Reversible vs. Irreversible Conversion of Hydrogen: How to Store Energy Efficiently?


  • Karsten Müller,

    Corresponding author
    1. Lehrstuhl für Thermische Verfahrenstechnik, Institut für Chemie- und Bioingenieurwesen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen (Germany), Fax: (+49) 9131-8527441
    • Lehrstuhl für Thermische Verfahrenstechnik, Institut für Chemie- und Bioingenieurwesen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen (Germany), Fax: (+49) 9131-8527441
    Search for more papers by this author
  • Jin Geng,

    1. Lehrstuhl für Thermische Verfahrenstechnik, Institut für Chemie- und Bioingenieurwesen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen (Germany), Fax: (+49) 9131-8527441
    Search for more papers by this author
  • Prof. Dr. Wolfgang Arlt

    1. Lehrstuhl für Thermische Verfahrenstechnik, Institut für Chemie- und Bioingenieurwesen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstraße 3, 91058 Erlangen (Germany), Fax: (+49) 9131-8527441
    Search for more papers by this author


Chemical conversion is a key technology for hydrogen storage. Two basic options exist in this field: reversible storage using organic carrier materials (energy carrying compounds) and irreversible conversion into hydrocarbon fuels (gas-to-fuel). It has been shown that reversible storage exhibits significantly higher overall storage efficiency for the electricity-to-electricity storage process. The high storage density and existing infrastructure for hydrocarbon fuels are advantages of the Fischer-Tropsch-process whereas process complexity favors reversible conversion.


The world is currently facing a situation where growing energy demand meets shrinking resources for fossil fuels. Improvements in energy efficiency make it possible to reduce the growth of the demand, but in the long term, a gap between the demand and availability of resources will become unavoidable. To close this gap, the use of alternative energy resources is necessary. For some forms of renewable energy a conversion into electricity or mechanical work is possible on demand. Power plants that run on biogas for example could be used as production sites for electric power with variable capacity. However, the efficiency of the energy production per unit area for biomass is not sufficient to cover the total energy demand of the world′s population.1 To overcome this low areal efficiency a more direct use of the energy derived from the sun should be employed, for example, by utilizing photovoltaics or wind power. These energy sources suffer from the severe disadvantage that power generation does not happen on demand but more or less randomly with only limited predictability. Consequently, medium- and long-term energy storage units will become crucial elements of the energy systems of the future.

Many potential storage technologies are either already available or under development today. However, only chemical energy storage seems appropriate for large-scale, long- and medium-term storage, because other options suffer from low storage densities, limited availability, or losses during the storage period.2 To store electrical energy, hydrogen can be produced from electrolysis and subsequent power generation can be performed by re-oxidation of the hydrogen to water. The main challenge in this concept is the storage of the hydrogen. The gravimetric storage density of hydrogen is excellent because of the low molar mass, but the volumetric storage density at ambient conditions is unacceptably small. To increase the volumetric density, measures must be taken to reduce the specific volume, which can be achieved by either cooling or pressurizing. A temperature difference of approximately 270 K is created if hydrogen is cooled down to liquefaction. Heat transport into the tank therefore is unavoidable even in the case of very good insulation. To prevent the build-up of pressure, a permanent release of evaporated hydrogen is necessary, which causes continuous losses of hydrogen during the storage period. The alternative of high-pressure storage only reaches lower volumetric storage densities, and the heavy, pressure-stable tanks significantly lower the overall gravimetric storage density.

To overcome these drawbacks, it is reasonable to convert the hydrogen into another chemical form. This type of conversion can be differentiated into two principal types: reversible and irreversible conversion (comparison in Figure 1). In the case of reversible conversion, an unsaturated organic compound is hydrogenated and subsequently dehydrogenated when energy is needed, making the carrier compound effectively recyclable. These compounds are called liquid organic hydrogen carriers (LOHC) and are a subclass of energy carrying compound. A number of different compounds are under research for use as LOHCs such as toluene,3 azaborines4, 5 and N-ethylcarbazole.2, 6 For irreversible conversion the hydrogen reacts with a carbon source (usually carbon dioxide) to form a fuel that is combusted without having previously released the hydrogen. Typical products of such conversions are methane,7 methanol,8 and long-chain alkanes.9 A border case is the reforming of methane or methanol before power generation to regain the hydrogen, and a significant difference to the LOHC concept is that the produced carbon dioxide is not recycled but instead released to the atmosphere. The concentrated carbon dioxide stream that was used for the conversion of the hydrogen is therefore lost.

Figure 1.

Conceptual diagram of energy storage by reversible and irreversible processes for the conversion of hydrogen.

In this study, we compare the two general approaches for chemical energy storage using thermodynamic modeling. We chose hydrogen storage using N-ethylcarbazole6 as a model compound for the reversible conversion and the production of liquid alkanes (Fischer–Tropsch synthesis) as a model process for irreversible conversion. It has to be noted that renewable-energy-based Fischer–Tropsch processes can also be based on biomass conversion, as has been demonstrated by various research groups.1013 The use of biomass-derived synthesis gas has some advantages over reacting hydrogen (produced from electrolysis) with carbon dioxide, such as eliminating the need for the reverse water–gas-shift (RWGS) reaction. Nevertheless, we focused on the later approach, since only this is an option for the storage of electrical energy.

Aspects of the Two Concepts

Both types of chemical hydrogen conversion for energy storage purposes have their advantages and disadvantages (summarized in Table 1).

Table 1. Comparison of the principle criteria for hydrogen storage by chemical conversion.

Reversible conversion

Irreversible conversion

  1. [a] N-ethylcarbazole. [b] gasoline.

most probable power generation technology

fuel cell

Otto/Diesel engine

energy density

1.9 kWh kg−1 1.5 kWh L−1[a]

12 kWh kg−1 8.8 kWh L−1[b]

additional feedstocks

LOHC recycled in a closed cycle

concentrated carbon oxide stream

limitation of capacity

amount of produced LOHC

availability of concentrated carbon oxide stream

CO2 emission

zero emission

local: high global: neutral

CO/NOX emission

no CO or NOx emission

depends on the combustion process

usability of existing infrastructure

minor adjustments required

no significant adjustments required

main field of application

storage option for electrical power grids

suitable for mobility applications

Conversion and storage density

Electrical or mechanical power can be produced by different methods in both cases. Since LOHCs release molecular hydrogen, the most important technology in this field will most likely be the fuel cell. For the case of a Fischer–Tropsch reaction, the power generation will most likely be carried out using an Otto or Diesel engine, and this must be kept in mind if the storage densities are compared. The amount of energy per mass and volume stored by gasoline exceeds the amount stored in LOHCs by approximately a factor of 5–6. The efficiency of fuel cells is usually higher than that of engines, and thus, the effective difference becomes slightly smaller. If, for example, methane is produced by irreversible conversion, the volumetric storage density becomes smaller than in the case of gasoline or diesel.

Additional feedstock

A significant difference between the two approaches concerns the feedstock needed for the process. In a LOHC-based reversible-conversion process, the carrier material is recycled and hydrogenated again during the next storage cycle. In an irreversible conversion process, a concentrated stream of a carbon oxide is required. Carbon monoxide is ultimately necessary for this purpose, but usually only carbon dioxide is available (e.g., from processed biogas or flue gas). The carbon dioxide must be converted to carbon monoxide by a RWGS reaction. In this step, 1 mol hydrogen is consumed for 1 mol carbon dioxide and additional energy input is necessary because of the endothermic nature of the reaction.

Storage capacity

The maximum storage capacity in both cases is theoretically unlimited, but certain constraints exist with regard to the feedstocks. In the case of LOHCs, the absolute amount of energy that can be stored is limited by the amount of carrier that has been produced, and the rate of energy storage (energy per unit time) is only somewhat limited by technical constraints. In the case of irreversible conversion, the maximum amount of energy that can be stored is nearly unlimited (as long as the required storage tanks are available). The amount of energy that can be stored per time in this case is limited by the availability of concentrated CO2. As the separation of CO2 from the air is not reasonable because of thermodynamic reasons, this can constitute a severe constraint.


From an overall perspective, both approaches are carbon neutral. The LOHC concept does not cause any CO2 emissions, and the irreversible conversion causes localized CO2 emissions but is carbon-neutral if the entire system is considered as a whole. Additionally, the emission of carbon monoxide (CO) and nitrogen oxides (NOx) should be considered, depending on the combustion process. In the fuel cells proposed in the LOHC concept neither CO nor NOx are produced, whereas the combustion of hydrocarbons in an engine leads to the formation of CO as well as NOx. Nevertheless, if an appropriate flue-gas treatment is employed, the respective emissions can be viewed as being negligible for both cases.


The energy carriers in both cases are liquids and thus easy to handle. In the case of methane production for irreversible storage, the product is a gas. However, decades of technical experience with natural gas are likely to make this a tolerable disadvantage for the industry. For reversible conversion using organic carriers, carrier solidification is an issue for some of the potential carriers. Therefore, a current topic of research focuses on maintaining the liquid state of the carrier, for example by modification of the molecular structure.

Field of application

The primary differentiating factor between the two processes will be in the fields of application. The major application of LOHC-based energy storage would likely be the stationary storage of surplus electric power with the intent to feed power back to the electrical grid in times of surplus demand. Systems for the use of LOHCs in small, mobile applications like cars are currently still under development. The high energy density and the already-existing infrastructure for gasoline make irreversible conversion the most likely option for the mobility sector. The use of Fischer–Tropsch products as a storage system for power grids would likely be unfavorable because the high demand of the mobility sector will constitute a strong competition for the limited resources of liquid hydrocarbons. Thus, only irreversible conversion by methanization could have the potential to significantly contribute to energy storage for stationary applications.

Process Simulation of Overall Efficiency

Process simulations have been performed to evaluate the overall efficiency of the processes using Aspen Plus V 7.3. Energy demand for the transportation of the LOHCs or the Fischer–Tropsch products has not been taken into account in this simulation, because this amount of energy is small compared to the total energy transported in both cases considered.

Reversible conversion

The reversible conversion process consists of two separate steps: one for hydrogenation and one for dehydrogenation (as illustrated in Figure 2). In the hydrogenation step, the dehydrogenated carrier is mixed with hydrogen and preheated to 179 °C by cooling the hot stream exiting the hydrogenation reactor. Because of the strong exothermic nature of the reaction, it is proposed that the missing heat to reach the reaction temperature of 200 °C could be supplied by feeding the stream directly into the reactor. The reactor still produces about 50 kJ of waste heat per mol hydrogen, and because of the high temperature, this waste heat could be used for domestic heating, for example. Integration of this heat into the endothermic dehydrogenation process was not considered, because the temperatures do not match, and significant gaps in time can occur between the two process steps. Hydrogenation is carried out at elevated pressures of between 30 and 50 bar. Since electrolysis can be carried out at elevated pressures without a loss of efficiency we did not consider additional energy demand for compression.

Figure 2.

Process flow sheet for the reversible conversion of hydrogen.

In the dehydrogenation step, the hydrogen carrier is preheated (preheater 1 in Figure 2) by cooling the hot liquid stream coming out of the dehydrogenation reactor. A second preheater is used to heat the stream to the reaction temperature of 250 °C. Because of the endothermic nature of this process, additional energy must be supplied to the reactor. Dehydrogenation is carried out at 1 bar. The dehydrogenated carrier is later removed by partial condensation and the hydrogen is converted to water in a fuel cell.

An efficiency of 70 % relative to the lower heating value of hydrogen was assumed for the electrolysis and 55 % for the fuel cell. The heat demand of the dehydrogenation step is 86.0 kJ per mol hydrogen. The overall efficiency of the process thus can be calculated to be approximately 30.8 %. Similar results are obtained for reversible hydrogen conversion using other organic hydrogen carriers. Using toluene as a carrier compound, we calculated the overall efficiency to vary between 30.4 % and 31.2 % (depending on the process used). For azaborines we could show in previous work that an overall efficiency of up to 30.7 % can be reached.4

Irreversible conversion

For the irreversible-conversion storage process, a carbon resource is required, and CO2 instead of CO has been considered here to achieve a CO2-neutral process. An appropriate model must take the differences between these feedstocks into account, particularly the negative effect on catalytic performance, and the influence on the product selectivity.14, 15

The products of Fischer–Tropsch synthesis are very complex. More than a hundred different components, including paraffins, olefins, alcohols, and some aromatic compounds can be contained in the product stream. However, alkanes normally dominate hydrocarbon production in the Fischer–Tropsch process. As a simplification we therefore considered alkanes to be the only products of the reaction, because our focus was not a detailed modeling of the reaction mechanism but of the respective energy balance. Additionally, the energy demand for further byproduct removal is low compared to the total energy demand, and separation is not necessary because traces of alcohols or aromatics in the gasoline are acceptable and can even produce positive effects.

Generally, the product distribution of hydrocarbons formed during the Fischer–Tropsch process follows an Anderson–Schulz–Flory distribution (ASF)16 (see Figure 3), which can be expressed as(1)

equation image(1)
Figure 3.

Product distribution in Fischer–Tropsch reactions as a function of the chain-growth probability according to Ref. 12.

with Wn being the weight fraction of hydrocarbon molecules, containing n carbon atoms and α being the chain growth probability.

Although CO2 would be the preferred feedstock for the irreversible chemical storage of hydrogen, the direct conversion from CO2 as the only feedstock to liquid fuels is not the state-of-the-art technology. For this case many experiments show a dominancy of methane production instead of long chain alkane growth.17, 18

To achieve a realizable process, we considered a two-stage process. Carbon dioxide is converted to CO and H2O by using a high-temperature RWGS reactor and the produced gas mixture is fed to the Fischer–Tropsch reactor after removing the water.

In the presence of CO, CO2 is negligibly hydrogenated; it has only a minor effect on product distribution and behaves as an inert species.18 Synthesis using a gas mixture of H2/CO/CO2 has already been researched extensively, because it is also desirable for a conventional—that is, coal based—Fischer–Tropsch process to use the raw feed stream after coal gasification, without employing an additional CO2 separation process.

A description of the process simulation is shown in Figure 4. H2 and CO2 are mixed with recycling streams and compressed to a pressure of 30 bar. The RWGS reaction is carried out in reactor R1, which was modeled by an R-Gibbs Reactor from Aspen Plus. The produced syngas stream is cooled in a heat exchanger block and then flashed in block F1 to remove the water.

Figure 4.

Process flow sheet for the irreversible hydrogen conversion.

A syngas mixture with a CO2/CO/H2 molar ratio of 0.65/2.82/6.52 enters the Fischer–Tropsch Reactor (R2), which is modeled as an R-Yield reactor. Fortran user block models were used, which contained the yield calculation of conversion processes based on the ASF-function [Eqn. (1)].

The chain growth probability (α) is assumed to be 0.9, and the conversion of CO to be 70 %, which is a typical value for the low-temperature Fischer–Tropsch reaction. The product stream contains a hydrocarbon mixture of various alkanes, unreacted CO, H2, and CO2.

The product stream is flashed at different temperatures (F2 and F3). The resulting liquid stream, which consists of a water-rich- and an organic phase, is separated by decantation. The organic phase, with wax as the dominant product, is converted to a heavy-hydrocarbon product downstream in a hydrocrack reactor (R3).

The gaseous product is cleaned by using two separate steps to remove the H2, CO, and CO2. The hydrogen is recovered by pressure swing adsorption (S1). CO and CO2 are separated from the hydrocarbon gas by using a cryogenic treatment at −65 °C.

The hydrocarbons derived from the separation units S2 und D1 are separated in a distillation column (C1) into light- and heavy-hydrocarbon products. A part of the light hydrocarbons (mainly C1–C4) is considered here as exhaust, which could also be condensed by using high pressure or low temperature. However, because they only occur in small amounts, they have not been taken into account. Heat integration has been assumed for the whole process by using a pinch point analysis.

The overall efficiency of the irreversible-conversion process has been evaluated by assuming an efficiency of 70 % for the electrolysis step as before and an energy demand of 2 MJel kg−1 CO2 for the separation of CO2 from the flue gas.7 The efficiency of the irreversible conversion can be calculated as the ratio of the energy output divided by the energy input:(2)

equation image(2)

for which the symbols are defined in the section below.

We have modeled the reconversion of the fuel into work using common combustion engines such as those used in cars. The light hydrocarbons (gasoline) are used in an Otto motor with an efficiency of ηgasoline=30 % and the heavy hydrocarbons (diesel) in a Diesel motor with an efficiency of ηdiesel=40 % (both values for full load operation according to Ref. 19). The overall efficiency of the whole process thus can be calculated to be approximately 16.7 %. This is in the same order of magnitude as the efficiency of methanization with reconversion in a gas motor (21.4 %).7 For methane, other options for electricity production are possible (e.g., a combined-cycle power plant) and therefore higher overall efficiencies are conceivable. However, the efficiencies of irreversible conversion processes remain lower than those of reversible conversion.

Results of Energetic Evaluation

The overall efficiency for electricity-to-electricity storage using reversible hydrogen conversion based on LOHCs was estimated to be approximately 30%. The overall efficiency for the irreversible conversion process was approximately 20 % (16.7 % for Fischer–Tropsch; 21.4 % for the Sabatier process to produce CH4 from H2 and CO2). At first, both values seem rather poor, but certain aspects have to be kept in mind. First, the recovery of waste heat (storage-heat coupling) enables the process to reach a higher total efficiency than that of pure electricity-to-electricity storage. Furthermore, hydrogen-based energy storage (with its inherent efficiency limitations) is unavoidable if the share of renewable energies is expanded to a significant level.

Nevertheless, the efficiency of the storage process might become a crucial parameter during the transition from the current energy systems. The reversible storage process can be regarded as the more efficient alternative because of its lower losses. The LOHCs therefore seem to be the most reasonable approach for the compensation of fluctuating power production from renewable energies. This is primarily caused by the higher electricity-to-electricity efficiency. Moreover, recovery of heat (e.g., for residential heating) from decentralized hydrogenation units is more practical than from large, and thus centralized, Fischer–Tropsch facilities thereby leading to higher efficiencies for the combined heat-and-electricity process.


Both the reversible and irreversible processes for the conversion of hydrogen exhibit different advantages, and thus, a conclusion concerning the best option must distinguish between the different fields of application.

For mobile applications, storage density is a crucial criterion. This criterion, and the fact that the existing infrastructure for mobility is currently based on hydrocarbons, makes the irreversible conversion of hydrogen to gasoline or diesel currently the best option for mobility. There are many new developments in the field of LOHCs, whereas no groundbreaking advances are on the horizon for carbon-based cars. The situation for the mobile use of stored energy might therefore change in the future.

In stationary applications, the overall efficiency is of considerably higher importance than storage density. However, the reversible conversion process seems clearly superior for energy storage to compensate the fluctuating power generation from renewable energy sources. The low efficiency of Fischer–Tropsch or related processes puts these storage options at a disadvantage for such applications, whereas organic hydrogen carriers appear to be a more suitable solution for stationary applications because of their higher efficiency.



  chain growth probability




  lower heating value


  liquid organic hydrogen carrier


  amount of substance




  weight fraction