Electrified methane reforming: A compact approach to greener industrial hydrogen production

See allHide authors and affiliations

Science  24 May 2019:
Vol. 364, Issue 6442, pp. 756-759
DOI: 10.1126/science.aaw8775

More-efficient heating

Large-scale production of hydrogen through steam reforming directly produces CO2 as a side product. In addition, the heating of reactors through fossil-fuel burning contributes further CO2 emissions. One problem is that the catalyst bed is heated unevenly, which renders much of the catalyst effectively inactive. Wismann et al. describe an electrical heating scheme for a metal tube reactor that improves the uniformity of heating and catalyst usage (see the Perspective by Van Geem et al.). Adoption of this alternative approach could affect CO2 emissions by up to approximately 1% of global emissions.

Science, this issue p. 756; see also p. 734


Electrification of conventionally fired chemical reactors has the potential to reduce CO2 emissions and provide flexible and compact heat generation. Here, we describe a disruptive approach to a fundamental process by integrating an electrically heated catalytic structure directly into a steam-methane–reforming (SMR) reactor for hydrogen production. Intimate contact between the electric heat source and the reaction site drives the reaction close to thermal equilibrium, increases catalyst utilization, and limits unwanted byproduct formation. The integrated design with small characteristic length scales allows compact reactor designs, potentially 100 times smaller than current reformer platforms. Electrification of SMR offers a strong platform for new reactor design, scale, and implementation opportunities. Implemented on a global scale, this could correspond to a reduction of nearly 1% of all CO2 emissions.

The synthesis of important chemicals such as hydrogen and ammonia has a substantial CO2 footprint because the heating of the processes often relies on the combustion of hydrocarbons. One of the largest endothermic processes is the production of hydrogen by steam-methane reforming (SMR), which accounts for ~50% of the global hydrogen supply, where all hydrogen production is estimated to account for 3% of global CO2 emissions (1, 2). In this strongly endothermic reaction, natural gas reacts with steam according to the following equations:


Where ΔHr° is standard reaction enthalpy. Heat is typically supplied to the reaction by combustion of a mixture of natural gas and potential off-gases from the synthesis. In total, conventional SMR produces 6.6 to 9.3 metric tons of CO2 per metric ton of H2, of which 17 to 41% is the direct product of hydrocarbon combustion (2, 3).

Today, a large-scale industrial SMR reformer consists of an array of more than 100 10- to 14-m-long tubular reactors in a large furnace, with gas burners positioned for an optimal distribution of heat among the reactor tubes (46). The combustion must occur considerably above the reaction temperature to generate the necessary inward heat flux, as illustrated by the temperature profile in Fig. 1A (5, 7). Because of limited thermal conductivity across the SMR catalyst and reactor walls, transporting the heat necessary to drive the reaction is a natural limitation (Fig. 1A), and typically less than 2% of the furnace volume contains catalyst (5, 8). Intrinsic catalytic activity is typically not a limiting factor for industrial reforming (9). Instead, the low thermal conductivity combined with a strongly endothermic reaction creates steep temperature gradients across the catalyst, leading to poor catalyst utilization and increasing the risk of detrimental carbon formation (1012).

Fig. 1 Heating principles.

(A) Conventional fired reactor. (B) Electric resistance–heated reactor. Characteristic radial length scales and temperature profiles are shown across the heat source, reactor wall (gray), and catalyst material (green). In (B), the heat source and reactor wall are one. Illustrations are not to scale.

For decades, thermal conductivity of SMR has been the subject of research. Efforts include using catalysts with higher thermal conductivity (13), lowering the temperature of SMR by shifting the equilibrium (1417), obtaining shorter characteristic length scales through μ-reactors (18, 19), performing room-temperature reactions using plasma (20), or employing direct heating of magnetic catalysts by induction (21). Alternatively, electrical heating of an integrated catalytically coated heating element enables reactor temperatures exceeding what is feasible in conventional reactors (22), and allows substantially improved temporal response, pushing start-up times to within minutes (23). However, despite decades of research, no alternatives with lower CO2 emissions have been implemented at the industrial scale.

This work describes a high-performing, fully electrically driven reformer based on direct resistive (ohmic) heating (Fig. 1B), which is scalable to industrial conditions and capacities. The intimate contact between the electric heat source and the catalyst enables energy to be supplied directly to the catalytic sites, removing thermal limitations and providing well-defined control of the reaction front. Electrification removes the fired section, substantially reducing reactor volume, CO2 emissions, and waste-heat streams. This provides a disruptive advantage to existing industrial reformers, enabling the production of “greener” hydrogen for the large-scale synthesis of indispensable chemicals such as methanol, ammonia, and biofuels (24, 25).

For this work, we prepared a laboratory-scale reactor based on an FeCrAl-alloy tube, which was chosen for its temperature-independent electrical resistance and coated with an ~130-μm nickel-impregnated washcoat on its interior (26). Copper sockets were mounted at opposite ends of the external surface of the reactor tube, and resistive heating was accomplished by applying an AC current along the tube (Fig. 2A). This allowed a direct heat supply to the catalytic washcoat (Fig. 2B). A section of the coat was removed at both ends of the reactor to obtain a quantified length of the catalyst (Fig. 2C) and to prevent reverse reaction toward the outlet. However, a thin residual layer of catalytically active coat (<5 μm) was present at the lower section of the reactor (fig. S1) as a result of the impregnation method of the material. Temperature profiles were measured with multiple thermocouples spot welded to the tube (Fig. 2D). The entire reactor was encapsulated in high-temperature insulation material.

Fig. 2 Laboratory-scale resistance-heated reactor.

(A) Resistance-heated reactor setup. The illustration is not to scale. (B) Cross-sectional illustration of the reactor in the coated region. (C) Axial cross-section of the reactor after experiments, showing the well-defined edge of the coat. (D) Axisymmetrical reactor cross-section, outlining the most relevant domains and thermocouple positions.

A feed mixture of CH4, H2O, and H2 (30/60/10) was preheated to 100°C to prevent condensation before entering the reactor. The experiments were operated 50 mbar above ambient, as the reactor was not prepared for pressure-bearing application.

A computational fluid dynamics model (CFD), including calculation of electric currents, thermal energy, fluid dynamics, mass transport, and reaction kinetics, was implemented to further understand the experiments and to extrapolate results to industrially relevant conditions. The computational model accurately describes the measurable values, such as external temperature and methane conversion (figs. S2 to S4).

Figure 3 shows experimental and computational data. The reactor can be divided into three sections. The first section, at the inlet, yields a rapid increase in temperature between the copper socket and coated zone (Fig. 3A) as the entire heat supply is used for heating the process gas (Fig. 3B). In the second section, the coated zone, the temperature initially drops because the endothermic reaction consumes more heat than supplied for the process (Fig. 3A). Hereafter, the temperature profile is close to linear, with a substantially smaller slope than in the first zone, as the endothermic reaction consumes large amounts of heat. In the third section, the outlet (Fig. 3A), the temperature increases again more rapidly, reaching a maximum of 800°C, before dropping promptly to 100°C. Near the end of the reactor, the copper sockets exchange heat with ambient conditions, facilitating rapid cooling.

Fig. 3 Experimental results and model predictions at ambient pressure.

(A) Axial temperature profile and methane conversion at 1.7 NL/min. The equilibrium temperature is the temperature at which a given gas composition is in thermodynamic equilibrium with respect to the SMR reaction (Eq. 1). (B) CFD-modeled thermal contours across the reactor. (C) Methane consumption rate for the innermost 50 μm of the coat (out of 128 μm), evaluated near inlet and outlet; compare with (A). (D) Measured exit temperature against methane conversion for the resistance-heated reformer for different gas flows.

Owing to the uniform supply of heat to the process, the nearly constant heat flux (fig. S5) ensures that the gas mixture is kept close to equilibrium throughout the entire catalytic length (Fig. 3A), as opposed to what is observed for conventional reformers (6, 8). This results in better utilization of the reactor volume and limits detrimental side reactions such as carbon formation (figs. S6 and S7). Radial thermal gradients (Fig. 3B) primarily arise from the convection in the reactor. Temperature difference across the coat does not exceed 2°C along the linear section of the temperature profile. There is no discernible temperature gradient across the reactor wall, a substantial benefit compared with a fired reformer, in which the temperature difference between the inner and outer wall of the tubular reactor can cause thermal stress, detrimental to mechanical strength and reactor lifetime (27).

Although internal diffusion limits the utilization of the catalyst, as the reaction quickly approaches equilibrium across the coat (Fig. 3C), the average catalyst utilization is 20% at the conditions shown in Fig. 3C, i.e., up to an order of magnitude higher than that reported for a heterogeneous catalyst for SMR (6, 28, 29). The most effective utilization of the catalyst is near the inlet, as lower temperature generates lower reaction rates (figs. S8 and S9). At the outlet, equilibrium is reached within the innermost 50 μm of coat, equivalent to 39% of the coat thickness. The improved catalyst utilization is primarily due to the absence of thermal gradient in the catalyst. Further optimization of the catalyst utilization is feasible; as shown in Fig. 3C, only 40 to 50 μm of a uniform coat is required for full conversion, increasing catalyst utilization up to 65%.

Because the flow is always completely laminar in the given process design [Reynolds number ≪2100 (30)], radial mass transport occurs solely by molecular diffusion to the surface of the catalyst, resulting in an external mass transport limitation (bulk to surface) that is correlated to the gas velocity (fig. S10). The external diffusion limit can be seen by the increased temperature required to reach equivalent conversion as the flow rate increases (Fig. 3D). Higher conversion may be achieved by increasing the reactor temperature, at the expense of increasing the temperature difference relative to equilibrium (Fig. 3D). As the process gas approaches full conversion, a vertical asymptote is observed because of increasing kinetic hindrance of the reaction. The vertical asymptote occurs at lower conversion for higher flow rates, thus limiting the maximum conversion achievable without altering the geometry or operational conditions, such as the pressure or steam-to-carbon ratio. For reference, an industrial SMR rarely operates above 90% conversion of methane.

An important benefit of the resistance-heated design is the possibility for exceptionally compact reactors (23). If we use the model developed in this work for a single tube and extrapolate it to several parallel reformer tubes matching the capacity of an SMR, we find that a conventional 1100-m3 side-fired reformer producing 2230 kmol H2/hour can be replaced with an ~5-m3 resistance-heated reformer (Fig. 4). Operating at similar conditions, the resistance-heated reformer has no risk of carbon deposition (fig. S11). The substantial volume reduction obtained for the resistance-heated reformer is achievable because integration of the heat source makes the furnace obsolete, thus removing a substantial portion of the reactor volume. Further volume reduction is envisioned if the geometry or operation conditions are optimized; however, this was not pursued in this study. It should be noted that the comparison is based on the SMR furnace, and does not include essential equipment such as combustion air blowers and waste heat section [6]. For the resistance-heated reformer, wiring and power supply are equivalently omitted for this comparison.

Fig. 4 Scaling opportunities for industrial production.

Modeled methane conversion for a resistance-heated reformer scaled to a capacity of 2230 kmol H2/hour compared with a side-fired SMR operating at 75.4% methane conversion. The comparison was done at industrial operating conditions (Tin = 466°C, Tout = 920°C, S/C = 1.8, Pout = 26.7 barg). The model is limited to a 20°C difference from the equilibrium temperature (26).

The electrification, uniform heating, and potential for exceptionally compact reactors present a disruptive approach to resolving CO2 emission issues and current constraints regarding design, operation, and process integration for hydrogen production by SMR. In addition to reducing CO2 emissions, implementation of the resistance-heated reactor into existing plants could offer alternative operation conditions, reducing the steam-to-carbon ratio, or operate at increased methane conversion, typically limited by carbon deposition and temperatures (i.e., material constraints). High methane conversion coupled with an alternative purification technology could even provide a local source of CO2 for other processes. With less need for heat recovery, resistance-heated reforming is efficient and applicable at many different sizes, promoting delocalization designs by using the existing and well-developed infrastructure of natural gas and potentially also biogas. Low thermal mass can also lead to reformers optimized for intermittent operation, following the fluctuations in availability of excess renewable energy with possible startup times in seconds (23, 26). The operating costs for an electrified reformer are directly related to the cost of electricity, natural gas, and CO2 taxes. Preliminary estimates indicate that a resistance-heated reformer would be on par with current fired reformers in regions with a high production of renewable electricity.

This work illustrates the disruptive opportunities achievable by electrification of fundamental industrial processes. With the swiftly decreasing cost of electricity from renewable sources, resistive heating is an environmental—and economically appealing—solution for providing the necessary heat for strongly endothermic industrial processes on the path toward a more sustainable society.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S12

Table S1

References (3141)

References and notes

  1. Materials and methods are available as supplementary materials.
Acknowledgments: Funding: This work was supported by Innovation Fund Denmark (IFD) under file no. 5160-00004B and research grant 9455 from Villum Fonden. Author contributions: All authors contributed substantially to this work. Competing interests: The authors declare no conflicts of interest. Data and materials availability: The catalyst can be made available under a material transfer agreement for Haldor Topsoe A/S. All other data are available in the main text or the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article