Content uploaded by Jose Manuel Andujar Marquez
Author content
All content in this area was uploaded by Jose Manuel Andujar Marquez on Aug 10, 2022
Content may be subject to copyright.
REVIEW PAPER
Sun, heat and electricity. A comprehensive study of non-
pollutant alternatives to produce green hydrogen
Julio José Caparr
os Mancera
1,2
| Francisca Segura Manzano
2
|
José Manuel Andújar
2
| Eduardo L
opez
1
| Fernando Isorna
1
1
Instituto Nacional de Técnica Aeroespacial, INTA (Spanish National Institute of Aerospace Technology), Mazag
on, Spain
2
Centro de Investigaci
on en Tecnología, Energía y Sostenibilidad (CITES), University of Huelva, Huelva, Spain
Correspondence
Julio José Caparr
os Mancera, Instituto
Nacional de Técnica Aeroespacial, INTA
(Spanish National Institute of Aerospace
Technology); Ctra. San Juan del Puerto,
Km. 33, Mazag
on 21130, Spain.
Email: jcapman@inta.es,julio.caparros@
diesia.uhu.es
Funding information
European Union Regional Development
Fund, Grant/Award Number: P20_00730;
European Union Regional Development
Fund 2014/2020, Grant/Award Number:
UHU-1259316; Spanish State Program of R
+D+I Oriented to the Challenges of
Society, Grant/Award Number:
PID2020-116616RB-C31; Funding for
Open Access charge, Grant/Award
Number: University of Huelva (UHU)/
CBUA
Summary
Water-based hydrogen production is currently an attractive research field, as it
provides a greener method to produce hydrogen than existing alternatives. Green
hydrogen is expected to progressively replace fossil fuels, which are highly harm-
ful environmentally. This paper presents a critical analysis over time of the main
water splitting technologies currently in use for sustainable hydrogen production.
As a result of the critical analysis, all the studied techniques have been ordered
chronologically in the way that it is possible to understand how new materials
have driven to new techniques, more efficientandlessexpensive.Thisallowshav-
ing a complete vision of these technologies. A high level of maturity has been
reached in electrolysis, while other techniques still have a long way to go,
although many improvements and relevant advancements have been made over
the years. The paper offers a global and comparative vision of each technology.
From this, it is possible to identify the different paths where efforts are needed to
make water-based hydrogen production a mature, stable and efficient technology.
KEYWORDS
chronological review, electrolysis, green hydrogen production, photolysis, technical study,
thermolysis, water splitting techniques
Abbreviations: AEM, Anion Exchange Membrane; CCM, Catalyst Coated Membrane; CNT, Carbon Nanotubes; FC, Fuel Cell; FCH JU, Fuel Cell
and Hydrogen Joint Undertaking; GDL, Gas Diffusion Layer; GHG, Green Houses Gases; HER, Hydrogen Evolution Reaction; HyCon, Hydrogen
Concentrator; LHD, Layered Double Hydroxide; LSF, Lanthanum Strontium Ferrite; LSGM, Strontium and Magnesium Co-Doped Lanthanum
Gallate; LSM, Lanthanum Strontium Manganite; MEA, Membrane Electrode Assemblies; MFCI, Multilayer Flow Channel Inserts; MWNT,
Multiwalled Carbon Nanotubes; NSTF, Nanostructured Thin Film Electrodes; OER, Oxygen Evolution Reaction; PBI, Polybenzimidazole; PE,
Polymer Electrolyte; PEFC, Polymer Electrolyte Fuel Cell; PEM, Proton Exchange Membrane / Polymer Electrolyte Membrane; PFSA, Poly
Perfluorosulfonic Acid; PS, Polystyrene; PSF, Polysulfone; PTFE, Polytetrafluoroethylene; RAFM, Reduced Activation Ferritic/Martensitic; SMR,
Steam Methane Reforming; SOEC, Solid Oxide Electrolytic Cell; SOFC, Solid Oxide Fuel Cell; SPE, Solid Polymer Electrolyte; SPEEK, Sulfonated
Polyetheretherketone; URFC, Unitized Regenerative Fuel Cell; WGS, Water-Gas Shift; YSZ, Yttria-Stabilized Zirconia.
Received: 24 March 2022 Revised: 21 July 2022 Accepted: 26 July 2022
DOI: 10.1002/er.8505
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any
medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2022 The Authors. International Journal of Energy Research published by John Wiley & Sons Ltd.
Int J Energy Res. 2022;1–30. wileyonlinelibrary.com/journal/er 1
1|INTRODUCTION
Currently, several alternatives can be selected to produce
hydrogen according to the raw material used in produc-
tion.
1
From an annual global production of 70 Mt,
2
most
hydrogen today is produced from fossil fuels and emits
CO
2
(grey hydrogen)
3-5
while sustainable hydrogen pro-
duction, generally from renewable energy, without pol-
lutant (green hydrogen),
6,7
is still in the process of
growth
8-10
Figure 1, shows that 76% of the current
world’s hydrogen production is derived from natural gas,
via steam methane reforming (SMR), gaining much
prominence in the last decade, as it corresponded to 48%
in 2010.
11
Other processes for hydrogen production are
from coal gasification (22%), and water electrolysis only
contributes with 2%.
12,13
Interest in hydrogen production methods arises in
view of the fact that hydrogen plays a very important role
in industrial processes.
14-19
Figure 2shows that 42% of
the pure hydrogen produced is used to manufacture
ammonia (NH
3
). This has been the main pure hydrogen
demand for decades. However, from 2010 more than half
demand is employed in hydrotreating processes in refin-
eries. The remaining 6% is divided between other uses,
including methanol production, hydrogenation of fats,
synfuel production and, of course, fuel cells feeding for
both stationary and transport applications.
12,13,20-24
The direct consequences of hydrogen production
methods based on conventional fossil fuels are: green-
house gases production,
25
inherent dependence of fossil
fuels,
26
and the requirement of hydrogen purification
stages to guarantee hydrogen purity degrees about 98%
(depending on the hydrogen application).
27,28
Based on the above, the hydrogen production from
water is becoming increasingly important
29-31
; to the
point that, following European directives and reports of
the Clean Hydrogen Joint Undertaking, framed in EU
Green Deal and Hydrogen Strategy,
32-34
research on the
production of hydrogen from water becomes a priority,
among other alternatives such as biomass, within the
technologies proposed to make a transition to more sus-
tainable energy systems.
35-39
Then, according to Figure 3, the amount of energy
required for the water splitting depends on the tempera-
ture in the sense that below 100C (water in liquid state),
the energy supply must be mostly in the form of electric
power. Above this point, as the temperature increases,
the water turns into steam and the contribution of energy
in the form of heat counteracts the amount of electrical
energy necessary for water splitting.
40
Therefore, for the
hydrogen production from water, the required energy
can be supplied in different ways like solar radiation,
thermal energy or electric energy.
Consequently, water-based hydrogen production pro-
cesses can be classified according to the nature of the
energy being used, Figure 4:photolysis, when energy
comes from solar radiation; thermolysis, when the source
is the applied temperature; and electrolysis, when the
division of water is done by electrical energy.
Several methods have been studied to generate hydro-
gen using water splitting techniques.
29
Although electrol-
ysis has been used for industrial applications from the
XIX century,
41,42
most water splitting techniques have
been out of scientific and industrial priority attention
until practically the beginning of the nineties of the last
century. Consequently, many of these techniques still
have little production efficiency, so there is much room
for improvement.
43
After carrying out a previous analysis of the main
review articles, which are developed in the state of the
art of hydrogen production techniques from water,
Table 1is presented, highlighting their contributions and
motivating the completion of this review. Some are
reviews of photolysis, including works focused on
technology,
44-47
and those that also include the chronol-
ogy of the advances.
48,49
The literature on relevant ther-
molysis reviews is limited as well, with some work
FIGURE 1 World’s hydrogen production classified by raw
material
12,13
FIGURE 2 Hydrogen uses classification
12,13,20-24
2CAPARRÓS MANCERA ET AL.
focused on the technical field.
50-52
The reviews regarding
electrolysis are numerous, mainly technical,
53-62
but also
historical.
29,63,64
As for works that consider various tech-
nologies, there is a relevant review on thermolysis and
photolysis,
65
as well as few relevant reviews that also
include electrolysis,
66,67
most of them from a broader
view that includes polluting hydrogen production
methods without analysing water splitting in such
detail,
68-74
all of them focused on technical analysis.
The novelty of the authors' proposal is visualized
through Table 1, since it carries out an important chrono-
logical analysis of all the main technologies, as well as
updates previous works. This is of great relevance due to
the strong impulse that is being given to research for the
production of large quantities of green hydrogen.
Compared to previous publications, in main water
splitting technologies currently in use for sustainable
hydrogen production, no previous works have been
found than include the analysis of the three renewable
sources (sun, heat and electricity), from a chronological
and technical perspective, being the main novelty of this
review. For this reason, this paper analyses the three
main techniques for green hydrogen production, compar-
ing their advantages and disadvantages from a qualitative
and quantitative point of view. As a result, the studied
techniques have been ordered over the time with the aim
to understand how new techniques arise from the devel-
opment of new materials and the commitment to over-
come the drawbacks of the earlier techniques. In
chronological order, the main drawbacks that each tech-
nology have presented, and the challenges overcome
today will be revealed. This analysis, compared to the
previous literature, provides a comprehensive view of the
three main technologies, so that the advances, chal-
lenges, and how they have been reached can be visual-
ized graphically and specifically. In this way, the main
objective is to establish a source of extensive information,
correspondingly classified and ordered in time, which
serves as a basis for the improvements of the technologies
analysed here. This entails taking not only the references
FIGURE 3 Energy demand for water
splitting
FIGURE 4 Sources for hydrogen
production based on water splitting
CAPARRÓS MANCERA ET AL.3
of the same technology, but of the advances in all of
them, which can well be extrapolated between them, as it
has happened historically and is reflected in this review.
The main scientific databases used in this review have
been Wiley Online Library, Elsevier ScienceDirect and
Scopus, Google Scholar and ResearchGate.
In Section 2, a chronological review of the main
water splitting techniques for hydrogen production is
made, starting with the techniques based on photoly-
sis, continuing with the techniques that use the heat
for decomposition and ending with the electrolysis
processes. In Section 3, the different techniques stud-
ied are discussed and they are compared from the
point of view of efficiency and the amount of hydrogen
production. Finally, the conclusions are compiled in
Section 4.
TABLE 1 Author’s contributions
regarding literature review
Ref. Author Sun
a
Heat
b
Electricity
c
Chronology Technical
44 Liu et al ✓-- - ✓
45 Basheer et al ✓-- - ✓
46 Giri et al ✓-- - ✓
47 Al-Ahmed et al ✓-- - ✓
48 Bak et al ✓-- ✓✓
49 Maeda ✓-- ✓✓
50 Mao et al - ✓-- ✓
51 Mehrpooya et al - ✓-- ✓
52 Safari et al - ✓-- ✓
53 Shiva et al - - ✓-✓
54 Ursua et al - - ✓-✓
55 Wang et al - - ✓-✓
56 Chi et al - - ✓-✓
57 Babic et al - - ✓-✓
58 Hyung et al - - ✓-✓
59 Liu et al - - ✓-✓
60 Xiang et al - - ✓-✓
61 Cossar et al - - ✓-✓
62 Zhou et al - - ✓-✓
29 Carmo et al - - ✓✓ ✓
63 Paidar et al - - ✓✓ ✓
64 Zeng et al - - ✓✓ ✓
65 Pietro et al ✓✓ -- ✓
66 Idriss et al ✓✓ ✓ -✓
67 Dutta et al ✓✓ ✓ -✓
68 Singla et al ✓✓ ✓ -✓
69 Epelle et al ✓✓ ✓ -✓
70 Faye et al ✓✓ ✓ -✓
71 Nnabuife at al ✓✓ ✓ -✓
72 Hermesmann et al ✓✓ ✓ -✓
73 Zhou et al ✓✓ ✓ -✓
74 Younas et al ✓✓ ✓ -✓
Author’s proposal ✓✓ ✓ ✓ ✓
a
Photolysis.
b
Thermolysis.
c
Electrolysis.
4CAPARRÓS MANCERA ET AL.
2|SUN, HEAT AND ELECTRICITY
FOR WATER SPLITTING-BASED
HYDROGEN PRODUCTION.
UNDERSTANDING THE TIMELINE
The first source used for water splitting was electricity; in
1789, Adriaan Paets van Troostwijk and Jan Rudolph
Deiman published the results of their experiments on the
decomposition of water by static electricity.
75
They used
an electrostatic machine to generate electricity that was
discharged by golden electrodes in a Leyden jar
(a primitive capacitor manufactured using a glass jar with
layers of metal foil on the outside and inside) filled with
water. This experiment could be considered the first in
demonstrating the water electrolysis process, however,
traditionally the authorship of water electrolysis has been
given to William Nicholson and Anthony Carlisle in
1800, but it seems that their only discovery was the used
electricity source, the voltaic pile invented by Alessandro
Volta that same year. The controversial fact of who first
discovered water electrolysis has been discussed over
years. Finally, the van Troostwijk and Deiman author-
ship has been agreed by de Levie
76
and Trasatti.
77
One
century later, in 1888, the first industrial water electroly-
sis process was developed by Santos et al.
78
Since then
and until now, new water splitting techniques have
emerged to reach a solution for hydrogen production
without harmful emissions, reduced energy consump-
tion, cost, and maintenance and, all this, with high effi-
ciency, durability and safety.
2.1 |The Sun: Inexhaustible resource
There is an important research line focused on trying to
take advantage of the solar radiation to produce hydro-
gen.
79
Thermal and photonic processes are the most
promising.
80
Thermal processes imply solar energy being
converted to heat, which can either be used directly,
stored or converted to another energy type. By contrast,
in solar photonic processes, photons are absorbed directly
by a photo-sensible material, without complete conver-
sion to heat.
81
There are four basic subcategories: (1) sun-
light is absorbed by isolated molecules in a solution
(photocatalysis); (2) sunlight is absorbed by photoelectro-
chemical cells based on semiconductors, usually shaped
like photovoltaic cell; (3) sunlight is absorbed by a biolog-
ical system sensitive to light (photobiological system),
usually algae and (4) a combination of all or some of the
above processes.
Unfortunately, none of these processes/
technologies currently have high efficiency, being
around 18%.
82
It seems clear that these technologies
still have a long way to go to be competitive commer-
cially speaking.
2.1.1 | Photocatalytic systems
From the point of view of hydrogen production, a photo-
catalyst system collects sunlight radiation to activate spe-
cific molecules that assist in the reactions for hydrogen
production, Figure 5.
The first papers found in the scientific literature
related to hydrogen production based on photocatalyst
technique date from 1970s, when several works attempts
to undertake direct photolysis for the hydrogen produc-
tion as Bockris compile in Ref. 27 One decade later, Bor-
gabello et al
83
demonstrate that depositing particles of
CdS (Cadmium Sulphide) onto the catalyst, the hydrogen
production is improved. In this sense, the search of new
materials for softening water and for accelerating the
redox reaction has predominated the working lines dur-
ing the 90s and the beginning of the XXI century.
84-86
Concerned by the efficiency of this technology, Bolton
noted in
87
that almost all of the previous studies calcu-
lated efficiency using the half reactions separately rather
than combining the two. There had only been one study
of the entire process, which recorded a 7% efficiency.
More recent works go over the improvement of existing
methods,
88,89
with the aim to increase the process effi-
ciency. In this sense, Li
30
proposes to use a technique
called junction: construction of a hetero-junction at a
photocatalyst interface, with the purpose of including a
built-in electrical field that aides with charge separation
and increases efficiency. Also, zeolite membranes were
introduced for improved gas separation.
About the undesirable electron-hole recombination,
what could also lead to low efficiency of photocatalysis,
Ag/reduced graphene oxide/TiO
2
nanocomposites are
proposed along with LaTiO
2
N, Ta
3
N
5
and Sm
2
Ti
2
S
2
O
5
as
relevant candidates for this issue during last decade.
90,91
FIGURE 5 Scheme of photocatalytic system for hydrogen
production
CAPARRÓS MANCERA ET AL.5
More recently, 2019, the Lawrence Berkeley National
Lab has been working on a photocatalyst research. They
claim to have found 12 new materials that will be useful
for the photoanode of an electrolytic cell.
92
Different
metal oxides were used during the experiments. They
explored 174 different compounds, and the result was
that structures composed of vanadium, oxygen and a
third element were the most useful because of their tune-
able results. In addition, the use of co-catalysts, along
with CdS, has been shown to achieve system efficiencies
of up to 12.8%,
93
depending on the production under dif-
ferent wavelength irradiation.
In this context, precious metals such as Rh, Pd, Ru, Pt
nanoparticles have been widely utilized as co-catalysts,
with several tests developed to find the suitable co-
catalyst to improve the charge carriers separation in
semiconductors.
90,94,95
However, the semiconductor mod-
ified with noble metals is not viable due to limited avail-
ability. Recently, it has been reported that the
photocatalytic performance of semiconductors (Ta
2
O
5
and TiO
2
) was prominently influenced by the non-noble
metal NiO core-shell structured co-catalyst. The reason is
that the NiO inhibition nature suppressed the undesired
back reaction.
96
Then, Ravi et al developed a system with
CuO-NiO hierarchical nanostructures as a co-catalyst
deposited on TiO
2
nanospheres for enhanced photocata-
lytic hydrogen generation. The formation of ultrathin
NiO shell over the CuO core was confirmed. A high rate
of hydrogen production of 26.1 mmol. h
1
g
1
was
showed, under direct sunlight.
97
Current studies high-
light in-situ co-catalyst formation and novel phosphoryla-
tion of NiAl-layered double hydroxide nanosheets as co-
catalyst.
98-102
2.1.2 | Photoelectrochemical cells based on
semiconductors
This type of cells is composed of a photoanode and a
cathode (metal) both immersed in an electrolyte and con-
nected to an external circuit,
103
Figure 6.
The ideal materials for the electrodes are the ones
that have a band gap of 1 to 2 V, because the ideal work-
ing voltage of the electrolysis cell is 1.229 V. Standard
semiconductors such as gallium arsenide and indium
phosphate have been shown to have high efficiencies in
the case of hydrogen production.
104
The first well-known semiconductor-based cell for the
photolysis (dissociation of molecules by the effect of light)
of water was studied in 1972 by Fujishima and Honda.
105
In their experiment, TiO
2
was used in the anode and plati-
num in the cathode. The cell was chemically biased due to
the different pH levels in each side of the system.
In a study published in 1983, Murphy et al
106
tested a
semiconductor comprised of two p-ncouples made of gal-
lium arsenide as a hydrogen generator using water pho-
tolysis. Up to that time, photoelectrochemical devices
had efficiency rates around 1%, while the two cells junc-
tion reached 8% conversion efficiency of sunlight to
hydrogen. Apart from these devices, it is possible to find
photo-aided electrolysis cells with efficiency of 10%,
higher than previous cases, but with the drawback of the
cost of electrical power required.
With experimental results, authors concluded that
low efficiency values, even in the two-cells junction
device, were due to both corrosion of the materials from
exposure to the highly alkaline electrolyte and because of
the recombination of hydrogen and oxygen, due to diffi-
culties of separating the two gases in an effective way.
Then, photolysis cells were not yet suitable for hydrogen
mass production.
With the aim to identify the factors which have effect
on the photolytic cell, during the 80s Bockris et al
attempt to model the hydrogen evolution reaction
derived from the photo-induced current.
107
They con-
clude that hydrogen production rate due to photocurrent
not only depends on semiconductor physical properties
(doping density, for example) but it also depends on the
surrounding electrolyte solution and the sunlight
wavelength.
During the last decade of the 20th century, the inter-
est of researching in this technology decreases and hardly
one half hundred published works have been found.
These works can be grouped in two: (1) those where
authors propose a photocatalytic cell implementation,
the effect of the etching treatments, the metal loading
and the mixing of two semiconductors,
108,109
; and
(2) those that focus on the use of alternative liquid elec-
trolytes like HCl,
110
or H
2
S.
111,112
FIGURE 6 Scheme of photolytic cell based on semiconductor
technology
6CAPARRÓS MANCERA ET AL.
Fortunately, the interest in photoelectrochemical sys-
tems reawakens in the last years as it can be seen in.
30
Then, contemporary authors recover research both in
new materials for photoanode
113
(BiVO
4
had a theoretical
efficiency of 9% and Ta
3
N
5
returned 15% theoretical effi-
ciency), and new nanostructures to build the photoelec-
trochemical cell
114
(a vertical nanorod based on Ta
3
N
5
to
form the photoelectrode). However, this material is
highly unstable and completely degrades in a few
minutes. To this problem, some solutions can be found
in
115
and,
116
where the hole storage layer based on ferri-
hydrite in the first case, and Ni(OH)
x
/MoO
3
in the sec-
ond, prolongs the lifespan of the photoelectrochemical
cell for 6 and 24 h respectively.
Not only photoanode in photoelectrochemical cells
have captured research attention but also photocathode
development. Then, in an attempt to approximate to the-
oretical photocurrent density value (14.5 mAcm
2
, this is
due to the incompatible light absorption of planar struc-
tures, and it equates to 18% efficiency), Paracchino
et al
117
and Luo et al
118
have recently proposed a novel
nanowired structure. On this, cuprous oxide (new mate-
rial) is growth over a layer of titanium oxide (old mate-
rial, remember original Fujishima and Honda design).
105
In each case, authors achieve, respectively, photocurrent
values above 7 mAcm
2
(7% efficiency) and 10 mAcm
2
(12.82% efficiency). Currently stand out layered seleno-
phosphites photocatalysts and improved Ag-sensitized
TiO
2
.
119-125
Photoelectrocatalysis report typical photoa-
nodes of BiVO
4
and Ta
3
N
5
, with still low efficiency, and
real pilot implementations with a short-term 100 kg/day
H
2
production.
126-128
2.1.3 | Photobiological cells
It seems that during history, practical application of pho-
tosynthetic processes for energy generation has not
received enough attention. In this sense, it is relevant to
remind that all fossil fuel reserves have their origin in
photosynthesis, and additionally photosynthesis is
responsible for all the energy stored in form of bio-
mass.
129
Then, as could not be otherwise, it is possible to
obtain hydrogen from a photosynthesis process.
Despite the belief that photobiological hydrogen pro-
duction is relatively recent, hydrogen metabolism by pho-
tosynthetic organisms was originally reported in 1940 by
Gaffron in.
130
It consists on employing microorganisms
such as green algae and cyanobacteria, which in the pres-
ence of sunlight, are capable to generate hydrogen.
131
Apart from green algae and cyanobacteria, there are
other microorganisms like purple non-sulphur bacteria
or dark fermentative bacteria which produces hydrogen.
However, the reaction depends on the physiological con-
ditions of microorganisms, for example, the presence of
some enzyme like hydrogenase or nitrogenase, some
anaerobic conditions, etc.
132
Despite the key role of the photosynthesis in the pres-
ence of fuels in the world, surprisingly, it is one of the
processes that less attention has received along the his-
tory, and it has always been considered as a long-term
challenge. After the theoretical formulation in 1940,
experimental tests developed during the following decade
demonstrate efficiencies near 10%, but with the drawback
that algae saturates with the light intensity at solar irradi-
ances above 0.03 suns (1 sun =100 Wcm
2
) and photobi-
ological hydrogen production ceases.
133
Thus, the main
goal of the most published works found in the scientific
literature has been to identify alternative species of algae
that also support hydrogen photobiological production
under a wider range of sun irradiance. This option has
been chosen by several researchers during years,
134-136
and more recently, it has appeared the genetic engineer-
ing as solution to obtain mutant algae.
137
For example,
Greenbaum et al
138
reduced the size of the antenna chlo-
rophyll pool, allowing irradiances of higher value. This
simply fact allows efficiencies of 15% to 20%.
In 2012, researches performed by the US Department
of Energy in the Pacific Northwest Laboratory found a
bacterium called cyanothece that is capable of producing
both hydrogen and oxygen for 100 h uninterrupted,
139
supposing a great improvement to previous technologies.
Recent studies distinguish Na
2
SO
3
as oxygen scavenger,
and enhanced production with Chlorella vulgaris and
Chlamydomonas reinhardtii.
140-145
Apart from new species or genetic engineering, other
solution must be approached when the problem is the
algae growth under nutrient deficiency. In this case, the
photosynthesis activity decreases to help the algae to sur-
vival, and consequently the hydrogen production rate
will also decline. Then, some authors begin to be con-
cerned in this last decade about this issue (probably due
to excess of contamination that sea water is suffering). In
these cases, when the problem is not to find new algae
species but looking for an alternative process that pro-
longs the life of the algae, the solution passes by intro-
ducing a new phase which consists on a dark anaerobic
incubation before to the phase of light illumination for
hydrogen production.
146
As a result, the hydrogen yield
increases.
Some authors, dedicated to photobiological hydrogen
production, guided their efforts toward new designs for
photosynthetic systems. These new designs began to
appear around the 90s, when photovoltaic, photocataly-
tic, photoelectrochemical and photobiological technolo-
gies had been established for several years. Wilner et al
CAPARRÓS MANCERA ET AL.7
propose in
147
an hybrid system where the counterparts of
the three different approaches (photocatalytic, photoelec-
trochemical and photobiological) come together in
assemblies and have effect on water the photodecomposi-
tion. The hybrid system assures the organization of light-
harvesting and subsequent hydrogen generation.
Another way to overtake the limitations that photobio-
logical hydrogen production presents (saturation at high
intensity light), passes by developing efficient photobior-
eactors. For example, a photobioreactor that is uniformly
illuminated at the optimum light intensity solves the prob-
lems of photoinhibition and dark hydrogen uptake can be
reduced. Additionally, a photobioreactor with high mass
transfer capacity can remove easily the oxygen produced
avoiding its further reaction with hydrogen.
148,149
Finally, analysing all the cited works, all of them
coincide that the maturation level of the sun-based
hydrogen production technology will increase with the
human concern about environmental problems; higher
concern will equal be to a more mature technology.
The advances discussed in this section for each of the
different techniques within the production of hydrogen
using the sun are synthesized in the scheme of Figure 7,
which represents the main advances and historical
achievements in photolysis technology.
2.2 |Heat. A way to reduce the amount
of required electricity
Thermal energy is an important resource when it comes
to obtaining water splitting. Thermolysis is defined as the
process at which water is heated to a high temperature
until decomposed to hydrogen and oxygen. In practice,
the temperature is required to reach 2500C. That is why
heat, as a source for water splitting, is more widespread
as a resource to reduce the electricity needed in high-
temperature electrolysis processes.
2.2.1 | Thermolysis
Direct thermal water splitting into hydrogen and oxygen
is not achieved until the temperature is very high, gener-
ally over 2500C.
150
For example, if it reaches 3000C, a
64% dissociation can be obtained at 1 bar.
The interest in thermolysis cycles increased between
the 70s and the 80s, during the oil crisis, and most of the
cycles proposed took a nuclear energy source for heat, so
operating temperature should remain below 900C. Sig-
nificant progress has recently been achieved regarding
solar heat collectors and concentrators, on a megawatt
scale. In this context, two-step solar heat cycles are char-
acterized by operational simplicity, with large-scale pro-
duction of H
2
from 1000C.
151
At recent years, Cu-Cl and Mg-Cl appear to be
most promising low-temperature thermochemical
cycles, without releasing any greenhouse gases (GHG)
to the atmosphere while requiring a temperature
higher than 550C.
152
Some drawbacks of this tech-
nology are the toxicity of the chemicals involved dur-
ing the cycles and their availability and cost. Another
main drawback of this process, to avoid an explosive
mix, comes from the requirement for an effective sep-
aration of H
2
and O
2
. Although some thermolysis
cycles produce both H
2
and O
2
in separate stages, they
contribute partially avoiding their recombination and
bypassing the need for costly gas separators.
151
To
avoid this recombination, semi-permeable membranes
based on ZrO
2
, along other high-temperature mate-
rials, can be used at a temperature up to 2500C.
153
This recombination can also be avoided when the
product gas is rapidly cooled (within just few millisec-
onds) through a sharp temperature decrease of 1500C
to 2000C. In that case, palladium membranes can be
integrated.
154,155
Nowadays, Cu-Cl has been identified as the most pro-
spective cycles, with lower cost and highest efficiency,
FIGURE 7 Main advances and historical achievements in photolysis technology
8CAPARRÓS MANCERA ET AL.
while new developed Zn–S–I cycle shows best exergy
efficiency.
51,156-160
2.2.2 | Medium and high temperature
electrolysis
Electrolysis processes above 100C are considered as
high temperature electrolysis, and when the process
reaches 2500C,itbecomesspontaneous,asthermoly-
sis. In general, the electrolysis thermal processes are
divided into medium temperature (100C to 300C)
and high temperature (600C to 1000C).
63
This type of
hydrogen production from water splitting is most use-
ful when there is already a heat source nearby, such as
the heat expelled from a nuclear power plant.
161
One of
the challenges in the electrochemical cells' design for
medium-high temperature electrolysis is the lack of liq-
uid electrolyte.
162
It was 1937 when this kind of cells,
called zero gap cells, was applied for the first time to
high temperature electrolysis cells - tests were run at
900C.
63
However, curiously after this first approach to
thermal electrolysis, zero-gap cells have become more
popular in polymer electrolyte membrane (PEM) cells
for electrolysis and their reverse, PE fuel cells, and
recently even for alkaline cells
163
(alkaline and PEM
electrolytic cells will be described in Sections 2.3.1 and
2.3.2 respectively).
After this early achievement, it was necessary to wait
until the late 70s to find works focused on studying the
hydrogen production from direct water splitting at high
temperature.
164,165
In these works, both Nakamura
164
and Ihara
165
agree in their theoretical formulation for the
process of hydrogen production using solar heat. Con-
temporaneously, other authors direct their efforts in
defining aspects related to the solar absorbers like the
cavity design
166
and the coating system
167
to increase its
solar absorption capability.
Regarding the scientific literature, it could be said
that the 80s would become the decade where the water
splitting techniques based on temperature spread out
toward both medium and high temperature. The first
one, water electrolysis at medium temperature attempts
to merge the advantages of materials stability employed
in the cell from conventional water electrolysis with
quick kinetic reaction from high temperature. As it is
known, cell voltage descends with temperature, but it
does not depend on operating pressure. However, com-
bining temperature and pressure, water can remain in
liquid phase up to 235C at a pressure of 30 bar.
63
This
was the aim of Abe et al,
168
when in 1983 contributed to
the Sunshine Project helping to resolve Japan’s energy
shortage problems. The authors proposed a high-
pressure, high-temperature new design for electrolytic
cells. The cells were made from porous polytetrafluor-
oethylene (PTFE) impregnated with potassium titanate
and they could operate above 120C and 20 bar with an
efficiency of around 90%. After this, other authors
focused their works on increasing the operating tempera-
ture but analysing previously the possibilities that each
cell technology offers.
169
Then, PEM technology for
medium-temperature electrolysis was discarded up to the
90s because of its cost. Other researches focused on the
modification of alkaline cells; for example, Divisek et al
put in work cells based on NaOH/LiOH molten mixtures
at 350C, with hydrogen yield of 100% and 0.5 Acm
2
.
170
Finally, it was also possible to operate with solid electro-
lyte cells giving rise to the so-called high-temperature
electrolysis.
Doenitz et al
171
studied the effects of high tempera-
ture on the production of hydrogen from a water
source in 1980. Their goal was to improve upon the
contemporary efficiency of electrolyzers in 1980—
which were 75% with respect to the electrical energy
input, using high-temperature methods. Their group
found out that for the anode of these high-temperature
electrolytic cells, it was necessary to use certain classes
of materials to prevent undesirable side reactions or
impurities in the hydrogen gas being produced. Their
study was carried out with tubular structure cells oper-
ating at 400C to 1000C, and the traditional liquid
electrolyte was replaced by yttria-stabilized zirconia
(YSZ), Figure 8.Thismaterialwaschosenforitsdesir-
able qualities such as cost and temperature-related
stability.
Doenitz et al developed this work in 1975 in collabo-
ration with the German Bundesministerium fiir For-
schung und Technologie in the German HOT ELLY
project, and this tubular design was used 10 years later
by Siemens to define the structure of its Solid Oxide Fuel
Cells (SOFCs).
172
The overall efficiency came upon 40%
to 50% considering co-generation; the heat produced by
the product gases was reused to warm the chamber and
this made the system to operate at its optimum
temperature.
FIGURE 8 Scheme of a solid oxide electrolytic cell
CAPARRÓS MANCERA ET AL.9
In the main reviews of the hydrogen production tech-
nologies of the 1980s,
173
high-temperature vapour elec-
trolysis stands out as a technique to be further developed
in the next decade. Data of very high efficiency, com-
pared to alkaline electrolysis, are collected, although it
still cannot for reliability as high-temperature vapour
electrolysis has only been tested in the laboratory. In the
documented tests, for the temperature of 1000C, a cur-
rent density of 5 kAm
2
is obtained, for a voltage of
around 1.2 V, and up to 7 kAm
2
, for a voltage of 1.4 V.
The goodness, from the efficiency point of view, of high-
temperature water splitting can be compared with alka-
line electrolysis, where, to obtain similar current densi-
ties, a voltage between 1.9 V and 2.3 V is required.
In 1991, the Research Institute for Scientific Measure-
ments of Japan carried out studies on the production of
hydrogen from high-temperature steam electrolysis,
using solar energy as a heat source for the process, with
temperatures of up to 1000C.
174
Given the high tempera-
tures, a solid electrolyte is used, manufactured in ZrO
2
+-
8 mol% Y
2
O
3
ceramic, for its stability with high
temperatures and its high ionic conductivity. It was con-
nected to the electrodes coated with porous platinum.
During the thermal electrolysis process, the water vapour
is delivered using Argon as a gas vector. The study con-
cludes that high-temperature electrolysis obtains an effi-
ciency of 98%, electrochemical efficiency of 71% and
overall efficiency of 20% to 28%, which is high in compar-
ison to the efficiency of the photovoltaic electrolysis of
those years, 10% to 15%.
175
High efficiency is achieved
from an affordable source of heat, and the development
of finer electrolytic cells is proposed as a challenge,
together with more efficient porous electrodes.
In subsequent studies, the question of the improve-
ment of the solid membrane for electrolysis is addressed.
In 1995, Naito showed experimental results of a system
that makes use of a ZrO
2
–TiO
2
–Y
2
O
3
(Ti-YSZ) mem-
brane.
176
Using high temperature, vaporized water was
dissociated while oxygen permeated through the mem-
brane, by the difference in partial pressure of oxygen. The
electrical properties of the system (Ti-YSZ) exhibit high
electronic conductivity at high temperatures under low
oxygen partial pressures.
177
Using the Ti-YSZ membrane
higher hydrogen production was achieved, compared to
the amount obtained with the YSZ membrane. This result
indicates that as Ti-YSZ has higher electronic conductivity,
then it is better as a membrane. The system is simplified,
as there’s no requirement of electrodes or electric power.
In the late 90s, studies for high-temperature water
splitting continue with ferrites membranes.
178
Taumaura
et al applied a mixed powder of MnFe
2
O
4
with different
mole ratios of CaO (or Na
2
CO
3
) to ferrite at 1000C. This
was considered as further progress in direct solar energy
conversion to hydrogen production, as the required tem-
perature was lower than the typical two-step water split-
ting by that time (1200C to 2000C).
Technology reviews of the 90s consider thermolysis as
one of the three main electrolysis techniques, with a
mature level. Main research focus was electrochemical
stability of the electrode material, that could give a better
context to make the technology commercially competi-
tive, with the use of metal silicides.
179
In the late 90s, in terms of improving this technology,
Gomez et al
180
classified the areas of interest in Solid
Oxide Electrolytic Cells (SOECs) for hydrogen production
based on water splitting in three groups. (1) Current den-
sity, because at that time cells suffered delamination
above 1 A/cm
2
; (2) stack design and material type; and
finally (3) manufacturing process, to make the whole pro-
cess more economically feasible.
In the 2000s, Colomban et al test perovskite ceramic
membranes for hydrogen production, with studies to
determinate their physical and chemical behaviour, as a
needed step to create successful commercial applications
for the industry.
181
This gives conclusions about the
importance to difference protonic species adsorbed on a
membrane surface and the bulk protons.
Industrial solutions are presented in
182
for high tem-
perature electrolysis (800C to 1000C), with Y
2
O
3
+-
ZrO
2
electrolyte, decrease in the electrical energy of up
to 40% and production flow rate of 2.5 to 3.5
kWh/Nm
3
H
2
. The expected cost varied around 800 and
1000 €/kW, with efficiencies up to 80%.
About SOEC components,
183
the studies show Yttria-
stabilized zirconia (YSZ) as the best option for the high
temperature electrolyte, thanks to high conductivity, low
cost and chemical compatibility with other component
materials. Meanwhile, for intermediate temperature, high
ionic conductivity makes strontium and magnesium co-
doped lanthanum gallate LaGaO3 (LSGM) as suitable for
the electrolyte, although nickel (Ni) is avoided due to reac-
tion with LSGM. For both cathode and anode, Ni-YSZ and
LSM-YSZ are the most used materials, with some recent
studies indicating that lanthanum strontium ferrite (LSF)
would be a better option than lanthanum strontium man-
ganite (LSM). The SOECs planar form is preferred because
of manufacturability and performance factor, although
tubular form presents mechanical strength.
Considering that manufacturing costs of SOEC
technology-based cells for hydrogen production are
impractical for commercial interests, last decade is char-
acterized by studies focused on improving the efficiency
of SOEC technology-based hydrogen production technol-
ogy. Petipas et al
184
discuss in 2014 the benefits of using
an outside heat source to provide much of the energy
needed for the solid oxide cells for water splitting to move
10 CAPARRÓS MANCERA ET AL.
forward. The study was based on heating the water into
steam and then fed it into the SOEC electrolyzers. Once
the 75% reaction is achieved, the hydrogen/water mix is
fed out the electrolyzers to be cooled down to 25C,
which then removes the rest of the unreacted water in
liquid phase, leaving behind the pure hydrogen. Next, the
hydrogen is compressed to 3 MPa and cooled to return to
room temperature. The achieved efficiency was 89% and
authors suggest the possibility to improve it near 100%, if
the waste heat produced by cooling the hydrogen is
reused in the next cycle (to heat the water into steam).
Late 2000s reviews consider the system efficient as the
main influence on the hydrogen production cost,
182
esti-
mating that the cost could be reduced from 50 $/GJ in
2010 to below 20 $/GJ in 2030,
185
to make it more consid-
erable for commercial purposes. Additional researches
are being focused on materials and form for SOEC
technology,
183
microstructural modification,
186
degrada-
tion and performance,
187
and new materials doping.
188
Recently, composite Cu-Cl medium-temperature elec-
trolysis membranes
189,190
and the use of ammonia based
thermochemical energy storage and layered-perovskite
oxides electrodes for high temperature electrolysis stand
out.
191-193
Heat and gas route improved schemes show
energy demand reduction, especially in SOEC.
194
The advances discussed for medium and high temper-
ature water splitting are synthesized in the scheme of the
main advances and historical achievements in thermoly-
sis technology, Figure 9.
2.3 |Electricity. The most used source to
obtain the highest hydrogen production
rates from water
Electrolysis is defined as the chemical decomposition of
water produced by passing an electric current through a
liquid or solution containing ions.
195
The history of water
electrolysis started in 1789, with van Troostwijk and Dei-
man research on the decomposition of water by static
electricity, followed in 1800 with Nicholson and Carlisle
using the voltaic pile to decompose water into hydrogen
and oxygen.
196
Depending on the electrolyte nature, it is
possible to talk about alkaline electrolysis or PEM (poly-
mer electrolyte membrane) electrolysis.
2.3.1 | Alkaline electrolysis
An electrolytic cell has both electrodes, anode and cath-
ode, submerged in a liquid alkaline electrolyte,
Figure 10.
197-200
Despite the fact that the discovery of
electrolytic water splitting occurred in acidic water, the
alkaline environment is preferred in industrial plants,
since corrosion is more easily controlled and construction
materials are cheaper than the ones used in in acidic
electrolysis.
196
Regarding the history, in 1902, more than 400 indus-
trial water electrolyzers were in operation, and in 1939,
the first large water electrolysis plant with a capacity of
10 000 Nm
3
H
2
/h went into operation, meanwhile in
1948, the first pressurized industrial electrolyzer, by
Zdansky-Lonza, was built.
196
Based on experimental results, scientists soon were
aware of the common problems that alkaline electrolysis
presents: bubble effect, material stability and current
density levels. Then, in 1976, Graziotti analysed how the
power level supplied to the cell had a clear influence on
the gas bubble effect: higher power, more hydrogen bub-
bles on the cathode, and consequently the cell efficiency
decreases.
201
As it can be deduced, this is an unsolved
problem that challenge to solve in.
64
Two years later,
Lecoz and Gras showed that using chrysolite as a dia-
phragm could be stable against corrosion up to 180C
using particular concentrations of soluble potassium sili-
cate. Then, at the next World Hydrogen Energy
FIGURE 9 Main advances and historical achievements in thermolysis technology
CAPARRÓS MANCERA ET AL.11
conference, Bailleux et al presented in 1980, that there
was an insignificant amount of change in the physical
properties of the diaphragm materials when run for
6400 h at 120C, which shows that at that time the mate-
rial was a desirable one.
201
In the 70s, nickel along with stainless steel and irid-
ium received attention for the first time to be used in the
cathode.
202
Appleby et al experimented an electrolyte cell
with a Teflon-bonded electrode as the anode, the separa-
tor was asbestos (after that, asbestos was dismissed
merely for not being particularly efficient besides its car-
cinogen nature) and a solution of 24% to 34% of KOH.
For the cathode, authors tested different materials like
nickel, stainless steel and iridium alloy. Obtained results
shown that iridium performed the best in alkaline solu-
tions (current densities around 400 mA/cm
2
, doubling
the values achieved up to then), but this material would
not be cost-effective. By contrast, stainless steel despite it
was a cheap material, its inability to resist corrosion in a
highly alkaline environment immediately eliminates it
from the list of possible solutions. This has involved that
nickel is being the most commercially extended material
in cathode for decades.
203
In a review of 2010, Zeng et al
have pointed out that there have been many advances in
terms of nickel based alloys for use as electrodes in order
to reduce the amount of damage caused with each usage
of the cell, and to therefore prolong the lifespan of the
electrodes.
64
Next decade of the 80s, researches were driven
towards the search of separator materials.
204
Potassium
titanate and polyantimonic acid were shown as high-
performance separators, from inorganic material, for
alkaline electrolysis, while oxide-coated metallic is well
considered, for long-term use. Between organic materials,
polysulfone and polyphenylene sulphide were promising
technologies, reaching stability temperature up to 120C
to 200C, with needed improvements in surface proper-
ties, to solve their poor wettability. Vandenborre et al
showed NiCo
2
O
4
and NiCo
2
S
4
as the best anode and
cathode electrocatalysts.
205
In 1983, Noranda Research Centre
206
shows that Tef-
lon with potassium titanate has the lowest resistance fac-
tor as separator, although the fragility and hydrophobicity
must be solved. Additionally, stability and good electroca-
talytic performance are shown with nickel electrodes,
plasma-sprayed with nickel/aluminium or nickel/stainless
steel powders, and the most promising anode type tested
to date was plasma-sprayed Ni/SS alloy.
Researchers from the late 80s show industrial
achievements in cells with ceramic diaphragms and
galvanically-deposited Raney nickel electrodes, with
operating temperature of 100C to 120C, pressure
between 1 and 5 bar, and energy consumption of 3.8
kWh/m
3
H
2,207
at 0.4 A/cm
2
and 100C.
Research during the 90s also checked important elec-
trodes improvements, at an affordable cost, using an elec-
trocatalytic surface layer. Also, the increase of surface
with Teflon-bonded diaphragms and Ni-layer alloys for
electrodes. This improvement represents an energy input
decrease of 0.96 kWh/m
3
H
2.208
In the mid 90's, there were studies about anodic and
cathodic behaviours of aluminium, iron, mercury steel
(HgSt), chrome-nickel steel (CrNiSt) and platinum.
209
In
1996, Hu et al developed a multilayer structure cath-
ode.
210
Meanwhile, Zirfon is also tested as a porous com-
posite separator material composed of a polysulfone
matrix and ZrO
2.211
New electrocatalysts are tested in the beginning of the
2000s, such as MmNi
3.6
Co
0.75
Mn
0.42
A1
0.27
alloy,
LaNi
4.9
Si
0.1
alloy and Ti
2
Ni alloy, and nickel-
molybdenum coatings. The results demonstrated that the
increase in molybdenum content causes activity for
hydrogen evolution to increase as well.
212
Electro-oxidation of ammonia on platinum
(Pt) electrode was studied by Zhou et al,
213
showing that
the cell efficiency can be up to 45%, while ammonia elec-
trolysis with Pt electrode. Electrolytic cell efficiency could
slightly be increased with higher KOH and ammonia
concentrations.
In the late 2000s, the long-term stability of NiCoZn
coating for hydrogen evolution reaction (HER) was
investigated. It was found that the NiCoZn coating had
a compact and porous structure.
214
Alternative studies
tested the properties of multiwalled carbon nanotubes
(MWNT) in anodes, obtaining enhanced exchange cur-
rent density when compared to graphite anodes. The
hydrogen production rate almost doubled that the rate
obtained with traditional graphitic carbon electrodes,
while the same overpotential was applied. This is caused
by the interaction of OH
ions with defects on the nano-
tubes. Therefore, the activation energy of dissociation of
OH
to O
2
is lowered, which significantly increases
energy efficiency.
215
FIGURE 10 Scheme of an alkaline electrochemical cell
12 CAPARRÓS MANCERA ET AL.
In the recent years, 2015-nowadays, researchers try to
fit together the different millstones that have been sepa-
rately achieved along the previous decades. Then, regard-
ing membrane in the cell structure, poly
perfluorosulfonic acid (PFSA) material was evaluated as
one of a few polymer membrane types that combine
excellent alkaline resistance with extreme hydrophilic-
ity.
216
Other studies also consider titanium (IV) oxide
composite membrane to limit the usage of an asbestos
separator, as it is hazardous for health among its low effi-
ciency against the materials studies in the last two
decades. This achieves 50% higher current density and
hydrogen purity against asbestos.
217
The uses of carbon
nano-structure composites for electrodes has been stud-
ied, obtaining the highest hydrogen production rate of
487 L/h-m
2
H
2
with the composite GC 73, composed by
70 wt% graphene and 30 wt% carbon nanotubes
(CNT).
218
About electrolytes, there have been new
improvements with the use of alkaline zinc hydroxide
solution, composed of sodium zincate and potassium zin-
cate in NaOH and KOH solutions, respectively. Results
show that the application of these solutions can improve
the hydrogen evolution rate minimally by a factor of 2.74
(with the use of sodium zincate) and 1.47 (when potas-
sium zincate is applied) in comparison to the typical alka-
line stacks.
Apart from the joined advanced on membranes and
electrodes, a combined-design cell was proposed by
Marini et al
219
in 2012. It consisted on two cell designs: a
zero-gap configuration and an anion exchange mem-
brane (AEM), similar to PEM cells from the point of view
of structural configuration. Additionally, the zero-gap cell
is immersed into a confined electrolyte, while the AEM
cell requires a liquid electrolyte flow,
220
Figure 11. The
electrodes are made of Raney-Ni doped with Mo (cath-
ode) and Fe (anode) filled between by a felt made of cel-
lulose meant to prevent the formation of gas bubbles.
The system is placed into a large, temperature-controlled
KOH bath to ensure smooth functionality. The anode is
filled with KOH electrolyte solution, transporting the
produced oxygen through electrolyte flow. The dry cath-
ode produces hydrogen from water permeating the mem-
brane from the anode.
221,222
Satisfactory results show
current densities around 500 mA/cm
2
and efficiency
above 60%. The first design mitigates the bubbles formed
in the internal space of the cell while the second one
reduces ohmic losses and make the system more stable.
In this last case, authors propose to eliminate the bubbles
in the AEM cell, in order to apply an overpressure on the
electrolyte compartment; which will cause an effect simi-
lar to a “sneeze”. These new proposals are being called
advanced alkaline electrolysis. In this context, researches
about AEM membranes using non-precious catalysts, dif-
ferent KOH electrolytes and operating temperature
ranges are showing potential for AEM electrolysis in the
next decade, 2020 onwards.
223-225
Alternatively, there has been growing interest in Mem-
braneless Electrolyzers (ME), where the use of a physical
membrane as such is avoided. The process occurs in a
chamber with an alkaline electrolyte and the reaction is
carried out with porous electrodes as the fundamental
axis, either with flow-by electrodes or flow-through elec-
trodes, which are the two most relevant types of ME today.
Recent reports prove better performance in flow-through
electrode MEs.
226
Membranes are one of the challenges in
electrolysis, due to their rapid degradation, which is why
MEs are planned as a prominent alternative with low cost
and enhanced lifespan. With first membraneless electroly-
sis cell patent in 2008, and scientific publications from
2015,
227
maximum efficiency of ME cells range between
64% and 82%, comparable to current alkaline electrolyzers.
The main challenge of the technology focuses on the elec-
trodes, and on finding a balance between their porosity
and conductivity.
228,229
Current novel alkaline approaches focus on compos-
ite membranes like polybenzimidazole incorporating gra-
phitic carbon nitride nanosheets (PBI/g-C
3
N
4
) and pore-
filling polytetrafluoroethylene/layered double hydroxide
(PTFE/LHD),
230-234
and Ru nanoclusters on nitrogen-
doped graphene as catalyst.
235
For AEM, Ni-based cata-
lysts and polysulfone (PSF) or polystyrene
(PS) hydrocarbon polymer backbone membranes they are
the most accepted in more recent developments.
236-239
2.3.2 | Solid Polymer Electrolysis (SPE) or
Proton Exchange Membrane (PEM) electrolysis
Solid Polymer Electrolysis (SPE) technology, also called
Proton Exchange Membrane Electrolysis (PEM), works
by a pair of electrodes (anode and cathode) being pressed
against a solid polymer electrolyte (SPE), which is typi-
cally Nafion, Figure 12. Water and electricity are fed
FIGURE 11 Scheme of an anion exchange membrane
(AEM) cell
CAPARRÓS MANCERA ET AL.13
through the cell to produce hydrogen at the cathode, and
oxygen at the anode. Usually, there are separators
between the electrodes and the solid electrolyte, which
are typically some kind of bipolar plate made of platinum
(for cathode) and iridium (for anode) or a metal similarly
resistant to corrosion.
240-243
Similar to previous water splitting techniques, SPE
technology began to receive attention in the middle of
the last century, 1959, when Grubb presented its work
where he developed an experimental study about ion-
exchange membrane.
244
At that time, the only applica-
tion found in this type of membranes was its use as mem-
branes for the purification of salt water by electrodialysis.
Fortunately, Grubb himself foreseen the potential of
these membranes as solid state battery electrolyte (which
then led to PE fuel cells).
After this, in 1966, Nuttall et al developed the concept
of SPE electrolyzer for General Electric.
245,246
The pro-
posal was presented as a way to replace the alkaline tech-
nology for large-scale hydrogen production.
247
The
authors, aware of the system limitations at the time, pre-
dicted that this kind of system would achieve its best per-
formance through the year 2000.
The positive aspects of PEM water electrolysis are
29
:
higher current densities than alkaline electrolysis (2 Acm
2
vs 500 mAcm
2
), low gas crossover rate, allowing the SPE
electrolyzer to work under a lower partial load range (0%
to 10% vs 20% to 40% for alkaline electrolyzers), and com-
pact design with solid structural properties that achieves
high operational pressures. By contrast, the main draw-
backs come from the SPE technology nature: special mate-
rials like noble catalysts to withstand adverse conditions
(pH 2, cell voltage 2 V and current density 2
Acm
2
), and thicker although more resistant membranes
to support high pressures. These two issues (catalyst and
membrane) have led the way in PEM electrolysis research
from the beginning of the 70s up to the present day. Addi-
tionally, other secondary aspects like current collector and
separator plates involved in the cell assembly have also
received attention from researchers.
Already in the 70s, the General Electric Solid Polymer
Water Electrolyzer,
248
that was operating at about 120C
to 150C, seemed most promising in achieving the high
energy efficiency and low capital cost goals, capable of
operation at high current densities (over 1 Acm
2
)ata
cell voltage of less than 1.8 V. As a consequence of the
highly acidic environment, noble metals must be used as
electrocatalysts. With funded programs, Nafion Du Pont
developed 5 MW prototypes.
249
In the 80s, as General Electric SPE system contuse its
development and improvement,
250
new research in mem-
branes and electrocatalysts are developed. New cation
exchange membranes were developed, produced by radi-
ation grafting of styrene groups on a polyethylene matrix,
followed by chemical sulphonation of the resulting poly-
mer. These membranes were supposed to have much less
cost than the perfluorinated-hydrocarbon-based polymers
which, so far, were the most used for SPE.
251
Reviews form the 80s show that SPE issues include
high cost of the Nafion membrane and poor stability of
inexpensive anodic current collectors and catalysts in the
acidic environment. An important phenomenon limiting
achievable efficiency is back-diffusion of product hydro-
gen to the anode, where it is oxidized. This resulted in
energy losses that were reported by General Electrics to
be high as 10% at 150C, and by Brown Boveri laborato-
ries to be around 6% at 130C.
252
This can explain why in
that decade, alkaline electrolysis was still considered as
the best option.
Next, in the 90s, new procedures for the preparation
of SPE electrocatalyst composites were developed, where
microparticles, based on noble metal electrocatalyst, were
precipitated simultaneously inside and outside near both
surfaces of a perfluorinated ion-exchange membrane,
applying chemical reduction of cationic precursor salts.
Nafion-based porous structures, with developments in
Grenoble, can be coated on both faces of the membrane,
and then catalysts is performed.
253
Nafion is the most
used electrolyte of the decade, with some progress in
develop alternative membranes, like the Dow Chemical
membranes.
254
Studies of the polymer electrolytes also tried to reach
higher temperature materials, while maintaining hydro-
lytic stability. Liquid crystal polyesters, polybenzimida-
zoles and some polyimides show stability at 200C.
Polyphenylene sulphides, polysulfones, polyketones and
some polyimides showed stability, in reasonable grade, at
300C, while no polymer was found stable enough at
400C.
255
In the early 1990s, around 1993, SPE technology
(introduced by General Electrics in the 70s), start to be
FIGURE 12 Scheme of a Solid Polymer Electrolyte (SPE)
electrochemical cell
14 CAPARRÓS MANCERA ET AL.
named as PEM (proton exchange membrane) by the sci-
ence literature, and the meaning is extended as a system
coating or pressing two electrodes onto a membrane used
as electrolyte being recognize as the most promising can-
didate for low temperature production.
256
New research
works go ahead with the improvement of the materials
stability, as membrane water electrolyzers remain more
expensive than alkaline, using IrO
2
/Ti electrodes on both
sides of Nafion membranes, what obtain much lower
anodic over voltages and better stability, due to lower
cathode sensitivity to poisoning, that Pt electrodes.
256
In the late 90s, studies about materials continue to
achieve higher temperature than the 125C to 150C
range, with polymers as polyether ketones, polyether sul-
fones, polybenzimidazoles and polyphenyl quinoxalines.
SPEEK (sulfonated polyetheretherketone), showing sta-
ble performance, up to 300C is comparable to existing
commercial ionomer membranes.
257
Great advanced in
SPE electrolyzers can be found in this decade, especially
focused on modelling and simulation.
258,259
In the 2000s, more of these simulations allow to study
electrolysis performance, showing that the high anode
overpotential is the limiting factor.
260
Reviews show that
a lot of R&D work was done in the field of PEM electroly-
zers, but high price has limited their mass production,
due to membrane, noble metals electrocatalyst (Ru, Ir,
Pt), water systems and constructional material (Ti). A
composition at 40% to 50% of RuO
2
is similar to activity
of pure IrO
2
, and parameters of electrolysis with
RuO
2
(30%)-IrO
2
(32%)-SnO
2
(38%) as anode electrocatalyst
with platinum 0.8 mg/cm
2
are almost similar to electroly-
sis with iridium anode electrocatalyst with 2.0 to
2.4 mg/cm
2
. Also, electrolysis at increased pressure of
30 bar, reached by PEM, shows an improvement of volt-
ampere characteristic in comparison with electrolysis at
atmospheric pressure.
Active surface area, electronic resistance, specific
activity and structure are main parameters properties to
be enhanced if high efficiency and performance of PEM
want to be achieved. Development in new electrocatalysts
showed the best result with an Ir
0.6
Ru
0.4
O
2
anode and
20 wt% Pt/C cathode, with 1.567 V of cell voltage at
1 Acm
2
and 80C using Nafion as membrane. This pro-
vided cell efficiency of 94.4% and an energy consumption
of 3.75 kWhNm
3
H
2
at 1 Acm
2.261
Also novel
manufacturing process for catalyst-coated membrane
(CCM) was developed to manufacture membrane elec-
trode assemblies (MEA) for PEM, revealing that the
sprayed Nafion layers are very effective for increasing the
reaction interface between PEM and the electrode cata-
lyst layer.
262,263
At the end of the decade, scientific works are focused
on polymeric membranes properties. Glassy polymers
were generally used for manufacturing hydrogen-
selective membranes. Size variations on these mem-
branes adjust the discrimination ability. Main commer-
cial polymers used for H
2
-selective membranes are Ethyl
cellulose or Polyetherimide. H
2
-selective membranes are
able to tolerate higher compression and temperature.
264
In last decade, 2010–2020, with 50 bar studies, non-
noble electrocatalysts cobalt-glyoximes. Although less
efficient than platinum and also less sensitive to poison-
ing, they can replace it, reducing costs by almost a factor
of two and paving the way for the large-scale develop-
ment of PEM technology.
265
Recent scientific works
53
show that the most com-
monly used membranes are perfluorosulfonic acid poly-
mer membranes such as Nafion, Fumapem, Flemion and
Aciplex. Nafion membranes (115, 117, and 212) have high
current densities (2 A/cm
2
), high proton conductivity, high
durability and good mechanical stability. Pours titanium
plates are being used as current collectors due to good
electrical conductivity, mechanical stability, and corrosion
resistant, with promising results. They act as current col-
lectors and gas diffusion layer (GDL) for both sides of the
MEA, enclosed by bipolar plates. Also, titanium grids/
meshes/felts, carbon current collectors and stainless-steel
grids are used, but the electro-chemical performance is
lower. Separator plates are typically made of titanium,
stainless steel and graphite but these materials are high
cost, with some operational drawbacks, so many studies
are focused on precious metal coatings and alloys to pro-
tect the titanium plates. Separator plates and current col-
lectors are responsible for the 48% of overall cell cost.
Noble metal-based electrocatalysts are used such as Pt/Pd-
based catalysts as cathode towards the hydrogen evolution
reaction (HER) and RuO
2
/IrO
2
catalysts as anode for oxy-
gen evolution reaction (OER). Then, PEM electrolysis con-
siderable achievements have been made in commercial
criteria, but overall water splitting-based hydrogen produc-
tion is resulting in only 4% of global industrial hydrogen
consumption, so for future research direction, it should be
proceed to achieve a more cost-effective solution. That is
why it is possible to find researches focused on catalyst
separation, recovery and recycling,
266
studies to analyse
the PEM degradations issues,
267
and also PEM cells that
can operate at higher current density, with typical 1 to
3A/cm
2
,upto10A/cm
2.268
In these years, the new concept of Hydrogen Concen-
trator (HyCon) yields high efficiencies combining multi-
junction solar cells with proton exchange (PE) membrane
water electrolysis, using a titanium hybrid fibre sinter
function both as a porous transport layer and flow field.
The cell shows high performance with a voltage of 1.83 V
at 1 A/cm
2
and the HyCon module is capable of achiev-
ing an efficiency of 19.5% from sunlight to hydrogen.
269
CAPARRÓS MANCERA ET AL.15
One of the newest works is published by Kaya et al in
2017.
270
Authors develop a numerical model validated
with experimental results where the effects on the cell
performance are analysed of two types of catalysts for the
anode (Pt and Pt-Ir). The results demonstrated that mem-
brane thickness affects the performance of the cell, the
thinner the better, and that the temperature increases the
production of H
2
within the cell.
Currently, one of the largest parts of the total cost of a
PEM electrolytic stack is the bipolar plates that must be
placed between each cell to evenly distribute the charge
throughout the entire stack. For now, they are often
made of titanium, which is very expensive, in comparison
to the manufacturing costs of other materials. The reason
other metals are not used is the production of oxides that
create ohmic resistance, which develop on the surface
due to the highly corrosive and acidic environment of the
typical PEM cell. Lædre et al tested in 2017 several mate-
rials in order to determine which functions the best to be
used as bipolar plates in a PEM electrolytic cell.
271
Even
three types of steel were tested for their functionality, but
it must be noted that they did not display an increase in
current density until 1.5 V and that could be caused
either by corrosion or by the undesired oxygen evolution
reaction.
Arrived at this point, it can be understood what hap-
pens inside a PEM electrolytic and how it is possible to
improve its performance making some changes in cata-
lyst, membrane, bipolar plates or operating conditions.
However, what about external factors like power supply,
what effects does the interrupted power supply have on
the cell? As it was mentioned, in alkaline electrolytic
cells, the lower partial load range must be higher than
20% to 40% (to beat the gas crossover rate in liquid elec-
trolyte), but in PEM electrolytic cell, this drops to 0%. In
this case, it would be interesting to know the changes
brought about by fluctuation in the electrical supply to
the cell. This issue has been studied between 2010 and
2014 by several authors who present systems that have in
common a PEM electrolyzer supplied by a photovoltaic
panels.
272,273
Nevertheless, these works disagree in the
degradation factor. Rakousky et al were the first who
identified an operating window for the PEM electrolytic
cell with minimum degradation rate.
274
He analysed the
volatile current density effect that mimics the behaviour
of the solar cells, from which these electrolytic cells are
expected to get their energy. The testing of the degrada-
tion lasted for about six weeks and was run between zero
and 2 A/cm
2
, to discover the causes of cell degradation.
The metric used to mark cell degradation was the
increase in cell voltage. As it was expected, the cell run
under the constant current of 2 A/cm
2
was the one that
experienced the most degradation (0.19 mVh
1
). The
good thing about these tests was that, compared to the
cell kept at constant high current density, the cells sub-
jected to a dynamic current density profile experienced
very little damage (0.06 mVh
1
).
Apart from these research works focused on improv-
ing the different parts that integrate the structure of a
PEM electrolytic cell, recently there are arising other
papers dedicated to exploring new perspectives for water
electrolysis. Then, in 2016 Schalenbach et al developed
an exhaustive comparative experimental study about effi-
ciency in alkaline and PEM electrolysis. Authors con-
clude that modified alkaline cells, with thinner separator
diaphragm and microporous electrodes enable better effi-
ciency than those cells with acidic Nafion membranes.
275
These modified structure alkaline cells consist on the
AEM cells.
In the same way and also in 2016, in an attempt to
estimate the round-trip energy efficiency of a hydrogen
redox battery, Lamy proposed two alternatives for a Unit-
ized Regenerative Fuel Cell (URFC) for storage of inter-
mittent energy.
276
The URFC consists of a (PEM
electrolyzer +PE fuel cell), and a (SOEC electrolyzer +
SOFC fuel cell). As it can be deduced, both systems have
in common their high cost in comparison with alkaline
technology. Against what might be expected, experimen-
tal results show better efficiency for PEM URFC (40% to
50%) than for SOC URFC (20% to 35%).
It is important to note that, as happened in the 90s,
when the scientific community accepted the most wide-
spread use of the PEM nomenclature (proton exchange
membrane) to the detriment of the odder SPE terminol-
ogy, introduced by General Electrics in the 1970s,
256
a
new transition takes place in its terminology in the
decade of 2010. It is verified in the scientific literature
that the extended term PEM, still very widespread, now
refers to polymer electrolyte membrane,
277-279
with some
mixed nomenclature meaning referred as polymer
exchange membrane,
280
instead of its previous typical use
of proton exchange membrane, given that it is this char-
acteristic that sets it apart from other hydrogen produc-
tion technologies. In addition, the term membrane is also
excluded from the nomenclature associated with the
technology, being referenced only simply as PE (polymer
electrolyte), with extended use of PEFC (polymer electro-
lyte fuel cell),
281-283
where it differs from the term PEM,
which remains used in this case to speak only of the poly-
meric membrane electrolysis,
284
and not of the devices or
technology, as it has been done until now.
These days, minimum membrane thickness and max-
imum operating temperature are investigated with great
potential in hydrocarbon membranes. Iridium is mature
and stable as a catalyst, with improvements demon-
strated mainly in mixtures with ruthenium, as well as
16 CAPARRÓS MANCERA ET AL.
cobalt or nickel. Titanium suboxide and antimony oxide
show promise. Nanostructured thin film electrodes
(NSTF) stand out.
285-290
The main advances throughout the history of alkaline
electrolysis and PEM have been synthesized in the
scheme of Figure 13.
3|DISCUSSION
According to the chronological-technical review carried
out about water splitting techniques, the main advances
and milestones achieved in the different technologies
throughout the different decades of scientific and
technological development for hydrogen production are
summarised in Table 2.
It should be clarified that these data refer to the most
current references analysed in this review, in terms of
more extended values and ranges. That is why, although
there are case studies where higher values are obtained,
as mentioned in the document, these cases constitute
examples of an experimental prototype that are not con-
sistent with the level of maturity and efficiency required
by the different water splitting techniques nowadays. It
should also be noted that not enough scientific literature
has been found referring to some of the values of the
specifications of water splitting technologies, such as
temperature and pressure in photolysis, as well as the
FIGURE 13 Main advances and historical achievements in electrolysis technology
TABLE 2 Technical characteristics of the water splitting-based hydrogen production alternatives
Water Splitting Technique Ref.
Cell
Prod.
(Nm
3
/h)
a
Cell
Volt.
(V)
b
Cell
Curr.
(A/cm
2
)
c
Cell Temp.
(C)
d
Cell
Pres.
(bar)
e
Cell
eff.
(%)
f
Photolysis Photocatalitic systems 93,291,292 - - - - - 0.5 to 12
Photoelectro chemical
semiconductor-based
cells
104,293,294 - 1.2 14.5 - - 2 to 18
Photobiological 138,295,296 - - - - - 0.1 to 20
Thermolysis Thermolysis cycle 152,297,298 <27 - - 550 to 2500 - 17 to 55
Medium temperature
electrolysis
63,293,299 <10 0.95
to 1.29
0.3 to 1 100–300 <30 30 to 50
High temperature
electrolysis (SOEC)
293,299,300 <10 0.95
to 1.29
0.3 to 1 900 to 100 <30 40 to 70
Electrolysis Alkaline cells 29,64,236,298 <1400 1.85 0.2 to 0.5 60 to 100 <30 59 to 82
PEM cells 29,64,236,298 <400 2 1 to 2
g
/23 50 to 90 <35 65 to 84
a
Cell Production rate.
b
Cell Voltage.
c
Cell Current Density.
d
Cell Temperature.
e
Cell Pressure.
f
Cell Efficiency.
g
Commercial units.
CAPARRÓS MANCERA ET AL.17