportation of compressed hydrogen, which is more suited to transport over short distances using the existing pipeline infrastructure (25). Methanol has also been reported to be economical for long-distance transport, with some studies suggesting that methanol can be a more cost-effective option than LOHCs (66). The same applies to ammonia (41). The Committee on Climate Change has reported that using ammonia as a hydrogen carrier for long-distance transport could be economically viable (25, 67). One of the key disadvantages of the LOHC technology is additional costs required for transporting of the LOHC loaded with hydrogen to end users followed by a transport of the dehydrogenated LOHC back to a chemical plant for loading with hydrogen.

To improve economic viability of the LOHC technology, additional developments into optimising the LOHC technology might be needed. Reducing energy intensity during loading and particularly unloading of the LOHCs with hydrogen and efficient system integration can contribute to the cost reduction of the LOHC technology. In our following work, we discuss efficient system integration and potential deployment of the LOHC technology within different industries (28) and provide a detailed analysis of the most promising LOHC candidates, catalysts used for hydrogenation as well as dehydrogenation of LOHCs along with operating conditions (29). Efficient system integration of the LOHC technology as well as selection of the most suitable LOHC system, optimisation of reaction conditions and catalysts might improve economic viability and facilitate widespread commercial deployment of the LOHC technology.

Reactor configuration can also contribute significantly to the overall efficiency of the LOHC technology. The particular challenge for the LOHC technology is the hydrogen release from hydrogen-rich LOHC systems during the endothermic dehydrogenation process which is combined with the reaction mixture volume expansion. As an example, 1 ml of fully hydrogenated dibenzyltoluene can release more than 650 ml of hydrogen (68). A range of reactor types have been proposed for dehydrogenation of LOHCs, such as fixed-bed, continuous stirred tank reactor batch-type, tubular, spray-pulsed, pressure-swing and three-dimensional (3D) structured monolith reactors, among others (Figure 3) (68).

Fig. 3.

Tubular-type reactors are often discussed for the LOHC technology. To optimise the process, different orientations of the tubular reactors have been evaluated (69). Tubular reactors can have either a horizontal or vertical orientation and have two possible operational modes: (a) LOHC flowing through the central tube and heat transfer fluid in the annulus; (b) heat transfer fluid in the central tube and the LOHC in the annulus (70). The heat transfer fluid enables external heating of the LOHC (69). In comparison to their horizontal equivalents, vertical tubular reactors have the advantage of a more even heat distribution, as a result of rising bubbles mixing the LOHC (70). Moreover, several vertical tubular reactors are suitable for both the hydrogenation and dehydrogenation reaction. Often, this is not the case for horizontal tubular reactors, where the interfacial region between the gaseous hydrogen and LOHC is often too small to achieve the mass transfer rates required for a hydrogenation process (69). Vertical dehydrogenation tubular reactors can also be constructed to produce either a co-flow or counter-flow between the LOHC and released hydrogen, offering more flexibility in a system design. In a study using N -ethylcarbazole as the carrier, different orientations of the reactor (horizontal/vertical) were studied, revealing that the maximum power densities are similar in both orientations, but radial heat transfer is improved compared to that in a horizontal orientation (69). Despite this, several examples of horizontal tubular reactors can be found, with advantages including convenient removal of hydrogen from the top of the reactor and prevention of a multi-phase flow (71).

Helical reactors are more complex and comprise the LOHC flow in the central tube, with the heat transfer fluid in the surrounding shell space (70). A more effective heat transfer can be achieved with a helical reactor than a tubular reactor, as a result of a ‘shear’ formation, while the rate of reaction (hydrogen release) is also increased. A double helical reactor is also discussed for the LOHC technology (70). While this enables a significant saving in terms of space, the construction and assembly is typically complex and expensive. The deployment of a hot pressure-swing reactor was demonstrated for stationary hydrogen storage in which hydrogenation and dehydrogenation were performed within the same reactor using the same catalyst with dibenzyltoluene as the LOHC (72). This demonstrates the benefit of potential capital expenditure and operational expenditure savings.

Structured reactors, such as monolithic, 3D printed or foam reactors, are an attractive research area within the LOHC technology as they provide high heat conductivity, allowing for a good heat input for the endothermic dehydrogenation reaction (73, 74). Furthermore, the high porosity of such reactor systems lowers pressure drop, on the contrary to fixed-bed reactors where catalyst pellets are typically used, and facilitates an efficient hydrogen removal. This is important for dehydrogenation of hydrogen-rich LOHCs, during which significant reaction mixture volume expansions take place upon hydrogen release (for example, 1 ml of fully hydrogenated dibenzyltoluene can release more than 650 ml of hydrogen) (68, 73).

Depending on the intended application, the hydrogen purity released from LOHCs during dehydrogenation should be considered. Given that high temperatures are required for dehydrogenation of LOHCs, side products might be formed during hydrogen release, contaminating the hydrogen stream. In such cases, separation systems are required to purify the hydrogen, which increases the cost of the LOHC technology. The use of membrane reactors, which combine the benefit of shifting the equilibrium of dehydrogenation to the product side with integrated purification of the released hydrogen, have been reported in the literature for the LOHC technology (71, 75–80). Byun et al . performed the technoeconomic assessment of methylcyclohexane dehydrogenation in both a packed-bed reactor and a membrane reactor (78). This study demonstrates the cost effectiveness of an membrane reactor: the unit hydrogen production cost of a packed-bed reactor is US$11.76, US$9.50, US$8.50 and US$8.08 and that of an membrane reactor is US$9.37, US$7.43, US$6.58 and US$6.23 for hydrogen production capacities of 30 m³ h^–1, 100 m³ h^–1, 300 m³ h^–1 and 700 m³ h^–1, respectively (78).

A lower temperature of operation for the dehydrogenation reaction, deploying a reactive distillation column under reduced pressure, could also allow for an increase in the efficiency of the LOHC system (81). This was demonstrated for perhydrobenzyl toluene dehydrogenation. A lowering of reaction temperature reduces the energy intensity of the overall process and facilitates a simpler and more effective integration of heat with waste heat sources or from subsequent hydrogen utilisation steps (81). Another recent study demonstrated that an electrochemical hydrogen compression (EHC) unit, which is connected to the LOHC dehydrogenation unit, reduces hydrogen pressure and shifts the thermodynamic equilibrium towards dehydrogenation (82). This accelerates the hydrogen release for the perhydrodibenzyltoluene LOHC and lowers dehydrogenation temperatures to 240°C. In addition, the EHC unit produces high value compressed hydrogen and purifies hydrogen contaminated with impurities such as traces of methane (82).

Several approaches to effective hydrogen storage technologies must be explored in parallel to facilitate a smooth transition to the hydrogen economy. LOHCs for hydrogen storage and transportation are an attractive option for storing and transporting green hydrogen. Key advantages of the LOHC technology are: high storage capacity compared to alternative hydrogen storage technologies such as MOFs; no hydrogen losses during extended storage periods; ambient storage pressures and compatibility with existing fossil fuel infrastructure (pipes, ships, trucks). LOHCs ideally have high, reversible hydrogen-loading capacities, enabling large quantities of hydrogen to be stored in the liquid carrier. Importantly, LOHCs facilitate the storage and intercontinental transport of hydrogen and can capitalise on infrastructure originally constructed for fossil fuels. Thus, over long distances, the transport of hydrogen using the LOHC technology has the potential to be economically viable in comparison to other hydrogen storage and transportation technologies. Though costs required for transporting of the LOHC loaded with hydrogen to end users followed by a transport of the dehydrogenated LOHC back to a chemical plant for reloading with hydrogen should be considered. Furthermore, additional developments and research into optimising the technology might be needed to improve economic viability of the LOHC technology. Reducing energy intensity during loading and particularly unloading of the LOHCs and efficient system integration can contribute to cost reduction of the LOHC technology.