comment

Electrochemical synthesis of ammonia as a

potential alternative to the Haber–Bosch process
The preeminent Haber–Bosch process has been feeding humankind for more than one hundred years.
Are electrochemical pathways for ammonia synthesis able to compete with it in the future? Electrocatalysts,
electrolytes and novel cell design may be key.

Grigorii Soloveichik

I
n 1909 Fritz Haber invented a process Water Improved
converting nitrogen and hydrogen to electrolysis Haber–Bosch
ammonia (equation (1)). With amazing
speed his bench reactor was converted by
Carl Bosch in 1913 into the first commercial
ammonia plant in Oppau, Germany, making
twenty tonnes of ammonia per day1.
3H 2 + N2 → 2NH 3 (1) Electrochemical NH3
synthesis

Currently, ammonia is produced at world

scale of about 170 million tonnes per year, or
465 thousand metric tons per day (MTD).
A hundred years of development resulted in a
great decrease in production cost and energy
consumption, from more than 110 kWh kg–1 Fig. 1 | Pathways to renewable ammonia. Water electrolysis combined with the improved Haber–Bosch
to 7.7 – 10.1 kWh kg–1, due to thorough process (scenario A) and direct electrochemical synthesis (scenario B).
engineering and an increase in production
scale. The typical ammonia plant size is 2,000
to 3,000 MTD. For instance, plants recently
built by Thyssenkrupp Industrial Solutions plants)2. In addition, to reap the benefits of water electrolysis (equation (5)), followed
in Saudi Arabia and the United States spew economies of scale, the current huge Haber– by the conventional Haber–Bosch process
3,300 tons per day, and the biggest plants Bosch plants require billions of dollars of at elevated temperature and pressure
developed by Casale and Haldor Topsoe can capital and several years to build. (scenario A, Fig. 1); direct electrochemical
deliver up to 6,000 MTD. Besides nitrogen reduction of dinitrogen in the presence of
separated from air, ammonia plants use two Renewable ammonia water (equation (6)), first demonstrated by
major feedstocks — natural gas (the most To completely eliminate carbon emissions Stoukides and co-workers4, where elevated
economical and least polluting) and coal, and from ammonia production, it is necessary to pressure is not necessary (scenario B, Fig. 1);
the ammonia production costs are mostly deliver both hydrogen and the power needed and a combination of renewable hydrogen
due to the cost of hydrogen produced from to run the process from renewable sources. and the electrochemical synthesis of
these feedstocks. In addition to producing In this case, hydrogen may be produced by ammonia (equation (1) and scenario C), the
hydrogen (equations (2) and (4)), fossil fuels commercially available (alkaline or proton benefits of which are not yet evident.
are used to run the Haber–Bosch process exchange membrane, PEM) or emerging
resulting in higher emissions (equation (3)), (anion exchange membrane, AEM, or H 2O ↔ H 2+0.5O2 (5)
with the highest emissions produced by coal solid oxides electrolyser cells, SOEC)
based plants. water electrolysis. Several processes are N2+3H 2O ↔ 2NH 3+1.5O2 (6)
being discussed as potential competitors
3CH 4 + 6H 2O + 4N2 → 8NH 3 + 3CO2 (2) to the Haber–Bosch process. Besides All scenarios assume that the scale of
temperature and pressure, different types ammonia production should match the
0.88CH 4 + 1.24H 2O + 1.25air of plasma can also be used to accelerate the scale of renewable energy sources that
(3) reaction in equation (1) (ref. 3). However, currently vary from a couple of megawatts
→ 2NH 3 + 0.88CO2 the ammonia production rate by plasma (a standalone wind turbine), to 100–200
synthesis so far is low, and it is unclear how megawatts (solar or wind farms). A
3C + 6H 2O + 2N2 → 4NH 3+3CO2 (4) much energy is wasted for non-productive typical Haber–Bosch plant is an order
plasma generation, or how expensive it of magnitude larger, so the building of
Ammonia synthesis constitutes 1.44% of will be to run the plasma generators at similar scale renewable plant eliminates its
global CO2 emissions (the global average is reasonable scale. This leaves three plausible major advantage — avoiding the expense
2.86 tons of CO2 per ton of NH3 and 1.6 tons scenarios for the potential production of of powerline building. Major renewable
of CO2 per ton of NH3 in the most efficient renewable ammonia: generation of H2 by energy sources — wind and solar — are
Nature Catalysis | VOL 2 | MAY 2019 | 377–380 | www.nature.com/natcatal 377
comment

intermittent, so an ammonia plant has of following the load, as has already been with temperature, and above 390 oC HER
to match the electrical power input proven by water electrolysis. Both routes becomes thermodynamically favourable.
unless a massive energy and feedstock require lower pressures and temperatures in The proximity of potentials for HER
storage are installed to compensate for comparison to the Haber–Bosch process, and and NRR creates a fundamental problem
the intermittency; and this substantially their energy efficiency could, in principle, at for electrochemical ammonia synthesis,
increases the associated costs. practical current densities, reach efficiencies which results in low FE of the reactions in
demonstrated for water electrolysis (about equations (1) and (6). This low efficiency
Improving the Haber–Bosch process 70% for scenario B and 50% for scenario C). leads to high energy consumption and
Scenario A is definitely an achievable Efficiencies for scenario B and C are higher unwanted hydrogen generation. In all
goal, due to the relatively advanced than for A. Efficiency here is defined as cases, the rate of HER increases faster with
development stage of both production the sum of the efficiencies of three steps — the electrode potential than the rate of
steps. Major research efforts are focused on water electrolysis (about 70%), hydrogen NRR. Because of this, the FE of ammonia
reducing operational and capital costs by pressurization (about 90%) and ammonia electrosynthesis reaches a maximum at a
decreasing pressure and temperature via synthesis (equation (1)), which itself is a relatively low cell voltage (usually about
the development and deployment of more function of many parameters, including –0.2 V versus RHE, but it may be more
active catalysts and novel methods for the pressure, temperature, recirculation rate, negative when the proton concentration is
separation of ammonia from the reaction ammonia isolation and so on. low, for example, in ionic liquids). Therefore,
mixture, for example, by absorption. Several different electrolytes have been with constant potential electrolysis, which
Commercial Haber–Bosch catalysts used in electrochemical ammonia synthesis is generally less preferred industrially than
are iron activated with potassium or other (equation (6)): solid electrolytes such as constant current electrolysis, accurate
metals with low ionization energies. A more PEM8, AEM, high temperature proton- potential control is needed to maximize
expensive ruthenium-based catalyst is used in conducting9 and oxygen-conducting10 solid ammonia production.
the commercial Kellog Advanced Ammonia oxide membranes; and liquid electrolytes such There are two approaches to suppress
Process, but the gain in productivity was as aqueous basic and acidic electrolytes, ionic HER. The first is the development of
not big enough to justify the catalyst cost liquids11 and the more practical molten alkali selective electrocatalysts that do not catalyse
difference to enable its wide implementation. hydroxides12 or chloride salts13. Electrolysis HER at potentials practical for NRR. For
The catalyst activity is defined by complex in acidic electrolytes results in accumulation example, addition of ZrO2 to single-atom Ru
interplay between nitrogen, hydrogen and of ammonium salts in the solution and the catalysts suppresses HER, affording a higher
ammonia surface coverage (the dissociative alteration of the solution’s pH, which makes a NH3 FE of 21% (ref. 16). The second approach
Mars–van Krevelen mechanism)5. The continuous process unlikely, so this process is is to reduce the availability of protons at the
electron transfer assisted dissociation of used mostly for electrocatalyst screening. In catalyst surface, by decreasing proton activity
dinitrogen is the rate determining step, addition, electrolysis in solutions is hampered there, for example by using basic electrolytes
followed by its reaction with adsorbed by the low solubility of nitrogen. Therefore, with a high pH. From this standpoint, the
hydrogen atoms or metal hydrides. In recent design of gas diffusion electrodes is critical for use of AEM should provide better efficiency
years, two new promising catalyst types with high production rates. than PEM. In liquid electrolytes, it is possible
different reaction mechanisms have been A variety of electrocatalysts have been to decrease proton activity by increasing the
developed. For ruthenium on a 12CaO·7Al2O3 employed in electrochemical ammonia concentration of salts to an extremely high
electride, the rate controlling step of ammonia synthesis, such as supported and molarity (for example, the use of a 21 M
synthesis is not the dissociation of the N2 unsupported precious metals and metal solution of LiTFSI in water reduces the HER
triple bond, but the subsequent formation oxides, or metal nitrides, depending on potential by more than 1 V)17, or using an
of N–Hn species6. A similar effect can be the pH and electrolyte physical state (solid, organic proton donor with a low dissociation
achieved by applying an electric field. A polymer or liquid)14. Transition-metal-free constant, such as ethanol or quaternary
combination of various transition metals catalysts, such as nitrogen-doped carbons, ammonium18.
(in addition to Ru and Fe) and group i and ii BN and black phosphorus, have also been An alternative approach is to separate the
group metal hydrides allowed the separation reported. Designing catalysts that are active nitrogen reduction and ammonia formation
of N2 dissociation and hydrogenation sites, and selective towards the nitrogen reduction steps, realized in a multistep synthesis
and resulted in high catalytic activity at low reaction (NRR) is highly desirable as they including electrochemical (equation (7))
temperatures (200–350 oC) and pressures will enable both high faradaic efficiency and chemical NRR, and ammonia formation
(1–10 bar). The best catalyst so far, 10% (FE) and high NH3 production rate. reactions (equations (8) and (9)). This
Co-CNT-3Ba(NH2)2, gave a rate of ammonia Standard electrochemical cell potentials method, avoiding electrochemical NRR,
synthesis above 10 mmol g–1 h–1 at 350 oC for the reactions in equations (1) and (6) was patented by Ceramatec19 and studied by
and 10 bar7. are –0.057 V and 1.172 V, respectively, research groups at Stanford20.
and dependent on the state of the product
Electrochemical ammonia synthesis (gas, solution or protonated ion) and the 6LiOH → 6Li + 3H 2O + 1.5O2 (7)
The latest successes in reducing pressure and pH (the former shifts to +0.27 V for NH4+
temperature during NH3 synthesis — due in acidic media)15. One can see that the 6Li + N2 → 2Li 3N (8)
to development of novel catalysts — still latter potential is very close to that of water
didn’t eliminate the major disadvantage electrolysis (1.23 V, equation (5)). Therefore, 2LI 3N + 2H 2O → 2NH 3+2LiOH (9)
of the Haber–Bosch process when run by the hydrogen evolution reaction (HER)
renewable energy (scenario A) — the lack of is the most competitive reaction to NRR Electrolysis of aqueous Li salts at the
ability to follow the electrical power input. in scenarios B and C. NRR is favoured anode, and in propylene carbonate at the
The electrochemical synthesis of ammonia at low temperatures, but the difference cathode separated by LISICON, yields
(scenarios B and C), where the main reactor between the theoretical cell potentials of the electrodeposited Li metal that is then
is an electrolyser, would be perfectly capable reactions in equations (5) and (6) decreases washed with Me-THF and allowed to react

378 Nature Catalysis | VOL 2 | MAY 2019 | 377–380 | www.nature.com/natcatal

 comment

with nitrogen. An ammonia synthesis rate higher are required), and the energy consumption and associated footprint,
of 1.88x10–9 mol cm–2 s–1 was demonstrated consumption is unacceptably high due scalability, lower process pressures and
for Li, with a capacity of 1.39 mAh cm–2 to a low FE and a high cell voltage at temperatures, the ability to follow the
and total FE about 50% (ref. 21). This path substantial current densities. So far, intermittent electrical power input and
requires large quantities of lithium metal the best FE results for the reaction in use nitrogen with reduced purity justify
(equations (7), (8) and (9)), and much more equation (5) have been demonstrated further research. In 2016, ARPA-E initiated
energy than the direct electrochemical in molten electrolytes — alkali metal the Renewable Energy to Fuels through
(scenario B) and even electrolysis-Haber– chlorides with added Li3N (70–80% FE)26 Utilization of Energy-dense Liquids
Bosch (scenario A) pathways20. Therefore, and hydroxides (35% FE)12. (REFUEL) programme that aims, in part,
this technology makes sense only in the case to develop both types of technologies
of very cheap electricity and high current Conclusions according to scenarios A and B, using the
density, which will reduce the capital cost. The conventional highly optimized ammonia production rate and cost
The major advantage of electrochemical Haber–Bosch process uses about 7.9 kWh as metrics29. ❐
ammonia synthesis compared to scenario of energy derived from fossil fuels per kg
A is the use of only one electrolyser NH3 at the scale of 1,000 tonnes per day, Grigorii Soloveichik
stack instead two major pieces of capital where 2.0 kWh kg–1 (including process Advanced Research Program Agency – Energy
equipment (this advantage is negated when heat credit from ammonia synthesis)27 is (ARPA-E), US Department of Energy,
H2 is used for the electrochemical process used for pressurization, heating, pumping Washington, USA.
in scenario C), employment of ambient and so on — this number will be higher e-mail: [email protected]
or slightly elevated pressures and low or at a smaller scale due to the increased
intermediate temperatures, potentially lower heat losses. In scenario A, the energy Published online: 13 May 2019
energy consumption, and, most importantly, needed for water electrolysis varies from https://doi.org/10.1038/s41929-019-0280-0
the ability to adapt to the intermittent 54 to 44.7 (current and DOE 2020 target
electrical power input. In addition, the use of levels, respectively) kWh kg–1 of H2, which References
wet air instead of pure nitrogen with a Nafion corresponds to 9.9 to 11.5 kWh kg–1 1. Hager, T. The Alchemy of Air: A Jewish Genius, a Doomed
membrane results in only a mild reduction NH3, possibly even less if the hydrogen Tycoon, and the Scientific Discovery that Fed the World but Fueled
the Rise of Hitler (Random House, New York, 2009).
in performance (1.1x10–9 mol cm–2 s–1 versus is produced via a solid oxide electrolyser 2. Industrial Efficiency Technology Database (Institute of
3.1x10–9 mol cm–2 s–1)8. In the Haber–Bosch cell that utilizes the process heat. Scenario Productivity, 2018); https://go.nature.com/2VhsYwK
process, even traces of oxygen dramatically B, theoretically, has the smallest energy 3. Peng, P. et al. A review on the non-thermal plasma-assisted ammonia
synthesis technologies. J. Clean. Prod. 177, 597–609 (2018).
decrease the catalyst performance, so the consumption if both voltage and coulombic 4. Marnellos, G. & Stoukides, M. Ammonia synthesis at atmospheric
requirements for N2 purity are very high. efficiencies are higher than 70%. pressure. Science 282, 98–100 (1998).
If possible, air, or even nitrogen of reduced Currently, the major part of the ammonia 5. Mars, P. & van Krevelen, D. W. Oxidations carried out by
purity, and a single chamber cell22, should Haber–Bosch production cost is the means of vanadium oxide catalysts. Chem. Eng. Sci. 3,
41–59 (1954).
be employed to substantially reduce the cost of the feedstock and the energy 6. Kitano, M. et al. Electride support boosts nitrogen dissociation
nitrogen feedstock cost. consumed, that is, US$90 ton–1 for natural over ruthenium catalyst and shifts the bottleneck in ammonia
NRR in molten NaOH–KOH (180–260 gas (at US$3 million British thermal units–1) synthesis. Nat. Commun. 6, 6731 (2015).
7. Gao, W. et al. Barium hydride-mediated nitrogen transfer and
°C) at porous nickel electrodes with a and US$232 ton–1 for hydrogen by water hydrogenation for ammonia synthesis: a case study of cobalt.
nano-Fe3O4 catalyst generates ammonia electrolysis (at US$0.0235 kWh–1) (ref. 28). ACS Catal. 7, 3654–3661 (2017).
at ambient pressure with a FE of 14.2% However, with renewable electricity 8. Lan, R., Irvine, J. T. S. & Tao, S. Synthesis of ammonia directly
from air and water at ambient temperature and pressure. Sci. Rep.
and rate of 4.4×10−9 mol cm–2 s–1(ref. 23). prices falling fast, the difference between 3, 1145 (2013).
Increasing the FE further is possible if energy costs from renewable sources 9. Sik Yun, D. et al. Electrochemical ammonia synthesis from
the hydrogen produced by the reaction and fossil fuels will be smaller, and the steam and nitrogen using proton conducting yttrium doped
barium zirconate electrolyte with silver, platinum, and
in equation (4) is recycled back to the significance of capital costs will increase. lanthanum strontium cobalt ferrite electrocatalyst.
cathode24, where it reacts with nitrogen via It is difficult to estimate the capital costs J. Power Sources 284, 245–251 (2015).
the conventional reaction (mixed scenarios of new technologies at different scales, but 10. Jeoung, H. et al. Electrochemical synthesis of ammonia from
water and nitrogen using a Pt/GDC/Pt Cell. Korean Chem.
B and C). it is clear that reducing operation pressure Eng. Res. 52, 58–62 (2014).
Combining plasma nitrogen activation and temperature will substantially reduce 11. Suryanto, B. H. R. et al. Rational electrode–electrolyte design for
and electrochemical ammonia synthesis associated costs. efficient ammonia electrosynthesis under ambient conditions.
resulted in a high FE (close to 100%) In summary, water electrolysis, ACS Energ. Lett. 3, 1219–1224 (2018).
12. Licht, S. et al. Ammonia synthesis by N2 and steam electrolysis in
at high current density (200 mA cm–2) combined with an improved smaller- molten hydroxide suspensions of nanoscale Fe2O3. Science 345,
without any catalyst.25 It should be noted scale Haber–Bosch process, seems to be a 637–640 (2014).
that this approach was demonstrated in a short-term solution for the generation of 13. Murakami, T. et al. Electrolytic synthesis of ammonia from water
and nitrogen under atmospheric pressure using a boron-doped
beaker cell with a very small Pt electrode ammonia from renewables at a matching diamond electrode as a nonconsumable anode. Electrochem. Solid
(1 mm2) and low energy consumption scale. There is a major research focus State Lett. 10, E4–E6 (2007).
due to the plasma generation being very on reducing the NH3 synthesis reaction 14. Wang, K., Smith, D. & Zheng, Y. Electron-driven heterogeneous
catalytic synthesis of ammonia: current states and perspective.
high (2,270 kWh kg–1 NH3). A transition pressure and temperature, while keeping Carbon Resour. Conver. 1, 2–31 (2018).
to a practical reactor design may lead the catalyst cost and means to follow the 15. Harris, D. C. in Quantitative Chemical Analysis
to a substantial decrease of the energy intermittent electrical power input. These (W. H. Freeman, 2010).
16. Tao, H. et al. Nitrogen fixation by Ru single-atom electrocatalytic
consumption, making this a more technologies are anticipated to be developed reduction. Chem 5, 204–214 (2019).
competitive technology. within next few years. Electrochemical 17. Suo, L. et al. “Water-in-salt” electrolyte enables
Currently, the reported rates of ammonia production is much more distant high-voltage aqueous lithium-ion chemistries. Science 350,
938–943 (2015).
electrochemical ammonia production from commercialization, however, potential 18. Zhang, L. et al. A theoretical study of the effect of a non-aqueous
(10–9 to 10–11 mol cm–2 s–1) are too low benefits of electrochemical ammonia proton donor on electrochemical ammonia synthesis.
(at least two orders of magnitude production, such as reduced energy Phys. Chem. Chem. Phys. 20, 4982–4989 (2018).

Nature Catalysis | VOL 2 | MAY 2019 | 377–380 | www.nature.com/natcatal 379

comment

19. Joshi, A. V. & Bhavaraju, S. Ammonia synthesis using lithium the catalyst. Faraday Discuss. 182, 353–363 (2015). 26. Murakami, T., Nohira, T., Ogata, Y. H. & Ito, Y. Electrolytic
ion conductive membrane. US8916123 patent 8,916,123 (2018). 23. Bicer, Y. & Dincer, I. Performance assessment of electrochemical ammonia synthesis in molten salts under atmospheric pressure
20. McEnaney, J. M. et al. Ammonia synthesis from N2 and H2O ammonia synthesis using photoelectrochemically produced using methane as a hydrogen source. Electrochem. Solid State Lett.
using a lithium cycling electrification strategy at atmospheric hydrogen. Int. J. Energ. Res. 41, 1987–2000 (2017). 8, D12–D14 (2005).
pressure. Energ. Environ. Sci. 10, 1621–1630 (2017). 24. Vasileiou, E. et al. Electrochemical enhancement of ammonia 27. Benner, J., van Lieshout, M. & Croezen, H. Identifying
21. Kim, K. et al. Electrochemical synthesis of ammonia from water synthesis in a BaZr0.7Ce0.2Y0.1O2.9 solid electrolyte cell. Solid State Breakthrough Technologies for the Production of Basic Chemicals
and nitrogen: a lithium-mediated approach using lithium-ion Ion. 288, 357–362 (2016). (University of Delft, 2012).
conducting glass ceramics. ChemSusChem 11, 120–124 (2018). 25. Hawtof, R. et al. Catalyst-free, highly selective synthesis of 28. Pfromm, P. H. Towards sustainable agriculture:
22. Lan, R., Alkhazmi, K. A., Amar, I. A. & Tao, S. Synthesis of ammonia from nitrogen and water by a plasma electrolytic Fossil-free ammonia. J. Renew. Sustain. Energ. 9, 034702 (2017).
ammonia directly from wet air using Sm0.6Ba0.4Fe0.8Cu0.2O3−δ as system. Sci. Adv. 5, eaat5778 (2019). 29. Refuel. ARPA-E https://go.nature.com/2KfWT7k (2016).

380 Nature Catalysis | VOL 2 | MAY 2019 | 377–380 | www.nature.com/natcatal