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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5

Available online at www.sciencedirect.com

ScienceDirect

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

Hydrogen production using methane: Techno-

economics of decarbonizing fuels and chemicals

Brett Parkinson a,b, Mojgan Tabatabaei a, David C. Upham c,

Benjamin Ballinger a, Chris Greig a, Simon Smart a, Eric McFarland a,d,*
a
Dow Centre for Sustainable Engineering Innovation, School of Chemical Engineering, University of Queensland, St
Lucia, QLD, 4072, Australia
b
Imperial College London, School of Chemical Engineering, Kensington, London SW7 2AZ, United Kingdom
c
Department of Chemistry and Biochemistry, University of California, Santa Barbara, CA 93106-9510, USA
d
Department of Chemical Engineering, University of California, Santa Barbara, CA 93106-5080, USA

article info abstract

Article history: In the near-to-medium future, hydrogen production will continue to rely on reforming of
Received 21 September 2017 widely available and relatively low-cost fossil resources. A techno-economic framework is
Received in revised form described that compares the current best practice steam methane reforming (SMR) with
10 December 2017 potential pathways for low-CO2 hydrogen production; (i) Electrolysis coupled to sustainable
Accepted 12 December 2017 renewable electricity sources; (ii) Reforming of hydrocarbons coupled with carbon capture
Available online 6 January 2018 and sequestration (CCS) and; (iii) Thermal dissociation of hydrocarbons into hydrogen and
carbon (pyrolysis). For methane pyrolysis, a process based on a catalytic molten Ni-Bi alloy
Keywords: is described and used for comparative cost estimates. In the absence of a price on carbon,
Hydrogen production SMR has the lowest cost of hydrogen production. For low-CO2 hydrogen production,
Methane pyrolysis methane pyrolysis is significantly more economical than electrochemical-based processes
Decarbonization using commercial renewable power sources. At a carbon price exceeding $21 t
1 CO2
Low-CO2 equivalent, pyrolysis may represent the most cost-effective means of producing low-CO2
Techno-economic hydrogen and competes favorably to SMR with carbon capture and sequestration. The
current cost disparity between renewable and fossil-based hydrogen production suggests
that if hydrogen is to fulfil an expanding role in a low CO2 future, then large-scale pro-
duction of hydrogen from methane pyrolysis is the most cost-effective means during the
transition period while infrastructure and end-use applications are deployed.
© 2017 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

10 trillion kilograms of carbon in low cost fossil fuels annually.

Introduction
Fossil hydrocarbons possess enormous stored chemical po-
tential from prehistoric photosynthetic processes. Humanity
has globally prospered because of their widespread availability
and relatively low cost. At present we combust approximately
Their use, however, is ultimately limited by the fact that they
are a finite resource. In the shorter term, before the prices in-
crease due to depletion, the risk of unacceptable environ-
mental consequences from increased atmospheric CO2 may
drive society towards alternatives to ensure sustained human

* Corresponding author. Department of Chemical Engineering, University of California, Santa Barbara, CA 93106-5080, USA
E-mail address: ewmcfar@engineering.ucsb.edu (E. McFarland).
https://doi.org/10.1016/j.ijhydene.2017.12.081
0360-3199/© 2017 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
prosperity [1]. The immense challenge of shifting away from
carbon dioxide intensive processes to more sustainable alter-
natives will require many massive projects and investments
spanning decades and generations [2].
Of the fossil resources, methane in natural gas provides
the most energy per unit mass of carbon dioxide produced
and is potentially the largest fossil reserve (if hydrates are
included) [3]. Substitution of coal by methane would achieve
significant emissions reductions. However, for the long term
moving to fuels that produce no CO2 is highly desirable.
Hydrogen, burns cleanly producing only water and can be
used in fuel cells; it has long been a candidate for a major
future fuel. Today, 96% of global hydrogen is produced by
fossil fuel reforming technologies (Eqs. (1) and (2)) and is
predominately used in the production of ammonia-based
fertilizers for agriculture [2]. The remaining 4% is produced
by electrolysis that splits water with electricity that could
potentially be CO2 free (Eq. (3)).
The most common hydrocarbon based hydrogen produc-
tion technology is steam methane reforming (SMR) which
satisfies approximately 48% of global demand. SMR produces
stoichiometric amounts of process-based CO2 from the com-
bination of steam reforming and the water-gas shift reactions
and additional combustion-based CO2 when supplying the
heat required to drive the overall endothermic reaction [4].
CH4 þ 2H2 O 4CO2 þ 4H2 DH0 ¼ 165 kJ mol
1 ;
DG0 ¼ 114 kJ mol
1
C þ 2H2 O/ CO2 þ 2 H2 DH0 ¼ 90 kJ mol
1 ; DG0 ¼ 63 kJ mol
1
2H2 O/ O2 þ 2H2 DH0 ¼ 572 kJ mol
1 ; DG0 ¼ 47 kJ mol
1
A typical SMR hydrogen plant produces approximately
9e10 tonnes of CO2 equivalent (t CO2e) per tonne of hydrogen
[5]. By contrast, if hydrogen were to be produced by coal
gasification (currently 18% of global H2 production), the CO2
emissions per tonne of hydrogen product would double
compared to SMR. Put simply, hydrogen produced from fossil
hydrocarbons using today's fossil-based technology has
diminished value as a fuel in a carbon constrained world.
Alternate processes for hydrogen production from non-
fossil resources are commercially available for small mar-
kets, but to be economically competitive for large-scale ap-
plications the technologies need to compete with technologies
that utilize low-cost and abundant hydrocarbons [6]. In the
most ambitious vision, all the hydrogen required would be
liberated from water using sustainable energy sources and
utilized in fuel cells or burned in hydrogen-based thermal
engines [7]. At present, renewable energy-based processes for
hydrogen production (electrolysis of water coupled with
photovoltaic cells or wind turbines, photo-catalysis, photo-
biological water decomposition, etc.) cannot produce
hydrogen at a price or scale that is competitive with fossil
fuels [8]. As such, the use of fossil hydrocarbons such as
methane for power, heat and electricity production is likely to
continue until they become less economically competitive
than alternatives like hydrogen [8].
Low-cost methane can be used to produce hydrogen
without producing carbon dioxide by pyrolysis (Eq. (4)) which
produces hydrogen and solid carbon [9]. Previous attempts at
2541
commercializing pyrolysis technologies for hydrogen pro-
duction have been limited by the need to use a solid catalyst,
which rapidly deactivates, decreasing hydrogen yields. How-
ever, for production of more valuable solid carbon products
(e.g. carbon black) pyrolysis has been used commercially
[10,11].
CH4 4CS þ 2H2 DH0 ¼ 75:6 kJ mol
1 ; DG0 ¼ 50:8 kJ mol
1 (4)
Numerous studies concerning future hydrogen demand
have been published [12e14]. We are interested in the pro-
duction of large-volumes of low-cost hydrogen with very low
associated CO2 emissions, which would increase the rate of
adoption of power generation using hydrogen as well as
hydrogen vehicles. Applying rigorous economic analyses to
proof-of-concept research and development technologies can
provide critical guidance on future resource allocation,
research and development priorities and technology pro-
jections. Accordingly, in this study we develop a conceptual
design for a novel methane pyrolysis process and describe a
techno-economic framework for comparing low-carbon
hydrogen production options using the dominant pathways
actively discussed in literature [8]. The pathways are: (i)
Electrolysis coupled to sustainable electricity sources; (ii)
Reforming of hydrocarbons coupled with carbon capture and
sequestration (CCS) and; (iii) Thermal dissociation of hydro-
carbons into hydrogen and carbon (pyrolysis). These com-
(1)
parisons can help us to address the following questions and in
doing so guide research, development and demonstration
priorities:
(2)
(3) (i) What are the comparative economics of hydrogen pro-
duction from water electrolysis, SMR, SMR with CCS,
and methane pyrolysis?
(ii) How does a price on carbon (i.e. $ t
1 CO2e) influence the
overall relative economics?
(iii) What innovations might make methane pyrolysis
compete directly with SMR without a carbon price?
Current commercial hydrogen production
Steam methane reforming
Current costs and carbon emissions
In conventional SMR, methane is reacted with steam using a
catalyst at relatively high temperature, 650e1000  C, and a
pressure of 5e40 bar to produce carbon monoxide and
hydrogen. Additional hydrogen is produced by reacting car-
bon monoxide with water in the water-gas shift reaction [15].
The final stage of the process separates high-purity hydrogen
(99.99%) from CO2 using pressure swing adsorption. The CO2
generated exits in two streams, a diluted stream (stack gases
with CO2 concentration 5e10% vol.) and a concentrated
stream (approximately 50% by vol. after PSA) [16]. Approxi-
mately 9.5 t CO2e t
1 H2 are produced, 60% of which is
generated from the process chemistry, while the remaining
40% is the product of the combustion of the additional flue gas
required by the steam reformer [17]. The pure CO2 stream is
readily amenable for low-cost CCS if appropriate geological
storage reservoirs are located nearby. If carbon dioxide
2542 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
emissions must be reduced from the entire process, then an
amine solvent (MEA) might be used to capture up to 90% of the
CO2 contained in the stack gases for sequestration as a pure
stream [18].
Hydrogen production costs vary with the price of the nat-
ural gas feedstock, the relative scale of the facility, and the
specific project economics. Gray and Tomlinson [19] among
others have developed a relationships for the hydrogen cost of
production in large-scale facilities without CCS:
   
$ $
Hydrogen Cost ¼ 0:137x NG Price þ 0:1123 (5)
kg GJ
This relationship is applicable to facilities with production
capacities of approximately 100 million standard cubic feet
per day (SCFD) where the capital cost is approximately
$275e339 per kg per day. Penner [20] developed a relationship
of similar form without specifying plant size or operating as-
sumptions for the price of hydrogen without CCS:
   
$ $
Hydrogen Cost ¼ 0:271x þ 0:15
kg GJ
Using current long-term natural gas forecast prices of
$4 GJ
1 [21] and the LHV of H2, adjusting the prices to 2016
dollars using the Chemical Engineering Cost Plant Index
(CEPCI), yields hydrogen costs of $7.55 GJ
1 and $10.46 GJ
1
($0.91 kg
1 H2 and $1.26 kg
1 H2) from the two relationships
respectively. These cost estimates are in agreement with
other models stating hydrogen produced from SMR typically
costs two to three times the cost of natural gas per unit of
energy produced, depending on plant size, illustrating the
effect of different operating assumptions and financial as-
sumptions used in model calculations.
Only three sites around the world have demonstrated the
integration of CO2 capture with transport and storage at an
SMR facility. These include the Air Liquide Port Arthur Project
in the U.S., the Shell Quest Project in Canada and the Toma-
komai Project in Japan [18]. None have publicly released pro-
cess costs. Likewise, although several studies concerning
costs of CCS for power generation are available [22e25], only
one has costed SMR coupled with CCS. The International En-
ergy Agency (IEA) 2017 evaluation of a standalone SMR based
hydrogen plant with CCS, places CO2 avoidance costs between
U.S. 2016 $52e78 t
1 [18]. Importantly, this study represents an
ideal scenario where CO2 storage sites are available at mini-
mal cost.
Model development
A typical SMR process schematic is shown in Fig. 1. For
modelling SMR equipment capital costs, data from the Na-
tional Renewable Energy Laboratory [26], Molburg et al. [27],
the U.S. Department of Energy (DOE) H2A Model [28], the
Parsons Infrastructure and Technology Group Inc. [29] and
Bartels et al. [30] were used. Costs were adjusted to 2016
using CEPCI and scaled using a process cost correlation
exponent of 0.79 for SMR [31]. Factored cost estimates, as
applied in engineering, procurement and construction for
process industries from the Association for the Advance-
ment of Cost Engineering were used to estimate other inside
and outside battery limits (ISBL and OSBL) costs for an all-
fluid process [32].
Operating costs were based on feedstock, utility usage and
water consumption and crosschecked with values from the
DOE H2A model for raw material consumption per kg H2
product (Table 1). Operating and maintenance labour cost data
from the Parsons Infrastructure and Technology group Inc.
Report [29] for the DOE was used to calculate annual labour
requirements. High-pressure SMR feedstock steam was priced
at $29.97 t-1 [33]. Carbon emissions intensity was calculated
based on natural gas consumption per tonne of H2 product.
Net steam production is used to generate 12.4 MWe (after
parasitic loads) exported to the grid at $0.06 kWh-1. For CCS
costs, this study presents the economics of the 90% capture
scenario for a standalone merchant SMR facility of $ 78 t
1 CO2
captured. No by-product credits for CO2 value (e.g. for
enhanced oil recovery) are applied and reduced electricity
export for CCS is considered.
Electrolysis hydrogen production costs
(6) Current costs and carbon emissions
Water electrolysis processes split water into hydrogen and
oxygen through the application of electrical energy. There are
three main types of water electrolysers available; alkaline,
polymer electrode membrane (PEM) and solid oxide electrol-
ysis cells (SOEC) [34]. Of these technologies, the most
commercially mature technology is alkaline water electrol-
ysis, but more PEM and SOEC units are under development
[35]. In all three technologies, the rate of hydrogen generation
is directly proportional to the current density. High current
densities result in higher power costs per unit of hydrogen
produced but reduce capital costs. Alkaline cells have the
lowest capital costs but also the lowest electrical efficiency
resulting in the highest electrical energy costs. SOEC electro-
lysers are the most electrically efficient, but are the least
developed technology and suffer from corrosion issues and
other technical pitfalls [36]. PEM electrolysers have higher
electrical efficiency than alkaline cells and do not have the
corrosion and seal issues of SOEC, but currently cost more
than alkaline cells. Of the available electrolyser technologies,
PEM electrolysis is both energy efficient (the largest cost driver
for at scale production) and has reducing capital costs. As
such it is generally considered the most prospective future
renewable electrolysis hydrogen production route [4].
The cost of hydrogen produced by electrolysis of water can
be described by a linear function of electricity cost. The
theoretical minimum energy input is approximately
39.3 kWh kg
1 H2 (at room temperature), but in practice is
closer to 50e60 kWh kg
1 H2. Using current industrial elec-
tricity grid prices of $0.06 kWh-1, minimum hydrogen costs
based on the electricity cost alone are USD $3 kg
1 H2
excluding capital charges, replacement costs, ancillary costs
or other operational expenditures.
The end-use application of hydrogen as a fuel produces
almost no environmental impact, but the production via
electrolysis from a fossil fuel dominated electrical grid needs
to be considered [34]. If power were supplied by a natural gas
combined cycle turbine following the EPA emissions guide-
lines of 467 kg CO2 MWh
1 of electricity produced, a minimum
emission of approximately 23 kg CO2 kg
1 H2 would be pro-
duced, more than double the CO2 produced from SMR. If
 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5 2543

Fig. 1 e Schematic of typical process flows and heat integration in a SMR facility.

the levelized cost of electricity (LCOE) of carbon-free nuclear

power ($0.1 kWh-1 [41]). Today, pressurized electrolysers
Table 1 e SMR raw material operating costs. operating at 30 bar discharge pressure are market established,
Raw material operating costs and have therefore been selected as the basis for this study.
Pipeline natural gas price $ 4 GJ
1 Factored cost estimates and vendor quotes sourced by the U.S
Electricity price $0.06 kWh
1 (grid) DOE for the H2A model for electrode replacement and un-
High pressure steam (30 bar) $ 29.87 t
1 planned maintenance schedules were applied [17]. Other
Process water $ 22.79 ML
1 OSBL costs not included in capital costs were sourced from the
Catalyst costs $ 358,400 yr
1 [18] AACE for an all-fluid process [32]. By-product O2 credits have
CO2 Avoidance cost $ 78 t
1 CO2 [18]
not been included in this analysis.

electricity from renewable resources such as PV or wind are

used, the emissions footprint is substantially reduced, but to
Hydrogen production technology by methane
be economically competitive electrolysers would require high
pyrolysis
capacity factor, dispatchable carbon-free electricity only
widely available from hydro or nuclear power [37], or a
Technology development
strongly interconnected grid with intermittent renewables
and large-scale energy storage.
Conversion of methane to hydrogen and solid carbon by
pyrolysis faces significant challenges associated with man-
PEM electrolysis model development agement of the solid carbon produced which fouls (cokes)
A typical PEM electrolysis process is shown schematically in
solid catalysts including metals [42e44], metal oxides [45,46]
Fig. 2. For modelling the capital costs of a PEM electrolyser
and carbon-based catalysts [47,48]. Proponents claim these
system, data from Saur et al. [28], Bertuccioli et al. [38] and
limitations can be managed by appropriate process and
Jacobs et al. [39] were used for direct capital costs ($ kW
1).
reactor design [35,49]. In some processes, the catalyst is re-
Costs were adjusted to 2016 using the CEPCI, resulting in an
generated by combustion of carbon, but this results in
average cost of $800 kW
1 ($1.34 W-1 total capital investment).
considerable CO2 emissions [50]. Solid catalysed methane
The breakdown of PEM electrolyser capital cost is summarized
decomposition has been investigated extensively for the for-
in Table 2.
mation primarily of more valuable carbons as the primary
Cost projections for electrolyser cell stacks are based on
products, including graphitic turbostratic carbon, carbon
limited large-scale experience. Developed market costs for
nanotubes and carbon filaments [51]. The preferred temper-
‘nth of a kind’ production plants ordinarily extrapolated for
ature ranges, carbon products formed and catalysts are re-
larger capacities using a power law function (3) are not
ported in detail elsewhere and are not discussed here [8].
applicable to electrolyser systems. For electrolysers the
When producing hydrogen at relatively small-scale, the solid
limiting unit is the cell stack itself, with the area of each
carbon by-product may provide additional revenue, however,
electrode limited by manufacturing and fluid dynamics, and
to have a significant impact on global carbon emissions,
total number of cells in a stack limited by tolerances in
massive scale pyrolysis deployment would saturate the
manufacturing and minimizing excessive voltages [40]. For
known carbon product markets and eventually the solid car-
greater capacities, several parallel cells are installed with
bon would need to be stored in perpetuity. A detailed analysis
some sharing of utilities and other balance-of-plant compo-
of the reverse carbon chain (relative to coal) for large scale
nents. For these reasons, direct capital costs of electrolyser
methane pyrolysis in a global energy context is discussed by
cells have been scaled linearly (exponent of n ¼ 1) to higher
Kreysa [52].
capacities.
Recent developments indicate that large-scale, practical
Costð$Þ ¼ Capacityn (7) and viable applications of methane pyrolysis without CO2
Operating costs are based on feedstock preparation, utility, emissions may be possible through the use of molten liquid
maintenance, labour, electrode replacement and other OPEX systems with catalytic metals [2,50]. Lab-scale tests suggest
factors summarized in Table 3. Electricity costs are based on that the limited solubility of carbon in molten metals may
2544 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5

Fig. 2 e Schematic of a typical PEM electrolysis process.

achieved at 900  C in a suspension of solid silica carbide in

liquid tin; compared to 51% conversion in pure molten tin.
Table 2 e PEM electrolyser capital cost breakdown. Plevan et al. [66] also compared a molten tin bubble column
Component Cost breakdown and a blank tube reactor at similar temperatures of
Total installed cost $ 800 kW
1 700e950  C. A 1 m bubble column of molten tin was used at
Electrolyser stacks 47% 1175  C to achieve 78% methane conversion [53]. Although the
Balance of systems total 53% prior molten metal systems investigated provided excellent
Balance of systems breakdown heat transfer properties and effective carbon separation
Hydrogen gas management system 9%
through buoyancy, limited catalytic effect resulting in low
(Cathode system side)
methane conversion had been challenging for commercial
Oxygen gas management system 3%
(Anode system side) development [53]. The only report of the use of molten metals
Water reactant delivery management system 5% as a catalyst for methane pyrolysis utilized liquid magnesium,
Thermal management system 5% achieving 30% methane conversion at 700  C [67]. Higher
Power electronics 21% temperatures and conversions were not investigated due to
Controls & sensors 2% magnesium evaporation.
Mechanical balance of plant 5%
Ultimately, the usefulness of pyrolysis as a means of using
Other 3%
methane for carbon dioxide free production of hydrogen will
depend on the specific process economics [68]. The technology
allow continuous removal of carbon and eliminate catalyst is still in development and lacks the significant investment
deactivation, which previously had prevented the develop- required to develop large-scale commercial processes.
ment of methane cracking processes on the industrial-scale Hydrogen production costs are strongly dependent on sepa-
[53]. Liquid metals have also been observed as catalysts in ration of the products. In SMR units, the single pass natural
other applications, including alcohol dehydrogenations gas conversions are typically 70e85% [5]. For most applica-
[54e59], hydrogen transfer reactions [60] and for waste py- tions, hydrogen must be separated from the carbon oxides as
rolysis [61,62]. well as unreacted methane to prevent catalyst poisoning in
Steinberg et al. [63,64] was the first to propose using molten ammonia production and fuel cell applications. Pyrolysis
metals for methane pyrolysis to facilitate the decarbonization would only require separation of the unconverted methane,
of hydrocarbons. Methane pyrolysis using low-melting point for which pressure swing absorption (PSA) is most commonly
metals (lead and tin), heated solids (alumina, silica carbide) used. Both fuel cells and ammonia synthesis are relatively
and combinations of both have since been extensively inves- insensitive to small amounts of methane and thus the sepa-
tigated [65]. A maximum methane conversion of 57% was ration requirements are less stringent than for the carbon
oxides in SMR. The complexity and cost of product separation
for a pyrolysis process depends on the conversion and the
application dependent hydrogen purity requirements.
Table 3 e Key PEM electrolyser specific modelling
assumptions.
Methane pyrolysis model development
Component Factor
Theoretical minimum energy usage 39.3 Process design
(kWh kg
1 H2) A conceptual process model for pyrolysis was developed and
Electricity cost ($ kWh
1) 0.1
is shown schematically in Fig. 3. Methane pyrolysis is cata-
Cell voltage efficiency 75%
Dryer loss 3% Electricity
lysed by a molten metal in a bubble column. An immiscible
Permeation loss 0.7% Electricity molten salt is used on top of the molten metal column to
Total balance of plant loads (kWh kg
1 H2) [28] 5.04 provide heat and remove carbon from the reactor. Hot salt is
Electrical usage (kWh kg
1 H2) [28] 54.3 added to the reactor and cool salt is removed along with the
DI Water cost ($ ML
1) [28] 477 carbon produced in the pyrolysis reaction. Recent experi-
Electrode replacement Cost (% of Purchased 15%
mental data [69] at the University of California, Santa Barbara
equipment cost)
was used to size the catalytic molten metal component of the
Capital scaling factor 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
bubble column using a Ni-Bi alloy (27:73mol%). Methane
bubbles were generated at a 2 mm gas orifice and at 1065  C
and a methane partial pressure of 1.6 atm, and methane
conversion was 95%, (equilibrium conversion is 98% under
these conditions). The catalytic activity of the Ni-Bi melt re-
sults in higher methane conversions than other molten metal
systems such as lead or tin.
To reduce the reactor volume and the sizes of the down-
stream separation units it is desirable to operate at relatively
high pressures. A multi-variate optimization of the pyrolysis
temperature and overall system pressure was performed to
explore the minimum overall process cost and its relation-
ships to pressure and temperature. The parameters for the
optimization of reactor size were system pressure, pyrolysis
temperature, equilibrium conversion and materials of con-
struction. With increasing system pressure, the equilibrium
conversion is lower, however smaller unit sizes (in particular
reactor size) and less downstream compression for product
separation are required. Increasing the reactor temperature
results in higher equilibrium methane conversion, however,
more expensive materials of construction are required.
The process description and economics presented here are
based on the catalytic thermal pyrolysis of methane at 1000  C
reactor temperature and 35 bar feed natural gas pressure for
100 kilotonnes per annum (kta) of product hydrogen. Kinetic
data reported for liquid Ni-Bi [69] was used to determine the
required surface area at lower temperatures and higher
pressure operation for equilibrium conversion. The operating
conditions were determined through minimizing the levelized
cost of hydrogen (LCOH) in the multivariate optimization of
process design. In the absence of experimental methane
conversion data in Ni-Bi melts at higher pressures and volu-
metric holdups, an additional 33% of reactor height was added
as a conservative safety factor to reach equilibrium conver-
sion at higher pressures. This results in a total molten Ni-Bi
liquid height of 2 m.
Reactor design concept
A pyrolysis reactor design concept is shown schematically in
Fig. 4. A multi-layered system using molten metal and molten
salt has been utilized. The specific role of the additional
molten salt layer is explained in detail below. For vertical
bubble column reactors, achieving high volumetric
throughput within superficial velocity constraints, results in
length to diameter ratios significantly less than one for any
reasonable number of reactor vessels. To minimize the
number of required reactors through increasing pressure, this
drastically increases vessel cost to remain below maximum
allowable stress conditions for materials of construction.
Maximizing gas volumetric hold-up for a fixed reactor height
of 2 m (short residence time) is a primary design consider-
ation. A horizontal vessel design maximizes gas volumetric
throughput for allowable superficial velocities of less than
150 m h
1 [70] through extending vessel length to create a
larger cross sectional area for gas injection. The reactor vessel
was sized based on these constraints:
1) A minimum Ni-Bi molten reactor height of 2 m to approach
equilibrium conversion under operating temperature and
pressure;
2545
2) Superficial gas velocity and gas holdup (maximum 25%)
volume at the bottom of reactor vessel for a given natural
gas feed pressure;
3) Superficial gas velocity and gas holdup volume at the sur-
face of the molten media for a given natural gas feed
pressure, accounting for volumetric expansion due to
reduced static head pressure and reaction stoichiometry;
4) A maximum vessel diameter and length within appro-
priate engineering design guidelines and to be ubiquitous
with horizontal bullet sizes used in industry.
The carbon formed in a catalytic molten metal bubble
column can be a fine graphitic powder that can be entrained in
the gas stream exiting the reactor. Separating the fine powder
using cyclones, filters, and/or electrostatic precipitators is
possible; however, when a layer of less dense, immiscible,
molten salt is supported on the surface of the molten metal,
insoluble carbon aggregates are observed to accumulate on
the surface of the salt. More importantly, a layer of molten salt
on top of a catalytic molten metal provides a liquid heat
transfer fluid that can be circulated with the solid carbon out
of a reactor. The controlled transport of liquid metal still poses
many challenges in engineering design and materials of con-
struction. Continuous molten metal streams using electro-
magnetic pumps are described in academic and patent
literature [71,72]. Others have proposed the use of bubble-lift
pumps for molten metal circulation [68,73]. Due to tempera-
ture and pressure limitations of mechanical pump design and
the prohibitively high density of Ni-Bi melts (8.7 specific
gravity) for bubble lifts, an additional layer of molten MgCl2-
NaCl salt above the Ni-Bi media is proposed. The molten salt
layer on top of the molten metal serves several important and
distinct purposes:
1) To collect carbon at the salt surface, preventing fines from
flowing downstream;
2) To allow continuous separation of the solid carbon from
the molten metal reactor as a low-concentration slurry.
The carbon can be removed from the salt media at lower
temperatures in a separate vessel where physical handling
and batch like operations are capable;
3) To allow the necessary heat addition required to supply the
endothermic pyrolysis reaction to be achieved by direct
contact heating from a high heat capacity, low vapour
pressure molten salt under hot service;
4) A proven commercial equipment application for molten
salt transportation in solar thermal and nuclear industries;
5) The Mg-Cl2 molten salt is a cheap media allowing some
solvent losses in carbon separation;
6) Carbon is insoluble in molten salts, so precipitation of
carbon on heat exchanger walls would not occur, unlike
molten metals.
The salt streams are designed such that the molten salt
pumps contact salt streams at a lower temperature (<650  C)
than the reactor. The molten salt pumps have been sized and
their cost estimated based on molten salt pumps currently
utilized by concentrated solar thermal designs under similar
conditions [74]. Alternative direct contact heating schemes
were considered, but due to the low heat capacity of natural
2546 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5

Fig. 3 e Methane pyrolysis model flowsheet.

Fig. 4 e Pyrolysis reactor design concept.

 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
gas an additional 800  C temperature rise per mole of methane
would be required to supply the necessary heat for the
endothermic reaction. The large volumes of molten metal and
high heat input also make inductive heating of the molten
media prohibitively expensive [75]. The internal weir design
shown in Fig. 4 for continuous carbon separation incorporates
learnings from three-phase emulsion separations in hori-
zontal pressurized drums used in petroleum processing.
The carbon-rich molten salt slurry leaving the pyrolysis
reactor is cooled by four shell and tube heat exchangers in
parallel, incorporating heat recovery by heating the returning
carbon-free molten salt and preheating the natural gas to
625  C. The natural gas preheat temperature of 625  C is the
minimum heat extraction required from the molten salt loop
to fall within operability limits of molten salt pumps. The
carbon-rich molten salt is sent to the carbon removal drum
where the residence time is sufficient for the carbon particle
settling velocity to separate carbon from the bulk salt phase.
Through removing the carbon-rich molten salt layer as a low-
solids content slurry, sensible heat from the carbon is also
recoverable. At 600  C, physical filtration or skimming as
routinely performed for slag removal can remove entrained
carbon particles [76]. The carbon-free molten salt is pumped
from the carbon removal drum sump through the heat re-
covery molten salt exchanger and additional heat supplied in
a gas fired heater prior to returning to the pyrolysis reactor.
The heater could use natural gas, hydrogen, or potentially be a
solar thermal collection unit, however, for this analysis it is
assumed to be natural gas.
As the molten salt is the sole heat source for the reactor
vessel, the flow rate within the external pump-around loop
was determined by the heat requirements of the endothermic
pyrolysis reaction and the direct contact heating of the natural
gas feed from 800  C to 1000  C in the reactor. Preheating the
natural gas feed beyond 800  C was deemed infeasible due to
concerns of premature pyrolysis in the feed tubes. Other
design constraints include an upper temperature limit of the
molten salt loop at 1100  C for nickel-chromium fired heater
tube materials, and a carbon-rich molten salt concentration
not exceeding 10% by volume. The gas fired heater operates
with a conservative net thermal efficiency of 80% [77].
Pyrolysis economic modelling
The molten salt heat exchangers have been sized and their
costs estimated using correlations from Ludwig et al. [78] and
Couper et al. [79]. Additional material factors have been
incorporated into the costing and design for higher tempera-
ture and pressure operation of the exchangers. The Idaho
National Laboratory for advanced high temperature nuclear
reactor design discusses shell and tube heat exchangers with
Ni alloy materials of construction, similar operating temper-
atures and significantly higher operating pressures in detail
[80]. Heat transfer properties for the KCl-MgCl2 molten salt
have been investigated extensively by the Oak Ridge National
Laboratory for application in nuclear reactors [81].
The molten metal pyrolysis reactor and molten salt carbon
separation drum vessels are constructed of carbon steel lined
with a 375 mm layer of MgO-C refractory brick for high
2547
temperature (1500  C) operation [82] at a cost of $361.3 t-1 [83].
The design parameters of the horizontal pyrolysis reactor are
summarized in Table 4.
The processes downstream from the reactor are designed
to maximize heat recovery from the 1000  C pyrolysis exhaust
gases and minimize overall process costs. Pyrolysis reactor
product gases are cooled to separate product hydrogen from
unreacted methane feedstocks for recycling. The heat recov-
ered is used to pre-heat the feed natural gas from 625  C as to
800  C in a stainless steel shell and tube heat exchanger prior
to injection into the pyrolysis reactor. The remaining heat
from the exhaust gases is recovered through generating
approximately 21 kg s
1 of high-pressure superheated steam
in a waste-heat boiler. The steam is utilized to generate power
for the overall process and export excess to the grid at a value
of $0.06 kWh-1. Parasitic loads (i.e. natural gas recycle
compressor) and a 5% heat loss applied through electrical
trace heating reduces the net grid export from 15.4 MWe (peak
generated) to 5.35 MWe.
The product gases are cooled to 60  C with an air cooler
prior to separation of hydrogen to pipeline specification
product of 99.9% from unreacted methane. Cryogenic sepa-
ration, hydrogen membranes, and PSA were considered as
options and PSA selected due to industry acceptance and
wide-scale deployment for similar production capacities in
SMR. The PSA unit required for product separation and recy-
cling has been costed by scaling the throughput capacity from
a SMR facility [26]. 22% of the unreacted methane is burned in
the natural gas fired heater to provide the necessary 110 MW
of process heat required. The remaining natural gas is re-
compressed and returned to the pyrolysis pre-heaters.
For calculating equipment capital costs, sizing correlations
from Towler et al. [84], Ludwig [78], Couper et al. [79], Loh et al.
[85], Chauvel et al. [86], Turton et al. [33] and Ulrich et al. [87]
were used. Costs were escalated to 2016 dollars using the
CEPCI. Factored cost estimates for standard equipment items
as applied in engineering, procurement and construction for
process industries from the Association for the Advancement
of Cost Engineering were used to estimate other ISBL and OSBL
costs for an all-fluid process [32].
The operating costs were based on feedstock, utility and
water consumption calculated for the process design. Oper-
ating and maintenance labour cost data from the Parsons
Infrastructure and Technology group Inc. Report [29] for the
DOE was used to calculate annual labour requirements.
Table 4 e Design parameters of the horizontal drum
pyrolysis reactor.
Reactor parameter Dimension
Number of reactors 3
Reactor length 18 m
Reactor diameter 4.35 m
MgO-C Brick lining 375 mm
Reactor L/D ratio 4.1
Reactor wall thickness 144 mm
Vessel weight 143.4 t
Total vessel cost 1.98 $M each
2548 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5

Associated CO2 emissions from burning natural gas in the

Table 5 e General economic assumptions.
fired heater contribute 2.5 t CO2 t
1 H2.
Fixed financial input
parameters
Economic modelling Year of dollar value analysis 2016
Construction period 2 Years
Start-up period 6 months
Methods
Start-up capacity 50%
Plant lifetime 25 years
A consistent basis using the same economic assumptions, On-Stream factor 91%
operating lifetimes and plant capacity is required for direct Inflation 2%
comparison of hydrogen production costs from each tech- Weighted average cost of 10% (Nominal, After Tax)
nology. For this analysis, published capital costs and con- capital (WACC)
ceptual process models of centralized hydrogen production Finance 100% Equity finance
Depreciation MACRS GDS Schedule 49.222
facilities using PEM electrolysis and SMR are used. The cost of
Depreciation period 7 Years
the methane pyrolysis process is estimated as a first-ofea- Decommissioning cost 10% of depreciable capital
kind (FOAK) plant using large contingencies (Lang Factor of 10 Plant salvage value 10% of depreciable capital
compared with 6 for the other more mature technologies) so Company tax rate 38.9% (including state taxes)
as to compare the technical and economic opportunities for Location U.S. Gulf coast
each technology as they currently stand. The design basis for
each technology is for 100 kta of hydrogen. Delivery and
storage of product hydrogen is not included in this analysis.
Based on industry standards for other fixed and variable Table 6 e Fixed operating costs.
non-capital items, estimates were made of the operating ex-
Fixed operating costs
penses (OPEX). A discounted cash flow analysis was used to
Insurance/Local taxes/Royalties 2% Full capital
determine the estimated return on invested capital for each
Maintenance 2% Direct capital [29]
process. The economic assumptions used in the analysis
Labour Calculated based on Parsons
across each model are summarized in Table 5 and Table 6. All Infrastructure Technology Group
calorific values used throughout the analysis are based on the Assessment [29]
lower heating value of the fuel (methane and H2). All costs are Operating supplies 10% Maintenance
in 2016 US dollars. Supervision 25% Labour
Processes that emit CO2 incur an additional operating cost, Lab/QC/QA 10% Operating labour
Location U.S. Gulf coast
calculated as multiplying the mass rate of vented CO2 and the
prevailing CO2 emissions charge, so that CO2 emissions are
fully reflected in the levelised production cost of the hydrogen. costs. All heating requirements for the SMR process are sup-
plied from natural gas, resulting in the production of
approximately 9.3 t CO2e t
1 H2. The SMR facility exports
Results and discussion 12.6 MWe to the grid.
For PEM electrolysis, the PCE for a facility producing 100 kta
Discounted cash flow analysis of hydrogen product is approximately $496 million. Electro-
lyser cell stacks are packaged units incorporating large por-
The total capital investment estimated for SMR is lower than tions of the ISBL components into the PCE value, often
the estimate for methane pyrolysis and significantly less than reported by vendors in a total $ kW
1 installed. To separate
for PEM electrolysis (Table 7). The purchased cost of equip- included and excluded equipment costs in $ kW
1 installed
ment and operating costs for SMR, PEM electrolysis, and values, vendor quotes sourced by the DOE were used with
methane pyrolysis all producing 100 kta of hydrogen product additional OSBL and indirect cost factors applied from the
are summarized in Table 7. For SMR, this results in a total AACE. This results in a TCI of $829 million for a 100 kta facility.
purchased cost of equipment of $ 42 million. Using factored Despite poor size scaling of electrolyser facilities, it is ex-
cost estimates for ISBL and OSBL factors from the AACE re- pected that volume production and supply chain development
sults in a Lang Factor of 6 for a nth-of-a-kind process defined as will occur in parallel with increased deployment of systems,
fixed capital cost (FC) divided by purchased cost of equipment which will enable the use of standardized low cost compo-
(PCE). The total capital investment (TCI) was then determined nents in mass manufacture. Capital cost reductions for PEM
to be $252 million, an estimate that agrees well with a similar electrolyser systems are discussed in the Section PEM
sized SMR facility scaled to the same hydrogen output [88]. electrolysis capital and electricity costs.
The annual operating cost of the SMR facility is approxi- The annual operating cost for a PEM electrolysis facility
mately $88 million, which is less than methane pyrolysis and producing 100 kta of hydrogen is approximately $582 million.
significantly less than PEM electrolysis. The dominate por- This cost is dominated by the $543 million in electricity
tions of this cost are $63 million in natural gas costs for charges required to supply 54.3 kWh kg
1 H2 with carbon-free
feedstock and heating requirements (~329 kta). For an IRR of electricity at $0.10 kWh-1. To achieve an IRR of 10%, a LCOH of
10%, the LCOH is $1.26 kg
1. The remaining costs are associ- $7.1 kg
1 is required. The electrolysis facility produces no
ated with other feedstock, utilities, maintenance and labour direct CO2 emissions from the battery limits of the process.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
Table 7 e Summary of capital, operating and production costs for SMR, PEM electrolysis and methane pyrolysis producing
100 kta of hydrogen product per annum.
Component
Purchased cost of equipment (Million USD 2016)
Total capital investment (Million USD 2016)
Operating costs (Million USD 2016 yr
1)
Raw material cost (Million USD 2016 yr
1)
Carbon emissions intensity (kg CO2e kg
1 H2)
H2 Sale price IRR ¼ 10% ($ kg
1 H2)
CO2 Charge to equate SMR IRR10% ($ t
1 CO2e)
a
PEM electrolyser cell stacks are a grouped cost encompassing a large portion of ISBL and OSBL costs.
b
Based on $0.10 kWh
1 electricity price.
c
Based on $150 t
1 carbon product value.
The PCE for a methane pyrolysis facility producing 100 kta
of hydrogen product is approximately $41 million. For stan-
dard equipment that is ubiquitous in industry, factored cost
estimates for ISBL and OSBL from the AACE have been used to
result in a Lang Factor of 6. For FOAK equipment including
pyrolysis reactors, carbon removal drums, molten salt heat
exchangers and molten salt pumps a Lang Factor of 10 has
been applied to reflect the higher project risk associated with
FOAK projects. This results in a TCI of $350 million.
The annual operating cost for the pyrolysis facility pro-
ducing 100 kta of hydrogen and 300 kta of carbon is approxi-
mately $122 million. The largest portion of this cost is the
$93million for feedstock and heating requirements. As ex-
pected, the natural gas consumption of 492 kta per annum for
methane pyrolysis is considerably higher than SMR largely
due to the stoichiometry of the reaction for an equivalent
hydrogen production. This is based on an idealized stoichio-
metric evaluation of 4 t CH4 t
1 H2 and the secondary output of
3 t C t
1 H2. To achieve and IRR of 10% with a carbon sale price
of $0.15 kg
1, a LCOH of approximately $1.4 kg
1 is required.
Through supplying all heating and electricity requirements
for the pyrolysis facility from burning natural gas (direct fired
or recovered steam), the process generates approximately 2.5 t
CO2 t
1 H2. The pyrolysis facility is a net generator of power,
exporting 5.4 MWe to the grid.
For natural gas pyrolysis to compete with SMR, a CO2 price
of approximately $21 t
1 is required. This corresponds to an
equivalent LCOH of $1.45 kg
1 for both SMR and pyrolysis. For
PEM electrolysis to compete with SMR at $0.1 kWh-1 electricity
supply and current electrolyser capital costs, a CO2 price
exceeding $600 t
1 would be required to be economically
competitive. For SMR with CCS to be economically competi-
tive with pyrolysis at 90% capture and CO2 capture and
sequestration costs of $78 t
1 CO2, a CO2 price of approxi-
mately $300 t
1 CO2 is required.
Sensitivity analysis
The LCOH values presented are highly sensitive to the as-
sumptions made in the model. The sensitivity of the hydrogen
price in process to variations in the major input parameters
are shown in Fig. 5. Unsurprisingly, in the absence of a CO2
price, at a natural gas price of $4 GJ
1 and carbon free elec-
tricity costing $0.1 kWh-1 SMR has more favourable economics
than any of the alternative technologies. It is apparent that
2549
SMR PEM Electrolysis Methane pyrolysis NG fired
a
$ 42 $ 495.8 $ 40.8
$ 252.1 $ 829.1 $ 349.7
$ 94.5 $ 582.4 $ 122
$ 63.2 $543.8 $ 92.8
9.3 0 2.5
$1.26 $ 7.13b $ 1.39c
N/A $585 $18
there is no large-scale, continuous production scenario where
electrolyser technologies are competitive with traditional SMR
or methane pyrolysis. However, electrolyser technologies will
undoubtedly continue to develop and generate substantial
value for niche applications in the immediate and long-term
future, particularly in isolated locations or where off-peak
free or negative value electricity is available. It should be
noted, however, that if very low cost electricity were available
it could be applied in electric heaters at high efficiency to
provide heat for pyrolysis.
Varying the natural gas feedstock and fuel price from a U.S
2016 average of $2 GJ
1 up to $6 GJ
1 has a significant effect on
the economic potential of methane pyrolysis due to the
increased natural gas consumption. At any substantial carbon
price (>$30 t
1 CO2e), the significantly reduced CO2 footprint of
a pyrolysis process compared to traditional SMR economically
favours pyrolysis. However, the increased natural gas con-
sumption of a methane pyrolysis process would likely incur
additionally economic penalties associated with upstream
methane leakages under such a scenario, a factor discussed in
detail in Section Natural gas fugitive emissions. Alternatively,
a substantial portion of the hydrogen product (~35%) itself
could be used to supply the necessary heat without any CO2
production. This would further increase the required capital
and natural gas utilization, and only provide an economic
benefit when the carbon prices exceeded $50 t
1 CO2e.
For methane pyrolysis, capital estimates were also varied
by 20% but continue to utilize conservative Lang Factors of 10
to reflect the higher risk and uncertainty of process design.
Given that pyrolysis has had little development, it is antici-
pated that capital costs will be significantly reduced through
research, innovation and widespread deployment. Innovation
in reactor design, carbon removal schemes and high temper-
ature processing equipment could substantially reduce capital
costs and enhance cost competitiveness of pyrolysis tech-
nologies. Reactor designs that can achieve high conversions
and facilitate product separation will be important in opti-
mizing a commercial pyrolysis process [68]. The challenge is
to perform pyrolysis, add heat, and remove carbon all in a
‘single-pot’, removing the need for extensive molten salt
handling and heating.
Carbon product value
The most sensitive factor in the overall economics of methane
pyrolysis is the product value (if any) of the solid carbon.
2550 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
Fig. 5 e Sensitivity analysis of key model parameters for SMR, PEM electrolysis and natural gas pyrolysis on the minimum
sale price of hydrogen to achieve a NPV of zero and IRR of 10%.
Initial process deployments will likely find receptive markets
for high value carbon products. However, massive scale
implementation of pyrolysis technologies is needed to prog-
ress global decarbonization efforts. Over 60 million tonnes of
hydrogen per year are currently produced by steam reforming
hydrocarbons, over 80% of this demand is supplied by natural
gas [89]. If the hydrogen produced by SMR were replaced by
methane pyrolysis, roughly 0.15 Gt of solid carbon would be
produced. The only solid manufactured product used in
comparable volumes is concrete.
Global cement production exceeds 1.6 Gt yr
1, which when
combined with stone aggregate produces approximately
5 Gt yr
1 of concrete [82]. The World Steel Organization places
2017 global steel production at 1.5 Gt yr
1 [90], with other
studies [91] estimating more than 0.5 Gt yr
1 of the steel pro-
duction already utilized in infrastructure construction. Other
available carbon markets and approximate sizes are shown in
Fig. 6. One could also envisage the substitution or blended
addition of pyrolysis carbon for metallurgical coke in the
carbothermic reduction of iron oxide in steel manufacturing, a
market which in 2016 exceeded 0.6 Gt yr
1 [92]. Carbon bal-
ance calculations show that 3 t C from pyrolysis could replace
3.69 t of coking coal, not including emissions from extraction
and transportation coal from the earth's surface [93].
The physical and electrical properties of the solid carbon
will strongly influence its value as a co-product. Most mate-
rials are more valuable than fuels and the carbon product
may, in some relatively small markets, be more valuable than
the hydrogen. Control of the solid carbon product properties
has been demonstrated in many hydrocarbon pyrolysis pro-
cesses [8], thus it is plausible that a high-value fibrous carbon
could be produced from molten metal systems for composite
reinforcements as discontinuous fibres, fibrils or whiskers. For
the base case scenario, a pure carbon product value of $150 t
1
has been used. This carbon price is approximately 75% of the
price of petroleum coke of moderate purity used in metallurgy
processes [93]. The value of carbon product has been varied
from $500 t
1 to zero to investigate the effect of first produc-
tion with receptive markets and the nth-of-a-kind process
assuming no carbon value.
PEM electrolysis capital and electricity costs
The primary cost drivers for PEM electrolysis are utility elec-
tricity charges and electrolyser capital costs. The variable cost
of electricity for the process can be minimized through higher
electrolyser operating efficiencies and identification of the
lowest cost electric power supplies. An assumed base-case
scenario electrolyser electrical efficiency of 72.3% (54.3 kWh
kg
1 H2) is projected to increase to closer to 90% (45 kWh kg
1
H2) with future technology developments. This represents a
reduction of $93 M yr
1 in utilities charges at an electricity cost
of $0.1 kWh-1.
Location-specific scenarios where renewable electric
power is in excess during off-peak times can support low-cost
electricity prices of the order $0.02e0.03 kWhe1. However,
often solar power output approximately matches demand for
near peak loads including air conditioning and in those in-
stances would be too valuable for hydrogen production. The
utilization of off-peak electricity supply from nuclear, wind, or
hydropower or any renewable source where production does
not match electric load profile may lower hydrogen produc-
tion costs significantly [39]. To be competitive in these sce-
narios, electrolysers would also require relatively high
utilization factors (several thousand hours per year).
Electrolyser capital costs are anticipated to be reduced with
increased deployment through benefits of high-volume
 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
Fig. 6 e Approximate size and cost of major carbon markets.
production and supply chain development enabling the use of
lower cost, mass produced components [39]. Capital cost re-
ductions from $800 kW
1 ($1.34 W-1 TCI) used in the base-case
scenario to a potential future cost of $400 kW
1 ($0.67 W-1 TCI)
could reduce the hydrogen production costs to $6.20 kg
1. This
analysis is consistent with the NREL sensitivity analysis,
which showed in all cases tested that electricity price is the
single most important economic factor for hydrogen produc-
tion by electrolysis.
Natural gas fugitive emissions
In the base case scenario, the greenhouse gas emissions from
fugitive methane emissions from upstream processing of
methane was not included. However, methane itself is a more
potent greenhouse gas than CO2 and when assessed over a
100-year period it is 34 times more potent than CO2 on a mass
basis [94]. Higher fugitive emissions associated with greater
production of natural gas reduces the potential GHG savings of
any hydrogen production technology that uses methane.
Fugitive emissions of natural gas along the production and
supply chain can never be zero and today are approximately
0.93% (~6 g CO2e MJ
1 CH4 (LHV)) [95] of total natural gas
extracted (roughly 6.5 million tonnes of natural gas) or 162
million t CO2e yr
1 [94,95]. The effect of accounting for fugitive
natural gas emissions in upstream processing on the LCOH to
achieve an IRR of 10% is shown in Fig. 7. This assumes the
worst-case where fugitive emissions increase linearly with
natural gas consumption.
A recent study by Weber and Clavin [96] suggested that the
EPA reported values were too low and GHG footprints are
higher for the industry as a whole. Through use of a Monte
Carlo uncertainty analysis of recent studies, they determined
2551
95% confidence intervals of 11e21 g CO2e MJ
1 CH4 (LHV)
(~2.3% of total natural gas extracted). Including the Weber and
Clavin leakage data, the equivalent CO2 price is only margin-
ally increased to $30 t
1. It is also clear that the industry par-
ticipants with strict adherence to best practice standards emit
well below 1% of their total extracted natural gas. Fugitive
emissions are unlikely to drastically influence the economic
potential or overall GHG footprint of hydrogen processes.
Discussion of sensitivity results
It is apparent that there is no reasonable carbon price (i.e. < $
600 t
1 CO2e) where electrolytic processes can be economi-
cally competitive for the large-scale, centralized, hydrogen
production facilities necessary for supplying industrial de-
mand for molecular hydrogen. Alternative scenarios for
small-scale distributed hydrogen generation and consump-
tion might be competitive when the costs of compression and
transportation from centralized facilities, $29.2 GJ
1 H2
(~$3.50 kg
1), are included [97].
In the public debate and discussions surrounding the Paris
Agreement and future carbon pricing, the large-scale
deployment of CCS is the only method widely considered to
be available to substantially mitigate CO2 emissions from
fossil power generation and industrial processes including
SMR or gasification. The deployment of CCS at-scale has so
far failed to meet expectations, and continued delays of
demonstration projects cause considerable uncertainty about
the role CCS can play in carbon mitigation. If societies
enforce a cost for CO2 emissions pyrolysis will represent an
economically attractive alternative to SMR with CCS for
hydrogen production. Even when considering upstream
fugitive emissions from natural gas extraction (which can be
2552 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
Fig. 7 e Effect of upstream fugitive natural gas emissions on hydrogen sale price under various CO2 emissions equilancy
charges.
minimized), pyrolysis processes represent substantial cost
reductions for low-CO2 hydrogen compared to SMR with CCS.
Further, electricity produced by combustion or conversion in
a fuel cell of hydrogen produced by methane pyrolysis is the
lowest cost low-CO2 source of dispatchable and reliable
power generation.
The shale gas revolution has underpinned the abundance
of low-cost natural gas prices in the U.S., a situation antici-
pated to remain low for decades. It is possible that natural gas
prices could also fall in other international locations. Research
and development of methane pyrolysis systems may reduce
the overall process costs and construction of facilities at scale
will allow improved deployment cost estimates. The engi-
neering challenge of ‘one-pot’ pyrolysis, with in-situ carbon
and hydrogen separation may result in a competitive
hydrogen cost of production without carbon prices or policy
interventions. Further investigation into control of the carbon
product formation and innovative new uses for massive
quantities of carbon could see the carbon co-product generate
higher revenue than considered here. Early market entrants
can take advantage of the small, but high-value, markets for
carbon products improving the economics of the scale-up
process. The development of low-cost, low CO2 emissions,
hydrogen technologies at scale encourages investment in the
infrastructure necessary to facilitate a future hydrogen
economy, and potentially increases the rate of adoption.
Summary and conclusions
The production of hydrogen using technically, environmen-
tally and economically justifiable processes is fundamental to
further deployment of future hydrogen infrastructure. Today,
in the absence of a price on carbon, steam methane reforming
is the most cost-effective process for producing hydrogen.
However, for low CO2 hydrogen production methane pyrolysis
was found to be significantly more competitive than electro-
chemically based processes using any known low-carbon
power source. Although not presently competitive with SMR
(without a by-product carbon sales value > $ 200 t
1), using the
techno-economic framework developed here, molten metal
pyrolysis technologies were found to have specific areas
where improvements and innovations in reactor design and
improved catalytic activity of molten systems could signifi-
cantly improve the process economics. In the absence of such
improvements, carbon prices or policy intervention (which
would benefit both pyrolysis and SMR with CCS), there exists
no cost-effective pathway for reducing CO2 emissions from
hydrogen production in industry.
The current cost disparity between renewable electricity
powered electrolysis and fossil-based hydrogen production
suggest that if hydrogen is to fulfil an expanding role in a future
low-CO2 energy economy, production from methane pyrolysis
could be the most cost-effective bridging technology. Use of
pyrolysis at large-scale would provide competitively priced
hydrogen over a transition period during which the required
infrastructure and end-use applications necessary to sustain a
large penetration of hydrogen into the energy economy are
deployed. Currently, the large-scale deployment of pyrolysis
processes represents an interesting, potentially low-cost, GHG
abatement opportunity for the fertilizer and petrochemicals
industries. The development of markets for the massive carbon
by-product quantities from pyrolysis could also generate sub-
stantial value and employment in new carbon applications.
 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
Acknowledgements
The authors would like to acknowledge the technical advice and
direction provided by Dr. Howard Fong and Mr. Phillip Grosso
throughout the process development and capital cost esti-
mating procedures. The authors would also like to thank Dr.
Diego Schmeda for his input and advice for the technical
development of the conceptual reactor design presented here.
Funding was provided by the Dow Centre for Sustainable Engi-
neering Innovations at the University of Queensland, Australia,
and by the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences, grant #DE-FG03-89ER14048.
references
[1] Muradov NZ, Vezirog  lu TN. “Green” path from fossil-based to
hydrogen economy: an overview of carbon-neutral
technologies. Int J Hydrogen Energy 2008;33:6804e39.
[2] Weger L, Aba nades A, Butler T. Methane cracking as a bridge
technology to the hydrogen economy. Int J Hydrogen Energy
2017;42:720e31.
[3] Mahajan D, Taylor CE, Mansoori GA. An introduction to
natural gas hydrate/clathrate: the major organic carbon
reserve of the Earth. J Pet Sci Eng 2007;56:1e8.
[4] Koroneos C, Dompros A, Roumbas G, Moussiopoulos N. Life
cycle assessment of hydrogen fuel production processes. Int
J Hydrogen Energy 2004;29:1443e50.
[5] Bossel U. Does a hydrogen economy make sense? Proc IEEE
2006;94:1826e37.
[6] Logan BE. Peer reviewed: extracting hydrogen and electricity
from renewable resources. Environ Sci Technol
2004;38:160Ae7A.
[7] Schrope M. Which way to energy utopia? Nature 2001;414:
682e4.
[8] Muradov NZ, Veziroǧlu TN. From hydrocarbon to
hydrogenecarbon to hydrogen economy. Int J Hydrogen
Energy 2005;30:225e37.
[9] Tang P, Zhu Q, Wu Z, Ma D. Methane activation: the past and
future. Energy Environ Sci 2014;7:2580e91.
[10] Sevilla M, Mokaya R. Energy storage applications of activated
carbons: supercapacitors and hydrogen storage. Energy
Environ Sci 2014;7:1250e80.
[11] Pandolfo AG, Hollenkamp AF. Carbon properties and their
role in supercapacitors. J Power Sources 2006;157:11e27.
[12] Muradov N. Decarbonization at crossroads: the cessation of
the positive historical trend or a temporary detour? Energy
Environ Sci 2013;6:1060e73.
[13] Dumortier M, Tembhurne S, Haussener S. Holistic design
guidelines for solar hydrogen production by photo-
electrochemical routes. Energy Environ Sci 2015;8:3614e28.
[14] Shaner MR, Atwater HA, Lewis NS, McFarland EW. A
comparative technoeconomic analysis of renewable
hydrogen production using solar energy. Energy Environ Sci
2016;9:2354e71.
[15] Wang Y-N, Rodrigues AE. Hydrogen production from steam
methane reforming coupled with in situ CO2 capture:
conceptual parametric study. Fuel 2005;84:1778e89.
[16] Collodi G, Wheeler F. Hydrogen production via steam
reforming with CO2 capture. Chem Eng Trans 2010;19:37e42.
[17] U.S. Department of Energy. Hydrogen and fuel cells program:
production case studies. In: Melaina M, editor. Current
central hydrogen production from natural gas without CO2
2553
sequestration version 3.101. U.S. Department of Energy; 2012.
Online.
[18] IEAGHG. Techno-economic evaluation of SMR based
standalone (merchant) hydrogen plant with CCS. 2017.
Cheltenham, UK.
[19] Gray D, Tomlinson G. Hydrogen from coal. Mitretek
Technical Paper MTR 2002;31:2002.
[20] Penner S. Steps toward the hydrogen economy. Energy
2006;31:33e43.
[21] Botterill AH, Henry, Lesley Mitchell, Jonathan Listoe. Price
forecast. Canada: Deloitte Resource Evaluation & Advisory;
2015.
[22] Global CCS Institute. Economic assessment of the carbon
capture and storage technologies: 2011 update. 2011.
[23] Boot-Handford ME, Abanades JC, Anthony EJ, Blunt MJ,
Brandani S, Mac Dowell N, et al. Carbon capture and storage
update. Energy Environ Sci 2014;7:130e89.
[24] Rubin ES, Chen C, Rao AB. Cost and performance of fossil fuel
power plants with CO 2 capture and storage. Energy Pol
2007;35:4444e54.
[25] Rubin ES, Zhai H. The cost of carbon capture and storage for
natural gas combined cycle power plants,. Environ Sci
Technol 2012;46:3076e84.
[26] NREL. Equipment design and cost estimation for small
modular biomass systems, synthesis gas cleanup, and
oxygen separation equipment. In: Ibsen K, editor. National
renewable energy laboratory, Colorado, US; 2006. p. 33e7.
[27] Molburg JC, Doctor RD. Hydrogen from steam-methane
reforming with CO2 capture. In: 20th annual international
Pittsburgh coal conference; 2003. p. 1e21.
[28] Saur G, Ramsden T, James B, Colella W. Hydrogen and fuel
cells program: production case studies. In: Melaina M, editor.
Current central hydrogen production from PEM electrolysis
version 3.101. U.S. Department of Energy; 2012. Online.
[29] Rutkowski MD, Klett M, White JS, Schoff RL, Buchanan T.
Hydrogen production facilities plant performance and cost
comparisons. Final Report. March 2002. p. 2002.
[30] Bartels JR, Pate MB, Olson NK. An economic survey of
hydrogen production from conventional and alternative
energy sources. Int J of Hydrogen Energy 2010;35:8371e84.
[31] Sinnott RK. Chemical engineering design: SI Edition. Elsevier;
2009.
[32] Prasad R, Uppal KB, Aaron AL, Bredehoeft PR, Dysert LR,
Whiteside JD. Development of factored cost estimateseas
applied in engineering, procurement, and construction for
the process industries. 2011.
[33] Turton R, Bailie RC, Whiting WB, Shaeiwitz JA. Analysis,
synthesis and design of chemical processes. Pearson
Education; 2008.
[34] Bhandari R, Trudewind CA, Zapp P. Life cycle assessment of
hydrogen production via electrolysis e a review. J Clean Prod
2014;85:151e63.
[35] Holladay JD, Hu J, King DL, Wang Y. An overview of hydrogen
production technologies. Catal Today 2009;139:244e60.
[36] Zhu W, Deevi S. A review on the status of anode materials for
solid oxide fuel cells. Mater Sci Eng A 2003;362:228e39.
[37] Philibert C. The present and future use of solar thermal
energy as a primary source of energy. Paris, France:
International Energy Agency; 2005.
[38] Bertuccioli L, Chan A, Hart D, Lehner F, Madden B, Standen E.
Development of water electrolysis in the European union.
2014. p. 160.
[39] Jacobs JD. Economic modeling of cost effective hydrogen
production from water electrolysis by utilizing Iceland's
regulatory power market. 2016.
[40] Genovese J, Harg K, Paster M, Turner J. Current (2009) State-
of-the-Art hydrogen production cost estimate using water
2554 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
electrolysis: independent review. National Renewable Energy
Laboratory; 2009.
[41] EIA. Levelized avoided cost of new generation resources in
the annual energy outlook 2014. US Energy Information
Administration; 2014.
[42] Liao M-S, Zhang Q-E. Dissociation of methane on different
transition metals. J Mol Catal A Chem 1998;136:185e94.
[43] Aiello R, Fiscus JE, zur Loye H-C, Amiridis MD. Hydrogen
production via the direct cracking of methane over Ni/SiO2:
catalyst deactivation and regeneration. Appl Catal A
2000;192:227e34.
[44] Zhang T, Amiridis MD. Hydrogen production via the direct
cracking of methane over silica-supported nickel catalysts.
Appl Catal A 1998;167:161e72.
[45] Muradov NZ. CO2-free production of hydrogen by catalytic
pyrolysis of hydrocarbon fuel. Energy Fuels 1998;12:41e8.
[46] Muradov NZ. How to produce hydrogen from fossil fuels
without CO2 emission. Int J Hydrogen Energy 1993;18:211e5.
[47] Lee EK, Lee SY, Han GY, Lee BK, Lee T-J, Jun JH, et al. Catalytic
decomposition of methane over carbon blacks for CO 2-free
hydrogen production. Carbon 2004;42:2641e8.
[48] Muradov N. Catalysis of methane decomposition over
elemental carbon. Catal Commun 2001;2:89e94.
[49] Guo Y, Fang W, Lin R. Zhejiang daxue xuebao (gongxue ban).
J Zhejiang Univ Sci 2005;39:538e41.
[50] Abbasi T, Abbasi SA. Decarbonization of fossil fuels as a
strategy to control global warming. Renew Sustain Energy
Rev 2011;15:1828e34.
[51] Shah N, Panjala D, Huffman GP. Hydrogen production by
catalytic decomposition of methane. Energy Fuels
2001;15:1528e34.
[52] Kreysa G. Climate protection by an alternative use of
methanedthe carbon moratorium. Chem Sus Chem
2009;2:49e55.
[53] Geißler T, Aba  nades A, Heinzel A, Mehravaran K, Müller G,
Rathnam RK, et al. Hydrogen production via methane
pyrolysis in a liquid metal bubble column reactor with a
packed bed. Chem Eng J 2016;299:192e200.
[54] Ogino Y. Catalysis by molten metals and molten alloys. Catal
Rev Sci Eng 1981;23:505e51.
[55] Saito Y, Miyashita F, Ogino Y. Studies on catalysis by molten
metal: VI. Kinetics and the reaction scheme for the
dehydrogenation of isopropyl alcohol over the liquid indium
catalyst. J Catal 1975;36:67e73.
[56] Miyamoto A, Ogino Y. Studies on catalysis by molten metal.
V. Kinetics of the dehydrogenation of sec-butyl alcohol over
the liquid indium catalyst. J Catal 1972;27:311e21.
[57] Miyamoto A, Okano K, Ogino Y. Studies on the catalysis by
molten metal: VII. A simple MO treatment on the
dehydrogenation of alcohols and amines over the molten
metal. J Catal 1975;36:276e84.
[58] Miyamoto A, Ogino Y. Studies on the catalysis by the molten
metal: IX. A comparison between the isotope effect of two-
dimensional dehydrogenation and the isotope effect of
three-dimensional dehydrogenation of methyl alcohol. J
Catal 1976;41:212e7.
[59] Miyamoto A, Ogino Y. Studies on the catalysis by the molten
metal: VIII. Kinetic isotope effects measured by using a pulse
reaction technique. J Catal 1975;37:133e41.
[60] Miyamoto A, Ogino Y. Studies on catalysis by molten metals:
X. Hydrogen transfer reactions between alcohols and
ketones on liquid indium catalyst. J Catal 1976;43:143e53.
[61] Nagel CJ, Chanenchuk CA, Wong EW, Bach RD. Catalytic
extraction processing: an elemental recycling technology.
Environ Sci Technol 1996;30:2155e67.
[62] Miller CB, Malone DP. Molten metal decomposition
apparatus. Google Patents US5435814, 1995
[63] Steinberg M. Fossil fuel decarbonization technology for
mitigating global warming. Int J Hydrogen Energy
1999;24:771e7.
[64] Steinberg M. Production of hydrogen and methanol from
natural gas with reduced CO2 emission. Int J Hydrogen
Energy 1998;23:419e25.
[65] Serban M, Lewis MA, Marshall CL, Doctor RD. Hydrogen
production by direct contact pyrolysis of natural gas. Energy
Fuels 2003;17:705e13.
[66] Plevan M, Geißler T, Aba  nades A, Mehravaran K, Rathnam R,
Rubbia C, et al. Thermal cracking of methane in a liquid
metal bubble column reactor: experiments and kinetic
analysis. Int J Hydrogen Energy 2015;40:8020e33.
[67] Wang K, Li W, Zhou X. Hydrogen generation by direct
decomposition of hydrocarbons over molten magnesium. J
Mol Catal A Chem 2008;283:153e7.
[68] Parkinson B, Matthews JW, McConnaughy TB, Upham DC,
McFarland EW. Techno-economic analysis of methane
pyrolysis in molten metals, decarbonizing natural gas. Chem
Eng Technol 2017;40:1022e30.
[69] Upham DC, Agarwal V, Khechfe A, Snodgrass ZR, Gordon MJ,
Metiu H, et al. Catalytic molten metals for the direct
conversion of methane to hydrogen and separable carbon.
Science 2017;358:917e21.
[70] Kantarci N, Borak F, Ulgen KO. Bubble column reactors.
Process Biochem 2005;40:2263e83.
[71] Ando T, Ueno K, Taniguchi S, Takagi T. Induction pump for
high-temperature molten metals using rotating twisted
magnetic field: thrust measurement experiment with solid
conductors. IEEE Trans Magn 2002;38:1789e96.
[72] Mirovics I, Frost MT, Koenig RL, Tait PJ. Calcination using
liquid metal heat exchange fluid. Google Patents US6482366,
2002
[73] Areaux LD, Scowden AD, Corio PJ. Inert gas bubble-actuated
molten metal pump with gas-diffusion grid. Google Patents
US6068812, 2000
[74] Alinta Energy. Port Augusta solar thermal generation
feasibility study stage 1-Pre-feasibility study options study
report. 2014.
[75] Sumper A, Baggini A. Electrical energy efficiency:
technologies and applications. John Wiley & Sons; 2014.
[76] Naro R, Williams D, Satre P. Control of slag and insoluble
buildup in ladles, melting and pressure pour furnaces.
Ductile Iron News; 2004.
[77] GPSA. Engineering data book, gas processors suppliers
association, vol. 2; 2004. p. 16e24.
[78] Ludwig EE. Applied process design for chemical and
petrochemical plants. Gulf Professional Publishing; 1997.
[79] Couper JR, Penney WR, Fair JR. Chemical process equipment
revised 2E: selection and design. Gulf Professional
Publishing; 2009.
[80] Sabharwall P, Kim ES, McKellar M, Anderson N, Patterson M.
Process heat exchanger options for the advanced high
temperature reactor. INL/EXT-11e21584. June, 2011.
[81] Williams D. Assessment of candidate molten salt coolants
for the NGNP/NHI Heat-Transfer Loop. ORNL/TM-2006/69.
Oak Ridge, Tennessee: Oak Ridge National Laboratory;
2006.
[82] Taddeo M. Assessing the effect a refractory insulation lining
has on EAF energy consumption. SEAISI Q 2010;39:53.
[83] Gerrard A. Guide to capital cost estimating. IchemE; 2000.
[84] Towler G, Sinnott RK. Chemical engineering design:
principles, practice and economics of plant and process
design. Elsevier; 2012.
[85] Loh H, Lyons J, White CW. Process equipment cost
estimation. Report No. DOE/NETL-2002/1169. National
Energy Technology Center; 2002.
 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 3 ( 2 0 1 8 ) 2 5 4 0 e2 5 5 5
[86] Chauvel A, Fournier G, Raimbault C. Manual of process
economic evaluation, Editions Technip. 2003.
[87] Ulrich GD, Vasudevan PT. Chemical engineering: process
design and economics; a practical guide. Process Publ;
2004.
[88] Glose J. Air products to build $400M steam methane reformer
in Texas. Lehigh Valley Business; 2016.
[89] Kool A, Marinussen M, Blonk H. LCI Data for the calculation
tool feedprint for greenhouse gas emissions of feed
production and utilization. 2012. Netherlands.
[90] World Steel Assocation. Short range outlook. 2017.
[91] Li VC. Large volume, high-performance applications of fibers
in civil engineering. J Appl Polym Sci 2002;83:660e86.
[92] Jones A. Global Coke Production in 2016 and Beyond: is there
enough coal tar for everybody?. International Tar
Association; 2016. Online.
2555
[93] Machhammer O, Bode A, Hormuth W. Financial and
ecological evaluation of hydrogen production processes on
large scale. Chem Eng Technol 2016;39:1185e93.
[94] Environmental Protection Agency. Inventory of US
greenhouse gas emissions and sinks: 1990e2015. In: EPA
(Ed.), US EPA, Washington DC, USA; 2017.
[95] Petron G, Frost G, Miller BR, Hirsch AI, Montzka SA, Karion A,
et al. Hydrocarbon emissions characterization in the
Colorado Front Range: a pilot study. J Geophys Res Atmos
2012;117.
[96] Weber CL, Clavin C. Life cycle carbon footprint of shale gas:
review of evidence and implications. Environ Sci Technol
2012;46:5688e95.
[97] Lemus RG, Martı́nez Duart JM. Updated hydrogen production
costs and parities for conventional and renewable
technologies. Int J Hydrogen Energy 2010;35:3929e36.