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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5

Available online at www.sciencedirect.com

ScienceDirect

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

Critical assessment of the production scale

required for fossil parity of green electrolytic
hydrogen

Joris Proost*
Division of Materials and Process Engineering, Universit
e Catholique de Louvain, 1348 Louvain-la-Neuve, Belgium

highlights

Electrolytic H2 is fossil-free when produced from renewable electricity.

Today's price of H2 from fossil-based SMR processes (fossil parity) is 2,0 V/kg.

This can be reached for electrolytic H2 at a renewable electricity price of 30 V/MWh.

The required electrolyser CAPEX of 500 V/kW can be reached already at a few MW.

Such small-scale allows for decentralised H2 production in line with local H2 demand.

article info abstract

Article history: Hydrogen produced from renewable electricity through Power-to-Hydrogen can facilitate
Received 9 March 2020 the integration of high levels of variable renewable electricity into the energy system. An
Received in revised form electrolyser is a device that splits water into hydrogen and oxygen using electricity. When
27 April 2020 electricity is produced from renewable energy sources, electrolytic hydrogen can be
Accepted 29 April 2020 considered to be green. At the same time, electrolysers can help integrate renewable
Available online 27 May 2020 electricity into power systems, as their electricity consumption can be adjusted to follow
wind and solar power generation. Green hydrogen then also becomes a carrier for
Keywords: renewable electricity. Key green hydrogen production technologies, mostly PEM and
Power-to-hydrogen alkaline electrolysers, are still further maturing, both in technical (efficiency), economical
Electrolyser (CAPEX) and durability (lifetime) performance. Nonetheless, we will show in this contri-
Renewable electricity bution how fossil parity for green hydrogen, i.e. a Total Cost of Ownership (TCO) similar to
Fossil parity grey H2 coming from todays CO2 intensive SMR processes, can already be achieved today.
Moreover, this can be realised at a scale which corresponds to the basic units of renewable
electricity generation, i.e. a few MW.
© 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

this context, low-carbon electricity from renewables may

Introduction become the preferred energy carrier. The share of renewable
electricity in all of the energy consumed by end users world-
The global energy system has to undergo a profound trans- wide would need to increase to 40% in 2050 (from about 4% in
formation to achieve the targets of the Paris Agreement. In

* Corresponding author.
E-mail address: joris.proost@uclouvain.be.
https://doi.org/10.1016/j.ijhydene.2020.04.259
0360-3199/© 2020 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
17068 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5

2015) to achieve the decarbonised energy world envisaged by 2015 a total estimated value of 115 billion V, corresponding to
the agreement [1]. In absolute terms, this implies that the total a hydrogen demand of about 56 Mton/yr. By dividing the total
installed renewable power capacity should increase from estimated market value by the total worldwide hydrogen de-
about 1.500 GW in 2015 to more than 15.000 GW in 2050, i.e. a mand at the same year, a reasonable first-order estimation of
10-fold increase [2]. However, the total decarbonisation of the “average” market price for fossil H2 can then be obtained
certain sectors, such as transport, industry and applications as 115/56 y 2,0 V/kg. In the current paper, we aim at critically
that require high-grade heat, may be difficult purely by means assessing the production scale that would be required to reach
of electrification. This challenge could be addressed by green such fossil parity using electrolytic hydrogen.
hydrogen produced electrochemically from renewables (so- While doing so, it is important to acknowledge that sig-
called Power-to-Hydrogen or P2H [3]), allowing large amounts nificant regional differences may still exist on a micro-
of renewable electricity to be channeled from the power sector economical level. This is not only due to geographical varia-
into these end-use sectors [4]. Renewable electricity can be tions in the production price of SMR H2, but also depending on
used to produce green hydrogen via water electrolysis, a well- the availability of sufficient and low-cost renewable elec-
known process splitting acidified or alkalised water into ul- tricity. It is for instance well-known that the production cost
trapure (up to 99.998%) H2 and O2 [5]. Such electrolytic H2 can of hydrogen from SMR is significantly influenced by natural
then further be used downstream as a green and clean gas prices, which account for 45%e75% of the total SMR pro-
chemical feedstock material in sectors otherwise difficult to duction cost. As a result, the low gas prices in the Middle East,
decarbonise through electrification. The latter include both the Russian Federation, and North America give rise to some
the chemical industry itself, as well as new applications in the of the lowest hydrogen SMR production costs, sometimes
transport sector [6]. As to the first, hydrogen is currently even down to 1.5 V/kg [15]. On the other hand, gas importers
already widely used in several industrial sectors (refineries, such as Japan, Korea, China and India have to contend with
ammonia production, bulk chemicals, etc.), with the majority higher gas import prices, which inevitably results in higher
of it being produced from natural gas by steam-methane hydrogen production costs. As a result, it will be much more
reforming (SMR), a vast CO2-intensive process [7]. Green feasible for electrolytic hydrogen produced from renewable
hydrogen from renewables could replace such fossil fuel- electricity to compete effectively with SMR in countries
based feedstocks in high-emission applications. For the relying on natural gas imports and characterised by good
transport sector, fuel cell electric vehicles (mainly cars and renewable resources.
busses) provide already today an attractive low-carbon
mobility option when the hydrogen is produced from renew-
able energy sources, and offer driving performances compa- Hydrogen production
rable to conventional vehicles. On the longer run, H2-based
electrofuels, i.e. liquid fuels produced from renewable power, Hydrogen production today
can also replace fossil fuels in the freight sector (including
aviation and heavy-duty rail and trucks), without the need to As of today, hydrogen is being used as a specialty chemical in
change end-use technologies [8]. a number of applications. These are generally classified into 4
Although water electrolysis is already a well-established H2 main categories [14], as illustrated in Table 1: (1) the chemical
production technology for almost a century [9], its large-scale industry, where H2 is a basic building block for the synthesis of
implementation for the production of green H2 has been ammonia, methanol and a number of technical polymers; (2) a
hampered mainly by cost issues. In a recent review [10], the number of downstream refining processes, like hydro-
production cost of hydrogen from electrolysis has been cracking and hydro-treating; (3) iron, steel and glass
extracted from a large amount of litterature data, resulting in manufacturing, where H2 is the preferred reducing gas during
a very wide range of cost values, ranging from about 2 V/kg to
20 V/kg. This was attributed to the large variability of the
underlying assumptions and working parameters of the Table 1 e Today’s major industrial sectors using H2. For
different sources, the production scale being the most each sector, a typical unit plant size is given (in Nm3/h2
important one [11]. Moreover, when evaluating the potential demand), as well as the corresponding equivalent
and economic viability of such green hydrogen production by electrolyser power.
water electrolysis, the current price of fossil SMR-based H2 Industry Key Applications Unit plant size Equivalent
often appears as a rather challenging benchmark [12]. For a Sector (in Nm3/h2 electrolyser
fair comparison though, it was recently pointed out [13] that demand) power
one should always keep in mind that the industrial SMR pro- Chemical Ammonia (NH3) 80.000 400 MW
duction price is usually considered for immediate use, while Methanol (CH3OH) 10.000 50 MW
Refining Hydrocracking 400.000 2.000 MW
often additional storage is necessary to meet fluctuations in
Hydrotreating
demand and delivery as well. On top of that, hydrogen from
Iron & Annealing 400 2 MW
SMR still needs to be purified for most applications in order to Steel Blanketing gas
reach the same grade as electrolytic one. Moreover, in such Forming gas
cost comparisons, the potential valorisation of ultra-pure General Semiconductor 50 0,25 MW
electrolytic oxygen (8 kg for each kg of H2) is totally neglec- Float glass production
ted. In any case, on a macro-economical level, according to Propellant fuel
Cooling of generators
Ref. [14], the global hydrogen feedstock market represented in
 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5 17069

annealing, blanketing and forming processes; (4) other spe- that todays local H2 consumers, besides becoming local (on-
cialty applications, like the semiconductor industry, the use as site) producers of renewable electricity, also need to become
a propellant fuel or the cooling of generators. The first two local (on-site) producers of electrolytic H2, at a production
categories represent with 65% by far the largest contribution scale which still allows to meet the stringent requirement of
to the total H2 demand, followed by refining, iron, steel and fossil parity at about 2,0 V/kg. On the longer run, with the
glass manufacturing (all together about 25%) and the projected 10-fold increase in renewable power to 15.000 GW in
remaining 10% for the other specialty applications. 2050, a mere 2% use of this capacity would be required to
An important difference between each of these 4 categories satisfy the equivalent 300 GW water electrolysis demand. This
is the scale of the so-called unit process or production size, i.e. can be considered to be within the range of grid balancing
the typical individual plant or reactor capacity required to services, making such green electrolytic hydrogen production
generate the appropriate amount of H2 feedstock in each on the long run an even more viable and attractive alternative
application. Table 1 gives in this respect some indicative hydrogen production technology.
numbers (in Nm3/h of H2 demand) for each of these 4 cate- The above suggested transformation from centralised (off-
gories. In its last column, it also provides the equivalent site) fossil-based H2 production to a decentralised (on-site)
electrolyser capacity that would be required to satisfy these green electrolytic H2 production provides a significant para-
unit size H2 feedstock demands by on-site electrolytic H2 digm shift, allowing local consumers to become local pro-
production (assuming a state-of-the art electrolyser efficiency ducers as well. Up to now, in an industry largely governed by
of 70% [16], corresponding to a renewable electricity need of CO2-intensive chemical processes, such a local H2 production
47,1 kWh/kg H2). Large variations in production scale can be in line with the local H2 consumption was simply not feasible,
noticed across the different sectors, ranging from about because of the minimum production scale required for both
250 kW at the low-end (typical for float glass production) to a SMR and coal or oil gasification. The latter typically starts at a
few GW at the high-end (typical for H2 demand in refineries). few 10.000 Nm3/hr (about 8000 ton/yr) for the smallest unit
As of today, the great majority of all of the above H2 is being size installations, equivalent to a 50 MW electrolyser. As can
delivered by a centralised, off-site hydrogen production, be seen in Table 1, this largely exceeds the industry needs in a
dominated by 2 large-scale chemical processes: steam number of applications (iron & steel, as well as general in-
methane reforming (SMR) and coal gasification. According to dustry). Moreover, additional CO2-intensive logistics (incl.
Ref. [15], these processes made up about 76% and 23% Transport, compression and storage) are required in these
respectively of the total H2 production in 2018. Unfortunately, applications as well.
as shown in Fig. 1, both of these processes are heavily CO2
intensive, SMR emitting up to 8 tons of CO2 per ton of H2 Green hydrogen production scale-up
produced. Therefore, with the objective of reaching the CO2
emission targets in todays fossil-based H2 production, the part Contrary to the intrinsically large-scale SMR, water electrol-
of green electrolytic hydrogen production from renewable ysis is intrinsically small-scale, as illustrated in Fig. 2. Both the
electricity (which represents less than 4% today) can be ex- geometrical area of the electrodes (a few m2 at most) and the
pected to significantly increase over the coming years. In order number of electrodes that can be compiled in series in a single
to meet the current global H2 demand of around 60 Mton/year, stack is relatively limited. As a result, the unit size of water
a total of 300 GW installed electrolyser capacity would be electrolysers has long been limited to the kW-range, a typical
needed. As this represents today about 20% of the total on-site containerized production unit being a few 100 kW at
installed renewable power capacity, such massive electrolyser most. However, in order to be able to realise the mandatory
deployment is currently not very realistic. As a result, a se- coupling to renewables, mainly wind and solar, the power
lection of technologically feasible market penetrations for scale of water electrolysers needs to become of the same order
electrolytic H2 needs to be made. Such selection also implies

Fig. 1 e Summary of today’s main hydrogen production Fig. 2 e A typical unit size configuration for an industrial
technologies. alkaline water electrolyser.
17070 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5

Fig. 3 e Projected cost reduction associated with a scale-up

of green electrolytic hydrogen production. The blue
diamonds represent real CAPEX data for PEM electrolysers,
except for the ones between 100 and 1000 Nm3/hr, which
are cost projections. The red squares and green triangles
are real SMR CAPEX data for small-scale (on-site) and large-
scale (centralised) reformers, respectively. Taken from
Refs. [17]. (For interpretation of the references to color in
this figure legend, the reader is referred to the Web version Fig. 5 e Electrolytic H2 production cost as function of
of this article.) electrolyser’s annual operating time (1 yr ¼ 8760 h),
simulated according to Eq. (1) for two different CAPEX/
electricity cost combinations. The electrolyser’s efficiency
of magnitude as the renewable electricity source itself. As was fixed at 70%, corresponding to 47,1 kWh/kg. The
illustrated in Fig. 3 (reproduced from Ref. [17]), this requires a additional dashed trendlines for the 2 data sets include,
major scale-up from the kW-scale, typical for state-of-the art besides the electricity cost, also an additional fixed OPEX
electrolysers about a decade ago, towards the multi-MW scale cost equal to 30% of CAPEX.
typical for state-of-the art on-shore wind turbines today. Fig. 3
also shows that this mandatory scale-up has the potential to

significantly reduce the investment cost (CAPEX) of electro-

lysers, potentially reaching the same order of magnitude as
small-scale SMR installations from the MW-level onwards.
Such an electrolyser scale-up has initially been realised by
increasing the number of cells per stack, as illustrated in Fig. 2.
However, from the state-of-the-art data that we recently
collected from a number of electrolyser manufacturers, such a
“keep-on-stacking” approach seems to have a practical limit
at around 100 cells/stack [18]. Beyond that number, other
balance-of-plant issues come into play, including the risk of
electrical shorts [19] and the technological complexity of a
safe large-scale gas collection [20]. There are also a number of
specific issues related to electrochemical reactor design, like
the increased risk of a non-homogenous electrolyte distribu-
tion when pumped through a larger stack [21], and a non-
homogeneous current distribution within the different cells
[22]. Note that this apparent 100-cell limitation is by no means
a stringent intrinsic limitation, but rather an empirical
observation based on the above cited industrial data. In other
words, it appears that for a number of electrochemical and/or
technical reasons, electrolyser manufacturers are currently
preferring to upscale production capacity modularly, rather
than increasing the capacity of a single stack.
As a result, for (multi-)MW applications, multi-stack elec-
trolyser systems are typically being used. As an example, Fig. 4
Fig. 4 e Examples of multi-stack alkaline electrolyser shows both an iconic historical illustration of a 135 MW
systems: (a) A 135 MW electrolyser plant from 1953 [23]; (b) alkaline electrolyser plant dating back already from 1953 [23],
Current multi-MW plant designs from a major electrolyser and a number of today's multi-MW plant designs from a major
manufacturer, based on a single stack electrolyser capacity electrolyser manufacturer, based on a single stack electrolyser
of 2,2 MW [24]. of 2,2 MW [24].
 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5 17071

Fossil parity for green hydrogen

The cost of electrolytic hydrogen

While Fig. 5 shows that it is technically feasible to produce

green electrolytic hydrogen at the multi-MW scale
(even > 100 MW), the critical question still remains at what
price/cost. Clearly, if green H2 is to become competitive with
today's “grey” SMR H2, it should be made available at its cur-
rent market price, i.e. around 2,0 V/kg. In this respect, Fig. 5
illustrates the effect of the 3 major parameters affecting the
electrolytic H2 production cost: the operational time of the Fig. 6 e Documented and extrapolated decrease in average
electrolyser (in full load hours or FLH), the cost of renewable auction prices for renewable electricity from solar PV and
electricity (ELCTR, in V/MWh), and the electrolyser CAPEX on-shore wind (from Refs. [26]).
(CPX, in V/kW). The basic equation used for this first-order
cost simulation is as follows
renewable electricity from both solar PV and on-shore wind.
 
ELCTR CPX 1 Clearly, prices on the order of 30 V/MWh can be expected to be
H2 costðV = kgÞ ¼ þ , ,ε (1)
1000 10 FLH realistic already as of 2020. At the same time, this very study
also projects load factors of combined wind and solar power to
where ε represents the electrolyser power consumption (in
exceed 50% in vast areas. A recent German field study
kWh/kg). With respect to the latter, the theoretical minimum
reporting on the operational experience of a 6 MW Power-to-
value εth for obtaining H2 through electrochemical water
Hydrogen demonstration plant seems to confirm these
splitting can simply be calculated based on a 2 electron
promising numbers [27]. During its initial testing phase, when
reduction step
operation time was limited to 8 h during working days, the
electrolyser load demand curve led to an average electricity
2Hþ þ 2e ¼ H2 (acid) (2)
cost (as purchased from the EPEX SPOT day-ahead auction
market) of about 36 V/MWh. After full automation of the plant
2H2O þ 2e ¼ H2 þ OH (alkaline)

With 1 kg of H2 requiring 103 F Coulomb (F being Faraday's

constant), 1 kg/h of H2 then corresponds to an electrical cur-
rent of (96487$103)/3600 ¼ 26802 A. Multiplied by the theoret-
ical water decomposition potential of 1.23 V, this then gives a
theoretical minimum power consumption εth ¼ 33 kWh/kg. A
typical electrolyser efficiency being 70% [16], a typical ε-value
to be used in Eq. (1) is therefore 33/0.7 ¼ 47.1 kWh/kg. Also
note that in Eq. (1), the factor 1/1000 in the first term serves to
convert MWh into kWh, while the factor 1/10 in the second
term comes from a linear depreciation for a 10 years electro-
lyser operation.
First of all, for the red set of parameters in Fig. 5, i.e. a
CAPEX of 1000 V/kW and a renewable electricity cost of 70 V/
MWh (as taken form [25], a reference which dates back already
from 2014), it is clear that producing H2 from water electrolysis
is not always economically viable with respect to the current
SMR benchmark price of 2 V/kg. In particular, before becoming
a realistic alternative production technology, there is a need
for cheap (er) renewable electricity (well below 70 V/MWh),
the investment cost of electrolysers needs to be brought down
(well below 1000 V/kWh), and there should preferably also be a
clear industrial commitment to CO2 reduction. The latter
might notably impose an additional tax/cost to SMR H2,
helping to further close the gap with electrolytic H2.
Luckily, with respect to the red parametric values used in Fig. 7 e State-of-the-art CAPEX data for alkaline
Fig. 5, significant progress has been made since 2014, both in electrolysers as a function of power input. The change in
reducing the price of renewable electricity and in reducing the slope for alkaline electrolysers corresponds to the use of
electrolyser CAPEX. As to the first, Fig. 6, taken from a recent multi-stack systems.
study from the International Energy Agency (IEA) [26], shows
the projected reduction in average auction prices for
17072 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5

to a 24/7 operation so that electricity could be bought in times

of low spot prices, additional cost savings of more than 15 V/
MWh could be realised.
Secondly, as to the electrolyser CAPEX, Fig. 7 shows state-
of-the-art data from NEL, one of the world's largest alkaline
electrolyser manufacturers. They clearly show a significant
decrease well below the value of 1000 V/kW used for the red
data set in Fig. 5. In particular, a CAPEX value of 750 V/kW,
considered by utility providers to be the capital cost for storing
renewable electricity, is already realistic today for a single
stack 2 MW system. Moreover, a significant further reduction
in CAPEX as low as 500 V/kW is projected for multi-stack
systems when scaling up to 50e100 MW. Also note from
Fig. 7 that single stack electrolysers are much more suscepti-
ble to CAPEX reduction than multi-stack systems when Fig. 9 e Total Cost of Ownership (TCO) of electrolytic H2
upscaled. (delivered at 35 bar) as a function of power input, obtained
Based on the above updated numbers, the green data set in from the Danish electrolyser manufacturer
Fig. 5 then allows to anticipate a significant reduction in the GreenHydrogen. Prices are based on a 10 years operation
electrolytic H2 production cost. Indeed, assuming the most (including a 10 year’s service and maintenance agreement)
favorable but still realistic CAPEX value of 500 V/kW in com- for a complete turnkey, containerized alkaline electrolyzer
bination with an electricity cost of 30 V/MWh and a state-of- unit (including inverter and water treatment), delivered
the-art electrolyser efficiency of 70% (i.e. 47,1 kWh/kg), green and installed in Europe. OPEX includes the use of
electrolytic H2 can indeed start competing with SMR from 4500 electricity, water and nitrogen (for purge), assuming a
operating hours onwards (i.e. a load factor of about 50%). Note renewable electricity price of 40 and 45 V/MWh,
that under these conditions, the total H2 cost calculated from respectively.
Eq. (1) comes down to 1,95 V/kg and is mainly determined by
the electricity cost, which represents 47,1*0.03 ¼ 1,41 V/kg or
72%. One should therefore be aware that any efforts to further
reduce the CAPEX of alkaline water electrolysers below 500 V/
kW will only have a minor overall effect. For instance, for a Instead, Eq. (1) indicates that it will be much more effective
CAPEX of 250 V/kW, Fig. 8(a) shows that the total H2 cost goes to focus technological efforts on improving the electrolyser's
down to 1,68 V/kg, of which only 15% would come from the electrochemical efficiency, since the related power con-
very stringent techno-economical measures needed to reach sumption ε is a common factor to both the OPEX and CAPEX
such low CAPEX value. part. The main effect of an efficiency increase, which is
equivalent to a lower kWh/kg H2 electricity consumption, is

Fig. 8 e (a) Simulated electrolytic H2 production cost as a function of electrolyser CAPEX, assuming a 70% efficiency (i.e.
47,1 kWh/kg) and a 50% load factor (i.e. 4380 h/year). Renewable electricity price was fixed at 30 V/MWh, corresponding to
47,1* 0.03 ¼ 1,41 V/kg. The right axis gives the resulting % contribution of the electrolyser CAPEX to the total H2 cost; (b)
Renewable electricity price (in V/MWh) as a function of electrolyser efficiency to arrive at an electrolytic H2 cost of 2.0 V/kg
for three different CAPEX values (500, 1000 and 2800 V/kW).
 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5 17073

that it allows to relax the sometimes rather stringent condi- installations (cfr. Fig. 3). As already suggested from Table 1, for
tions on the renewable electricity price needed to reach fossil some feedstock applications, like ammonia or methanol pro-
parity. In this respect, Fig. 8(b) shows, as a function of elec- duction, such a large unit size can be relevant even for an on-
trolyser efficiency, the renewable electricity price (in V/MWh) site, decentralised green H2 production. However, Fig. 7 in-
that would be needed to arrive at an electrolytic H2 cost of dicates that there might still be a much smaller production
2,0 V/kg for three different CAPEX values: 500, 1000 and scale for reaching such low CAPEX values. Indeed, when
2800 V/kW. The last one can be considered, according to extrapolating the CAPEX data of single-stack alkaline elec-
Ref. [15], to be a realistic CAPEX target value by 2030 for solid trolysers in Fig. 7, the level of 500 V/kW (dashed horizontal
oxide electrolysers (SOE). These are especially known for their green line) can already be reached around 3e4 MW. Such a
high intrinsic efficiency (up to 90%) resulting from high tem- significant reduction in the scale required for fossil parity is
perature operation (650e1000  C). Nonetheless, from the directly related to the much steeper reduction in CAPEX that
negative red data in Fig. 9(b), it can be seen that even in that can be realised for single-stack as compared to multi-stack
case the higher SOE efficiency will still not be able to provide systems. A straightforward consequence of the above obser-
electrolytic H2 at fossil parity. This would require SOE CAPEX vation is that the minimum investment cost needed to install
values to decrease even further, down to a level of 1000 V/kW. electrolytic hydrogen production units capable of delivering
We do acknowledge that the trends presented in Figs. 5 and green H2 at fossil parity goes down significantly as well: from
8 should be considered to be a first-order cost simulation, about 100$103*500 ¼ 50 MV to a mere 200 kV, a very realistic
based on the rather basic Eq. (1). For instance, most other number in view of a local, decentralised production.
economic analyses tend to add a yearly interest rate (typically Even more direct corroborating evidence for this relatively
7e10%) and in some cases an additional OPEX contribution on small-scale fossil parity for green electrolytic H2 is provided in
top of electricity cost (up to 30% of CAPEX). We have decided Fig. 9. The latter presents state-of-the-art industrial data for
here not to do so, as the exact numbers are often rather the Total Cost of Ownership (TCO) of electrolytic H2 as a
arbitrary chosen. Moreover, they barely change the principal function of power input, as obtained from GreenHydrogen, a
trends induced by varying the major parameters (i.e. elec- Danish electrolyser manufacturer who produces pressurized
tricity price, capacity factor, CAPEX and efficiency), which are stacks. Values are based on a 10 year operation, including a 10
already included in our basic Eq. (1). This has been explicitly years’ service and maintenance agreement, for a complete
illustrated in Fig. 5 by the two additional dashed trendlines, turn-key, containerized alkaline electrolyzer unit (including
which include an additional fixed OPEX cost equal to 30% of inverter and water treatment), delivered and installed in
CAPEX. Europe. The electrolyser power consumption is guaranteed at
Finally, for a fair comparison, an additional CO2 price for 46.7 kW/kg, and delivers H2 at 35 bar. The hydrogen produc-
SMR H2 should be taken into account as well [28]. In this tion cost in Fig. 9 also takes into account the OPEX part,
respect, the seminal IEA report “The Future of Hydrogen” [15] including the use of water, nitrogen (for purge) and electricity,
predicts an average increase of 50% in the hydrogen produc- assuming a renewable electricity price of 40 and 45 V/MWh,
tion cost from SMR when imposing a carbon price of 100$/ respectively. A number of striking observations can be made
tCO2. This might even trigger the large-scale implementation from this figure. First of all, these TCO data confirm the trend
of CCUS, which would become economically attractive if CO2 already observed in Fig. 7, namely the much steeper decrease
prices were above 50$/tCO2. Adding CCUS to SMR plants in unit price for hydrogen delivered from a single stack vs. a
would then lead in turn to cost increases of about 50% in terms multi-stack system. Secondly, until now, it was commonly
of CAPEX and 10% for fuel, the exact amounts depending on agreed that the only option to decrease the electrolytic H2
the plant design. It also leads on average to a doubling of OPEX production cost towards fossil parity was to increase the ca-
as a result of CO2 transport and storage costs [15]. Nonethe- pacity of the (multi-stack) system up to 50e100 MW, as
less, our above projections, without considering any carbon already discussed with Fig. 7. Fig. 9 now clearly shows that in
price, are sufficiently promising as such to stimulate already that case, even a decrease in renewable electricity cost (from
today a further penetration of water electrolyser technology 45 to 40 V/MWh) only slightly affects the scale of fossil parity,
for renewable energy storage purposes. At the same time, they due to the relatively small decrease in TCO with power typical
should also provide confidence for the ultimate consideration for such multi-stack systems. As a result, large-scale electro-
of electrolytic H2 as a basic chemical building block, enabling lyser systems would still be necessary to reach fossil parity.
direct coupling to renewable electricity production and hence However, as suggested by the dashed red line in Fig. 9 through
helping to green the chemical feedstock industry. the TCO data for single stack systems, there is another tech-
nological alternative. It consists of extending the power input
The scale of fossil parity for green hydrogen that can be taken up by a single-stack electrolyser up to a few
MW, the exact power depending on the electrolyser technical
A final issue then relates to the production scale that is characteristics. This then also corresponds to the scale of the
required for obtaining such fossil parity with electrolytic H2. basic units of renewable electricity generation. Also note that
Indeed, from the data in Fig. 7, one could wrongly conclude when considering an even more stringent SMR price level of
that reaching the required reduction in electrolyser CAPEX 1.5 V/kg (corresponding in Fig. 9 to the horizonal axis rather
down to 500 V/kW would require very large-scale electrolytic than the dashed green line at 2.0 V/kg), our conclusions on the
H2 production units, on the order of 100 MW. In that case, the required scale for reaching fossil parity do not fundamentally
minimum scale for economically viable electrolytic H2 pro- change. Indeed, in that case, extrapolation of the single-stack
duction would need to become similar to current SMR TCO data would arrive at 3 MW, instead of 1.5 MW for 2.0 V/kg.
17074 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5

Note that at this stage, there is not really a rigorous sci- fossil parity could potentially also be reached at a much
entific reasoning behind the single-stack cost line extrapola- smaller production scale, on the order of a few MW. Although
tions in Figs. 7 and 9. It is a mere empirical observation, but this still requires a further intensification of the water elec-
still a rather reliable one since based on two independent in- trolysis process, e.g. by extending the power range of a single-
dustrial data sets over a relatively large power scale. It is also stack electrolyser, such small-scale fossil parity provides an
important to realise that the obtained extrapolated single- important paradigm shift. Indeed, with respect to the current,
stack power for fossil parity is not a unique number, but large-scale fossil fuels (SMR) based centralised hydrogen
something that is specific to each stack geometry (e.g. the production, it has the important advantage of allowing a
number and area of electrodes used). For instance, the Nor- decentralised local H2 production. Renewables can then be
wegian HydrogenPro already has a 3.7 MW single-stack alka- harvested anywhere, and used directly for the local produc-
line electrolyser on the market, while the Belgian-Chinese tion and consumption of green electrolytic hydrogen, in line
Cockerill Jingli Hydrogen even sells 7.5 MW single-stacks, to with the small-scale local H2 demand.
the best of our knowledge the largest single-stack on the
market.
The challenge on the electrolyser level is then to try to
Acknowledgements
technologically implement this single-stack extrapolation in
order to arrive at higher single-stack power levels. As the
We cordially thank Eric Dane from NEL and Henrik Steen
number of cells/stack seems to have reached its limit [18], an
Pedersen from GreenHydrogen for providing their state-of-
alternative option is to increase the specific area of each in-
the-art CAPEX and TCO data on electrolytic hydrogen. This
dividual electrode, e.g. by replacing classical 2-D plates by 3-D
paper was realised within the framework of the Hydrogen
foams [29,30] hence allowing for a higher current density
Technology Collaboration Programme (TCP) of the Interna-
operation [31]. Incidentally, a recent European demonstration
tional Energy Agency (IEA), more specifically Task 38 on
project (Demo4Grid) showing the technical feasibility and
Power-to-Hydrogen. It also benefited from a number of
greening potential of a single-stack 4 MW alkaline electrolyser
dric Philibert
insightful discussions with Uwe Remme and Ce
has been launched in that sense [32]. Such small-scale fossil
at the IEA's Energy Technology Policy Division and the
parity has the important advantage of allowing a decentral-
Renewable Energy Division, respectively. Finally, financial
ised local H2 production. Renewables can then be harvested
support from the Public Service of Wallonia e Dept. Of Energy
anywhere, and used directly for the local production and
and Sustainable Building is gratefully acknowledged for
consumption of green electrolytic hydrogen. This will not only
allowing these fruitful interactions with the IEA.
allow to open up todays market to electrolytic H2 in a number
of small unit scale segments (like iron & steel and glass
manufacturing, cfr. Table 1), but also to widen the use of green references
electrolytic H2 to a number of new small-scale markets (like
the food industry targeted in Ref. [32]). This is a significant
paradigm shift with respect to the current large-scale fossil [1] Pursiheimo E, Holttinen H, Koljonen T. Inter-sectoral effects
fuels (SMR) based centralised hydrogen production, the latter of high renewable energy share in global energy system.
also requiring an additional cost to transport the H2, both in Renew Energy 2019;136:1119e29.
[2] reportIRENA report 2018, Global Energy Transformation: A
terms of V/kg and CO2 footprint.
Roadmap to 2050, (for download from www.irena.org/
publications).
[3] Go€ tz M, Lefebvre J, Mo€ rs F, McDaniel Koch A, Graf F, Bajohr S,
Conclusion Reimert R, Kolb T. Renewable Power-to-Gas: a technological
and economic review. Renew Energy 2016;85:1371e90.
In this paper, we have addressed the question what would be [4] Bockris JOM. The origin of ideas on a Hydrogen Economy and
an economically viable (minimum) production scale for green its solution to the decay of the environment. Int J Hydrogen
Energy 2002;27:731e40.
hydrogen, produced from water electrolysis using renewable
[5] L. Bertuccioli et al., reportStudy on development of water
electricity. A realistic benchmark to do so is the current price
electrolysis in the EU, FCH JU Report 2014, (downloadable
of grey hydrogen produced by fossil-based and thus CO2 from http://www.fch.europa.eu/node/783).
intensive processes (the so-called fossil parity), currently [6] Gahleiter G. Hydrogen from renewable electricity : an
estimated at 2,0 V/kg. Firstly, we acknowledged the promising international review of power-to-gas pilot plants for
market opportunities for such green hydrogen in todays H2 stationary applications. Int J Hydrogen Energy
markets. The latter represent about 60 Mt/yr, and can be 2013;38:2039e61.
[7] Holladay JD, Hu J, King DL, Wang Y. An overview of hydrogen
classified in 4 major applications, all of them having their own
production technologies. Catal Today 2009;130:244e60.
typical unit size in terms of equivalent H2 demand, ranging [8] Brynolf S, Taljegard M, Grahn M, Hansson J. Electrofuels for
from a few hundreds of kW up to several GW. Secondly, it was the transport sector: a review of production costs. Renew
shown how, based on current state-of-the-art CAPEX data for Sustain Energy Rev 2018;81:1887e905.
todays multi-stack electrolysers and using a renewable elec- [9] Zeng K, Zhang D. Recent Progress in alkaline water
tricity price of 30 V/MWh, such fossil parity can be reached electrolysis for hydrogen production and applications. Prog
Energy Combust Sci 2010;36:307e26.
already today at 50e100 MW. This is about the same scale as
[10] Maggio G, Nicita A, Squadrito G. How the hydrogen
the smallest SMR installations. Finally, using the most recent
production from RES could change energy and fuel markets:
TCO values for electrolytic hydrogen, it was concluded 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 5 ( 2 0 2 0 ) 1 7 0 6 7 e1 7 0 7 5 17075

a review of recent literature. Int J Hydrogen Energy [23] Langås HG. Large-scale hydrogen production, workshop on
2019;44:11371e84. renewable energy and hydrogen export, trondheim. 2015 (for
[11] Fragiacomo P, Genovese M. Modeling and energy demand download at, https://www.sintef.no/contentassets/
analysis of a scalable green hydrogen production system. Int 9b9c7b67d0dc4fbf9442143f1c52393c/9-hydrogen-production-
J Hydrogen Energy 2019;44:30237e55. in-large-scale-henning-g.-langas-nel-hydrogen.pdf.
[12] Ayodele TR, Munda JL. Potential and economic viability of [24] taken from NEL Hydrogen brochure, Atmospheric Alkaline
green hydrogen production by water electrolysis using wind Electrolyser A Series (available at: https://nelhydrogen.com/
energy resources in South Africa. Int J Hydrogen Energy products/).
2019;44:17669e87. [25] Beeker E.
Y a-t-il une place pour l’hydroge ne dans la
[13] Nicita A, Maggio G, Andaloro APF, Squadrito G. Green transition e
nerge

tique ? Note d'analyse N 15 de France
hydrogen as feedstock: financial analysis of a photovoltaic- Strate
gie; August 2014. p. 4 (downloadable from, http://www.
powered electrolysis plant. Int J Hydrogen Energy strategie.gouv.fr/sites/strategie.gouv.fr/files/atoms/files/201-
2020;45:11395e408. 08-06na-fs-hydrogene-hd.pdf.
[14] reportIRENA report 2018 : Hydrogen from renewable power : [26] Philibert C. Renewable energy for industry : from green
technology Outlook for the energy transition (for download energy to green materials and fuels. IEA Insight Series; 2017.
from www.irena.org/publications). p. 28 (downloadable from, https://www.iea.org/publications/
[15] IEA report 2019. The future of hydrogen, prepared by the IEA insights/insightpublications/Renewable_Energy_for_
for the G20 summit. June 2019. Japan (available for download Industry.pdf.
at, https://www.iea.org/topics/hydrogen/. [27] Kopp M, Coleman D, Stiller C, Scheffer K, Aichinger J,
[16] Leroy RL, Bowen CT, Leroy DJ. The thermodynamics of Scheppat B, Mainz Energiepark. Technical and economic
aqueous water electrolysis. J Electrochem Soc analysis of the worldwide largest Power-to-Gas plant with
1980;127:1954e62. PEM electrolysis. Int J Hydrogen Energy 2019;42:13311e20.
[17] reportIEA/HIA Task 33 report : local hydrogen production for [28] Stoll RE, von Linde F. Hydrogen - what are the costs ?
energy applications (http://ieahydrogen.org/pdfs/2016-06-15- Hydrocarbon processing November 2000 (downloadable
IEA-HIA-Task-33-Ulleberg.aspx). from, https://www.caloric.com/wp-content/uploads/2016/
[18] Proost J. State-of-the art CAPEX data for water electrolysers, 11/Caloric-Downloads-Wasserstoff.pdf.
and their impact on renewable hydrogen price settings. Int J [29] Pletcher D, Li X. Prospects for alkaline zero gap water
Hydrogen Energy 2019;44:4406e13. electrolysers for hydrogen production. Int J Hydrogen Energy
[19] Stojic DL, Marceta MP, Sovilj SP, Miljanic SS. Hydrogen 2011;36:15089e104.
generation from water electrolysis e possibilities of energy [30] de Radigue s Q, Thunis G, Proost J. On the use of 3-D
saving. J Power Sources 2003;118:315e9. electrodes and pulsed voltage for the process intensification
[20] Schiller M, Anderson E. Five considerations for large-scale of alkaline water electrolysis. Int J Hydrogen Energy
hydrogen electrolyser development. Gas for Energy 2019;44:29432e40.
2014;1:44e7. [31] Schalenbach M, Kasian O, Mayrhofer KJ. An alkaline water
[21] Walsh FC, Ponce de Le
on C. Progress in electrochemical flow electrolyzer with nickel electrodes enables efficient high
reactors for laboratory and pilot scale processing. current density operation. Int J Hydrogen Energy
Electrochim Acta 2018;280:121e48. 2018;43:11932e8.
[22] Jupudi RS, Zhang H, Zappi G, Bourgeois R. Modeling bubble [32] Demo4Grid : Demonstration for grid services production of
flow and current density distribution in an alkaline green hydrogen greening of industry (https://
electrolysis cell. J Comput Multiph Flows 2009;1:341e8. www.demo4grid.eu/).