i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Options for net zero emissions hydrogen from

Victorian lignite. Part 2: Ammonia production

M.A. Kibria a, D.E. McManus b,**, S. Bhattacharya a,*

a
Monash University, Department of Chemical and Biological Engineering, PO Box 36, Clayton Victoria 3800,
Australia
b
Australian Carbon Innovation, Building 4W-127 Federation University Gippsland, Churchill, Victoria 3842,
Australia

highlights graphical abstract

A detailed model for net-zero

emissions ammonia from Victo-
rian lignite proposed.

Gasification of lignite, production
of ammonia and electricity, with
CCS included.

Power consumption for liquefac-
tion of 32.4 t h1 hydrogen or
conversion to ammonia surpris-
ingly similar.

Production of ammonia from
Victorian lignite with net-zero
emissions is feasible.

Direct comparison with resource
requirements for the production of
green ammonia.

article info abstract

Article history: This the second of a two-part study investigating the feasibility of producing export
Received 22 January 2023 quantities (770 t/d) of blue hydrogen meeting international emissions standards, by gasi-
Received in revised form fication of Victorian lignite plus carbon capture and storage (CCS). Part 1 focussed on the
6 May 2023 resources, energy requirements, and greenhouse gas emissions associated with the pro-
Accepted 9 June 2023 duction of gaseous and liquefied hydrogen, while Part 2 focusses on the production of
Available online 26 June 2023 ammonia as an alternative hydrogen carrier for export.
In this study, an Aspen Plus simulation of a conventional 1500 t d1 iron-based catalyst
Keywords: Haber-Bosch ammonia synthesis process is developed and incorporated into the earlier
Ammonia lignite-to-hydrogen process model. Development of the simulation involves (i) estimation
Lignite of the instantaneous rate kinetics, (ii) calibration against test data for a reactor of known
Gasification dimensions, (iii) scaling up the bed dimensions to achieve a target production capacity, (iv)
Carbon capture selecting an appropriate feed gas composition to optimise performance, (v) adjusting purge

* Corresponding author.
** Corresponding author.
E-mail addresses: dmcmanus@acinnovation.com.au (D.E. McManus), sankar.bhattacharya@monash.edu (S. Bhattacharya).
https://doi.org/10.1016/j.ijhydene.2023.06.098
0360-3199/© 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY
license (http://creativecommons.org/licenses/by/4.0/).
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37167

Renewable energy gas flowrates to achieve stable operation, and (v) incorporation into the previous lignite-to-
Greenhouse gas intensity hydrogen simulation.
This study finds that 178.2 t h1 liquid ammonia and all electricity required to support
the process can be produced from 1050 t h1 Victorian lignite. Surprisingly, the simulation
results show that the electrical power requirement for ammonia synthesis (176.4 MW) is
essentially the same as that needed for liquefaction of an equivalent output of hydrogen
(175.5 MW). On this basis both options are equally attractive, although ammonia synthesis
is at a higher level of technological maturity than large-scale hydrogen liquefaction.
This is the first study to quantify the greenhouse gas emissions intensity of ammonia
production from lignite, accounting for the full production chain from lignite mining to CO2
sequestration. It is found that ammonia can be produced from Victorian lignite with very
low CO2 emission intensity (0.49 kgCO2-e kg1
NH3) equivalent to that of next-generation nat-

ural gas reforming with CCS processes. If required, the emission intensity can be reduced
to 0.05kgCO2-e kg1
NH3 with a post-combustion CO2 capture system, and then made carbon

neutral by co-gasification with 1.4% biomass.

For comparison, this study also examines the implications of producing the same
quantity of green ammonia using renewable energy alone. It is estimated that production
of 178.2 t h1 green ammonia would require 1946 MW renewable energy and associated
transmission infrastructure. In Victoria, this could be supplied by a wind farm with a
5.4 GW rated capacity, occupying an area of over 72,000 ha. This is highly unlikely to be a
viable option.
This analysis indicates that clean hydrogen in the form of ammonia, produced in Vic-
toria by lignite gasification with CCS, can be consistent with global emissions reductions
targets over the next few decades. The unique combination of low-cost lignite and high-
quality CO2 storage geology means that Victoria is well placed to become a significant
exporter of low-emissions ammonia to the world market. Further research is recom-
mended on recovery of energy from the low grade waste heat streams and opportunities
for additional electricity generation using the organic Rankine cycle.
© 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications
LLC. This is an open access article under the CC BY license (http://creativecommons.org/
licenses/by/4.0/).

Nomenclature N2 Nitrogen
NH3 Ammonia
Acronyms Selexol™ A proprietary blend of DEPG licensed by
ASU Air Separation Unit Honeywell UOP, and the process for its use
CCS Carbon Capture and Storage UAN Urea-ammonium nitrate
EOR Enhanced Oil Recovery
HESC Hydrogen Energy Supply Chain
LHV Lower Heating Value
PEM Proton Exchange Membrane
PSA Pressure Swing Adsorption
RO Reverse Osmosis
SEC Specific Energy Consumption Introduction
SEI Specific Emissions Intensity
SMR Steam Methane Reforming Hydrogen is anticipated to become a major source of clean
energy as the world transitions toward a low-emissions
Chemicals economy. As a chemical carrier of energy, hydrogen can play
CH4 Methane a crucial bridging role in linking sources of clean energy with
CO2 Carbon dioxide remote energy users, particularly in the transportation, in-
CO2-e Carbon dioxide equivalent dustrial, heating and power sectors. Hydrogen produced using
DEPG Dimethylether polyethylene glycol renewable energy or fossil fuels with carbon capture and
H2 Hydrogen storage (CCS) is environmentally benign and can be utilised
H2S Hydrogen sulphide via direct combustion, co-combustion and fuel cells [1].
MDEA Methyldiethanolamine The main challenge for hydrogen as a tradeable energy
MCH Methylcyclohexane commodity is its low energy density. Hydrogen is a gas at at-
mospheric temperature and must be converted to liquid form
37168 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

for bulk transport. Unlike natural gas, which can be trans- Ammonia is produced commercially using the Haber-
ported as a liquid at 169  C [2], hydrogen must be cooled to Bosch process, in which hydrogen is catalytically reacted
253  C which is very energy intensive [3]. Nevertheless, bulk with nitrogen at high temperatures (typically 400e450  C) and
transport of liquid hydrogen is being promoted by the Japa- high pressures (typically 15e25 MPa). The majority (72%) of
nese company Kawasaki Heavy Industries and its collabora- modern ammonia plants are based on steam methane
tors in the Hydrogen Energy Supply Chain (HESC) Project, reforming (SMR) of natural gas, 26% on coal gasification, about
which envisages a “CO2 free hydrogen chain” in which low 1% on oil products, and a fraction of a percentage point on
cost blue hydrogen is produced in Australia by gasification of electrolysis. There are currently around 550 ammonia plants
Victorian lignite with CCS, then liquefied and transported to operating globally, ranging in capacity from around 200 kt y1
Japan by ship [4,5]. During 2020e2021, the HESC Pilot Project to 1,200 kt y1, generating cumulative CO2 emissions of
produced hydrogen at 99.999% purity by gasification of lignite around 450 Mt y1 [12]. Zhang et al. [13] reported the specific
and lignite-biomass blends, which was liquefied and suc- emissions intensity (SEI) of the large ammonia plants in China
cessfully transported to Japan on the Suiso Frontier, the world's as 9.0 kgCO2-e kg1 1
NH3 for coal and 3.0 kgCO2-e kgNH3 for SMR.
first ocean-going liquid hydrogen carrier ship [6]. However, Stocks et al. [14] estimated that the total emissions
Another Japanese company, Chiyoda, has developed an intensity of SMR (including fugitive emissions) could be
alternative hydrogen supply chain concept based on methyl- reduced to 0.44e0.54 kgCO2-e kg1NH3 by employing CCS with 90%
cyclohexane (MCH) as a hydrogen carrier, which is a stable CO2 capture.
liquid at ambient temperature and pressure. In Chiyoda's In Part 1 of this study, we developed an Aspen Plus simu-
SPERA Hydrogen™ process, hydrogen is reacted with toluene lation model for production of blue hydrogen by gasification of
to produce MCH which, after transport, is dehydrogenated Victorian lignite plus CCS [15]. The individual unit operations
back to toluene and hydrogen using a proprietary catalyst. The were validated against published industrial data to ensure
SPERA Hydrogen™ technology has been successfully demon- that the simulation was as realistic as possible. It was shown
strated by transporting MCH by sea between a hydrogenation that 770 t d1 liquefied hydrogen and all required electricity
plant in Brunei Darussalam and a dehydrogenation plant in could be co-produced with a specific emissions intensity of
the Kawasaki Coastal Area in Japan [7]. A disadvantage of this 2.73 kgCO2-e kg1
H2, including upstream fugitive methane emis-
approach is that the hydrogenation reaction is highly sions. This conforms to the current EU Taxonomy limit of 3.0
exothermic, but there is limited scope to make use of the kgCO2-e kg1
H2 for ‘sustainable’ hydrogen. It was also shown that
excess energy at the hydrogenation plant. On the other hand, the SEI can be reduced to 0.3 kgCO2-e kg1
H2 by addition of a post-
the dehydrogenation reaction is endothermic, requiring a combustion CO2 capture unit, and to net-zero or net-negative
suitable source of renewable energy at the destination site. by co-gasification with biomass. The concept of producing
Also, the recovered toluene must be recycled by sea transport blue hydrogen by gasification of Victorian lignite plus CCS is
back to the hydrogenation plant, representing an additional thus consistent with policy settings of ‘net zero by 2050’.
energy penalty and source of greenhouse gas emissions [8]. In Part 2, we extend this analysis to investigate the relative
Ammonia is rapidly emerging is a more advantageous merits of converting the same quantity of gaseous hydrogen
hydrogen carrier than either liquefied hydrogen or MCH. to ammonia rather than liquefied hydrogen. In principle, if
Ammonia can be liquefied at 25  C when pressurized to less energy is required to convert the hydrogen to ammonia
1.0 MPa, or at 33.4  C at atmospheric pressure, and has an than for liquefaction, the CO2 emissions and cost would both
energy density approximately double that of liquid hydrogen be reduced. To conduct this study on a like-for-like basis, the
[9]. Ammonia can be directly used as fuel without CO2 emis- simulated ammonia synthesis process must be based on
sions, and NOX emissions in ammonia combustion can be industrially-validated design data, as in the previous study.
controlled. While ammonia has acute toxicity (but no chronic This is challenging, however, because publication of
toxicity) with strong smell, it is easy to detect and safety commercially-sensitive ammonia plant design and perfor-
measures are common practice [10]. In Japan, leading power mance data is typically restricted. As such, there is very little
companies, manufacturers and research institutes are relevant data available in the public domain.
collaborating toward commercialization of a CO2-free The key to ammonia reactor design is the catalyst reaction
ammonia value chain, with significant achievements in kinetics. The conventional Haber-Bosch process uses iron-
development of ammonia-fuelled gas turbine power plant, a based catalysts, which are equilibrium-limited with typically
direct ammonia-fuelled solid oxide fuel cell, and co-firing of 20e30% conversion efficiency per pass, requiring large reactor
ammonia at a commercial coal power plant [10]. volumes to achieve good economics. Alternative catalysts
Ammonia is already one of the most important industrial have been developed for operation under milder conditions,
chemicals in the world, with the lives of around half of hu- e.g., ICI's cobalt-modified magnetite catalyst and Kellog's
manity being dependent on ammonia-based fertilisers. In ruthenium catalyst, but they do not significantly improve the
2020, 185 million tonnes of ammonia was produced and process economics [16].
around 20 million tonnes was traded globally, so the infra- Some conceptual studies of potential low-emissions
structure to support safe and reliable storage, distribution and ammonia production schemes make the simplifying
export of ammonia is already highly developed. Around 70% assumption that equilibrium is achieved within the synthesis
of ammonia is used to make fertilisers, with the remainder reactor, e.g., Refs. [17e21]. In reality, however, ammonia
used for a wide range of industrial applications, such as synthesis is equilibrium-limited, so this approach over-
plastics, explosives and synthetic fibres [11]. estimates the reaction temperatures and underestimates the
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37169

energy requirement. A rate-based modelling approach is The main objective of Part 2 of this study is to extend our
needed to more accurately simulate the performance of a model to include the Haber-Bosch process to convert the
commercial process. produced hydrogen to ammonia, and to compare the resource
The most widely used rate expression used to characterise requirements and CO2 emissions intensity of export-scale
performance of iron-based catalysts is the modified Temkin production of either liquefied hydrogen or ammonia from
equation, which is a correlation based on laboratory rate data Victorian lignite with CCS. To accomplish this, an Aspen Plus
for the widely-used Haldor Topsøe KM1 catalyst. This equa- process model for ammonia synthesis is developed, using the
tion has been successfully used in one-dimensional analytical published design and test data for a commercial Egyptian
models [23e25] to simulate the performance of industrial ammonia plant [29]. The developed model, scaled for pro-
Haber-Bosch ammonia synthesis reactors. Araújo and Sko- cessing 32.4 t h1 hydrogen, is then incorporated into our
gestad [26] developed an Aspen Plus model using modified earlier lignite-to-hydrogen simulation model, allowing esti-
Temkin reaction kinetics to develop a control strategy for an mation of the overall resource requirements and CO2 emis-
industrial ammonia plant in Germany. While the model was sions intensity of lignite-to-ammonia production.
based on industrial data, no design details of the plant were
published. Yoshida et al. [27] used Aspen Plus simulation to
compare the economies of scale for ammonia production Model development, calibration and scaleup
using either iron- or ruthenium-based catalysts. Based on
laboratory rate kinetics, it was found that iron-based catalysts Overview of Aspen Plus simulation concept
offer lowest production costs at scales of more than 100 t d1,
while ruthenium-based catalysts are more economical at The lignite-to-hydrogen simulation previously developed [15]
smaller scales. Rossetti et al. [28] developed an Aspen Plus comprises lignite mining, lignite drying and milling, air sep-
simulation for a process to produce hydrogen from waste aration unit (ASU), dry-feed entrained flow gasification, gas
biomass and conversion to liquid ammonia, using a combi- cooling and cleaning, sour water-gas shift reaction, acid gas
nation of both iron-based and ruthenium-based catalysts. In removal, pressure swing adsorption (PSA) for hydrogen puri-
this case, the rate equations were validated against micro- fication, elemental sulphur recovery, CO2 compression for
pilot plant data. El-Gharbawy et al. [29] used the modified transport and injection, steam and gas turbines to generate all
Temkin equation in an Aspen HYSYS simulation of a com- process power, plus an optional post-combustion CO2 capture
mercial ammonia plant in Egypt. Industrial plant test data was step. In the present study, the simulation model is extended to
used to validate a novel correlation for the equilibrium con- include Haber-Bosch ammonia synthesis.
stant. This is the only known report where industrial The high-level Aspen Plus flowsheet for the entire lignite-
ammonia plant design details, plant test data and vendor to-ammonia process is depicted in Fig. 1. It can be seen that
simulation data have all been published. the AMMONIA Hierarchy block is integrated with the lignite-
While most ammonia is produced from natural gas using to-hydrogen process through two input streams and two
SMR, about one quarter is produced by gasification of coal. In output streams. The input streams are hydrogen from the PSA
1980, Haldor Topsøe announced development of a coal-to- and nitrogen from the ASU. The output streams are liquid
ammonia process based on a Texaco slurry-feed gasifier. In ammonia and a relatively small purge gas stream that is sent
this process, shift conversion was done in two stages, using a for combustion in the gas turbine block, to extract energy and
sulphur-tolerant shift catalyst. Hydrogen sulphide and carbon prevent atmospheric release of waste gases.
dioxide were then removed in an acid gas removal unit, and The Aspen Plus flowsheet of the AMMONIA Hierarchy
remaining carbon oxides were removed by methanation block (Fig. 2) comprises three main sections: the compression
before ammonia synthesis in an iron-based catalytic reactor system, the synthesis reactor, and the cooling unit. The feed
[30]. In 1984, the Japanese company Ube Corporation streams into the AMMONIA Hierarchy block are ultrapure
commissioned a 1000 t d1 coal-to-ammonia plant based on gaseous hydrogen and nitrogen. Hydrogen at 99.999% purity is
this process, which is still operational today [31]. The produced in the PSA unit 15  C, 2.75 MPa, at the rate of
biomass-to-ammonia simulation developed by Rossetti et al. 32.4 t h1. Nitrogen at >99.9% purity is extracted from the top
[28] is also based on the same process concept. Stork et al. [32] of the high pressure column of the ASU at 177  C, 0.58 MPa
examined the economic feasibility of co-producing ammonia [34,35]. To handle the feed of 32.4 t h1 hydrogen, the
and electricity from coal was assessed. However, no simula- ammonia synthesis system is conceived as three separate
tion modelling of this process was published. systems operating in parallel. Each individual system has an
In 2011, the U.S. National Energy Technology Laboratory ammonia production capacity of around 1500 t d1, which is
published a feasibility study on a coal-to-hydrogen process typical for large commercial ammonia facilities [16]. A single
incorporating CO2 capture [33]. The process included oxygen- such unit is represented in Fig. 2, with splitters, duplicators
blown slurry-feed gasification, sour water-gas shift reaction, and mixers replicating the other two units in parallel. To ini-
two-stage Selexol™ acid gas removal and hydrogen purifica- tialise the simulation, it is necessary for there to be a low but
tion using pressure swing adsorption (PSA). An Aspen Plus non-zero feed of NH3 into the system to prevent the reaction
model was developed to assist in sizing the equipment for cost rate equation (Eq. (2), Section 2.2 below) going to infinity.
estimation, but no details were published. This process was The core of the AMMONIA Hierarchy block is the ammonia
the basis for Part 1 of our study, where we developed a detailed synthesis reactor, representing the conventional Haber-Bosch
Aspen Plus simulation for production of 32.4 t h1 low emis- process with Haldor Topsøe iron-based KM1 catalyst. The
sions hydrogen from Victorian lignite [15]. exothermic reaction is conducted at 400e500  C temperature
37170 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

and 17.1 MPa pressure. Each synthesis reactor comprises three reactor pressure, while operation of FLASH2 at 2.5 MPa is
catalyst beds with two intercooling stages, which is the most convenient for transportation purposes.
efficient design in terms of ammonia production, energy The cooling unit also includes five heat exchangers which
savings, capital, and maintenance cost [16]. In Aspen Plus, the reduce the temperature of the gas stream prior to ammonia
catalyst beds are represented by RPlug adiabatic plug flow separation. WC1 uses water to raise steam and cool the gas
reactors. Within each RPlug block, the instantaneous reaction from 452  C to 232  C. Air cooler AIR cools the gas to 150  C,
rate is calculated as a function of position, temperature, and then WC2 uses water to cool the gas to 20  C. In ASU1, a
pressure. The procedure for calculating the instantaneous portion of nitrogen from the ASU at 177  C cools the gas to
reaction rate, gas composition and temperature is described in 5  C before FLASH1. In ASU2, a second portion of nitrogen
Section 2.2. The simulation is calibrated against industrial cools the gas to 3  C before FLASH2. After passing through
plant data to account for diffusion limitations and catalyst ASU1 and ASU2 the nitrogen is at 15  C, ready for compression
ageing (described in Section 2.3) and the catalyst bed di- in COMP1. A total of 123 MW of cooling energy is required,
mensions are scaled up to a capacity of 1500 t d1 (Section 2.4). with 52 MWth extracted in WC1 and 17 MWth extracted in
The ammonia synthesis reaction is inhibited by the high ASU1 and ASU2.
temperatures involved, so unreacted gases are recycled in a To prevent accumulation of the impurities in the H2 and
loop to boost the conversion efficiency. N2 feed streams under continuous recycle, it is necessary to
In the Egyptian industrial plant, temperature control of the purge a portion of the recycle stream. Due to the high purity
catalyst beds is achieved using two interbed heat exchangers, of the reactant gases, a purge of around 1% of the recycle gas
which preheat incoming feed gas against the hot gas exiting suffices. This is extracted from the vapour phase from
beds 1 and 2, plus controlled injection of feed gas directly into FLASH1 at splitter SPL3. The vapour phase outlet from
the reactor as quench streams [29]. This arrangement allows FLASH2, which is predominantly nitrogen, is also sent to the
the temperature profile within the reactor to be maintained gas turbine.
without any external heating or cooling load. To simplify In the compression system, COMP1 boosts the pressure of
calculations in the Aspen Plus model, the temperature profile the H2 and N2 feed streams to the reaction pressure of
is adjusted using two inter-bed heat exchangers, with dummy 17.1 MPa. COMP2 is used to boost the recycle gas stream back
heat loads that do not connect with the rest of the plant. up to reaction pressure. COMP3 is a small compressor used to
Reaction gas exits the synthesis reactor at 452  C and must match the pressure of the purge gas from FLASH2 (2.5 MPa)
be chilled to condense and separate the ammonia product. with that of FLASH1 (16 MPa), before they are combined and
There is limited information in the literature on the condi- sent to the gas turbine for combustion. At the gas turbine, the
tions used industrially to condense ammonia. Araújo and gas stream is preheated and expanded to synchronize with
Skogestad [26] describe an industrial plant where ammonia is the turbine operating conditions. Further details of the ele-
condensed at 40.5  C and 19.6 MPa, but this pressure is higher ments within the AMMONIA Hierarchy block are presented in
than that used in the present model. Zhang et al. [36] provide Table 1, and a description of each of the numbered streams is
details of a state-of-the-art Chinese methane-to-ammonia provided in Table 2.
plant in which ammonia condensation is accomplished Quantifying the energy intensity of ammonia synthesis
through two flash cooling steps. The first stage operates at requires a realistic model of a typical industrial-scale syn-
5  C and 16 MPa, and the second operates at 3  C and thesis plant, validated by test data. The following sections
2.5 MPa, producing liquid NH3 product at 99.8 mol% purity. describe the process followed used to develop a representative
The same approach is adopted in the present model. Opera- model based on published data, involving (i) estimation of the
tion of FLASH1 at 16 MPa is readily synchronised with the instantaneous rate kinetics, (ii) calibration against test data

Fig. 1 e High level Aspen Plus process flowsheet for ammonia production from lignite.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37171

Fig. 2 e Aspen Plus model of the AMMONIA Hierarchy block.

for a reactor of known dimensions, (iii) scaling up the bed

log10 Ka ¼  2:691122 log10 T  5:519265x105 T
dimensions to achieve a target production capacity, (iv)
þ 1:848863x107 T2 þ 2001:6=T þ 2:6899 (4)
selecting an appropriate feed gas composition to optimise
performance, (v) adjusting purge gas flowrates to achieve The activity of each component is given by:
stable operation, and (v) incorporation of the developed model
into the previous lignite-to-hydrogen Aspen Plus simulation. ai ¼ Xi fio (5)

where: Xi is the mole fraction of each component.

Catalyst rate kinetics
Goi is the pure component fugacity at the temperature and
The key to ammonia synthesis is the catalyst, which facili- pressure of the system
tates the reaction of hydrogen with nitrogen to produce
ammonia (Eq. (1)). The fugacity of each pure component is given by:

N2 þ 3 H2 42 NH3 DHo298 ¼ 46 kJ:mol1 NH3 (1) fio ¼ gi P (6)

It is assumed that the catalyst for this reaction is the widely
used commercial Haldor Topsøe iron-based KM1 catalyst. The where: P is the pressure, atm
instantaneous reaction rate, rNH3, for ammonia synthesis with h
gH2 ¼ exp eð3:8402T þ0:541Þ P  eð0:1263T 15:980Þ P2
0:125 0:5

this catalyst can be estimated using the modified Temkin i

  (7)
expression, following the procedure detailed by Dyson and þ 300 eð0:011901T5:941Þ eP=300  1
Simon [22]:
0 " 3 #0:5 " 2 #0:5 1 gN2 ¼ 0:93931737 þ 0:3101804x103 T þ 0:295896x103 P
aH2 aNH3
rNH3 ¼ kr @K2a aN2  2   3 A (2)  0:2707279x106 T2 þ 0:4775207x106 P2 (8)
aNH3 aH2

where: rNH3 is the instantaneous reaction rate in kmol NH3.- gNH3 ¼ 0:1438996 þ 0:2028538x102 T  0:4487672x103 P
h1.m3 of catalyst bed.  0:1142945x105 T2 þ 0:2761216x106 P2 (9)
kr is a kinetic constant of the reverse reaction,
dimensionless. Using Eqs. (2)e(9), the reaction rate is calculated as a
Ka is the equilibrium constant of the reaction, function of position, temperature, and pressure in the catalyst
dimensionless bed. The resulting values represent the instantaneous reac-
ai (i ¼ H2, N2, NH3) is the activity of component i tion rate at each point, and have been shown to closely match
the performance of powdered Haldor Topsøe KM1 catalyst
kr ¼ 1:7698 x 1015 exp ð40765 = Rc TÞ (3) [16].
where: Rc is the gas constant, 1.987 cal K 1
mol 1 In practice, the apparent reaction rate is substantially
T is temperature, Kelvin lower than the instantaneous reaction rate, due to the
37172 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

Table 1 e AMMONIA Hierarchy block description and operating conditions.

Block ID Aspen Block Purpose Specification
SPL-1,2 Splitter To split incoming N2 and H2 Both 0.333
MIX-1-7 Mixer To mix incoming stream Adiabatic
COMP-1,2,3 Compressor To compress N2 and H2 mixture Discharge pressure 17.1 MPa(a)
HX-1-4 Heater block To cool material stream HX-1,2: T-394  C, P-0 MPa
HX-3: T-434  C, P-0 MPa
HX-4: T-413  C, P-0 MPa
WC1-2 Heater block To cool material stream WC1: T-232  C, P-0 MPa
AIR WC2: T-20  C, P-0 MPa
ASU1-2 AIR: T-150  C, P-0 MPa
ASU1: T-(-3  C), P-0 MPa
ASU2: T-(-3  C), P-0 MPa
R-SEC-1,2,3 RPlug Kinetic reactor Adiabatic
Pressure drop - Ergun
Catalyst: Diameter 2.3 mm, Density 2.3 g/ml
FLSH1-2 Flash To flash the incoming stream by cooling and reducing pressure FLSH1: T-(-5  C), P-16 MPa
FLSH2: T-(-3  C), P-2.5 MPa
DUP-1-3 Duplicator To duplicate material stream Each 3

combined effects of diffusional limitations within the catalyst rate to the intrinsic reaction rate, is used as a correction factor
pellets and deactivation of catalyst over time. A pore effec- for engineering design of ammonia converters [16]. Dyson and
tiveness factor, x, defined as the ratio of the actual reaction Simon [22] derived an expression for x from first principles,
and then used the results of numerous calculations to develop
an empirical expression for x in terms of P, T and h, with h
Table 2 e Key to AMMONIA Hierarchy block stream defined as the conversion of nitrogen as measured for a
numbering. reference mixture (h ¼ 0) containing 3:1H2/N2 ratio and 12.7%
Stream Number Description inert. Elnashaie et al. [37,38] found that the empirical
expression provided the best fit to plant data, but it was
7-2(IN) Ultrapure H2 at 32.4 t h1
N2-AMMON(IN) Ultrapure N2 from the ASU necessary to further modify the reaction rate by a factor of
NH3 Nominal non-zero flow of NH3 to 0.75 to properly account for deactivation of aged catalyst [37].
stop reaction rate going to infinity In the present case, the concentration of inert gases is very
S-1 Blended stream of 1/3 x 7-2(IN), 1/3 x low, so the empirical expression of Dyson and Simon is not
N2-AMMON(IN) and NH3 applicable. The results of [37] suggest that, provided suitable
1 Reactant stream entering the first
plant data is available for validation, it is not necessary to
catalyst bed
2 Product stream exiting the first
calculate x and then estimate a second factor to account for
catalyst bed catalyst deactivation. Instead, simulation data based on the
3 Reactant stream entering the instantaneous reaction rate, rNH3, can be directly correlated
second catalyst bed with plant data to estimate an apparent effectiveness factor, b,
4 Product stream exiting the second which includes the effects of both diffusional limitations and
catalyst bed
catalyst aging. This is the approach adopted in this study, as
5 Reactant stream entering the third
described below.
catalyst bed
6 Product stream exiting the third
catalyst bed Calibration of model using industrial plant data
V Vapour phase exiting the first flash
separator The instantaneous reaction rate, rNH3, for ammonia synthesis
L Liquid phase exiting the first flash is dependent on the composition of the gas stream and tem-
separator
perature at each point in the catalyst bed (Eqs. (2)e(9)), varying
RECY-1 Purge stream from ammonia
across the bed width. To emulate the performance of an
reactor recycle loop
RECY-4 Recycle loop stream returning to operating industrial plant, the instantaneous reaction rate is
synthesis reactor modified by an apparent effectiveness factor, b, to match
S-3 Vapour phase exiting the second published plant test data. In this case, the industrial data is for
flash separator a state-of-the-art Uhde ammonia synthesis reactor
V-1 Vapour phase purge S-3 after comprising three catalyst beds with inter-bed cooling. The
compression from 2.5 MPa to 16 MPa
reactor is a cylindrical vessel, about 20 m high with an internal
T-GT(OUT) Total purge to gas turbine from 3
diameter of 2.8 m, containing three beds of 1.5e3 mm iron-
synthesis reactors
L-1 Liquid phase exiting the second based catalyst. The catalyst beds are of radial-flow configu-
flash separator ration, with an external diameter of approximately 2.5 m and
LIQUDNH3(OUT) Total liquid ammonia product from bed depth of 0.602 m, 0.694 m and 0.916 m for the 1st, 2nd and
3 synthesis reactors 3rd beds, respectively [29]. To simulate radial-flow conditions,
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37173

Table 3 e Catalyst bed dimensions and apparent effectiveness factor, b

Reactor Bed Volume, m3 Bed Depth, m Equiv. Diameter, m Void Fraction Residence Time, s b
1 17.46 0.602 6.077 0.241 8.18 0.5
2 13.43 0.694 4.964 0.241 5.28 0.7
3 20.96 0.916 5.398 0.225 8.68 0.7

each bed is modelled as a short, large diameter cylinder, with investigated by calculating the gas composition profile as a
a cumulative gas path length of 2.21 m through the reactor. function of distance across Bed 1, maintaining a constant total
The catalyst bed dimensions are shown in Table 3, including molar flowrate (26,409 kmol h1, as in Table 4) and average
estimates of the average bulk gas residence time in each operating conditions of 450  C and 17.1 MPa. The mole fraction
catalyst bed, derived from the single-pass industrial field test of NH3 in the feed gas is constant at 4.8%. The resulting data
results shown in Table 4. The simulation model is calibrated to on H2 conversion efficiency and NH3 production as a function
mimic the plant performance by varying the value of b by trial- of inlet H2:N2 ratio is presented in Fig. 4.
and-error to achieve the best match with the experimental The H2 conversion efficiency is seen to rise steeply to a
data. The estimated values of b for each bed are also included maximum of 26.5% at a H2:N2 ratio of 0.4, and then decrease
in Table 3. progressively as H2:N2 increases. The concentration of NH3
The best-match temperature estimates at the inlet and rises to a maximum of 62.7 mol% at a H2:N2 ratio of 1.5, and
outlet of each bed, and the final ammonia concentration, are then falls slowly as H2:N2 increases. Based on this data, a H2:N2
compared with the trial test results in Table 4. Vendor simu- ratio of 1.0 is selected to achieve both a high H2 conversion
lation data for the same test results [29] are compared with the efficiency and a high NH3 product concentration. This ratio is
present simulation results in Fig. 3. The close match between used in subsequent simulation designs.
the ammonia concentrations and temperature profiles within
each bed serves to validate the accuracy of the developed Scaleup of catalyst beds
simulation model. It can be seen that, under the operating
conditions used, the reactor comes close to achieving equi- Having calibrated the Aspen Plus model to simulate real in-
librium in all beds. The single-pass conversion of hydrogen to dustrial plant performance, the next step is to scale the
ammonia in this case is 27%, consistent with typical com- reactor to a standard commercial scale of 1500 t d1 NH3. This
mercial experience [16]. involves adjusting the catalyst bed dimensions to maintain
constant gas residence time.
Effect of feed composition on ammonia production By stoichiometry, an ammonia production rate of 1500 t d1
requires a hydrogen feed rate of 11 t h1. Nitrogen is provided
From Eq. (1), reaction stoichiometry suggests that a H2/N2 as both fresh feed and recycle, with both streams combining
molar ratio of 3:1 would be optimal for NH3 production. at the reactor inlet with a H2:N2 molar ratio of 1:1. Adopting a
However, Eq. (2) indicates that, for constant temperature and simplified scenario, reactor outlet gas is separated into a liquid
pressure, the rate of NH3 formation is related to the mole ammonia stream and a recycle stream, with 10 mol% of the
fraction of H2, N2 and NH3 in a complex fashion. The influence recycle stream being purged. The fresh nitrogen feed rate and
of the feed H2:N2 molar ratio on the reactor performance was the catalyst bed equivalent diameters are both adjusted
incrementally to achieve a H2:N2 ratio of 1.0 at the reactor
inlet, while maintaining constant gas residence times in each
Table 4 e Comparison of industrial field test data and bed. In this scenario, the required N2 feed rate is 57.3 t h1
Aspen Plus simulation results. (feed H2:N2 molar ratio of 2.69:1), and the total bed volume
being is increased from 51.85 m3 (Table 3) to 80.16 m3 (Table 5).
Data Field Data [29] Simulation Data
The resulting NH3 concentration and temperature profiles
Feed composition, mol. frac.
within each catalyst bed are shown in Fig. 5. The profiles are
N2 0.1986
seen to be very similar to those in Fig. 3. This confirms that the
H2 0.6230
Ar 0.0589 scaled-up catalyst beds exhibit similar performance to the
NH3 0.0270 commercial plant.
C1 0.0943
Molar flow, kmol h1 26409.44 Ammonia synthesis reactor performance
Pressure drop across bed, MPa 0.17
Inlet temperature, oC 265 15
The final step in the design process is to incorporate the
Inlet pressure, MPa(abs) 17.1 17.1
scaled-up ammonia synthesis reactor into the complete
Bed 1 inlet temperature, oC 381, 371, 391 394
Bed 1 outlet temperature, oC 519, 519, 512 509 AMMONIA Hierarchy block (Fig. 2), integrated with the lignite-
Bed 2 inlet temperature, oC 430, 434, 437 437 to-hydrogen flowsheet (Fig. 1). The design flowrate of
Bed 2 outlet temperature, oC 471, 475, 479 476 hydrogen from the lignite-to-hydrogen plant is 32.4 t h1, and
Bed 3 inlet temperature, oC 411, 411, 416 414 the ammonia synthesis reactor is designed to handle 11 t h1
Bed 3 outlet temperature, oC 439, 443, 448 448 hydrogen, so three ammonia reactors operating in parallel are
NH3 concentration out, vol% 15.84 15.92
needed to provide the necessary capacity.
37174 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

(a)

0.18

0.16

0.14

NH3 Mole Fraction

0.12

0.10

0.08

0.06

0.04

0.02
BED 1 BED 2 BED 3
0.00
0.0 0.5 1.0 1.5 2.0
Reactor Length (m)

(b)

550

500
Reactor Temperature ( C)

450

400

350

300

250
BED 1 BED 2 BED 3
200
0.0 0.5 1.0 1.5 2.0
Reactor Length (m)

Fig. 3 e Comparison between industrial test and simulation data for (a) ammonia concentration and (b) temperature profiles
in each catalyst bed (d Simulation results; B Vendor data from Ref. [29]).

The feed rate of H2 to each ammonia synthesis reactor is production of 32.4 t h1 low-emissions hydrogen from lignite,
set at 10.8 t h1, and the N2 feed rate was set at 57.3 t h1, as in with all required electricity co-generated on site and conver-
Section 2.5 above. Initially, the catalyst beds were scaled up sion of the hydrogen to liquid ammonia. The assumptions
using a simplified scenario in which 10% of the recycle stream involved in estimating the energy intensity of this process are
was purged. In the final design, a purge of only 0.7 mol% of the summarised in Table 6.
recycle gas is found to be sufficient to prevent accumulation of
argon in the system, due to the high purity of the reactant
gases (H2 at 99.999% purity and N2 at >99.9% purity). However, Results
the simulation is unstable under these conditions and does
not converge. This problem is solved by also purging the Considering the performance of a single ammonia synthesis
vapour phase outlet from the second flash unit, which is sent unit, the simulation results for the main streams associated
to the gas turbine for combustion of residual H2 and NH3. This with the AMMONIA Hierarchy block are presented in Table 7.
increases the total purge volume to 1.3 mol% and allows the Based on this data, Fig. 6 illustrates a simplified schematic of
simulation to converge. the system inputs and outputs.
The methodology described in Ref. [15] and its Supple- The mass flows of fresh hydrogen and nitrogen in the
mentary Information is used to estimate the total resource combined input stream (S-1) are 10.8 t h1 and 57.3 t h1,
requirements and CO2 emissions intensity associated with respectively, giving a feedstock H2:N2 molar ratio of 2.64:1.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37175

70.0 at 176.4 MW and 175.5 MW respectively. It was expected that

ammonia synthesis would be less energy-intensive than
60.0 hydrogen liquefaction, but new designs for large-scale
hydrogen liquefaction offer significant efficiency improve-
50.0 ments [43]. The energy required by COMP1 to compress
hydrogen and nitrogen to the Haber-Bosch reaction pressure
40.0 is 171.9 MW, largely determined by the high mass flow of ni-
Percent, %

trogen. The two other compressors, COMP2 and COMP3, are

30.0 tiny in comparison, consuming 2.9 MW and 0.7 MW
respectively.
20.0
Since the power requirements for hydrogen liquefaction
and ammonia synthesis are the same, the lignite feed rate was
10.0
set as equal in each case to better highlight the differences
between the two scenarios. Consequently, the mass flows of
0.0
0.0 1.0 2.0 3.0 4.0 inputs and outputs associated with gaseous hydrogen pro-
H2:N2 Molar Ratio duction are also the same.
The LH2 scenario is designed so that a lignite feed rate of
Fig. 4 e Conversion efficiency of H2 (✕) and mol% of NH3 () 1050 t h1 is sufficient for production of 32.4 t h1 hydrogen
as a function of H2:N2 ratio. and enough electricity to just satisfy the needs of the process
(see Table S1 for details). In the NH3 scenario, the same lignite
feed rate supports an excess of 26.9 MW electricity production
After mixing with the recycle loop (RECY-4), the H2:N2 molar due to combustion of purge gas from the ammonia reactors in
ratio at the reactor inlet (1) is 0.9:1. Liquid ammonia at 3  C, the gas turbine. This extra electricity could potentially be used
2.5 MPa is produced at a purity of 99.1 mol%, at a rate of within the process to slightly reduce the amount of lignite
59.4 t h1 or 1426 t d1. feed required and the overall emissions intensity of the
Of the 10.8 t h1 H2 entering the system, 0.3 t h1 is lost in process.
the combined purge streams and none is present in the liquid In both scenarios, the heat that can be recovered from the
ammonia product. Thus, 10.5 t h1 H2 is converted to NH3, steam cycle is sufficient to match the endothermic load within
representing an overall conversion efficiency of 97.2%. The the hydrogen production process (see Table S2 for details). In
flow of H2 is 28.2 t h1 at the reactor entry and 17.7 t h1 at the the NH3 scenario, there is an excess of 148 MWth associated
exit, representing a conversion efficiency of 37.2% per pass. with waste heat from the inlet gas compression system.
This value is higher than anticipated from Fig. 4, which was Normal industrial practice is to use water as the compressor
based on 10 mol% purge from the recycle stream. In this case cooling medium, producing hot water that can be used for
the purge is only 1.3%, so the additional H2 recycled to the heat integration [44]. Alternatively, some of this waste heat
reactor results in higher conversion efficiency. could be converted to electricity using an organic Rankine
The designed ammonia synthesis reactor thus achieves a cycle system. This is an established industrial process [45], but
representative industrial-scale capacity and a high conversion a more detailed study is required to assess the potential for
efficiency. The design concept involves three such reactors extra electricity generation in this particular situation.
operating in parallel, to convert 32.4 t h1 H2 and 171.9 t h1 N2 Since both the LH2 and NH3 scenarios involve the same
into 178.2 t h1 NH3 at 99.1 mol% purity. lignite feed rate, the quantities of CO2 produced and captured
Integration of the AMMONIA Hierarchy block into the are also the same (see Table S3 for details). As previously
larger lignite-to-hydrogen simulation allows the resource re- described [15], CO2 capture in the hydrogen production pro-
quirements and emissions intensity of ammonia production cess is mainly accomplished using a two-stage Selexol™
to be estimated. Drawing upon the equivalent data from our process, which achieves an apparent CO2 capture efficiency of
earlier study [15], a direct comparison can be made between 91.7%. If required, an additional MDEA capture system can be
liquefied hydrogen (LH2) and ammonia (NH3) production. The installed to capture CO2 from the gas turbine flue gas,
respective data is presented for comparison in Table 8, with increasing the apparent capture efficiency to 99.2%. The same
additional details provided in the Supplementary capture efficiencies also apply in the NH3 scenario because
Information. combustion of purged H2 and NH3 in the gas turbine does not
The first, and most surprising, point to note from these produce CO2.
results is that the electrical power requirements for hydrogen The specific emissions intensity (SEI) of producing liquefied
liquefaction and ammonia synthesis are essentially identical, hydrogen using only the Selexol™ system for CO2 capture is

Table 5 e Scaled up bed dimensions for conversion of 11 t h¡1 hydrogen to ammonia.

Reactor Bed Volume, m3 Bed Depth, m Equiv. Diameter, m b Residence Time, s
1 35.46 0.602 8.66 0.5 8.12
2 17.40 0.694 5.65 0.7 5.28
3 27.30 0.916 6.16 0.7 8.68
37176 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

(a)

0.18

0.16

0.14

NH3 Mole Fraction

0.12

0.10

0.08

0.06

0.04

0.02
BED 1 BED 2 BED 3
0.00
0.0 0.5 1.0 1.5 2.0
Reactor Length (m)

(b)

600

500
Reactor Temperature (oC)

400

300

200

100
BED 1 BED 2 BED 3
0
0.0 0.5 1.0 1.5 2.0
Reactor Length (m)

Fig. 5 e Simulation of scaled-up reactor showing (a) ammonia concentration and (b) temperature profile in each catalyst bed.

estimated as 2.73 kgCO2-e kg1 H2, including fugitive methane Haber-Bosch ammonia synthesis process for incorporation
emissions during lignite mining. This is consistent with the EU into our earlier lignite-to-hydrogen process model.
Taxonomy specification for ‘sustainable hydrogen’, currently Despite the widespread industrial use of the Haber-Bosch
set at 3.0 kgCO2-e kg1
H2 [15]. Adding the optional MDEA unit, to process, this is the first time that an Aspen Plus simulation
conform with a ‘net zero by 2050’ policy, would further reduce has been developed using modified Temkin rate kinetics
the SEI to 0.30 kgCO2-e kg1
H2. The results of this study (Table 8) calibrated to the popular iron-based catalyst, and scaled to
indicate that production of ammonia using a similar process representative industrial dimensions using actual plant data.
would have a SEI of 0.49 kgCO2-e kg1 NH3 using the Selexol™ This study is also unusual in that the H2:N2 ratio in the reactor
system alone, and 0.05 kgCO2-e kg1 NH3 with the optional MDEA feed was selected to maximise H2 conversion efficiency while
unit installed. minimising the recycle loop volume., rather than simply using
the stoichiometric ratio of 3:1. The scaled-up reactor design,
calibrated against industrial test data, provides a reasonable
Discussion simulation of a typical industrial-scale ammonia synthesis
plant. Incorporation of this simulation into the larger lignite-
Significance of this study to-hydrogen process model enables estimation of the
resource and energy requirements of a lignite-to-ammonia
The objective of this work is to extend our previous study on process, as well as the associated CO2 emissions intensity.
production of low emissions hydrogen from Victorian lignite Surprisingly, the simulation results show that the electrical
to investigate the relative merits of producing ammonia power requirement for ammonia synthesis (176.4 MW) is
rather than liquefied hydrogen for export. This comparison essentially the same as that needed for liquefaction of an
required the development of an Aspen Plus simulation of the equivalent output of hydrogen (175.5 MW). Given that
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37177

Table 6 e Electrical and thermal energy requirements of individual unit operations.

Unit operation Energy requirement Source
Lignite mining
Total electrical energy 5.61 kWeh t1 lignitea Plant data from AGL Loy Yang
Lignite drying
Electrical energy for compression 36.34 MWeh t1 lignitea Model estimate [15]
Additional thermal energy for drying 0.077 MWthh t1 lignitea Model estimate [15]
Specific milling energy
Lignite 6 kWeh t1 lignitea Model estimate [15]
Air Separation Unit
Electrical energy (air input basis) 0.060 MWeh t1 air Model estimate [15]
Electrical energy (O2 output basis) 0.263 MWeh t1 O2 (@ 95%) Model estimate [15]
Electrical energy (N2 output basis) 0.077 MWeh t1 N2 (@ 98%) Model estimate [15]
Acid Gas Removal (Selexol™)
Electrical energy required 0.012 MWeh t1 CO2 Model estimate [15]
Reboiler duty 0.099 MWeh t1 CO2 Model estimate [15]
Ammonia chiller
Cooling load 0.098 MWthh t1 CO2 Model estimate, based on [39]
Reboiler duty 0.147 MWthh t1 CO2 Model estimate, based on [39]
Sour water stripping
Reboiler duty 0.0245 MWthh t1 water Model estimate [15]
Post-combustion MDEA
Reboiler duty 1.333 MWthh t1 CO2 Replicated from [40]
CO2 dehydration (TEG)
Reboiler duty 1.660 MWthh t1 water Using gas stripping [41]
CO2 compression
Compressor duty 0.075 MWeh t1 liquid CO2 Replicated from [1]
Pressure Swing Adsorption
Compressor duty 0.5 MWeh t1 H2 Reported by [42]
Hydrogen pipeline transport
Compressor duty 1.59 MWeh t1 H2 Model estimate [15]
Hydrogen liquefaction
Compressor duty 6 MWeh t1 H2 Linde design [43]
Ammonia synthesis
Compressor duty 0.98 MWeh t1 NH3 Model estimate (current)
a
Based on as-mined wet lignite.

hydrogen liquefaction plants at the required scale are still comparable to the present study because it includes the en-
only at the conceptual stage, this finding suggests that con- ergy consumption during mining and preparation of the coal
version of hydrogen to ammonia presents a lower technical feedstock. Most reports on energy consumption in ammonia
risk for hydrogen export from Victoria with no additional production relate to SMR of natural gas, where the SEC ranges
energy penalty. In fact, this option may provide opportunities from 35 to 41 GJ t1 [47]. However, these values does not take
for additional electricity production from waste process heat, into account the energy involved in extraction, purification
but a detailed heat integration study is required to quantify and transport of natural gas, or the associated fugitive
this, which is beyond the present scope. methane emissions.
This study finds that 1050 t h1 lignite (LHV 24.91 MJ kg1) This is the first time that the greenhouse gas emissions
can produce 178.2 t h1 liquid ammonia, equating to a specific intensity of ammonia production from lignite have been
energy consumption (SEC) of 146.8 GJ t1. This compares well quantified, accounting for the full production chain from
with the value of 165.9 GJ t1 reported for the Indian fertiliser lignite mining to CO2 sequestration. Following the approach of
industry during the 1990s [46]. This figure is directly our previous study [15], two CO2 capture scenarios are

Table 7 e Stream data for each ammonia synthesis reactor.

Data S-1 1 6 V RECY-1 L V-1 L-1 RECY-4
o
Temperature, C 15.2 392.7 452.2 5 5 5 3 3 393
Pressure, MPa 2.75 17.1 17.0 16 16 16 25 25 17.1
Mass flow, t h1 68.1 491.2 491.3 427.1 4.3 64.0 4.6 59.4 423.0
H2 mass flow, t h1 10.8 28.2 17.7 17.6 0.2 0.0 0.1 0 17.4
N2 mass flow, t h1 57.3 443.7 394.9 390.2 3.9 4.7 3.8 0.9 386.3
NH3 mass flow, t h1 0 19.3 78.7 19.3 0.2 59.3 0.7 58.5 19.3
Ar mass flow, t h1 0.00002 0.001 0.001 0.001 0 0 0 0 0
37178 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

RECY-1 V-1

RECY-4

Ammonia
synthesis
S-1 1 reactor 6

L-1

Fig. 6 e Simplified mass flows around each ammonia synthesis reactor (t h¡1).

considered. In the baseline scenario, all CO2 capture is There is currently no agreed specification set for emissions
accomplished in the main Selexol™ acid gas recovery plant, from blue ammonia production, but the hydrogen used as
operating at an apparent capture efficiency of 91.7%. The SEI feedstock meets the EU Taxonomy specification for ‘sustain-
of ammonia production in this scenario is 0.49 kgCO2-e kg1
NH3, able hydrogen’ and no further emissions are produced in
including fugitive methane emissions during lignite mining. ammonia production, so the proposed lignite-to-ammonia
process should qualify as ‘sustainable’ or ‘low-emission’.
In the second CO2 capture scenario, the emissions intensity
Table 8 e Summary of process inputs and outputs for is reduced further by fitting a MDEA adsorption unit to the gas
each scenario. turbine flue gas stream. It is anticipated that this option may
Scenario become necessary as allowable emissions limits are progres-
sively lowered over the coming decades. The MDEA unit al-
Process inputs LH2 NH3
lows hydrogen to be produced with a SEI of 0.30 kgCO2-e kg1 H2,
As-mined lignite feed rate, t h1 1050 1050
Pulverised lignite feed rate, t h1 466.7 466.7 and ammonia at 0.05 kgCO2-e kg1 NH3. As previously shown, the

Oxygen flowrate, t h1 345 345 process could then be made net-zero by co-gasification with
Nitrogen flowrate, t h1 e 171.9 1.4 wt% biomass [15].
Process outputs There is a lack of industrial data that can be directly
Liquefied H2 production, t h1 32.4 e compared with the proposed lignite-to-ammonia process. The
NH3 production, t h1 e 178.2
SEI of large coal-based ammonia plants in China is reported to
CO2 captured (Selexol™ only), t h1 1170.9 1170.9
be 9.0 kgCO2-e kg1NH3 [13], but CCS is not used. In Japan, Ube
CO2 captured (with MDEA), t h1 1249.6 1249.6
Sulphur production, t h1 1.1 1.1 Corporation produces 1000 t d1 ammonia from coal but does
Water from lignite dryer, t h1 584 584 not capture CO2 emissions. In the USA, Coffeyville Resources
Slag production, t h1 10.8 10.8 produces 1,300 t d1 ammonia from petroleum coke, and
Fugitive CH4 emissions, t h1 0.012 0.012 captures 650,000 t y1 CO2 [48], but the emissions intensity has
Electrical power not been reported.
Power for H2 production, MWe 304.9 304.9
However, it has been reported that the SEI of SMR is 3.0
Power for H2 liquefaction, MWe 176.4
kgCO2-e kg1
e
Power for NH3 synthesis, MWe e 175.5 NH3 [13] and that this could be reduced to 0.44e0.54

Total power required, MWe 481.3 480.1 kgCO2-e kg1 NH3 with 90% CO2 capture (including fugitive

Total power generated, MWe 481 507 methane emissions [14]. On this basis, the estimated 0.49
Excess capacity, MWe 0.3 26.9 kgCO2-e kg1NH3 for the proposed lignite-to-ammonia process,
Thermal power with 91.7% CO2 capture using Selexol™, compares favourably
Thermal power required, MWth 549 549
with next-generation SMR processes.
Thermal power recovered, MWth 554 554
Victorian lignite is actually more advantageous than nat-
Extra thermal power available, MWth 0 148
Apparent capture efficiency of CO2 in syngas, % ural gas for low-emissions ammonia production, because it
Selexol™ only 91.7 91.7 has inherently low fugitive methane emissions and a stable,
Selexol™ plus MDEA 99.2 99.2 low cost of production. It also has the advantage of proximity
Emissions intensity, kg CO2-e kg1 H2 or NH3 to massive, high quality CO2 storage reservoirs. Thus, the
Selexol™ only 2.73 0.49 proposed lignite-to-ammonia process seems to offer signifi-
Selexol™ plus MDEA 0.30 0.05
cant advantages for the emerging clean hydrogen market.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37179

Table 9 e Renewable energy required for production of 32.4 t h¡1 H2 and conversion to liquid hydrogen or ammonia.
Unit Operation Energy Intensity G-LH2, MW G-NH3, MW Reference
RO seawater desalination 4 kWh t1 purified H2O 2.6 2.6 [50,51]
Electrolysis 1.08 * 50 MWh t1 H2 1749.6 1749.6 [52]
N2 production by PSA 110 kWh t1 N2 18.9 18.9 [53]
H2 liquefaction 6 MWh t1 H2 176.4 e [43]
H2, N2 compression 0.98 MWh t1 NH3 e 175.2 This study
Totals 1947.5 1946.3

Application to green hydrogen and ammonia production 600 MW, in mostly small and individually manufactured
plants [56]. The industry is not configured for production
Given public scepticism around production of low emission volumes beyond several thousand parts per year [57], and the
hydrogen from fossil fuels with CCS, it of interest to compare exotic materials required (e.g. iridium, yttrium, platinum,
the findings of this study with the option of producing the strontium, and graphite) are in limited supply [58]. It therefore
same quantity of ammonia using renewable energy. seems likely that the global capacity to produce green
Using renewable energy, pure H2 is produced by electrol- ammonia will be constrained over the next several decades.
ysis of water which is typically produced by desalination of In contrast, gasification of coal and lignite is a very well
seawater using reverse osmosis (RO). By stoichiometry, 1 established, conventional technology with a proven track re-
tonne of H2 requires 9 tonnes water, and the typical water cord for reliable production of valuable industrial chemicals
recovery from seawater desalination is 45% [49], making a and fuels. Gasification technologies are available from
total water requirement of 20 t water t1 H2. Typical energy numerous vendors, and are well suited for production of
intensity of seawater desalination by RO is around 4 kWh t1 hydrogen at the scale investigated in this study. Likewise, CCS
purified water [50,51]. Water electrolysis using proton ex- is now well established. There are currently 114 active com-
change membrane (PEM) technology is expected to have a SEC mercial CO2 injection enhanced oil recovery (EOR) projects in
of 50 kWh t1 H2 by 2030, but about 8% extra energy is required the United States that collectively inject over 1.8 Mt y1 CO2
to account for the auxiliary load which must be met before and produce over 280,000 barrels of oil per day [59]. There over
hydrogen can be generated, and the losses through the power 40 sites where EOR has been used to safely and securely store
electronics [52]. Pure N2 can be extracted from air using PSA, captured CO2 underground, and continuous monitoring has
with a SEC of 110 kWh t1 N2 [53]. Next-generation hydrogen verified that the wells are intact and secure. There are
liquefiers with capacities of hundreds of tonnes per day are currently 30 full-scale projects operating globally and 164 new
expected to have SEC close to or below 6 kWh kg1 LH2 [43]. The projects in development, with a total storage capacity of
present study has found that the SEC for compression of H2 241.6 Mt y1 [60]. CCS has demonstrably been proven to be a
and N2 to the Haber-Bosch reaction pressure is 0.98 MWh t1 NH3 safe and effective climate mitigation tool.
produced. Using this data, Table 9 summarises the electrical In Victoria, the CarbonNet Project has validated that safe
energy required to produce 32.4 t h1 H2 via electrolysis and CO2 storage sites are available in the near-offshore region of
convert it to either liquefied hydrogen (G-LH2) or liquid the Gippsland Basin [61]. In addition, ESSO Australia recently
ammonia (G-NH3). The total energy required to produce announced the South Eastern Australia CCS Hub, which from
178.2 t h1 green ammonia is 1946.3 MW, equating to a SEI of 2024 will use existing gas extraction infrastructure to store
10.9 MWh t1NH3. This is consistent with a previously published CO2 in the depleted Bream reservoir, located 46 km offshore in
estimate of 10.6 MWh t1 NH3 [54]. Bass Strait. As ESSO's oil and gas fields in Bass Strait reach the
These estimates of electrical energy consumption are end of their working lives, they can be converted to CO2
based on 100% supply availability. A significantly larger storage reservoirs, with potential capacity of 50e300 billion
installed capacity of renewable energy generation is needed to tonnes [62].
account for supply intermittency. For example, the availability Based on resource requirements and availability, produc-
of wind energy in Victoria averages 36% throughout the year tion of blue ammonia by lignite gasification with CCS in Vic-
[55], so provision of 1946.3 MW for ammonia production toria has significant advantages compared to green ammonia.
would require a wind farm with a total nameplate capacity of Gasification of lignite with CCS has much greater capacity
5.4 GW. Currently, the largest wind farm in Victoria is the than electrolysis to be deployed at large scale, so blue
420 MW Macarthur installation, which has 140 turbines over ammonia from Victorian lignite is well placed to play a role in
an area of 5,500 ha. Installation of 5.4 GW wind turbine ca- the global transition to low emissions fuels.
pacity for green ammonia production in Victoria would
require the equivalent of 13 Macarthur wind farms, poten-
tially requiring an area of over 72,000 ha. Conclusion
Apart from the very large wind turbine capacity needed to
support this output of green ammonia, the feasibility of This study extends our previous study, on production of low
sourcing sufficient electrolyser capacity is questionable. The emissions hydrogen from Victorian lignite, to examine pro-
estimated global installed capacity of electrolysers in 2021 was duction of ammonia rather than liquefied hydrogen as a
37180 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

hydrogen carrier for export. The objective of this study is to

evaluate the resource requirements and emissions intensity Appendix A. Supplementary data
associated with ammonia production by lignite gasification
with CCS. This involves development of an Aspen Plus simu- Supplementary data to this article can be found online at
lation of the Haber-Bosch ammonia synthesis process, and https://doi.org/10.1016/j.ijhydene.2023.06.098.
running it within our earlier lignite-to-hydrogen process
model. references
Surprisingly, the simulation results show that the electrical
power requirement for ammonia synthesis is essentially the
same as that needed for liquefaction of an equivalent output [1] Hydrogen Council. Hydrogen scaling up. 2017. https://
of hydrogen. On this basis both options are equally attractive, hydrogencouncil.com/en/study-hydrogen-scaling-up.
although ammonia synthesis is at a higher level of techno- [Accessed 8 December 2022].
logical maturity than large-scale hydrogen liquefaction. [2] Mokhatab S, Mak JY, Valappil J, Wood DA. Handbook of
This study shows that blue ammonia can be produced from liquefied natural gas. Gulf Professional Publishing.; 2013.
Victorian lignite with very low carbon intensity (0.49 kgCO2-e p. 624. ISBN 0124046452.
[3] Juangsa FB, Irhamna AR, Aziz M. Production of ammonia as
kg1
NH3) equivalent to that of with next-generation SMR þ CCS
potential hydrogen carrier: review on thermochemical and
processes. If required, the emission intensity can be reduced electrochemical processes. Int J Hydrogen Energy
to 0.05kgCO2-e kg1 NH3 with a post-combustion MDEA system, 2021;46(27):14455e77. https://doi.org/10.1016/
and then made carbon neutral by co-gasification with 1.4% j.ijhydene.2021.01.214.
biomass. This suggests that Victoria is well placed to become a [4] Yoshino Y, Harada E, Inoue K, Yoshimura K, Yamashita S,
significant supplier of low emissions ammonia to the world Hakamada K. Feasibility study of “CO2 free hydrogen chain”
utilizing Australian brown coal linked with CCS. Energy Proc
market, consistent with global emissions reductions targets
2012;29:701e9. https://doi.org/10.1016/j.egypro.2012.09.082.
over the next few decades.
[5] Kamiya S, Nishimura M, Harada E. Study on introduction of
A limitation of this study is that the treatment of thermal CO2 free energy to Japan with liquid hydrogen. Phys Procedia
energy flows is restricted to satisfying process demands using 2015;67:11e9. https://doi.org/10.1016/j.phpro.2015.06.004.
high-value heat from the gasifier, and does not address uti- [6] HESC. In: Successful completion of pilot project report; 2022.
lisation of the excess heat generated by the exothermic https://www.hydrogenenergysupplychain.com/report-
ammonia synthesis reaction. This heat could potentially be successful-completion-of-the-hesc-pilot-project;. [Accessed
8 December 2022].
used to generate electricity for use within the plant, thereby
[7] Challenge Zero. Development of a technology for large-scale
reducing the load on the gasifier and gas treatment system, or storage and transportation of hydrogen, https://www.
for sale off-site. Further research is recommended on recovery challenge-zero.jp/en/casestudy/647 [accessed 8 December
of energy from the low grade waste heat streams and oppor- 2022].
tunities for additional electricity generation using the organic [8] Rao PC, Yoon M. Potential liquid-organic hydrogen carrier
Rankine cycle. (LOHC) systems: a review on recent progress. Energies
2020;13(22):6040. https://doi.org/10.3390/en13226040.
[9] Ghavam S, Vahdati M, Wilson I, Styring P. Sustainable
ammonia production processes. Front Energy Res 29 March
Role of funding source 2021. https://doi.org/10.3389/fenrg.2021.580808.
[10] Muraki S. Development of technologies to utilize green
Australian Carbon Innovation (ACI) is a private member-based ammonia in energy market. In: 2018 NH3 fuel conference.
company that provides funding for R&D of innovative carbon- Powerpoint presentation; 2018. https://www.
based technologies in Australia, and encourages public ammoniaenergy.org/wp-content/uploads/2021/02/AEA-Imp-
Con-01Nov18-Shigeru-Muraki-Keynote-Address.pdf;.
dissemination of the findings. As well as providing funding,
[Accessed 8 December 2022].
ACI also played a collaborative role in this project, assisting
[11] Erisman JW, Sutton MA, Galloway J, Klimont Z,
Monash University with the literature review and drafting of Winiwarter W. How a century of ammonia synthesis
the manuscript. changed the world. Nat Geosci 2008;1(10):636e9. https://
doi.org/10.1038/ngeo325.
[12] International Energy Agency. Ammonia technology
Declaration of competing interest roadmap. 2021. https://www.iea.org/reports/ammonia-
technology-roadmap. [Accessed 8 December 2022].
[13] Zhang W-f, Dou Z-x, He P, Ju X-T, Powlson D, Chadwick D,
The authors declare that they have no known competing
et al. New technologies reduce greenhouse gas emissions
financial interests or personal relationships that could have from nitrogenous fertilizer in China. Proc Natl Acad Sci USA
appeared to influence the work reported in this paper. 2013;110:8375e80. https://doi.org/10.1073/pnas.1210447110.
[14] Stocks M, Fazeli R, Hughes L, Beck FJ. Global emissions
implications from co-combusting ammonia in coal fired
power stations: an analysis of the Japan-Australia supply
Acknowledgements chain. J Clean Prod 2022;336:130092. https://doi.org/10.1016/
j.jclepro.2021.130092.
Funding for this project was provided by Australian Carbon [15] Kibria MA, McManus D, Bhattacharya S. Options for net zero
Innovation (ACI), a not-for-profit private member-based emissions hydrogen from Victorian lignite. Part 1: gaseous
and liquefied hydrogen. Int J Hydrogen Energy 2023. https://
company, with the support of the Victorian Government.
doi.org/10.1016/j.ijhydene.2023.04.213.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2 37181

[16] Appl M. Ammonia, 2. Production processes. Ullmann's [32] Stork K, Wiles CF, Mathur GK. Ammonia production using
Encyclopedia of Industrial Chemistry 2000. https://doi.org/ the Texaco coal gasification process. Nitrogen
10.1002/14356007.o17_o02. 1986;(160):39e50.
[17] Andersson J, Lundgren J. Techno-economic analysis of [33] Rath LK, Chou VH, Kuehn NJ. Assessment of hydrogen
ammonia production via integrated biomass gasification. production with CO2 capture volume 1: baseline state-of-the-
Appl Energy 2014;130:484e90. https://doi.org/10.1016/ art plants. National Energy Technology Laboratory (NETL),
j.apenergy.2014.02.029. Pittsburgh, PA, Morgantown, WV; 2011. https://www.osti.
[18] Arora P, Hoadley AF, Mahajani SM, Ganesh A. Small-scale gov/servlets/purl/1767148. [Accessed 8 December 2022].
ammonia production from biomass: a techno-enviro- [34] Jones D, Bhattacharyya D, Turton R, Zitney SE. Optimal
economic perspective. Ind Eng Chem Res 2016;55:6422e34. design and integration of an air separation unit (ASU) for an
https://doi.org/10.1021/acs.iecr.5b04937. integrated gasification combined cycle (IGCC) power plant
[19] Aziz M, Putranto A, Biddinika MK, Wijayanta AT. with CO2 capture. Fuel Process Technol 2011;92:1685e95.
Energy-saving combination of N2 production, NH3 synthesis, https://doi.org/10.1016/j.fuproc.2011.04.018.
and power generation. Int J Hydrogen Energy [35] Yu B-Y, Chien I-L. Design and economic evaluation of a coal-
2017;42:27174e83. https://doi.org/10.1016/ based polygeneration process to coproduce synthetic natural
j.ijhydene.2017.09.079. gas and ammonia. Ind Eng Chem Res 2015;54:10073e87.
[20] Darmawan A, Ajiwibowo MW, Tokimatsu K, Aziz M. Efficient https://doi.org/10.1021/acs.iecr.5b02345.
co-production of power and ammonia from black liquor. Int J [36] Zhang H, Wang L, Mare chal F, Desideri U. Techno-economic
Hydrogen Energy 2020;45:34437e48. https://doi.org/10.1016/ comparison of green ammonia production processes. Appl
j.ijhydene.2020.02.196. Energy 2020;259:114135. https://doi.org/10.1016/
[21] Ishaq H, Dincer I. Dynamic modelling of a solar hydrogen j.apenergy.2019.114135.
system for power and ammonia production. Int J Hydrogen [37] Elnashaie S, Mahfouz A, Elshishini S. Digital simulation of an
Energy 2021;46:13985e4004. https://doi.org/10.1016/ industrial ammonia reactor. Chem Eng Process: Process
j.ijhydene.2021.01.201. Intensif 1988;23:165e77. https://doi.org/10.1016/0255-
[22] Dyson D, Simon J. Kinetic expression with diffusion 2701(88)80013-8.
correction for ammonia synthesis on industrial catalyst. Ind [38] Elnashaie SS, Abashar ME, Al-Ubaid AS. Simulation and
Eng Chem Fundam 1968;7:605e10. https://doi.org/10.1021/ optimization of an industrial ammonia reactor. Ind Eng
i160028a013. Chem Res 1988;27:2015e22. https://doi.org/10.1021/
[23] Azarhoosh M, Farivar F, Ebrahim HA. Simulation and ie00083a010.
optimization of a horizontal ammonia synthesis reactor [39] Mansouri R, Boukholda I, Bourouis M, Bellagi A. Modelling
using genetic algorithm. RSC Adv 2014;4:13419e29. https:// and testing the performance of a commercial ammonia/
doi.org/10.1039/C3RA45410J. water absorption chiller using Aspen-Plus platform. Energy
[24] Shamiri A, Aliabadi N. Modeling and performance 2015;93:2374e83. https://doi.org/10.1016/
improvement of an industrial ammonia synthesis reactor. j.energy.2015.10.081.
Chemical Engineering Journal Advances 2021;8:100177. [40] Aspen Technology. Rate-based model of the CO2 capture
https://doi.org/10.1016/j.ceja.2021.100177. process by MDEA using aspen plus. Cambridge, MA, USA:
[25] Suhan MB, Hemal MN, Choudhury MS, Mazumder MA, Aspen Technology Inc; 2016. https://pdfcoffee.com/enrtl-rk-
Islam M. Optimal design of ammonia synthesis reactor for a rate-based-mdea-model-pdf-free.html. [Accessed 8
process industry. Journal of King Saud University- December 2022].
Engineering Sciences 2022 Jan 1;34(1):23e30. https://doi.org/ [41] Sakheta A, Zahid U. Process simulation of dehydration unit
10.1016/j.jksues.2020.08.004. for the comparative analysis of natural gas processing and
[26] Araújo A, Skogestad S. Control structure design for the carbon capture application. Chem Eng Res Des
ammonia synthesis process. Comput Chem Eng 2018;137:75e88. https://doi.org/10.1016/j.cherd.2018.07.004.
2008;32:2920e32. https://doi.org/10.1016/ [42] Lee D-Y, Elgowainy A. By-product hydrogen from steam
j.compchemeng.2008.03.001. cracking of natural gas liquids (NGLs): potential for large-
[27] Yoshida M, Ogawa T, Imamura Y, Ishihara KN. Economies of scale hydrogen fuel production, life-cycle air emissions
scale in ammonia synthesis loops embedded with iron-and reduction, and economic benefit. Int J Hydrogen Energy
ruthenium-based catalysts. Int J Hydrogen Energy 2021 Aug 2018;43(43):20143e60. https://doi.org/10.1016/
18;46(57):28840e54. https://doi.org/10.1016/ j.ijhydene.2018.09.039.
j.ijhydene.2020.12.081. [43] Krasae-In S. Optimal operation of a large-scale liquid
[28] Rossetti I, Tripodi A, Tommasi M, Ramis G. Conceptual hydrogen plant utilizing mixed fluid refrigeration system. Int
design of a process for hydrogen production from waste J Hydrogen Energy 2014;39:7015e29. https://doi.org/10.1016/
biomass and its storage in form of liquid ammonia. Int J j.ijhydene.2014.02.046.
Hydrogen Energy 2023. https://doi.org/10.1016/ [44] Kirova-Yordanova Z. Exergy analysis of industrial ammonia
j.ijhydene.2023.01.261. synthesis. Energy 2004 Oct 1;29(12e15):2373e84. https://
[29] El-Gharbawy M, Shehata W, Gad F. Ammonia converter doi.org/10.1016/j.energy.2004.03.036.
simulation and optimization based on an innovative [45] Tchanche BF, Lambrinos G, Frangoudakis A, Papadakis G.
correlation for (Kp) prediction. J Univ Shanghai Sci Technol Low-grade heat conversion into power using organic
2021;35:323e37. ISSN: 1007-6735. Rankine cycleseA review of various applications. Renew
[30] Dybkjaer I. Ammonia production processes. In: Nielsen A, Sustain Energy Rev 2011 Oct 1;15(8):3963e79. https://doi.org/
editor. Ammonia: catalysis and manufacture. Berlin, 10.1016/j.rser.2011.07.024.
Heidelberg: Springer; 1995. p. 199e327. https://doi.org/ [46] Schumacher K, Sathaye J. India's fertilizer industry:
10.1007/978-3-642-79197-0_6. productivity and energy efficiency, report LBNL-41846.
[31] Sueyama T. Features and practical use of Texaco coal Berkley, CA: Lawrence Berkeley National Laboratory; 1999.
gasification process. Coal science symposium and 91st https://www.osti.gov/servlets/purl/764326. [Accessed 22
special meeting on coke, osaka (Japan), 7-8 nov 1991. In: March 2023].
Proceedings of the joint meeting of the fuel society of Japan; [47] Rafiqul I, Weber C, Lehmann B, Voss A. Energy efficiency
1991. improvements in ammonia productiondperspectives and
37182 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 8 ( 2 0 2 3 ) 3 7 1 6 6 e3 7 1 8 2

uncertainties. Energy 2005 Oct 1;30(13):2487e504. https:// forecasting/ntndp/2016/report/2016-national-transmission-

doi.org/10.1016/j.energy.2004.12.004. network-development-plan.pdf?
[48] Barkley N. Petroleum coke gasification based ammonia plant. la¼en&hash¼6711A86D8DFD27BA8491A3B5A07A02DE.
AIChE 51st annual safety in ammonia plants and related [Accessed 8 December 2022].
facilities symposium, Vancouver, Canada 2006. [56] Odenweller A, Ueckerdt F, Nemet GF, Jensterle M, Luderer G.
[49] Gude VG. Energy consumption and recovery in reverse Probabilistic feasibility space of scaling up green hydrogen
osmosis. Desalination Water Treat 2011;36:239e60. https:// supply. Nat Energy 2022;7:854e65. https://doi.org/10.1038/
doi.org/10.5004/dwt.2011.2534. s41560-022-01097-4.
[50] Ghaffour N, Missimer TM, Amy GL. Technical review and [57] Ayers K, Danilovic N, Ouimet R, Carmo M, Pivovar B,
evaluation of the economics of water desalination: current Bornstein M. Perspectives on low-temperature electrolysis
and future challenges for better water supply sustainability. and potential for renewable hydrogen at scale. Annu Rev
Desalination 2013;309:197e207. https://doi.org/10.1016/ Chem Biomol Eng 2019;10:219e39. https://doi.org/10.1146/
j.desal.2012.10.015. annurev-chembioeng-060718-030241.
[51] Kim J, Park K, Yang DR, Hong S. A comprehensive review of [58] Badgett A, Brauch J, Buchheit K, Hackett G, Li Y, Melaina M,
energy consumption of seawater reverse osmosis et al. Water electrolyzers and fuel cells supply chain-supply
desalination plants. Appl Energy 2019;254:113652. https:// chain deep dive assessment. USDOE Office of Policy; 2022.
doi.org/10.1016/j.apenergy.2019.113652. https://www.energy.gov/sites/default/files/2022-02/Fuel%
[52] Haywood JA. The economics of producing H2 from 20Cells%20%26%20Electrolyzers%20Supply%20Chain%
electrolysis in Victoria. CSIRO, Australia. In: Options for 20Report%20-%20Final.pdf. [Accessed 8 December 2022].
production of low cost CO2-free hydrogen from Victorian [59] Saini D. Simultaneous CO2-EOR and storage projects. In:
Brown coal; 2018. https://www.acinnovation.com.au/aci- Engineering aspects of geologic CO2 storage. SpringerBriefs
reports. [Accessed 8 December 2022]. in petroleum geoscience & engineering. Springer; 2017.
[53] Morgan E, Manwell J, McGowan J. Wind-powered ammonia p. 11e9. https://doi.org/10.1007/978-3-319-56074-8_2.
fuel production for remote islands: a case study. Renew [60] Global CCS Institute. In: Global status of CCS 2022; 2022.
Energy 2014;72:51e61. https://doi.org/10.1016/ https://www.globalccsinstitute.com/resources/global-
j.renene.2014.06.034. status-of-ccs-2022. [Accessed 23 March 2023].
[54] Smith C, Hill AK, Torrente-Murciano L. Current and future [61] Earth Resources. The CarbonNet project. State Government
role of HabereBosch ammonia in a carbon-free energy of Victoria, https://earthresources.vic.gov.au/projects/
landscape. Energy Environ Sci 2020;13(2):331e44. https:// carbonnet-project [accessed 8 December 2022].
doi.org/10.1039/C9EE02873K. [62] The South East Australia carbon capture Hub, https://www.
[55] Australian Energy Market Operator. National transmission exxonmobil.com.au/energy-and-environment/energy-
network development plan. 2016 Dec. p. 110. https://aemo. resources/upstream-operations/the-south-east-australia-
com.au/-/media/files/electricity/nem/planning_and_ carbon-capture-hub [accessed 8 December 2022].