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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

Available online at www.sciencedirect.com

ScienceDirect

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

Techno-economic modelling of AEM electrolysis

systems to identify ideal current density and
aspects requiring further research

Laura J. Titheridge*, Aaron T. Marshall

Department of Chemical and Process Engineering, University of Canterbury, Christchurch, 8041, Canterbury, New
Zealand

highlights graphical abstract

This paper examines the impact of

current density on AEMWE per-
formance and stability.

The model estimates the optimal
current density with the lowest
LCOH.

The model highlights areas of
improvement with the greatest
potential to reduce LCOH.

The results highlight recommen-
dations for future directions in
AEMWE research.

article info abstract

Article history: Hydrogen produced by water electrolysis using renewable energy is a sustainable alter-
Received 26 June 2023 native to steam reformation. As a nascent commercial technology, performance, and
Received in revised form economic comparisons of anion exchange membrane water electrolyzers (AEMWE) to
30 July 2023 other electrolyzer technology benchmarks are not available. We present a techno-
Accepted 14 August 2023 economic model estimating AEMWE's baseline levelized cost of hydrogen (LCOH) at
Available online 28 August 2023 $5.79/kg, considering trade-offs between current density, efficiency, stability, capital, and
operating costs. The optimal current density is 1.38 A cm2, balancing stability and per-
Keywords: formance for the lowest LCOH. Using low-cost electricity and larger stack sizes, AEMWE
Levelized cost of hydrogen could achieve $2/kg low-carbon hydrogen. Technical improvements targeting system ef-
Anion exchange membrane water ficiency, particularly reducing overpotentials in hydrogen and oxygen evolution reactions,
electrolyzer could further reduce LCOH to $1.29/kg, approaching U.S. Department of Energy cost targets.
Green hydrogen There are hopes this model could raise the profile of AEMWE's economic potential to
produce green hydrogen and highlight its suitability for decarbonizing the energy sector.
© 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications
LLC. This is an open access article under the CC BY license (http://creativecommons.org/
licenses/by/4.0/).

* Corresponding author.
E-mail address: laura.titheridge@pg.canterbury.ac.nz (L.J. Titheridge).
https://doi.org/10.1016/j.ijhydene.2023.08.181
0360-3199/© 2023 The Author(s). Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY
license (http://creativecommons.org/licenses/by/4.0/).
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 519

metal-based catalysts [8]. Their suitability for coupling with

1. Introduction variable renewable energy sources and ability to use cheaper
materials is why research is actively addressing the chal-
In 2021, global hydrogen demand reached 94 Mt, with over 99% lenges hindering AEMWE commercialization.
of dedicated hydrogen production relying on fossil fuels, Several studies have investigated the cost of hydrogen pro-
resulting in 915 Mt of CO2 emissions, approximately 9.7 g of duced using PEMWE technologies. Franzmann et al. evaluated
CO2/g of hydrogen [1]. Projections indicate that if the current the cost for green hydrogen produced by PEMWE using wind and
rate of CO2 emission continue, the average global temperature solar, finding it could fall below V2.30/kg for solar-rich countries
will exceed the 1.5-degree Celsius global warming threshold [9]. Similar results were obtained for a PEMWE producing green
within the next decade [2], placing escalating pressure on the hydrogen in a combined hydroelectric-photovoltaic grid con-
energy sector to decarbonize. The production of low-carbon nected system, with hydrogen production costs at V2.25/kg [10].
hydrogen is essential for decarbonization as it is clean- Whilst techno-economic models have been applied for PEMWEs
burning, has a high-energy density, and is versatile [3], and [11e16], AWEs [13,17e19], and solid oxide electrolyzers [20,21],
will play a critical role in sectors where direct electrification is there is limited equivalent information available for AEMWE
challenging such as, steel, chemicals, long-haul shipping, systems. In particular, no literature exists that determines the
transport, and aviation [4]. From this perspective, hydrogen optimum current density and considers the clear trade-offs
must be low-carbon from the start of the transition, and between operating and capital cost.
eventually become green using electrolysis and renewable
electricity. The cost of green hydrogen production is
decreasing, owing to falling renewable prices, and once pro- 2. Critical current density
duced at competitive scale and cost, green hydrogen can be
used directly in fuel cells or converted into other energy car- Operating electrolyzers at higher current densities is benefi-
riers, such as ammonia, methanol, and synthetic fuels, which cial when prioritizing hydrogen productivity, as it allows for
are essential for dispatchable energy generation. the production of more hydrogen for a given amount of
Water electrolysis provides an effective method for green CAPEX. However, at elevated current densities, cell perfor-
hydrogen production using renewable electricity sources. The mance tends to decreases due to increased iR, overpotential,
flexibility of water electrolysis also allows it to be paired with and losses due to limited gas saturation and mass-transport
grid-connected or off-grid renewable electricity sources, above the critical current density [22]. As a result, the opera-
where dynamic variable loads can be utilized effectively, tional current density should fall below the critical current
providing peak-shaving benefits at times of high supply, and density, to ensure homogeneous gas distribution with uni-
energy storage solutions for future energy distribution in form current density values at catalyst sites, ultimately
times of low supply. The variety of grid services that elec- reducing the likelihood of hot spot formation and membrane
trolysis can supply are especially valuable in a world of dehydration. The presence of gas bubbles blocking areas of a
growing intermittent renewable energy generation. porous structure inhibits the triple phase boundary formation,
Proton exchange membrane (PEMWE) and alkaline water limiting the optimal use of the surface for electrochemical
electrolyzers (AWE) are the main commercially available reactions [23]. This effect becomes particularly significant for
technologies for water electrolysis. Among these, AWE is electrodes with high surface areas, as bubble detachment can
more established and has provided durable water electrolysis be influenced by the surface morphology, ultimately affecting
systems for decades [5]. Whilst they offer the advantage of the current density distribution [23]. Balancing hydrogen
low-cost components, the large internal resistance across the productivity and cell performance is thus largely influenced
porous diaphragm and electrolyte limits the current density by the current density, making it essential to establish re-
and efficiency. As a result, AWE require large stack designs to lationships between the current density, stability, energy ef-
accommodate the low current density, thus increasing capital ficiency, and the economics of AEM water electrolyzers.
expenditure (CAPEX) of these systems at given hydrogen
production rates [6], as the overall material requirement is
larger. Moreover, the AWE porous separator requires cautious
pressure balancing, posing challenges when operating in 3. Model objective
transient regimes, especially paired with intermittent
renewable generation systems like solar and wind [7]. On the The U.S. Department of Energy (DoE) has set ambitious
other hand, PEMWE allows rapid cycling in flexible operating hydrogen cost targets of < $2 kg1 for end uses and trans-
conditions, making it more suitable to pair with intermittent portation, and < $1 kg1 for industrial and bulk power gener-
renewables to provide grid-balancing services. However, the ation uses [24]. Achieving the second target would make water
use of platinum-group metals (PGM) for catalysts and flow electrolysis competitive with hydrogen produced via steam
components increases the CAPEX significantly. reformation [25]. Despite its promise in reducing system
The relatively new AEMWE systems, can combine and CAPEX, the AEMWE technology is currently available only at
improve upon the advantages offered by both AWE and small scale and is still in its early stages of development.
PEMWE systems. Alongside the ability to operate in a transient Presently, the current density in available AEMWE systems is
regime (due to the non-porous anion-exchange membrane limited to 0.2e2.0 A cm2 [4,26], with a 2050 target of
separator), the alkaline environment permits a large variety of 2.0 A cm2 [4]. Table S1 presents other key performance in-
catalyst material selection, allowing the use of non-PGM dicators for AEMWE systems.
520 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

As efficiency, stability, and cost depend on current den- 5.1.1. Reversible potential
sity, a multi-dimensional analysis is necessary to optimize The reversible thermodynamic cell voltage, ERev, is found
AEMWE system designs and operation. The levilized cost of using Eq. (2) [31], which considers the effect of temperature, T,
hydrogen (LCOH) metric provides a fair comparison that normalized partial pressures of the products, PH2 and PO2 , and
factors in all technical and economic parameters such as, the activity of water, aH2 O :
capital costs, operating costs (water, electricity, and main- pffiffiffiffiffiffiffi !
tenance), production efficiency, stack lifetime, and perfor- RT PH2 PO2
Erev ¼ Eorev;T þ ln (2)
mance degradation, allowing evaluation of the overall cost 2F aH2 O
per kilogram of hydrogen and helping to explore important
When electrolysis is performed in KOH electrolytes (which
trade-offs.
are often used in the anode and cathode chambers of
The primary objective of this model is to identify the
AEMWE), the activities of H2, O2 and H2O are influenced by the
optimal current density where the compromise between sta-
properties of the KOH electrolyte [32]:
bility and performance leads to the lowest LCOH. Additionally,
it aims to highlight the areas of improvement that offer the pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !
RT ðP  Pv;KOH Þ ðP  Pv;KOH Þ
greatest potential to reduce the LCOH, to provide recom- Erev ¼ Eorev;T þ ln (3)
2F aH2 O;KOH
mendations for future directions in AEMWE research.
where the ideal gas constant, R ¼ 8.314 J mol1 K1, Faraday's
constant, F ¼ 96,485C mol1, T is temperature (in K), P is the
4. Model approach operating pressure (in bar), Pv,KOH is the water vapour pres-
sure, and aH2 O;KOH is water activity in the KOH electrolyte. E0Rev;T
The cell specification and operating parameters are provided is the reversible voltage at standard pressure (1 bar), and is a
in Table S2, and the model follows several key assumptions. function of temperature [32,33]:

1. The complete AEMWE stack operates at constant pressure Eorev;T ¼ 1:5184  1:5421  103 ðTÞ þ 9:523  105 ðTÞlnðTÞ þ 9:84
and temperature.  108 ðTÞ2
2. Both the AEMWE and plant components operate at a con- (4)
stant capacity and electricity supply.
3. The cell operates at steady-state, ensuring that faradic ef- The vapour pressure and the activity of water in KOH can
ficiency and hydrogen production remain uniform during be found by Refs. [32,33]:
its lifetime. It is not sensitive to KOH concentration, and
Pv;KOH ¼ eð2:302aþblnðPv;H2 O ÞÞ (5)
operation is stable where no reactant or product crossover
exists.
4. The AEMWE is sized as a 1 MW electrolyzer. This a ¼ 0:0151m  1:6788  103 m2 þ 2:2588  105 m3 (6)
assumption implies that the physical size and hydrogen
generation from the stack change as the current density is b ¼ 1  1:2062  103 m þ 5:6024  104 m2  7:8228  106 m3
varied. (7)

 
7699:68
Pv;H2 O ¼ exp 81:6179   10:9lnðTÞ þ 9:5891  103 ðTÞ
T
5. AEM electrolyzer and operation (8)

5.1. Cell potential 3:177m  2:131m2

aH2 O;KOH ¼ 0:05192m þ 0:003302m2 þ (9)
T
For water electrolysis to proceed, it must overcome a variety of where m is the molar concentration of KOH (in mol L1).
barriers, which necessitates a sufficient electrical energy
supply. These barriers consist of physical and electrochemical 5.1.2. Ohmic potential
resistances. Physical resistances arise from bubble coverage The voltage drop due to ohmic resistance can be found using
on the gas-evolving electrodes, as well as ionic transfer re- [34]:
sistances in the electrolyte and membrane, which affect
 
transport phenomena [27] and result in heat generation in Lmem
Eohmic ¼ I  Rohmic ¼ I  Rcontact þ (10)
accordance with Joule's law [28]. Consequently, these factors Amem smem
lead to energy losses [29], thereby reducing system efficiency. where Rcontact is the cell resistance excluding the membrane,
On the other hand, electrochemical resistances result from and smem is the membrane ionic conductivity (in S m1) for a
charge transfer kinetics between the electrodes and the elec- solid alkaline polymer electrolyte with trimethylbenzyl qua-
trolyte. The overall cell potential is determined by the com- ternary ammonium functional groups. Which is a function of
ponents of the cell and the manner in which power is temperature, given as [33,35]:
exchanged between them [30]:
 
11190
Ecell ¼ Erev þ Eohmic þ hOER þ hHER þ Ebubble (1) RT
sM ¼ 198:3e (11)
 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 521

In a AEMWE, the potential increases over time due to catalyst 5.1.3. Overpotential
dissolution [36], corrosion [36], and chemical degradation of Activation overpotentials are due to charge transfer kinetics
the membrane [36e39]. Membrane stability is significantly between electrodes and the electrolyte, they can be expressed
affected by nucleophilic attack on cationic charge sites by using the Butler-Volmer equation, simplified to Ref. [30]:
hydroxide ions [36e39], leading to a loss of anion-exchange  
functional groups and a decrease in membrane conductiv- jrxn ¼ j0 eahF=RT  eð1aÞhF=RT (12)
ity. Consequently, this causes an increase in the cell voltage where j is current density, and j0 is current exchange density.
due to the ohmic drop and reduces the overall electrical ef- The charge transfer coefficient, a, for the anode and cathode
ficiency. The increased ohmic drop becomes a significant are both assumed to be equal to 0.5 [33]. From this, the acti-
issue when operating at high current densities. To account vation overpotentials for the oxygen evolution reaction (OER)
for this membrane degradation and the resulting loss in and hydrogen evolution reaction (HER) can be found using the
membrane conductivity (see Supplementary Information), Tafel equation [30]:
changes in conductivity were estimated from previous work
!
OER
[40]. This study found that the degradation was dependent RT jrxn
hOER ¼ ln (13)
on the local hydration number, which is related to bubble aOER F OER
j0
coverage and thus current density. The relationship between
current density and fractional bubble coverage highlights the !
HER
RT jrxn
impact of supplying the electrolyte at elevated flowrates, hHER ¼ ln (14)
aHER F HER
j0
especially in cells operating above 2 A cm2 (Fig. 1a). The
calculated degradation rate follows a worst-case scenario, OER HER
where j0 and j0 are half reaction current exchange den-
assuming that there is always degradation at sites with sities in 1 mol L1 KOH.
bubble coverage. The calculated rate at which Eohmic in-
creases over the lifetime of the electrolyzer is found to be 5.1.4. Potential losses from bubble evolution
strongly dependent on current density and aligns with At larger current densities, bubble evolution results in greater
literature reports (Fig. 1b). mass transport losses. Using a superficial current density, js

Fig. 1 e Variation between a stagnant and flowing electrolyte in an electrolyser operating with the conditions listed in Table
S2, and the impact on the a) fractional bubble coverage as calculated using Eqs. (15) and (16), and b) voltage decay rate as
calculated by Eq. 3 in the supplementary information. Where diamond markers indicate the performance of AEMWE
systems in literature [41e45].
522 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

(with units in A m2) an empirical relationship is imple- Separating the polarisation curve into activation and
mented to find the fractional bubble coverage of electrodes in ohmic potentials (Fig. 2a), provides further insight into which
a stagnant electrolyte [46]: factors dominate the overall performance. For current den-
sities below 3.6 A cm2, the OER overpotential is the domi-
q0 ¼ 0:023ðjs Þ0:3 (15) nating loss. However, above 3.6 A cm2, the impact of bubble
At larger current densities, a higher reactant flowrate is blocking becomes more significant as gas bubbles substan-
necessary to match the rate of reactant consumption [22] and tially affect the interelectrode resistance [52,53], inhibiting
facilitate fast transport, thus reducing limitations from mass- triple phase boundary formation at certain electrode areas,
transport and other failures [23,47,48]. The impact of fluid and so prevent effective surface use for electrochemical re-
velocity on fractional bubble coverage is given by Ref. [49]: actions [23]. According to Eq. (18), the total losses experienced
from the effect of bubble nucleation are due to kinetic losses at
   2
1 the electrolyte-catalyst interface from coverage of active OER
q ¼ q0 1 þ  1 v2k Þ (16)
q0 and HER reaction sites, and mass transfer losses at the
catalyst-membrane interface, with the latter having the
Here vk represents the mean cross-sectional velocity of forced
greatest impact. This suggests that removing evolved gas
flow to the electrode (in m s1). As fractions of the effective
bubbles from the catalyst-membrane interface may be a crit-
catalyst surface area (ECSA) are inhibited by bubbles adhering
ical factor for high current density operation in AEMWEs, in
to the electrode surface, the insulated fraction is inactive and
agreement with a similar model and findings by Liu et al. [46].
increases the real current density [50]. The real current den-
To validate the cell voltage and efficiency model, data has
sity can be distinguished from the superficial current density,
been taken from literature and overlaid with our model pre-
denoted as js [51]:
dictions (data points in Fig. 2). While the model predicts higher
js ¼ jð1  qÞ (17) cell voltages, and lower efficiencies compared to these litera-
ture reports, the cell behaviour vs. current density is similar to
The superficial partial current for OER and HER are calcu-
other AEMWE models [34,46,54e56]. Furthermore, the aim of
lated similarly. The total voltage losses from bubble nucle-
the model is not to precisely predict cell efficiency, but rather
ation can be found as the difference between the
investigate the impact of current density on LCOH and iden-
overpotentials and ohmic resistance at the superficial and real
tify which factors can be optimized to meet target costs for
current density:
AEMWE to be competitive with steam reforming.
OER
! HER
!
RT js;rxn RT js;rxn
Ebubble ¼ ln þ ln þ ARohmic ðjs  jÞ (18)
aOER F OER
jrxn aHER F HER
jrxn
6. AEM water electrolyzer economics
Finally, Eqs. (3), (10), (13), (14) and (18) can be substituted
into Eq. (1) to find cell potential at various current densities. The capital expenditure (CAPEX) and operational expenditure
(OPEX) costs are found in USD throughout the LCOH
5.1.5. Efficiency calculations.
The voltage efficiency of the AEMWE is dependent on current
density, which is found by Ref. [31]: 6.1. Operational capacity and lifetime

1:48
hV ¼ (19) The operational capacity is the ratio between the actual and
Ecell
maximum electrolyzer capacity, and it directly affects the size
where 1.48 is the thermoneutral cell voltage (V) required to of the electrolyzer system [26]. High operational capacities can
split liquid water without heat supplied from the environment be achieved with a grid-connected diversified renewable
[31]. The activation losses at the anode and cathode increase electricity supply [57]. However, for off-grid systems, electro-
with current density, and ohmic losses increase proportion- lyzer utilization is usually lower due to the daily and seasonal
ally as current density increases, resulting in lower effi- variability of renewable energy supply [58]. In addition to
ciencies. Whilst the rate of hydrogen evolution is clearly decreased operational capacity, coupling electrolyzers to
greater at higher current densities, the trade-off between highly variable renewable power sources require the elec-
production rate and efficiency must be considered. By trolysis systems to cope with flexible and dynamic operation.
employing the model described above, along with suitable Generally, PEM systems can tolerate loads below 10% of ca-
parameters taken from literature (Table S2), the polarisation pacity, whereas a conventional alkaline electrolyzer requires
curve and efficiency for an AEMWE, including the contribution almost continuous and stable power supply to prevent cross-
of the various losses, can be predicted (Fig. 2). Although our diffusion of gaseous products across the diaphragm under
model seems to underpredict AEMWE performance compared small loads [59]. While there are minimal accounts of how
to literature, the purpose of this model is not to perfectly AEM cells operate under varying power supply input, it is ex-
predict cell performance, but rather to enable the investiga- pected that AEM cells will perform in a similar manner to PEM
tion of the impact of various operating variables on the LCOH systems, as they utilize a gas-tight membrane.
using a general yet technically realistic model. The parame- To reduce inaccuracies from assuming that the electro-
ters used in this model can easily be optimized with adequate lyzer will survive a fixed period, the cost of hydrogen is
experimental data, allowing others to perform this analysis expressed over the operational lifetime in hours. The opera-
for specific cell stacks. tional lifetime used in this model is an estimate based on
 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 523

Fig. 2 e Simulated behaviour for a AEM water electrolyzer cell at 60  C using 1 M KOH as an electrolyte, operating at various
current densities, a) polarisation curve, b) efficiency losses. Where diamond markers indicate the performance of AEMWE
systems in literature [41e45]. The dashed line represents the predicted cell performance at the end of life.

current commercial AEMWE operational lifetimes, such as the BOP cost for a 1 MW system is assumed to be 2.5x that of
Enapter's AEMWE at 35,000 h [60]. Assuming a 90% allowable the stack costs. It is acknowledged that the BOP costing is
decrease to efficiency over time, a 20% capacity factor was difficult to estimate, given that individual manufacturers keep
selected as the base scenario to allow for a 20-year fixed period this information confidential. Therefore, bottom-up sizing
to evaluate the AEM performance and estimate the levelized and costing should be studied further to improve economic
cost of hydrogen, without requiring a stack replacement. estimates for AEMWE systems.
Since this techno-economic model is based on a 1 MW
6.2. Capital investment electrolyzer, choosing to operate the electrolyzer at low current
density (to minimize operating cost and reduce membrane
While capital costs for AEM water electrolyzers are influential degradation) would require a larger electrolyzer stack compared
[61,62], AEM systems benefit from utilizing cheaper earth- to a 1 MW electrolyzer operating at higher current density. The
abundant materials compared to PEM systems [63], which stack CAPEX will be larger for stacks operating at lower current
should reduce the CAPEX requirements associated with direct densities as the electrode and membrane areas will need to be
material costs. However, at the moment, the low economies of larger to ensure the same power across the stack.
scale and technology readiness of AEM electrolyzers are likely The CAPEX per unit area of the AEM electrolyzer stacks was
to result in higher CAPEX than otherwise expected. Addi- estimated using a fitted relationship (Eq. (20)) derived from
tionally, limited open literature regarding the cost of goods for commercial and literature electrolyzer size and cost data
AEM technology means that estimates for stack and BOP costs (Fig. S1):
carry a significant degree of uncertainty. Without more spe-
cific information. In the absence of more specific information, 107:6887
CAPEXstack ¼ (20)
the model assumes a linear relationship between stack size P0:8976
stack

and associated specific system costs. where CAPEXstack has units $ m2. The total cost of the stack
The core components for the BOP include power supply can be found using the required area for the electrolyzer at
and control, process components, water treatment and cool- each current density (Eq. (21)).
ing systems, and hydrogen processing and compression and
they can be benchmarked against AWE and PEMWE systems
SCE
to account for the actual cost of the system setup [6]. As such, Aelectrode ¼ (21)
j
524 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

Pstack jAelectrode
SCE ¼ (22) nH2 ¼ hF (23)
Ecell 2F
where SCE is the single cell equivalent current (in units A). 6.7. Levelized cost of hydrogen

6.3. Maintenance
The levelized cost of hydrogen (calculated in $ kg1) is used as
the focal economic metric in this analysis, the equation re-
Data surrounding the cost of maintenance and operation of
quires the total annual investment and operational costs, and
AEMWE is has not been widely reported, and the cost is
the annual hydrogen production to calculate the cost per kg.
assumed as a percentage of the system capital cost, taken as 5%,
The estimate does not include product storage or trans-
similar to PEM technologies as reported by CSIRO Energy [58].
portation costs, taxes, or the inflation rate, and is calculated as
[74]:
6.4. Electricity
Py
Sn ð1  rÞn
LCOH ¼ Pn¼0
y n (24)
n¼0 Pn ð1  rÞ
The cost of hydrogen is directly sensitive to variability in the
electricity source, regardless of whether it is from a grid-
The system cost in year n(Sn) and the total hydrogen pro-
connected source or off-grid sources or a renewable or non-
duction in year n(Pn) are used to calculate the Levelized Cost of
renewable source [64e70]. The availability and price of elec-
Hydrogen (LCOH) over the 20-year economic lifetime (y)
tricity are important to techno-economic analysis predictions,
[57,58,74]. The LCOH calculation takes into account the time-
as variable renewable sources impact the operation of green
value of money, as costs and hydrogen production are dis-
hydrogen systems and consequently the volatility of
counted based on the investment's cost of capital, which is
hydrogen costs [69]. It should also be noted that fluctuations in
assumed to be a 6% discount rate [8]. This discount rate re-
operation imposed by electricity availability can reduce the
flects the idea that costs incurred or the value of hydrogen in
electrolyzer lifetime [70]. To simplify the analysis, the model
the future will have less importance on the viability of the
assumes a fixed electricity supply to focus on the impact of
investment decision made in the present. The value of LCOH
operational current density, while avoiding expensive elec-
reveals the average price that hydrogen must be sold at in
tricity prices during times of peak demand.
order for the process to break-even financially [75]. A sum-
Assuming the AEM system is operated at 20% capacity, the
mary of the economic analysis parameters and relationships
active operational hours are determined as 20% of 8760 h a
can be found in the Supplementary Information.
year (1752 h). The power requirements each year can then be
used to determine the annual electricity cost. The electricity
price is reported to be between $0.053e0.060 kWh1 [4,71]. As
7. Economic model results
this model assumes a dedicated, fixed, embedded renewable
energy source, the electricity price is assumed to be the higher
In this evaluation, the AEMWE performance relationships
value of $0.06 kWh1.
over a 20-year fixed period were utilized to estimate the LCOH
(Fig. 3). Table S3 presents the key input and operational pa-
6.5. Water
rameters used in the economic analysis model. Using the
evaluated CAPEX and operating costs, the optimal current
The consumption of water in an electrolyzer can be easily
density was found to be 1.38 A cm2, resulting in a LCOH of
calculated from the current or hydrogen production rate.
$5.79 per kg of hydrogen. Noteably, the LCOH is within a
Assuming no losses, the water consumption would corre-
similar range of magnitude as the estimate by Ionomr, who
spond to 9 kg of water per kg of hydrogen. While the gas
found the baseline LCOH from AEMWE to be $3.69 per kg of
produced will normally contain water vapour (and it some
hydrogen [6].
cases entrained water droplets), the majority of this additional
While decreasing electrolyzer current densities is generally
water loss from the stack can be recycled. Therefore, in this
expected to increase the efficiency of the stack due to lower
model it is assumed that the water consumed by the stack is
overpotentials, ohmic and bubble related losses, this model
only that required by the reaction stoichiometry (i.e. a total
reveals that at current densities below 1.38 A cm2, the LCOH
water to hydrogen mass ratio of 9:1). However, as AEM elec-
actually increases with decreasing current density. This in-
trolyzers require water with sufficient purity, some additional
crease is caused by the CAPEX associated with the stack and
water is lost during the purification process. For reverse
BOP increasing (per kg of hydrogen produced) faster than the
osmosis, a typical single pass conversion is between 50 and
corresponding decrease in OPEX due to lower current densities.
85% [72]. Using the least conservative RO conversion of 50%
Essentially at lower current densities, while the electricity ef-
suggests the total water to hydrogen mass ratio is 18:1, which
ficiency is better, the stack must be much larger and thus more
is in line with literature reports with give 18:1e24:1, depending
expensive to consume the 1 MW specified in this model.
on the water source [4,73].
Above 1.38 A cm2, the LCOH is seen to increase with
increasing current density as above this current density, the
6.6. Hydrogen production electricity cost becomes the dominating factor of LCOH, due to
declining stack efficiency. It should be noted that an advan-
The molar flow rate of hydrogen produced is directly calcu- tage with operating at lower current densities is that cell
lated from the total charge: lifetime should be increased as hot-spot formation,
 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 525

Fig. 3 e LCOH variation as a function of current density, including cost contributions towards LCOH. The red diamond
indicates the optimal current density for the lowest LCOH. (For interpretation of the references to colour in this figure legend,
the reader is referred to the Web version of this article.)

membrane dehydration, and resultant membrane degrada- economies of scale, as the capacity of the process reduces the
tion is less likely [23]. However, at high current densities, material and energy demand per unit of production. However,
membrane degradation increases causing the efficiency to fall a larger stack requires a larger electricity supply. To achieve
over time (and thus increasing operating costs) and conse- optimal cost-effective performance in larger stacks, a lower
quently increasing the LCOH. current density is favored, as losses from overpotentials are
lower. Achieving larger stack sizes requires for any technical
limits that may restrict the achievable cell size or current
8. Strategies to lower the LCOH for AEMWE density to be established, such as, mass transport limitations,
ohmic losses, water and heat management, mechanical
8.1. Economic strategies integrity, and manufacturing limitations.
It should be noted that other infrastructure elements, such
As expected, the model demonstrates that the optimum cur- as storage and distribution, also have potentially significant
rent density decreases when stack and BOP capital costs are impacts on the LCOH, but they have not been examined in the
reduced (Fig. 4a and b). In the scenario where AEMWE capital scope of this analysis and would benefit from further
is reduced by 50%, the critical current density falls by only research. The potential infrastructure needed for widespread
12%, whereas it falls by 18% when the same reduction is green hydrogen adoption should consider whether once the
applied to BOP costs. This is due to the capital cost of the stack hydrogen is produced if it is stored, transported, or potentially
and BOP becoming more cost effective at lower capacities and converted into other organic liquid energy storage forms,
not influenced by the system's performance. unless it would be consumed on site. The future cost
The drivers for capital cost reduction are impacted by a competitiveness of green hydrogen heavily depends on the
quickly maturing and growing market, characterised by interplay between the end-use and the proximity of produc-
increasing plant sizes, manufacturing scale-up, design im- tion, which in turn, exposes the storage and transportation
provements, and moving production to cheaper global re- associated with the given application. Transport and distri-
gions. The size of electrolyzer plants has noticeably increased bution considerations should include an economic analysis on
over the past decade, as manufacturing supply-chains the use of new and pre-existing pipeline infrastructure, and
become more mature. Between 2010 and 2017, AWE systems non-pipeline transportation (e.g., shipping and trucking). As
increase from 120 kW to 2 MW, and PEMWE systems increased the trade-off between capital costs, flexibility, efficiency, and
from 10 kW to 2.9 MW on average [1]. These projects are transport distance would provide insight into which infra-
anticipated to increase by three orders of magnitude over the structure is more suitable for local and distributed hydrogen
following decade [1,76]. production.
For AEMWE, the most significant opportunity for cost
reduction lies in increasing the size of electrolyzer stacks 8.2. Performance strategies
(Fig. 4c).This can be achieved by either increasing the number
of cells per stack or increasing the areas of cells. A larger active The following scenarios highlight potential benefits, both
area would allow for greater volumes of hydrogen production, economic and performance-wise, that can result from tar-
leading to reduced unit costs for materials, components, and geted research and development strategies for AEMWE. By
assembly, making the overall system more cost-efficient via decreasing ohmic losses in the system, a lower LCOH and
526 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

Fig. 4 e Theoretical reduction in LCOH at various current densities when a) the stack cost per m2 is reduced, b) the balance of
plant cost is reduced, c) the area size of the stack increases. Additional markers highlight the optimal current density and
LCOH.

higher critical current density can be achieved (Fig. 5a). The research adopts Sustainion membranes to achieve high
relationship is particularly noticeable at high current densities product selectivity and low cell potential at high current
as ohmic losses follow a linear progression with current densities [82e88].
density. One strategy to reduce ohmic losses is to reduce However, it is important to note that while some polymers
membrane thickness, thus lowering the ion transport resis- exhibit adequately high hydroxide ion conductivity, the sta-
tance [34], however the membrane serves as a separator be- bility and durability of anion-exchange membranes have not
tween the two electrodes, and thinner membranes can result yet met the threshold required for commercialization. Devel-
in inferior mechanical integrity under high-pressure opera- oping new polymers with stable functional groups and high
tion, and greater chemical instability, which will shorten the ion-exchange capacities could enable AEMWEs to reduce
lifetime and reduce the conductivity of the membrane stack size by 11%e68%, while maintaining hydrogen produc-
respectively [77,78]. There are various commercial mem- tivity and reducing the LCOH by 3%e14%. This cost reduction
branes available currently such as, Sustainion 37e50, Fuma- would be partly due to the system requiring smaller electro-
sep FAA-3, Tokuyama A201, and Aemion membranes [79]. Liu lyzers, but primarily from significantly lower electricity
et al. [80] and Pushkareva et al. [81] found Sustainion 37e50 to demand.
have the lowest resistance during electrochemical impedance Reducing the electrolyzer losses associated with the HER
studies, and both found Sustainion 37e50 to attain the highest and OER overpotentials could significantly improve the
performance in 60  C and 1 M KOH obtaining current densities AEMWE energy efficiency (Fig. 5b). In AEMWEs operating
of 3.5 A cm2 and 3.0 A cm2 respectively. Currently, much below 1.38 A cm2, the OER overpotential is the dominating

Fig. 5 e Theoretical reduction in LCOH at various current densities when a) the conductivity increases, b) the overpotentials
for HER and OER are reduced, c) the voltage decay rate is reduced and lifetime is proportionally improved. Additional
markers highlight the optimal current density and LCOH.
 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 527

loss. As current density increases, the kinetic overpotential the chemical stability of the catalyst and membrane, and the
owing to slow reaction kinetics at the electrode surface corrosion stability from the impact of bubble detachment. To
gradually increases (see Fig. 2a). As most AEMWE systems enhance mechanical stability, water management inside the
operate between 0.2 and 2.0 A cm2 [4,26], and the target for membrane is crucial, as higher water uptake can flood the ion-
2050 is 2.0 A cm2 [4], prioritizing the development of highly conducting channel, leading to membrane softening and
active OER electrocatalysts should take precedent if the target reduced ionic conductivity [96e98]. Methods to increase
key performance indicator is purely cell performance. How- chemical stability can be developed from an understanding of
ever, this model does not incorporate cost or stability analysis the polymer architecture and common degradation pathways
on various catalyst species or loading parameters, so the for AEMs including, polymer backbone degradation [99], SN2
compromise between performance and catalyst stability on substitution of cationic side-chains [100], and the Hoffman
the LCOH is not accounted for. The voltage loss from OER and elimination [101]. The corrosion effect by bubble nucleation
HER overpotentials contributes to increased electricity con- depends on the electrode geometry and surface morphology
sumption and LCOH, but has little impact on current density. [23]. Strategies for controlling bubble release include the use of
At present, the most widely used electrocatalysts in coatings [102], and geometry and pore size optimization for
AEMWEs are transition metals such as, cobalt, nickel, or iron the PTL and CL [103,104].
[89]. Most studies also focus on OER catalyst optimization, as The impact of temperature and pressure on LCOH has not
the OER is the limiting step in the process due to its lower been explored in the scope of this work. An investigation to
reaction kinetics and higher overpotential to HER [90,91]. For verify the complex relationship between temperature, pres-
instance, Xu et al. [92] systematically tested first-row transi- sure, performance, and stability would be beneficial, as the
tion metal (oxy)hydroxide powders for OER, with NiCoOx:Fe trade-off that exists between degradation and diffusion,
performing best, reaching 0.9 A cm2 at 2.4 V. They found the relating to water content [105], would highlight optimal
high performance for NieCo oxide catalysts owes to their AEMWE operating conditions. Whilst higher temperatures
higher electrical conductivity, compared to other anodes. incite higher degradation rates of the ionomeric materials
Additionally, AEMWEs benefit from their ability to utilize [106e109], it simultaneously improves water diffusivity,
non-PGM catalysts, concatenating the cost saving potential facilitating water transport from the cathode to the anode
associated with OER and HER catalysts. In PEMWEs, the cur- [105]. The operating pressure should be evaluated similarly.
rent anode PGM loading is 2e5 mg cm2, at the lowest limit, a An increased pressure offers reduced anode and cathode
100 MW PEMWE operating at 4 W cm2 would require 50 kg of activation, and ohmic overpotential, and reduced gas bubble
iridium [8]. At iridium's current price of $147 per gram [93], the size [110], reducing degradation rates, as it is dependent on
OER catalyst alone would cost $7.35 million, not to mention, local hydration [40]. However, increasing operating pressure
the scarcity of iridium in Earth's crust [94]. Minke et al. [95] also additionally increases the partial pressures which, in turn,
cautioned that if iridium loading in PEMWEs is not reduced increases the open circuit potential [111] (see Eq. (2)).
and not fully recycled, a supply bottleneck is to be expected as
installation grows by 2 GW each year.
Finally, lowering AEM degradation rates is especially 9. Roadmap to drive the LCOH toward $1
effective at high current densities (Fig. 5c). Reducing the
voltage degradation rates by 50% would double the lifetime of There are many potential avenues to reduce the LCOH to
the electrolyzer and reduce electricity demand over the elec- achieve the DoE target for hydrogen production cost using
trolyzer lifetime as efficiency decreases more gradually. AEMWE, whilst no distinct path for AWE or PEMWE to fall to
The durability of AEMWE systems is dependent on several $1.00 per kg can be seen in techno-economic analysis models
factors, including the mechanical stability of the membranes, in literature [6,26,58,112e114].

Fig. 6 e Waterfall chart highlighting the opportunities to drive AEMWE costs toward a LCOH in line with DoE targets. PEMWE
and AWE costs are taken from Ionomr [6].
528 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

The LCOH is most sensitive to the price of electricity (Fig. 6).

Decreasing electricity prices from $0.06 to $0.03 per kWh can Acknowledgements
reduce the LCOH by 50%, which is a realistic and achievable
goal given the cost of intermittent renewable energy is falling This work was supported by the New Zealand Ministry of
below $0.02 per kWh [4,6], it also allows for greater current Business, Innovation and Employment via Catalyst Strategic
densities to be utilized as the profit from hydrogen outweighs Fund funding (UOCX2118). L.J.T. thanks the MacDiarmid
the cost of electricity when efficiency is low. Reducing the Institute for Advanced Materials and Nanotechnology and the
stack and balance of plant costs by half would lower the LCOH Department of Process and Chemical Engineering for her
by 10%, this could be achieved by leveraging advanced non- Doctoral scholarship award, and Prof. Aaron Marshall for
PGM catalysts to replace expensive iridium-based anode cat- providing valuable feedback.
alysts, and increasing production scale to lower investment
cost through enhanced manufacturing processes and by
establishing a secure supply chain [6]. Furthermore, scaling up Appendix A. Supplementary data
system size to allow peripheral components to optimize ma-
terial utilization provides a further 30% cost reduction. Supplementary data to this article can be found online at
Energy consumption is directly proportional to cell voltage https://doi.org/10.1016/j.ijhydene.2023.08.181.
(Fig. 2a). The required potential is a summation of ohmic
losses, anode and cathode thermodynamic equilibrium po-
tentials, their kinetic overpotentials, and the overpotential references
associated with mass transport losses from bubble coverage.
Reducing overpotentials offers the greatest opportunity to
lower LCOH (Fig. 6), however it has the smallest improvement [1] IEA. Hydrogen projects database. 2022. url: https://www.iea.
org/account/licence/products.
to critical current density by 0.17 A cm2. In contrast,
[2] Hydrogen Council. Path to Hydrogen Competitiveness: a
increasing conductivity by 50% has the greatest increase in Cost Perspective. Tech. Rep. Publisher: Hydrogen
critical current density by 0.35 A cm2, but it has the smallest Knowledge Centre; Jan. 2020. url: https://www.
impact on the LCOH. h2knowledgecentre.com/content/policypaper1202.
[3] Katebah Mary, et al. “Hydrogen production using piston
reactor technology: process design and integration for CO2
10. Conclusion emission reduction”. Energy Nov. 2022;259:124999. https://
doi.org/10.1016/j.energy.2022.124999. issn: 0360-5442.
[4] IRENA. Green hydrogen cost reduction: scaling up
The techno-economic model conducted in this study evalu- electrolysers to meet the 1.5C climate goal. En Tech Rep
ated the feasibility of hydrogen production using AEMWE. 2020:106. url: https://www.irena.org/-/media/Files/IRENA/
Based on the initial electrolyzer operating conditions and Agency/Publication/2020/Dec/IRENA_Green_hydrogen_
performance, a baseline LCOH of $5.79/kg was found, which is cost_2020.pdf.
within a similar order of magnitude as the estimate by Ionomr [5] Bowen CT, et al. “Developments in advanced alkaline water
electrolysis”. Int J Hydrogen Energy Jan. 1984;9(1):59e66.
($3.69/kg) [6]. Further investigation into potential future
https://doi.org/10.1016/0360-3199(84)90032-6. issn: 0360-
AEMWE improvement scenarios show that the estimated cost
3199.
reductions are possible, and there is potential to meet the U.S. [6] IONOMR. Hydrogen production cost by AEM white paper.
DoE end uses and transportation target cost ($2.00/kg). With White Paper 2020. url: https://ionomr.com/wp-content/
additional improvements, the estimate for the LCOH could uploads/2021/02/FM-7024-A-Hydrogen-Production-Cost-by-
further decrease to $1.29/kg, making AEMWEs competitive for AEM-White-Paper-copy.pdf.
industrial and bulk power generation applications, with the [7] Schillings Jonathan, Olivier Doche, Deseure Jonathan.
“Modeling of electrochemically generated bubbly flow
target cost of $1.00/kg [24]. The results of the analysis high-
under buoyancy-driven and forced convection”. en. Int J
light the importance of focusing research and development
Heat Mass Tran June 2015;85:292e9. https://doi.org/10.1016/
efforts on reducing the overpotentials associated with the j.ijheatmasstransfer.2015.01.121. issn: 0017-9310.
oxygen and hydrogen evolution reactions. Overall, these re- [8] Chatenet Marian, et al. “Water electrolysis: from textbook
sults support that AEMWE is a developing technology with knowledge to the latest scientific strategies and industrial
promise to provide hydrogen competitively against steam developments”. Chem Soc Rev 2022;51.11:4583e762. https://
reformation, PEMWE, and AWE in the near future. However, doi.org/10.1039/D0CS01079K. Publisher: Royal Society of
Chemistry.
substantial efforts are needed to overcome current
[9] Franzmann D, et al. “Green hydrogen cost-potentials for
challenges. global trade”. issn: 360-3199. In: International journal of
hydrogen energy; May 2023. https://doi.org/10.1016/
j.ijhydene.2023.05.012.
Declaration of competing interest [10] Zghaibeh Manaf, et al. “Optimization of green hydrogen
production in hydroelectric-photovoltaic grid connected
power station”. issn: 360-3199. In: International journal of
The authors declare the following financial interests/personal
hydrogen energy; June 2023. https://doi.org/10.1016/
relationships which may be considered as potential
j.ijhydene.2023.06.020.
competing interests. [11] Sharma Rishabh, et al. “Optimizing green hydrogen
Laura Titheridge reports financial support was provided by production: leveraging load profile simulation and
Ministry of Business Innovation and Employment. renewable energy integration”. issn: 360-3199. 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 529

International journal of hydrogen energy; Apr. 2023. https:// [25] IEA. The future of hydrogen. En Tech Rep June 2022:203. url:
doi.org/10.1016/j.ijhydene.2023.03.179. https://iea.blob.core.windows.net/assets/9e3a3493-b9a6-
[12] Ali Kakavand, et al. “Techno-economic assessment of green 4b7d-b499-7ca48e357561/The_Future_of_Hydrogen.pdf.
hydrogen and ammonia production from wind and solar [26] Bertuccioli Luca, et al. Development of water electrolysis in
energy in Iran”. Int J Hydrogen Energy May the European union. Tech Rep Fuel Cells and Hydrogen Joint
2023;48(38):14170e91. https://doi.org/10.1016/ Undertaking Feb. 2014. url: https://refman.
j.ijhydene.2022.12.285. issn: 0360-3199. energytransitionmodel.com/publications/2020.
[13] Roos Thomas H. “The cost of production and storage of [27] Transport phenomena d Appl. Mech. Rev. d ASME Digital
renewable hydrogen in South Africa and transport to Japan Collection. url: https://asmedigitalcollection.asme.org/
and EU up to 2050 under different scenarios”. Special Issue appliedmechanicsreviews/article/55/1/R1/458969/
on HYPOTHESIS XV Int J Hdrogen Energy Oct. Transport-phenomena.
2021;46(72):35814e30. https://doi.org/10.1016/ [28] Keith Oldham, Jan Myland. Fundamentals of
j.ijhydene.2021.08.193. issn: 0360-3199. Electrochemical Science. Google-Books-ID:
[14] Hemauer Johanna, et al. “Performance and cost modelling CpBMUqxGgNwC. Elsevier; Dec. 2012. isbn: 978-0-323-
taking into account the uncertainties and sensitivities of 13963-2.
current and next-generation PEM water electrolysis [29] Belmont C, Girault HH. “Coplanar interdigitated band
technology”. Int J Hydrogen Energy Aug. electrodes for synthesis Part I: ohmic loss evaluation”. J
2023;48(66):25619e34. https://doi.org/10.1016/ Appl Electrochem June 1994;24(6):475e80. https://doi.org/
j.ijhydene.2023.03.050. issn: 0360-3199. 10.1007/BF00249845. issn: 1572-8838.
[15] Abbas Majid K, et al. “Techno-economic analysis for clean [30] Zeng Kai, Zhang Dongke. “Recent progress in alkaline water
hydrogen production using solar energy under varied electrolysis for hydrogen production and applications”. Prog
climate conditions”. Int J Hydrogen Energy Jan. Energy Combust Sci June 2010;36(3):307e26. https://doi.org/
2023;48(8):2929e48. https://doi.org/10.1016/ 10.1016/j.pecs.2009.11.002. issn: 0360-1285.
j.ijhydene.2022.10.073. issn: 0360-3199. [31] Lamy Claude, Millet Pierre. “A critical review on the
[16] Matute G, Yusta JM, Naval N. “Techno-economic model and definitions used to calculate the energy efficiency
feasibility assessment of green hydrogen projects based on coefficients of water electrolysis cells working under near
electrolysis supplied by photovoltaic PPAs”. Int J Hydrogen ambient temperature conditions”. J Power Sources Jan.
Energy Feb. 2023;48(13):5053e68. https://doi.org/10.1016/ 2020;447:227350. https://doi.org/10.1016/
j.ijhydene.2022.11.035. issn: 0360-3199. j.jpowsour.2019.227350. issn: 0378-7753.
[17] Di Micco S, et al. “Techno-economic analysis of a multi- [32] Ursúa Alfredo, Sanchis Pablo. “Staticedynamic modelling of
energy system for the co-production of green hydrogen, the electrical behaviour of a commercial advanced alkaline
renewable electricity and heat”. issn: 360-3199. In: water electrolyser”. English. Int J Hydrogen Energy
International journal of hydrogen energy; May 2023. https:// 2012;24(37):18598e614. https://doi.org/10.1016/
doi.org/10.1016/j.ijhydene.2023.04.269. j.ijhydene.2012.09.125. issn: 0360-3199.
[18] Schnuelle Christian, et al. “Dynamic hydrogen production [33] Sood Sumit, et al. “Bond Graph based Multiphysic Modelling
from PV & wind direct electricity supply e modeling and of Anion Exchange Membrane Water Electrolysis Cell”. In:
techno-economic assessment”. Int J Hydrogen Energy Nov. 2020 28th mediterranean Conference on Control and automation
2020;45(55):29938e52. https://doi.org/10.1016/ (MED). Saint-raphae €l. France: IEEE; Sept. 2020. p. 752e7.
j.ijhydene.2020.08.044. issn: 0360-3199. https://doi.org/10.1109/MED48518.2020.9183344. isbn: 978-1-
[19] Bahou Soufiane. “Techno-economic assessment of a 72815-742-9.
hydrogen refuelling station powered by an on-grid [34] An L, et al. “Mathematical modeling of an anion-exchange
photovoltaic solar system: a case study in Morocco”. Int J membrane water electrolyzer for hydrogen production”. Int
Hydrogen Energy July 2023;48(61):23363e72. https://doi.org/ J Hydrogen Energy Dec. 2014;39(35):19869e76. https://
10.1016/j.ijhydene.2023.03.220. issn: 0360-3199. doi.org/10.1016/j.ijhydene.2014.10.025. issn: 0360-3199.
[20] Alexander Buttler, et al. “A detailed techno-economic [35] Pan Jing, et al. “Self-crosslinked alkaline polymer electrolyte
analysis of heat integration in high temperature electrolysis exceptionally stable at 90  C”. Publisher: The Royal Society
for efficient hydrogen production”. Int J Hydrogen Energy of Chemistry Chem Commun Nov. 2010;46(45):8597e9.
Jan. 2015;40(1):38e50. https://doi.org/10.1016/ https://doi.org/10.1039/C0CC03618H. issn: 1364-548X.
j.ijhydene.2014.10.048. issn: 0360-3199. [36] Li Dongguo, et al. “Durability of anion exchange membrane
[21] Seitz M, et al. “Techno economic design of a solid oxide water electrolyzers”. Publisher: The Royal Society of
electrolysis system with solar thermal steam supply and Chemistry. In: Energy & environmental science 14.6; June 2021.
thermal energy storage for the generation of renewable p. 3393e419. https://doi.org/10.1039/D0EE04086J. issn: 1754-
hydrogen”. Int J Hydrogen Energy Oct. 5706.
2017;42(42):26192e202. https://doi.org/10.1016/ [37] Park Eun Joo, et al. “Chemically durable polymer
j.ijhydene.2017.08.192. issn: 0360-3199. electrolytes for solid-state alkaline water electrolysis”. J
[22] Lee Jason K, et al. “Critical Current Density as a Performance Power Sources Jan. 2018;375:367e72. https://doi.org/10.1016/
Indicator for Gas-Evolving Electrochemical Devices”. Cell j.jpowsour.2017.07.090. issn: 0378-7753.
Reports Physical Science 1 Aug. 2020;8:100147. https:// [38] Parrondo Javier, et al. “Degradation of anion exchange
doi.org/10.1016/j.xcrp.2020.100147. issn: 2666-3864. membranes used for hydrogen production by ultrapure
[23] Schalenbach Maximilian, Kasian Olga, Karl J, Mayrhofer J. water electrolysis”. Publisher: The Royal Society of
“An alkaline water electrolyzer with nickel electrodes Chemistry RSC Adv Feb. 2014;4.19:9875e9. https://doi.org/
enables efficient high current density operation”. Int J 10.1039/C3RA46630B. issn: 2046-2069.
Hydrogen Energy July 2018;43(27):11932e8. https://doi.org/ [39] Khan Niaz Atif, et al. “Effects of the operation mode on the
10.1016/j.ijhydene.2018.04.219. issn: 0360-3199. degradation behavior of anion exchange membrane water
[24] U.S. Department of Energy. Department of Energy Hydrogen electrolyzers”. J Power Sources Jan. 2021;481:229093.
Program Plan. Tech. rep. 2020. url: https://www.hydrogen. https://doi.org/10.1016/j.jpowsour.2020.229093. issn: 0378-
energy.gov/pdfs/hydrogen-program-plan-2020.pdf. 7753.
530 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

[40] Dekel Dario R, Rasin Igal G, Simon Brandon. “Predicting Electrochem Soc May 2022;169.5:054506. https://doi.org/
performance stability of anion exchange membrane fuel 10.1149/1945-7111/ac69c4. issn: 0013-4651, 1945-7111.
cells”. J Power Sources Apr. 2019;420:118e23. https://doi.org/ [55] Stanislaw Lauren N, Gerhardt Michael R, Weber Adam Z.
10.1016/j.jpowsour.2019.02.069. issn: 0378-7753. “Modeling Electrolyte Composition Effects on Anion-
[41] Jang Myeong Je, et al. “Efficient and durable anion exchange Exchange-Membrane Water Electrolyzer Performance”. ECS
membrane water electrolysis for a commercially available Trans July 2019;92.8:767e79. https://doi.org/10.1149/
electrolyzer stack using alkaline electrolyte. ACS Energy 09208.0767ecst. issn: 1938-6737, 1938-5862.
Lett Aug. 2022;7.8:2576e83. https://doi.org/10.1021/ [56] Lee Jaeseung, et al. “Modeling of gas evolution processes in
acsenergylett.2c01049. Publisher: American Chemical porous electrodes of zero-gap alkaline water electrolysis
Society. cells”. Fuel May 2022;315:123273. https://doi.org/10.1016/
[42] Zengcai Liu, et al. “An Alkaline Water Electrolyzer with j.fuel.2022.123273. issn: 0016-2361.
Sustainion™ Membranes: 1 A/cm2 at 1.9V with Base Metal [57] Rapha Julysses Enero Perez. “Analysis of the levelized cost
Catalysts”. ECS Transactions 77.9 May 2017:71e3. https:// of green hydrogen production for very heavy vehicles in
doi.org/10.1149/07709.0071ecst. issn: 1938-6737, 1938- New Zealand”. PhD thesis. Te Herenga WakadVictoria
5862. University of Wellington; 2020.
[43] Park Yoo Sei, et al. “High-performance anion exchange [58] Jim Hinkley, et al. Cost assessment of hydrogen production
membrane water electrolyzer enabled by highly active from PV and electrolysis. Australia: Tech. rep. CSIRO; 2016.
oxygen evolution reaction electrocatalysts: synergistic p. 44. url: https://arena.gov.au/assets/2016/05/Assessment-
effect of doping and heterostructure”. Appl Catal B Environ of-the-cost-of-hydrogen-from-PV.pdf.
Dec. 2022;318:121824. https://doi.org/10.1016/ [59] Carmo Marcelo, et al. “A comprehensive review on PEM
j.apcatb.2022.121824. issn: 0926-3373. water electrolysis”. Int J Hydrogen Energy Apr.
[44] Motealleh Behrooz, et al. “Next-generation anion exchange 2013;38(12):4901e34. https://doi.org/10.1016/
membrane water electrolyzers operating for commercially j.ijhydene.2013.01.151. issn: 0360-3199.
relevant lifetimes”. Int J Hydrogen Energy Jan. [60] Enapter. AEM Electrolyser EL 4.0. 2023. url: https://handbook.
2021;46.5:3379e86. https://doi.org/10.1016/ enapter.com/electrolyser/el40/downloads/Enapter_
j.ijhydene.2020.10.244. issn: 0360-3199. Datasheet_EL40_EN.pdf.
[45] Chen Nanjun, et al. “High-performance anion exchange [61] Gutierrez-Martı́n F, Confente D, Guerra I. “Management of
membrane water electrolyzers with a current density of variable electricity loads in wind e hydrogen systems: the
7.68 A cm 2 and a durability of 1000 hours”. Energy Environ case of a Spanish wind farm”. Int J Hydrogen Energy July
Sci 2021;14.12:6338e48. https://doi.org/10.1039/D1EE02642A. 2010;35.14:7329e36. https://doi.org/10.1016/
issn: 1754-5692, 1754-5706. j.ijhydene.2010.04.181. issn: 0360-3199.
[46] Liu Jiangjin, et al. “Elucidating the Role of Hydroxide [62] Amin Mohammadi, Mehrpooya Mehdi. “A comprehensive
Electrolyte on Anion-Exchange-Membrane Water review on coupling different types of electrolyzer to
Electrolyzer Performance”. J Electrochem Soc May renewable energy sources”. Energy Sept. 2018;158:632e55.
2021;168.5:054522. https://doi.org/10.1149/1945-7111/ https://doi.org/10.1016/j.energy.2018.06.073. issn: 0360-5442.
ac0019. issn: 0013-4651, 1945-7111. [63] Mohamed Nasser, et al. A review of water
[47] Fang Ya-Hui, Liu Zhi-Pan. Tafel kinetics of electrocatalytic electrolysisebased systems for hydrogen production using
reactions: from experiment to first-principles. ACS Catal hybrid/solar/wind energy systems. Environ Sci Pollut Res
Dec. 2014;4.12:4364e76. https://doi.org/10.1021/cs501312v. Int 2022;29.58:86994e7018. https://doi.org/10.1007/s11356-
Publisher: American Chemical Society. 022-23323-y. issn: 0944-1344.
[48] Tse Edmund CM, Gewirth Andrew A. Effect of temperature [64] Genç Gamze, Çelik Muhammet, Serdar Genç M. “Cost
and pressure on the kinetics of the oxygen reduction analysis of wind-electrolyzer-fuel cell system for energy
reaction. Publisher: American Chemical Society J Phys demand in Pınarbas‚ı-Kayseri”. In: International journal of
Chem Feb. 2015;119(8):1246e55. https://doi.org/10.1021/ hydrogen energy. 12th CHEC. 37.17; Sept. 2012. p. 12158e66.
acs.jpca.5b00572. issn: 1089-5639. https://doi.org/10.1016/j.ijhydene.2012.05.058. issn: 0360-
[49] Balzer RJ, Vogt H. “Effect of Electrolyte Flow on the Bubble 3199.
Coverage of Vertical Gas-Evolving Electrodes”. J [65] Genç Mustafa Serdar, Çelik Muhammet, Karasu Ilyas. _ “A
Electrochem Soc 2003;150.1:E11. https://doi.org/10.1149/ review on wind energy and windehydrogen production in
1.1524185. issn: 00134651. Turkey: a case study of hydrogen production via electrolysis
[50] Venczel Jozsef. “Ueber den Stofftransport an system supplied by wind energy conversion system in
gasentwickelnden Elektroden”. de. Artwork Size: 79 S. Central Anatolian Turkey”. Renew Sustain Energy Rev Dec.
Medium: application/pdf Pages: 79 S. PhD thesis. ETH Zurich 2012;16.9:6631e46. https://doi.org/10.1016/
1961. https://doi.org/10.3929/ETHZ-A-000091264. j.rser.2012.08.011. issn: 1364-0321.
[51] Vogt H, Balzer RJ. “The bubble coverage of gas-evolving [66] Greiner Christopher J, Korp As Magnus, Holen Arne T. “A
electrodes in stagnant electrolytes”. Electrochim Acta Mar. Norwegian case study on the production of hydrogen from
2005;50(10):2073e9. https://doi.org/10.1016/ wind power”. EHEC2005 Int J Hydrogen Energy July
j.electacta.2004.09.025. issn: 0013-4686. 2007;32(10):1500e7. https://doi.org/10.1016/
[52] Hine F, et al. “Hydrodynamic Studies of Bubble Effects on j.ijhydene.2006.10.030. issn: 0360-3199.
the IR-Drops in a Vertical Rectangular Cell”. J Electrochem [67] Ivy Levene Johanna, et al. “An analysis of hydrogen
Soc Sept. 1975;122.9:1185e90. https://doi.org/10.1149/ production from renewable electricity sources”. Solar
1.2134422. issn: 0013-4651, 1945-7111. Energy 81 June 2007;6:773e80. https://doi.org/10.1016/
[53] Vogt H. “The incremental ohmic resistance caused by j.solener.2006.10.005. issn: 0038-092X.
bubbles adhering to an electrode”. J Appl Electrochem Jan. [68] Linnemann J, Steinberger-Wilckens R. “Realistic costs of
1983;13(1):87e8. https://doi.org/10.1007/BF00615891. issn: wind-hydrogen vehicle fuel production”. Int J Hydrogen
0021-891X, 1572-8838. Energy. EHEC2005 July 2007;32(10):1492e9. https://doi.org/
[54] Liu Jiangjin, Weber Adam Z. “Ionomer Optimization for 10.1016/j.ijhydene.2006.10.029. issn: 0360-3199.
Hydroxide-Exchange-Membrane Water Electrolyzers [69] Loisel Rodica, et al. “Economic evaluation of hybrid off-
Operated with Distilled Water: a Modeling Study”. J shore wind power and hydrogen storage system”. Int J
 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2 531

Hydrogen Energy June 2015;40(21):6727e39. https://doi.org/ [85] Park Yoo Sei, et al. Hierarchical chestnut-burr like structure
10.1016/j.ijhydene.2015.03.117. issn: 0360-3199. of copper cobalt oxide electrocatalyst directly grown on Ni
[70] Mansilla C, et al. “Economic competitiveness of off-peak foam for anion exchange membrane water electrolysis. ACS
hydrogen production today e a European comparison”. Sustainable Chem Eng Feb. 2020;8(6):2344e9. https://
Energy June 2013;55:996e1001. https://doi.org/10.1016/ doi.org/10.1021/acssuschemeng.9b06767. Publisher:
j.energy.2013.03.022. issn: 0360-5442. American Chemical Society.
[71] Fu Ran, et al. U.S. Solar Photovoltaic System Cost [86] Meena Abhishek, et al. “Crystalline-amorphous interface of
Benchmark: Q1 2017. Technical Report NREL/TP-6A20- mesoporous Ni2P @ FePOxHy for oxygen evolution at high
68925. National Renewable Energy Laboratory (NREL) current density in alkaline-anion-exchange-membrane
2017:73. url: https://www.nrel.gov/docs/fy17osti/68925.pdf. water-electrolyzer”. Appl Catal B Environ June
[72] Albright Labs LLC. Puretec Industrial Water d Deionized 2022;306:121127. https://doi.org/10.1016/
Water Services and Reverse Osmosis Systems. enurl: j.apcatb.2022.121127. issn: 0926-3373.
https://www.puretecwater.com. [87] Park Yoo Sei, et al. “Commercial anion exchange membrane
[73] Shi Xunpeng, Liao Xun, Li Yanfei. “Quantification of water electrolyzer stack through non-precious metal
freshwater consumption and scarcity footprints of electrocatalysts”. Appl Catal B Environ Sept.
hydrogen from water electrolysis: a methodology 2021;292:120170. https://doi.org/10.1016/
framework”. Renew Energy July 2020;154:786e96. https:// j.apcatb.2021.120170. issn: 0926-3373.
doi.org/10.1016/j.renene.2020.03.026. issn: 0960-1481. [88] Lee Jooyoung, et al. “Corrosion-engineered bimetallic oxide
[74] Gunawan Tubagus Aryandi, et al. “Towards techno- electrode as anode for high-efficiency anion exchange
economic evaluation of renewable hydrogen production membrane water electrolyzer”. Chem Eng J Sept.
from wind curtailment and injection into the Irish gas 2021;420:127670. https://doi.org/10.1016/j.cej.2020.127670.
network”. Proceedings of ECOS 2019:17. issn: 1385-8947.
[75] U.S. Energy Information Administration. “International [89] Lo pez-Ferna ndez Ester, et al. Recent advances in alkaline
Energy Outlook 2019”:85. exchange membrane water electrolysis and electrode
[76] Aurora Energy Research. Hydrogen market attractiveness manufacturing. Molecules Oct. 2021;26(21):6326. https://
report (HyMAR). Tech Rep Nov. 2021. url: https:// doi.org/10.3390/molecules26216326. issn: 1420-3049.
nkro22cl16pbxzrpzy39bezk-wpengine.netdna-ssl.com/wp- [90] Tao Leiming, et al. “Highly efficient mixed-metal spinel
content/uploads/2021/05/Aurora-Hydrogen-Service- cobaltite electrocatalysts for the oxygen evolution
HyMAR-Press-Release-Slides.pdf. reaction”. Chin J Catal Dec. 2020;41(12):1855e63. https://
[77] Vincent Immanuel, Bessarabov Dmitri. “Low-cost hydrogen doi.org/10.1016/S1872-2067(20)63638-5. issn: 1872-2067.
production by anion exchange membrane electrolysis: a [91] Du Feng, et al. “Strained heterointerfaces in sandwichelike
review”. Renew Sustain Energy Rev Jan. 2018;81:1690e704. NiFe layered double hydroxides/Co1-xS for highly efficient
https://doi.org/10.1016/j.rser.2017.05.258. issn: 1364-0321. and superior longeterm durable oxygen evolution
[78] Pavel Claudiu C, et al. “Highly Efficient Platinum Group reaction”. J Catal Sept. 2020;389:132e9. https://doi.org/
Metal Free Based Membrane-Electrode Assembly for Anion 10.1016/j.jcat.2020.05.026. issn: 0021-9517.
Exchange Membrane Water Electrolysis”. Angew Chem Int [92] Xu Dongyu, et al. Earth-abundant oxygen electrocatalysts
Ed 2014;53(5):1378e81. https://doi.org/10.1002/ for alkaline anion-exchange-membrane water electrolysis:
anie.201308099. issn: 1521-3773. effects of catalyst conductivity and comparison with
[79] Henkensmeier Dirk, et al. Overview: state-of-the art performance in three-electrode cells. ACS Catal Jan.
commercial membranes for anion exchange membrane 2019;9(1):7e15. https://doi.org/10.1021/acscatal.8b04001.
water electrolysis. Journal of Electrochemical Energy Publisher: American Chemical Society.
Conversion and Storage Aug. 2020;18(2):2381e6872. https:// [93] Johnson Matthey. PGM prices and trading. en-GB. Nov. 2022.
doi.org/10.1115/1.4047963. issn. url: https://matthey.com/products-and-markets/pgms-and-
[80] Liu Zengcai, et al. “The effect of membrane on an alkaline circularity/pgm-management.
water electrolyzer”. Int J Hydrogen Energy Dec. [94] Vesborg Peter CK, Jaramillo Thomas F. “Addressing the
2017;42(50):29661e5. https://doi.org/10.1016/ terawatt challenge: scalability in the supply of chemical
j.ijhydene.2017.10.050. issn: 03603199. elements for renewable energy”. RSC Adv 2012;2(21):7933.
[81] Pushkareva IV, et al. “Comparative study of anion exchange https://doi.org/10.1039/c2ra20839c. issn: 2046-2069.
membranes for low-cost water electrolysis”. Int J Hydrogen [95] Minke Christine, et al. “Is iridium demand a potential
Energy. Progress in Hydrogen Production and Utilization bottleneck in the realization of large-scale PEM water
Oct. 2020;45(49):26070e9. https://doi.org/10.1016/ electrolysis?”. Int J Hydrogen Energy July
j.ijhydene.2019.11.011. issn: 0360-3199. 2021;46(46):23581e90. https://doi.org/10.1016/
[82] Chen Pengzuo, Hu Xile. “High-Efficiency Anion Exchange j.ijhydene.2021.04.174. issn: 0360-3199.
Membrane Water Electrolysis Employing Non-Noble Metal [96] Mandal Mrinmay, Huang Garrett, Paul A, Kohl. “Anionic
Catalysts”. Adv Energy Mater 2020;10(39):2002285. https:// multiblock copolymer membrane based on vinyl addition
doi.org/10.1002/aenm.202002285. issn: 1614-6840. polymerization of norbornenes: applications in anion-
[83] Koshikawa Hiroyuki, et al. Single nanometer-sized NiFe- exchange membrane fuel cells”. J Membr Sci Jan.
layered double hydroxides as anode catalyst in anion 2019;570(571):394e402. https://doi.org/10.1016/
exchange membrane water electrolysis cell with energy j.memsci.2018.10.041. issn: 0376-7388.
conversion efficiency of 74.7% at 1.0 A cme2. ACS Catal Feb. [97] Mandal Mrinmay, Huang Garrett, Paul A, Kohl. Highly
2020;10(3):1886e93. https://doi.org/10.1021/ conductive anion-exchange membranes based on cross-
acscatal.9b04505. Publisher: American Chemical Society. linked poly(norbornene): vinyl addition polymerization.
[84] Jang Myeong Je, et al. “Superior performance and stability of ACS Appl Energy Mater Apr. 2019;2(4):2447e57. https://
anion exchange membrane water electrolysis: pH- doi.org/10.1021/acsaem.8b02051. Publisher: American
controlled copper cobalt oxide nanoparticles for the oxygen Chemical Society.
evolution reaction”. Publisher: The Royal Society of [98] Chen Wanting, et al. Highly conducting anion-exchange
Chemistry J Mater Chem A Feb. 2020;8(8):4290e9. https:// membranes based on cross-linked poly(norbornene): ring
doi.org/10.1039/C9TA13137J. issn: 2050-7496. opening metathesis polymerization. ACS Appl Energy Mater
532 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 9 ( 2 0 2 4 ) 5 1 8 e5 3 2

Apr. 2019;2(4):2458e68. https://doi.org/10.1021/ 2018;375:351e60. https://doi.org/10.1016/

acsaem.8b02052. Publisher: American Chemical Society. j.jpowsour.2017.08.026. issn: 0378-7753.
[99] Arges Christopher G, Ramani Vijay. Two-dimensional NMR [107] Mohanty Angela D, Bae Chulsung. “Mechanistic analysis of
spectroscopy reveals cation-triggered backbone ammonium cation stability for alkaline exchange
degradation in polysulfone-based anion exchange membrane fuel cells”. J Mater Chem A 2014;2.41:17314e20.
membranes. Proceedings of the National Academy of https://doi.org/10.1039/C4TA03300K. issn: 2050-7488, 2050-
Sciences 110 Feb. 2013;7:2490e5. https://doi.org/10.1073/ 7496.
pnas.1217215110. Publisher: Proceedings of the National [108] Sturgeon Matthew R, et al. “Hydroxide based
Academy of Sciences. Benzyltrimethylammonium Degradation: quantification of
[100] Chempath Shaji, et al. Density functional theory study of Rates and Degradation Technique Development”. J
degradation of tetraalkylammonium hydroxides. Publisher: Electrochem Soc Jan. 2015;162(4):1945e7111. https://doi.org/
American Chemical Society J Phys Chem C July 10.1149/2.0271504jes. Publisher: IOP Publishing, F366. issn.
2010;114(27):11977e83. https://doi.org/10.1021/jp9122198. [109] Marino MG, Kreuer KD. “Alkaline Stability of Quaternary
issn: 1932-7447. Ammonium Cations for Alkaline Fuel Cell Membranes and
[101] Chempath Shaji, et al. Mechanism of tetraalkylammonium Ionic Liquids”. ChemSusChem 2015;8(3):513e23. https://
headgroup degradation in alkaline fuel cell membranes. doi.org/10.1002/cssc.201403022. issn: 1864-564X.
Publisher: American Chemical Society J Phys Chem C Mar. [110] Jang Dohyung, Cho Hyun-Seok, Kang Sanggyu. “Numerical
2008;112.9:3179e82. https://doi.org/10.1021/jp7115577. issn: modeling and analysis of the effect of pressure on the
1932-7447. performance of an alkaline water electrolysis system”. Appl
[102] Tiwari Prerna, et al. “A new class of bubble-free water Energy Apr. 2021;287:116554. https://doi.org/10.1016/
electrolyzer that is intrinsically highly efficient”. Int J j.apenergy.2021.116554. issn: 0306-2619.
Hydrogen Energy Sept. 2019;44(42):23568e79. https:// [111] Awasthi A, Scott Keith, Basu S. “Dynamic modeling and
doi.org/10.1016/j.ijhydene.2019.07.100. issn: 0360-3199. simulation of a proton exchange membrane electrolyzer for
[103] Ahn Sang Hyun, et al. “Effect of morphology of hydrogen production”. International Journal of Hydrogen
electrodeposited Ni catalysts on the behavior of bubbles Energy. Fuel Cell Technologies: FUCETECH 2009 Nov.
generated during the oxygen evolution reaction in alkaline 2011;36(22):14779e86. https://doi.org/10.1016/
water electrolysis”. Publisher: The Royal Society of j.ijhydene.2011.03.045. issn: 0360-3199.
Chemistry Chem Commun Sept. 2013;49(81):9323e5. [112] Nami Hossein, et al. “Techno-economic analysis of current
https://doi.org/10.1039/C3CC44891F. issn: 1364-548X. and emerging electrolysis technologies for green hydrogen
[104] Yu Cunming, et al. “Aerophilic Electrode with Cone Shape production”. Energy Convers Manag Oct. 2022;269:116162.
for Continuous Generation and Efficient Collection of H2 https://doi.org/10.1016/j.enconman.2022.116162. issn: 0196-
Bubbles”. Adv Funct Mater 2016;26(37):6830e5. https:// 8904.
doi.org/10.1002/adfm.201601960. issn: 1616-3028. [113] Mayyas Ahmad T, et al. Manufacturing Cost Analysis for
[105] Karam Yassin, et al. “Quantifying the critical effect of water Proton Exchange Membrane Water Electrolyzers. NREL/TP-
diffusivity in anion exchange membranes for fuel cell 6A20-72740 Tech. Rep Aug. 2019:1557965. https://doi.org/
applications”. J Membr Sci Aug. 2020;608:118206. https:// 10.2172/1557965. NREL/TPe6A20e72740, 1557965, http://
doi.org/10.1016/j.memsci.2020.118206. issn: 0376-7388. www.osti.gov/servlets/purl/1557965/. url:.
[106] Dekel Dario R, et al. “The critical relation between chemical [114] Lazard Lazard’s. Levelized cost of hydrogen analysis. Tech
stability of cations and water in anion exchange membrane Rep June 2021:16. url: https://www.lazard.com/media/
fuel cells environment”. J Power Sources Jan. 451779/lazards-levelized-cost-of-hydrogen-analysis-vf.pdf.