Plasma-Based CO2 Conversion

Item Type Book Chapter

Authors Bogaerts, Annemie;Snoeckx, Ramses

Citation Bogaerts, A., & Snoeckx, R. (2019). Plasma-Based CO2

Conversion. An Economy Based on Carbon Dioxide and Water,
287–325. doi:10.1007/978-3-030-15868-2_8

Eprint version Pre-print

DOI 10.1007/978-3-030-15868-2_8

Publisher Springer Nature

Rights Archived with thanks to Springer International Publishing;This

file is an open access version redistributed from:
https://www.differ.nl/sites/default/files/attachments/
biblio/2019_69788.pdf

Download date 2024-11-14 11:37:44

Link to Item http://hdl.handle.net/10754/668874

Book chapter for CO2Chem book “Carbon dioxide utilisation: from
fundamental discoveries to production processes”
Chapter title: Plasma-based CO2 conversion

Chapter authors: Annemie Bogaerts1, Xin Tu2, Gerard van Rooij3 and Richard van de Sanden3
Department of Chemistry,

1
Research group PLASMANT, Department of Chemistry, University of Antwerp, Universiteitsplein 1,
BE-2610 Antwerp, Belgium

2
Department of Electrical Engineering and Electronics, University of Liverpool, Liverpool L69 3GJ, UK

3
Dutch Institute for Fundamental Energy Research, De Zaale 20, 5612 AJ Eindhoven, The
Netherlands

1. Introduction
Plasma technology is gaining increasing interest for CO2 conversion. Plasma is an ionised gas,
consisting of a variety of different species, including electrons, various types of radicals, ions, excited
species, photons, besides neutral gas molecules. This reactive cocktail makes it useful for a myriad of
applications [1]. Furthermore, as plasma is generated by electrical power, and can easily be switched
on/off, this combination makes it suitable for using intermittent renewable electricity. Hence, it may
provide a solution for the current challenges on efficient storage and transport of renewable
electricity, i.e., peak shaving and grid stabilisation.
So let us consider in more detail what plasma is and which promises it carries for chemical
transformations in general and of CO2 in particular. In short, plasma is ionised gas and generally
sustained by the application of electric fields, as depicted in the cartoon in Figure 1.

Fig. 1: Plasma is ionised gas and generally sustained by acceleration of the light electrons in an electric
field (here indicated as microwaves). Ionisation is indicated by positively charged atoms/molecules (in
red) and free electrons (in yellow) and is a small fraction compared to the neutral particle density.

Energy transfer from the external electric field starts with acceleration of the free electrons.
Subsequent collisions with (blue) feedstock molecules passes their kinetic energy on. However, the
large mass difference between electrons and molecules makes momentum transfer extremely
inefficient. Instead, energy transfer occurs predominantly via excitation of internal degrees of
freedom, such as molecular vibration. On the microscopic scale, it means that the free electron
modifies the configuration of the bound electrons of the atom or molecule. Internal energy is
subsequently transferred to translational and rotational degrees of freedom, of which the rates are
highly dependent of molecular properties, cross sections, pressure and temperature. In effect, a
hierarchy in excitation of the different degrees of freedom of the system is typically found. The free
plasma electrons are hottest, typically 1-3 eV. Rotational and translational degrees are coldest whilst
molecular vibration temperatures are necessarily intermediate. It goes without saying that at all time
energy might be consumed (or released) in chemical reactions, which is the overall purpose and hence
to be optimised.

In the present context of plasma-based CO2 transformations, especially those cases in which the
strongest non-equilibrium between the different modes is found are highly interesting. These are
generally referred to as Non-Thermal Plasma (NTP). It is under the far from thermodynamic
equilibrium conditions that it is possible to intensify traditional chemical processes and to achieve the
highest values of energy efficiency [2]. In the most ideal situation, one would have room temperature
rotation and translation, whilst high vibration temperature still drives strongly endothermic reactions.
Simply spoken, this saves energy that is otherwise to be invested in these modes and likely to be lost
as heat to the environment. It has the additional advantage of inherent quenching of the reaction
products. Both aspects makes the approach particularly advantageous for thermodynamically
unfavorable or energy-intensive chemical reactions, such as CO2 splitting or dry reforming of methane
(DRM), to proceed in an energy-efficient way. The strong non-equilibrium situation is opposed to
Thermal Plasma (TP), in which all degrees of freedom are in thermal equilibrium.
The nature of the excitation process depends on the energy of the electrons. In the tail of the electron
energy distribution function, the energy is high enough to excite the heavy gas particles into higher
electronic states or even induce ionisation, as shown in Figure 2. Obviously, ionisation is required for
sustaining the plasma discharge. For efficient CO2 reforming it should not become a dominant pathway
as it is an energetically inefficient way of initiating chemical reactions. After all, ionisation of CO2
requires ~14 eV/molecule, whereas its “net” dissociation energy is ~3 eV (considering the “net”
reaction CO2  CO + ½O2). This simple consideration implies a maximum energy efficiency of at most
20% for each dissociation event via ionisation. In practice, due to the fact that only the high energy
tail would drive dissociation and all other energy input would be “lost”, it would limit efficiencies to
even lower values of ~5%.

Fig. 2: Cross sections for electron collisions with CO2 after [3]. Average plasma electron energies are
usually of few eV, exactly where the cross section of vibrational excitation peaks. This confirms the
hand waving picture of preferential vibrational excitation in low temperature plasma. Ionisation to
replenish plasma losses requires energies over 10 eV and is still possible by electrons in the high energy
tail of the electron energy distribution function. The different approaches to plasma generation vary
all in shape and mean of the distribution, which determines the balance between power deposited in
vibration versus power consumed by ionisation and other high energy excitations.

The majority of the electrons are however at lower energy, typically a few eV. These are responsible
for collisions that predominantly excite vibrational modes in the molecule. The resulting vibrationally
excited molecules will further interact with each other and exchange vibrational energy or convert
vibrational energy into translational. Of special interest here is the asymmetric stretch vibration of
CO2, which carries two important properties. Firstly, the vibrational quanta are too large to be easily
converted into translational energy in a low energy collision. Secondly, the vibration is anharmonic,
which means that the vibrational level spacing of highly excited molecules is smaller than that of
molecules at a lower level. This results in a slight preference of highly excited molecules gaining
additional quanta compared to losing it to (the majority) molecules at the first levels.

In effect, it is the asymmetric stretch vibrational mode that can be brought to the highest degree of
non-equilibrium and in which vibrational energy can be driven up along the energy scale to reach the
dissociation limit. In this ladder climbing scheme, illustrated in Figure 3, the electrons, that were
energetically "expensive" to create, are used many times to deliver energy to overpopulate the lower
asymmetric stretch levels and essentially to the bond that is to be broken up to the point where
dissociation of the molecule is achieved. It is this qualitative mechanism that has been put forward to
explain the ultimate energy efficiencies that have been demonstrated in the former Soviet Union in
the 1970s [4-9] for the net reaction CO2  CO + O2, i.e., with over 80% energy efficiency.

Fig. 3: Schematic illustration of some CO2 electronic and vibrational levels. Illustrated is stepwise
vibrational excitation, i.e., the so-called ladder-climbing process. It is initiated by plasma electron
excitation and carried by vibrational exchange up to the point of dissociation. Opposed to energetically
advantageous ladder climbing is dissociative excitation, which involves a large activation barrier.
Reproduced from [10] with permission.

The fraction of charge is usually small, often 10-5 or even less compared to the neutral species.
Ionisation is therefore not significant in the power balance and the plasma acts as a power transfer
medium, converting electric energy into internal energy of molecules.
Having explained why plasma is promising for CO2 conversion, we will briefly present the most
common types of plasma reactors with their characteristic features in the next section. Referred will
be to the non-equilibrium nature of the discharges to illustrate why some plasma types exhibit better
energy efficiency than others. Subsequently, we will discuss the state-of-the-art on plasma-based CO2
conversion, including the combined conversion of CO2 with CH4, H2O or H2. Finally, we will discuss the
major limitations and steps to be taken for further improvement.

2. Plasma reactor types for CO2 conversion

Plasma is created, in its simplest form, by applying an electric potential difference between two
electrodes, positioned in a gas. The gas pressure can range from a few Torr up to (above) 1 atm. The
potential difference can be direct current (DC), alternating current (AC), ranging from 50 Hz over kHz
to MHz (radio-frequency; RF), or pulsed. Furthermore, the electrical energy can also be supplied in
other ways, e.g., by a coil (inductively coupled plasma; ICP) or as microwaves (MW).
Three types of plasma reactors are most often studied for CO2 conversion, i.e., dielectric barrier
discharges (DBDs), microwave plasmas and gliding arc (GA) discharges. Below, we will briefly present
their working principles and typical operating conditions, to explain why they are particularly
interesting and what their current limitations are. Furthermore, besides these three major types of
plasma reactors, other plasma types are being explored as well for CO2 conversion, and they will also
be very briefly discussed. Finally, we will introduce the concept of plasma catalysis, for the selective
production of value-added chemicals.

2.1. Dielectric barrier discharge (DBD)

A dielectric barrier discharge is created by applying an AC potential difference between two
electrodes, of which at least one is covered by a dielectric barrier. The latter limits the amount of
charge transported between both electrodes, and thus it prevents that the discharge would undergo
a transition into a thermal plasma, which is a less efficient regime for CO2 conversion. The electrodes
can be two parallel plates, but a more common design for CO2 conversion is based on two concentric
cylindrical electrodes (cf. Figure 4(a)), in which the inner electrode is surrounded by a dielectric tube
with a mesh or foil electrode wrapped around it. The gap between inner electrode and dielectric tube
is in the order of a few millimeter. One of the electrodes is connected to a power supply, while the
other electrode is grounded. The gas flows in from one side, and is gradually converted along its way
through the gap between inner electrode and dielectric tube, and flows out from the other side.
A DBD typically operates at atmospheric pressure, which is beneficial for industrial applications.
Furthermore, it has a very simple design, making is suitable for upscaling, and thus industrial
implementation, as demonstrated already for ozone synthesis, by placing a large number of DBD
reactors in parallel [11].
On the other hand, a DBD has only a limited energy efficiency for CO2 conversion, typically around
10 %, with some exceptions up to 20 % [12]; see also next section. The reason is that the reduced
electric field (i.e., ratio of electric field over gas number density) is typically above 100-200 Td (1 Td =
10-21 V m²), creating high-energy electrons, which mainly give rise to electronic excitation, ionisation
and dissociation of CO2 molecules in the ground state, and this is not the most energy-efficient CO2
dissociation pathway (see below).
By introducing a packing of dielectric material in the discharge gap, the energy efficiency can in
principle be improved, due to polarisation of the dielectric packing beads resulting from the applied
potential difference. Indeed, this will enhance the electric field near the contact points of the packing
beads, and thus the electron energy [13], causing more electron impact excitation, ionisation and
dissociation, and thus more CO2 conversion for the same applied power. In addition, such a packed
bed DBD is very suitable for plasma catalysis, as will be discussed in section 2.5. However, it should be
noted that the CO2 conversion efficiency is not always enhanced in a packed-bed DBD [14, 15], due to
the competing effect of reduced residence time in the smaller discharge volume, when comparing at
the same gas flow rate, as well as the loss of electrons and reactive plasma species at the surface of
the packing material.

Fig. 4: Schematic illustration of the three plasma reactors most often used for CO2 conversion, i.e.,
dielectric barrier discharge (a), microwave plasma (b), and gliding arc discharge, in classical
configuration (c) and cylindrical geometry, called gliding arc plasmatron (GAP) (d). Reproduced from
[16]
with permission.

2.2. Microwave plasma

In a microwave plasma, electromagnetic radiation with frequency between 300 MHz and 10 GHz is
applied to a gas, without using electrodes. Depending on the configuration, there exist different types
of MW plasmas, i.e., cavity induced plasmas, free expanding atmospheric plasma torches, electron
cyclotron resonance plasmas and surface wave discharges. The latter are most frequently used for CO2
conversion. In this configuration, the gas flows through a quartz tube, which is transparent to MW
radiation, intersecting with a rectangular waveguide, to initiate the discharge (see Figure 4(b)). The
microwaves propagate along the interface between the quartz tube and the plasma column, and the
wave energy is absorbed by the plasma.
MW plasmas can operate in a wide pressure regime, ranging from very low pressure (e.g., 10 mTorr)
up to atmospheric pressure. The low pressure regime yields very efficient CO2 conversion. Energy
efficiencies up to 90 % were reported for very specific conditions, i.e., supersonic gas flow and
pressures around 100-200 Torr [17]. This is attributed to the role of the vibrational kinetics (discussed
in above in section 2) [2, 10, 12, 18]. Indeed, a MW plasma is characterised by typical reduced electric
fields below 100 Td. This yields electron energies around 1 eV, which are most beneficial for vibrational
excitation of CO2 [2, 10, 12]. Hence, the electrons populate the lower vibrational levels of CO2, which
collide with each other in so-called vibrational-vibrational (VV) relaxation, gradually populating the
higher levels. This so-called ladder climbing process requires 5.5 eV for CO2 dissociation, which is
exactly the C=O bond dissociation energy, while electronic excitation to a dissociative level, which is
the main process in a DBD (see above), would require 7-10 eV. As the latter is much more than the
C=O bond dissociation energy, this extra energy is just waste of energy. This explains why the energy
efficiency in a DBD is much more limited (see above).
Note, however, that the vibrational levels can also get lost by vibrational-translational (VT) relaxation.
This becomes especially important at high gas temperature, as revealed by computer simulations [19],
and it results in a vibrational distribution function (VDF) in (near) thermal equilibrium with the gas
temperature. When the MW plasma operates at atmospheric pressure, it exhibits a quite high gas
temperature (in the order of several 1000 K), resulting in a VDF that is indeed close to thermal [19, 20].
Deviation from a thermal distribution can be realised by increasing the power density, reducing the
pressure and the gas temperature [19]. At atmospheric pressure, it is not straightforward to realise a
low gas temperature. A solution could be to apply a pulsed power, so that the gas can cool down in
between the applied pulses, or to apply a supersonic gas flow, as demonstrated by Azivov et al.[17], so
that the gas has not enough time to be heated. On the other hand, the gas must have a sufficiently
long residence time for the conversion to take place as well.

2.3. Gliding arc discharge

A gliding arc discharge is a transient type of arc discharge. A classical (two-dimensional) GA discharge
is created between two flat diverging electrodes (see Figure 4(c)). The arc is initiated at the shortest
interelectrode distance, and it “glides” towards larger interelectrode distance under influence of the
gas, which flows along the electrodes, until it extinguishes and a new arc is created at the shortest
interelectrode distance.
The classical GA discharge yields only limited CO2 conversion, because only a limited fraction of the
gas passes through the arc. Therefore, other types of (three-dimensional) GA discharges have been
designed, such as a gliding arc plasmatron and a rotating GA, operating between cylindrical electrodes.
Figure 4(d) schematically illustrates the operating principle of a GAP. The cylindrical reactor body acts
as cathode (powered electrode), while the reactor outlet is the anode (grounded). The gas enters
tangentially between both cylindrical electrodes. When the outlet diameter is (significantly) smaller
than the diameter of the reactor body, the gas flows in an outer vortex towards the upper part of the
reactor body, followed by a reverse inner vortex towards the outlet, with a smaller diameter because
it has lost some speed, and therefore it can leave the reactor through the outlet. The arc is again
initiated at the shortest interelectrode distance, and expands till the upper part of the reactor, rotating
around the axis of the reactor until it stabilises in the center after about 1 ms, due to the vortex gas
flow. Ideally, the inner gas vortex passes completely through this stabilised arc, allowing most of the
gas to be converted. However, the fraction of gas passing through the arc is still too limited, thereby
limiting the CO2 conversion [21, 22].
The GA discharge operates at atmospheric pressure, which makes it suitable for industrial
implementation. Moreover, it shows a good energy efficiency, i.e., around 30 % for CO2 splitting [21]
and 60 % for DRM [22]. The reason is the same as in the MW plasma, i.e., due to the favorable reduced
electric field, creating electrons of about 1 eV, which mainly give rise to vibrational excitation of CO2,
and thus, the vibrational pathway of CO2 dissociation is again promoted. Nevertheless, the gas
temperature is also fairly high (typically a few 1000 K), which limits the energy efficiency, due to VT
relaxation, yielding a VDF too close to a thermal distribution, just like in a MW plasma (see above).
More efforts are thus needed to better exploit the non-equilibrium behavior of a GA plasma, by
reducing the gas temperature.

2.4. Other plasma types used for CO2 conversion

Besides these three types of plasma reactors explained above, other plasma types are also increasingly
being used for CO2 conversion, such as nanosecond (ns)-pulsed discharges [23], spark discharges [24],
corona discharges [25] and atmospheric pressure glow discharges (APGDs) [26].
Ns-pulsed discharges are basically generated by repetitive ns-pulsed excitation, leading to a high non-
equilibrium with very high plasma densities for a relatively low power consumption due to the short
pulse duration. The short pulses offer good control of the electron energy, depending on the pulse
length, so that more energy can be directed towards the desired dissociation channels.
Spark discharges consist of an initiation of streamers between two electrodes, developing into highly
energetic spark channels, which extinguish and reignite periodically, just as lightning, even without
pulsed power supply.
Corona discharges are created near sharp edges or thin wires used as electrode. Either a negative or
a positive voltage can be applied to the electrode, yielding a negative or positive corona discharge.
Corona discharges are non-uniform discharges, with a strong electric field, ionisation and luminosity
close to the sharp electrode, while the charged particles are dragged to the other electrode by a weak
electric field. Their performance towards CO2 conversion is similar as for DBDs.
The name “APGD” stands for a collection of several types of plasmas, including miniaturised DC glow
discharges, microhollow cathode DC discharges, RF discharges, as well as DBDs. They typically operate
at not too elevated temperature, and can exist either in stable homogeneous glow or filamentary glow
mode. They can exhibit a typical electron temperature around 2 eV, thus still suitable for vibrational
excitation of CO2, while the gas temperature is limited to about 900-1000 K, hence lower than for GA
and MW plasmas. This guarantees more pronounced thermal non-equilibrium, and makes them
promising for CO2 conversion.

2.5. Principle of plasma catalysis

As explained above, plasma on its own is very reactive, due to the cocktail of chemical species
(electrons, various types of molecules, atoms, radicals, ions and excited species), but for the same
reason, it is not selective in the production of targeted compounds. This problem can be solved by
so-called plasma catalysis, which combines the high reactivity of a plasma with the selectivity of a
catalyst [27-29]. Plasma catalysis is most straightforward in a DBD plasma, more specifically in a packed
bed DBD, because the packing beads can be covered by a catalytic material or they can have catalytic
properties from their own. This is called one-stage plasma catalysis, but the catalyst can also be placed
after the plasma reactor, in so-called two-stage plasma catalysis. In the first case, short-lived plasma
species, such as excited species, radicals, photons, and electrons, can interact with the catalyst,
providing more possibilities for synergy than in the latter case, where only long-lived species can
interact with the catalyst. On the other hand, the two-stage configuration can also be applied to other
plasma types, such as MW and GA discharge, where one-stage plasma catalysis is not so
straightforward, among others due to the high gas temperature in the plasma (cf. above).
Nevertheless, the latter may also provide other opportunities; it can open the way for thermal
activation of catalysts, either inside the discharge zone (if the temperature could be somewhat
reduced, and when using thermally stable catalysts) but also downstream, when the gas leaving the
MW or GA reactor is still hot, in two-stage plasma catalysis.
Although plasma catalysis is a quite promising combination, not only to improve the selectivity of
product formation, but also to enhance the overall plasma performance in terms of conversion and
energy efficiency, the underlying mechanisms, especially in one-stage plasma catalysis, are very
complicated and far from understood.
On the one hand, the plasma can affect the catalyst performance in several ways:
a) changes in the physicochemical properties of the catalyst, i.e., a higher adsorption probability [30],
 a higher surface area [31], due to reduced metal particle size and enhanced dispersion of metal
particles at the catalyst surface [32], a change in the oxidation state [33], reduced coke formation [34],
and a change in the work function due to the presence of a voltage and current (or charge
accumulation) at the catalyst surface [35];
b) the formation of hot spots, modifying the local plasma chemistry [36];
c) lower activation barriers, due to the existence of short-lived active species, such as radicals and
vibrationally excited species [33].
On the other hand, the catalyst will also affect the plasma performance, by:
a) enhancement of the local electric field in the plasma, because the catalyst is mostly present in a
structured packing (e.g., pellets, beads, honeycomb,…; so-called packed-bed reactor), or simply
due to the porosity of the catalyst surface [36-38];
b) change of the discharge type from streamers inside the plasma to streamers along the catalyst
surface, resulting in more intense plasma around the contact points [39-42];
c) formation of microdischarges in the catalyst pores, resulting in more discharge per volume,
increasing the mean energy density of the plasma [36, 43];
d) adsorption of plasma species on the catalyst surface, affecting the residence time and hence the
concentration of species in the plasma [44], while new reactive species might be formed at the
catalyst surface.
Figure 5 presents a schematic overview of some of these plasma-catalyst interaction processes, in
one-stage plasma catalysis [45]. Roughly speaking, we can distinguish two types of effects, i.e., physical
and chemical effects. While the physical effects, such as enhanced electric field, are mainly responsible
for gaining a better energy efficiency, the chemical effects can lead to improved selectivity towards
value-added products. In case of CO2 splitting, mainly CO and O2 are formed, so the primary added
value of the catalyst is to increase the energy efficiency, although the conversion can also be improved
by chemical effects, such as enhanced dissociative chemisorption due to catalyst acid/basic sites.
When adding a co-reactant (e.g., CH4, H2O, H2), the catalyst allows to modify the selectivity towards
value-added products.

Fig. 5: Schematic illustration of some plasma-catalyst interaction mechanisms. Adopted from [45] with
permission.

The plasma-catalyst interactions can lead to synergy in plasma catalysis, when the combined effect is
larger than the sum of the two separate under the same operating conditions, but this is not always
realised up to now. Indeed, a lot of research, by combined experiments and computer modeling, will
be needed to understand all these mechanisms and to fully exploit the possible synergy. Furthermore,
more dedicated research is needed to effectively design catalysts tailored for the plasma environment,
which make profit of the typical plasma conditions, instead of using commercial catalysts typically
used in thermal catalysis. Indeed, nowadays this is still too often the case, limiting the real potential
of plasma catalysis in selectively producing the desired products. Examples of successful plasma
catalytic CO2 conversion will be given later in this Chapter.

3. CO2 conversion processes: reactions, reactors and performance

3.1. CO2 splitting
CO2 splitting in the plasma phase was pioneered around the 70s for a two-step hydrogen production
process in the former Soviet Union. It remained largely unknown to the rest of the world until it was
summarised in the book by A. Fridman in 2008 [2]. The potential of vibrational excitation to intensify
chemical reactions is a recurring theme in this work and CO2 reduction forms its showcase. Since that
moment, the promise to address the current global challenges regarding CO2 emissions has been well
recognised by the international plasma chemistry community and a number of groups started
investigating the maximally achievable energy efficiency for the reduction of CO2 in plasma.

A graph summarising the great promise of plasma chemistry on the basis of work from the 80s [2] and
some first results since the revival of the field [18, 46] is shown in Figure 6. It is clear that the
unprecedented high energy efficiency of ~85% has been the food for inspiration and is extremely
promising for opening pathways to CO2 reuse. The explanation of the ultimate energy efficiencies has
been the preferential excitation of vibrational modes that drives ladder climbing of vibrational quanta
all the way to dissociation, as explained in section 2.

Other remarkable features of the early data are the superior performance of microwave discharges
compared to other approaches (here only RF shown) and an apparent trade-off between energy
efficiency and conversion. Both observations align with the mechanism of vibrational ladder climbing.
Microwave discharges are recognised to have an average electron energy that is optimal for
preferential vibrational excitation, although it requires sub-atmospheric pressures. Degrading
efficiency at higher input power levels can be expected due to gas heating that quenches the
vibrational non-equilibrium and reduces performance to thermal values [2].

Fig. 6: Energy efficiency of plasma assisted CO2 reduction as a function of the specific energy input. The
black and white markers are from the summary by Fridman [2]. The red markers show initial modelling
results of vibrational ladder climbing by Kozak [18] and the orange markers first experimental
microwave results by Bongers[46]. Dashed lines indicate contours of conversion degree.

Some first results since the revival have also been included in the figure and are topical for many more
findings since. Firstly, zero-dimensional modelling of vibrational kinetics in microwave plasma that
included 25 vibrational levels (of the asymmetric stretch mode) up to the dissociation limit could not
reproduce the record efficiencies but predicted a maximum efficiency of ~25% [18]. Secondly,
experimental characterisation of microwave plasma achieved higher efficiencies of ~50% [46], however,
temperature measurements in similar configurations revealed gas temperatures of typically 3500 K
[47]
. The latter means that thermal decomposition must have been of importance too whilst vibrational
ladder climbing cannot be expected to dominate at such high temperatures [19]. Noteworthy is that
also for the early experiments that yielded 85% efficiency a high temperature core was observed and
that vibrational dynamics were assumed to be important in the colder surroundings [2]. In other words,
ultimate thermal conversion performance of 50% energy efficiency has been achieved in recent
experiments. However, the older record of 85% has not been reproduced yet and it seems that
schemes in which vibrational excitation are dominant are probably non-uniform and involve transport
of power and species.

Mechanisms of CO2 dissociation

Let us summarise the main pathways that lead to dissociation in a plasma. Here, we start with the
electron driven processes. Important for the present is that these differentiate largely from each other
in their threshold energy Eth:

i. Electron impact ionisation followed by dissociative recombination,

e + CO2 → CO2+ + 2e  CO + O + e Eth = Eion = 13.8 eV (1)
ii. Electron impact dissociative excitation,
 e + CO2  CO + O(1S) + e Eth = 11.5 eV (2)
iii. Vibrationally enhanced electron impact dissociative excitation, including vibrational ladder
climbing,
e + CO2(v≥1)  CO + O(3P) + e 0 < Eth < 11.5 eV (3)

It is clear that the first two of these are a priori not beneficial for achieving high energy efficiency. The
high threshold energies have to be compared with the reaction enthalpy of the net elemental
dissociation reaction,

CO2  CO + O(3P) ΔH = 5.5 eV (4)

Thus, (i.) dissociative recombination of ions as well as (ii.) dissociative excitation are highly inefficient
dissociation mechanisms and convert 6-8.3 eV into heat and/or internal energy per event. In fact,
more than one excited state will probably contribute to dissociative excitation, with (slightly) lower
threshold energy, and 1D or 3P oxygen atoms being created. These excited states have bent structures
of which still little is known, which means that their Franck-Condon overlap with the electronic ground
state is unknown and their electron impact energy threshold cannot be predicted.

The third mechanism is evidently favourable, providing potentially the smallest threshold energy. Here,
we include within (iii.) vibrationally enhanced electron impact dissociative excitation also the
aforementioned vibrational ladder climbing mechanism. The latter was explained in detail in Figure 3,
invoking the potential energy diagram of CO2 along one O-CO coordinate. A subtle addition here is the
decrease of the activation barrier of dissociative excitation by more favourable Franck Condon overlap
of a vibrationally excited level. Dissociative excitation from the ground state produced in this example
roughly 2.5 eV of kinetic energy in the fragments and an 1S oxygen radical. A molecule excited in the
first vibrational level benefits not only from a lower threshold energy due to its initial vibrational
energy (0.3 eV for the asymmetric stretch vibration), but also from Franck Condon overlap with
reduced energy of the upper state. In the schematic representation of Figure 3, it means a reduction
in threshold energy of ~1 eV and in released kinetic energy of ~0.7 eV.

The energy threshold vanishes as molecules get into the highest vibrational (asymmetric stretch)
levels, close to the dissociation limit. The non-adiabatic transition 1Σ+  3B2 in the point of crossing of
the 1B2 and 3B2 terms opens the most effective dissociation pathway CO2(1Σ +)  CO (1Σ +)+O(3P). As
was briefly touched upon before, this can become a significant channel by virtue of the vibrational
ladder anharmonicity under strongly non-equilibrium conditions in which the kinetic energy of the
molecules remains low. Due to the anharmonicity, vibration-vibration (VV)-exchange is no longer
resonant and Treanor has shown (although neglecting effects of dissociation and VT relaxation [19], a
boundary condition that determines the exact shape of the vibrational distribution [Diomede2017])
that this results into strong deviation from a Boltzmann distribution, of overpopulation of high
vibrational levels. Such strong overpopulation is indeed observed in the aforementioned state-to-
state modelling of the asymmetric stretch manifold, which is shown in Figure 7. The depopulation of
the highest levels is due to dissociative excitation.

Fig. 7: Vibrational distribution functions of the asymmetric mode vibrational levels of CO2 in a
microwave discharge after 8.0ms of power input at a rate of 20, 25 and 30 W/cm3, reproduced from
[18]
with permission. The plateau behaviour around 10 < v < 17 reflects the Treanor-like overpopulation.
Dissociative excitation of the levels v>19 causes strong depopulation of these highest levels and
produces CO most efficiently.
Finally, atomic oxygen created in the plasma should be able to create a second CO molecule and
molecular oxygen in order to optimise the overall efficiency and to explain the observed ultimate
energy efficiencies close to 90% [2]. Again, this requires a vibrationally excited CO2* molecule as the
reaction is endothermic:

O + CO2*  CO + O2 ΔH = 0.3 eV, Eth = 0.5-3 eV (5)

One should notice the large range that is given for the activation energy of the reaction. Its
consequence is that it may well be limiting the overall efficiency, as has been discussed in [48, 49]. At
sufficiently high power density, the neutral gas temperature in the plasma reactor can become high
enough for thermal decomposition of CO2 to set in. This requires temperatures exceeding ~1700 K, as
is seen from the calculated equilibrium composition of a carbon dioxide mixture at 100 mbar in Figure
8. The thermal conversion optimum shown in the same graph is just over 50% at 3200 K, which
requires ideal quenching of the reaction products. In fact, also here the plasma phase can help by
providing a vibrational non-equilibrium and thus quenching atomic oxygen by producing additional
CO in reaction (5). This is referred to as super-ideal quenching and would bring the efficiency limit of
thermal conversion in plasma up to at least 60% [2].

Fig. 8: The equilibrium composition of CO2 and its dissociation products as a function of temperature
at a pressure of 100 mbar. Instantaneous quenching preserving all CO formed is assumed to calculate
the efficiency. Reproduced from [47] with permission.

Dissociation performance in different plasma approaches

Recently, Snoeckx and Bogaerts reviewed the state of the art of plasma chemistry concerning CO2
conversion [50]. Fig. 9 summarises the performance of the different plasma approaches in terms of
combinations of efficiency and conversion.

Fig 9: Comparison of all the data collected from the literature for CO2 splitting in the different plasma
types as collected by Snoeckx and Bogaerts[50]. It shows combinations of energy efficiency and
conversion grouping the data per discharge type.

Although DBDs have been widely researched in view of their strong non-equilibrium character, their
successful commercial application for O3 production and relatively ease of operation at atmospheric
pressure, their performance stays significantly behind that of the other types. Although conversion
can reach up to 40%, efficiency appears to be limited to ~18% (except for the DBD record of 23% by[51]).
The effect of changing the applied frequency, power, gas flow rate, discharge length, discharge gap,
reactor temperature, dielectric material, electrode material, mixing with gases, i.e. Ar, He, N2, and by
introducing (catalytic) packing materials has been studied extensively, as well as numerically
modelling has been applied to gain understanding in the underlying reaction pathways. In general, it
appears that conversion can well be controlled via specific energy input (or, equivalently, the
residence time of the gas in the reactor), however, this goes on the expense of energy efficiency. The
limited efficiency seems to be due to unfavourable plasma electron energy distribution (or E/n), which
causes vibrational non-equilibrium effects to be insignificant and reactions (1) and (2) the main
dissociation pathways. This is despite the fact that gas temperatures remain low and in this respect
the DBD discharge being non-equilibrium par excellence. The low gas temperatures, however, also
cause thermal conversion insignificant.

Microwave discharges clearly span the largest parameter range and reach the highest energy
efficiencies. This discharge type is known to be best suitable to channel most of the discharge power
to vibrational modes and is thus best equipped to benefit from vibrational excitation. However, the
best results (efficiency >60%) date from the early work [4-9] and have not been reproduced in recent
years. All recent work is within the range of thermal equilibrium conversion and gas temperature
measurements have indeed shown that 50% efficiency is well achievable up to high conversion values
[47, 52-57]
. As thermal regions have also been observed in the early work, it seems that a combination of
thermal conversion with non-equilibrium chemistry in the periphery is an interesting route to further
optimisation. It means that transport of power and particles in a complex 3D geometry is to be
optimised.

A drawback that is often put forward to MW discharges is their preference to operate at reduced
pressure. On the one hand, the lower operational pressure is compensated by high flow rates so that
in effect the reactor power density is of the highest possible. On the other hand, also atmospheric
pressure operation is well possible, but likely limits performance to thermal operational space.

GA discharges seem to succeed in exploiting vibrational excitation enhanced dissociation channels

while operating at atmospheric pressure. Energy efficiencies up to 50% are common[58] and also record
values of 65% have been reported[59]. Model calculations (e.g. [20, 60-63]) revealed that also GA discharges
induce elevated temperatures. Just like for MW discharges, preventing gas heating to operate at lower
gas temperatures might be the key for benefitting fully from the potential of the vibrational excitation
pathways to dissociation. At the same time, it is very much likely that also in GA thermal conversion
plays a significant role [20, 60-63].

3.2 Plasma conversion of CO2 with CH4

3.2.1 Plasma conversion

CO2 (g) + CH4 (g) → 2CO (g) + 2H2 (g) ΔH° =247 kJ mol-1 (6)

The conversion of CO2 with CH4, known as dry reforming of methane (DRM) has received significant
interest as this reaction uses two abundant greenhouse gases CO2 and CH4 in the form of different
sources (e.g., landfill gas, biogas and shale gas) to produce value-added fuels and chemicals, with
syngas (H2+CO) being the most common target product (6). Syngas is a vital chemical feedstock that
can be used to produce a variety of platform chemicals and synthetic fuels, including via the Fischer-
Tropsch process. However, both CO2 and CH4 are highly stable, therefore high temperature (>700 oC)
is always required for thermal catalytic activation of CO2 and CH4 with reasonable conversions and
syngas production due to the thermodynamic barrier of this process, resulting in high energy
consumption. In addition, the rapid deactivation of reforming catalysts at high temperatures due to
sintering and coke deposition remains a major challenge for the use of this process at a commercial
scale. Non-thermal plasmas provide a promising alternative to the thermal catalytic process for the
conversion of CO2 with CH4 into higher value chemicals and fuels at low temperatures and ambient
pressure. Significant efforts have been devoted to the synthesis of syngas using different plasma
systems with or without catalyst [64-68]. In addition to syngas production, noticeable amounts of higher
hydrocarbons are often produced in the plasma DRM process, especially in the presence of a catalyst
(e.g. zeolite) [69]. CHx radicals initially formed in the dissociation of CH4 play a key role in the production
of higher hydrocarbons[10, 19, 70]. Thus, the content of CH4 in the CO2/CH4 mixture is of primary
importance for the synthesis of higher hydrocarbons in the plasma DRM reaction. Eliasson et al.
investigated the synthesis of higher hydrocarbons from CO2 and CH4 using a DBD plasma. The
selectivity of C5+ and oxygenates was up to 41.2% at a discharge power of 500 W and a CO2/CH4 molar
ratio of 1:2[69]. A mixture containing mainly C2H2 and synthesis gas with a H2/CO ratio of 2:1 was
produced in the plasma reaction using a point-to-point pulsed discharge at a CO2/CH4 ratio of 1:2[71].

In addition, a few groups have reported the formation of trace oxygenates (e.g. alcohols and acids)
along with the production of syngas and hydrocarbons in plasma-based DRM[72, 73]. Zhang et al.[72]
reported the production of acetic acid, propanoic acid, ethanol and methanol in the plasma DRM using
a DBD reactor. Acetic acid was the major liquid product with the highest selectivity of 5.2% achieved
at CH4 and CO2 conversion of 64.3% and 43.1%, respectively[72]. Acetic, formic, butanoic and propanoic
acids were also formed, along with methanol and ethanol, in the plasma oxidation of CH4 with CO2
using a DBD[74]. Zhou et al.[75] developed a starch-enhanced plasma process for the conversion of CO2
and CH4 into a range of oxygenates, including primarily formaldehyde, methanol, ethanol, formic acid,
and acetic acid. The total selectivity of the oxygenates was about 10-40% with the conversion of CH4
and CO2 of about 20%. They found that a lower methane concentration was favourable for the
production of oxygenates, and a higher feed flow rate led to higher selectivity of oxygenates in the
presence of starch[75]. The direct conversion of CH4 with CO2 using a DBD plasmas was carried out by
Li et al. The product includes syngas, gaseous hydrocarbons (C2 to C4), liquid hydrocarbons (C5 to C11+),
and oxygenates[76]. Bogaerts et al. developed a 1D fluid model to understand the plasma chemistry of
the DRM process in a DBD reactor. Their modeling results showed that oxygenates, including methanol,
ethanol, acetaldehyde and ketene, can be formed in the plasma DRM reaction[10] [77]. Very recently,
，

Wang et al. have developed a water-electrode DBD plasma reactor for the direct, one-step reforming
of CO2 with CH4 into oxygenates (e.g. acetic acid, methanol, ethanol and acetone) at atmospheric
pressure (1 bar) and room temperature (30 oC). The total selectivity to oxygenates was approximately
50-60% without a catalyst, with acetic acid being the major liquid product at 40.2% selectivity[78]. Two
possible reaction pathways could contribute to the formation of acetic acid in this process (Figure 10).
CO can react with a CH3 radical to form an acetyl radical (CH3CO) with a low energy barrier of 28.77 kJ
mol-1, followed by recombination with OH to produce acetic acid with no energy barrier. Direct
coupling of CH3 and carboxyl radicals (COOH) could also form acetic acid based on density functional
theory (DFT) modelling [74].

Fig. 10: Possible reaction pathways for the formation of oxygenates in plasma DRM using a DBD.
Reproduced from [78] with permission.

Carbon nanomaterials are often produced as a by-product in the plasma dry reforming reaction. Tu
and Whitehead reported the production of multi-wall carbon nanotubes (MWCNTs) and spherical
carbon nanoparticles with a diameter of 40-50 nm in the DRM reaction using an AC gliding arc
discharge with knife-shaped Al electrodes (Figure 11(a))[66]. Chung and Chang reported the synthesis
of MWCNTs via plasma DRM using a spark discharge. They found that the stainless-steel electrodes of
the spark discharge acted as a substrate for the deposition of MWCNTs (Figure 11(b))[79]. Carbon
nanomaterials have a variety of applications and are higher value products in the plasma DRM process,
which can further reduce the energy cost of the overall plasma DRM process and make this process
more attractive.

Fig. 11: (a) The formation of carbon nanotubes using a gliding arc plasma; (b) The formation of
MWCNTs using a spark discharge. Reproduced from [66],[79] with permission.

The conversion of CO2 with CH4 has been explored using different plasma systems. Most of previous
works have mainly focused on the production of syngas via plasma DRM [80-85]. The reaction
performance of the plasma dry reforming process has been affected by a range of operating
parameters, such as plasma input power, total gas flow rate, SEI, CH4/CO2 molar ratio, and dielectric
material. The plasma power is one of the most important parameters determining the effectiveness
of the plasma DRM process. Increasing discharge power enhances the conversion of CO2 and CH4
regardless of the type of plasma system used[65-67, 86, 87]. A higher discharge power generates more
energetic electrons and reactive species (e.g. O and OH radicals), which can activate the reactants and
promote the conversion[69]. In addition, increasing discharge power would increase the temperature
of the plasma reaction, which also contributes to the enhanced conversion of CO2 and CH4. In a DBD
plasma reactor, increasing discharge power by changing the applied voltage at a fixed frequency
increases the number of microdischarge and creates more reaction channels for chemical reactions,
resulting in higher conversion of CO2 and CH4. This effect can be demonstrated by the increased
magnitude and number of current pulses of the DBD plasma at a higher plasma power[88]. However,
the discharge power can also affect the distribution of gas products produced in the plasma DRM
process. Previous results showed that increasing discharge power decreases the selectivity of lower
hydrocarbons (e.g., C2) but increases the selectivity of higher hydrocarbons (e.g., C4 and C5) [89]. By
contrast, the change of discharge power has a limited effect on the selectivity of syngas and the H2/CO
molar ratio, although the yield of syngas is enhanced at a higher discharge power [90].

Increasing the total feed flow rate decreases the conversion of CO2 and CH4 due to the decrease of the
residence time of the reactants in the discharge region, which reduces the possibility of the reactant
molecules colliding with energetic electrons and reactive species [66]. A lower gas flow rate is beneficial
for producing more syngas and reducing the selectivity of higher hydrocarbons[89]. The increase of the
residence time resulting at a lower feed gas flow rate increases the chance for C2-C4 hydrocarbons to
be further dissociated via electron impact reactions and converted to produce more CO and H2 [91]. By
contrast, a high total feed flow rate is preferred for the production of C2-C4 hydrocarbons. In addition,
the change of the feed flow rate does not significantly change the H2/CO molar ratio [64, 82]. Although
the conversion of the reactants decreases when increasing the feed flow rate, the energy efficiency of
the plasma process increases as the total amount of reactants converted increases and more electric
energy could be converted to chemical energy stored in the products [66].

Specific energy input (SEI) is a major determining factor for the conversion and energy efficiency in
plasma chemical processes, as it combines the effect of power and gas flow rate. The variation of the
SEI can be achieved by changing the discharge power and/or gas flow rate. However, previous findings
showed that manipulating the SEI by changing the gas flow rate has a more pronounced effect on the
conversion of the reactants compared to the change of discharge power[88]. Increasing the SEI at a
constant gas ratio and frequency results in a higher conversion of CO2 and CH4 but with a decreased
energy efficiency of the plasma process. The trade-off between the conversion and energy efficiency
was often reported in previous studies[67, 87]. Therefore, both discharge power and gas flow rate should
be considered when pursuing a suitable SEI to achieve higher conversion and energy efficiency
simultaneously.

The reactant conversion and the H2/CO molar ratio, along with the product yields and selectivities, are
significantly affected by the molar ratio of CO2/CH4 in the feed[81, 92]. Increasing the CO2/CH4 molar
ratio significantly enhances the conversion of CH4 but only weakly decreases the conversion of CO2.
At a higher CO2 content in the feed, oxygen atoms generated from the dissociation of CO2 can also
react with CH4, enhancing the CH4 conversion. The CO2/CH4 molar ratio also significantly affects the
yield of CO and H2. Mei reported that the yield of H2 and CO was more than doubled when increasing
the CO2/CH4 molar ratio from 1:4 to 4:1 in the plasma DRM using a DBD (Figure 12) [93]. Zhang et al.
found that increasing the CO2/CH4 molar ratio from 2:3 to 3:1 significantly increased the H2 yield from
11.4% to 20.4% and the CO yield from 7.3% to 31.3% in a DBD reactor[64]. The CO2/CH4 molar ratio
plays a key role in determining the H2/CO molar ratio in the produced syngas. Thus, syngas with a
desired H2/CO molar ratio for the further synthesis of chemicals or fuels can be controlled by tuning
the CO2/CH4 molar ratio in the feed.

Fig. 12: Effect of CO2/CH4 molar ratio on the yield of syngas and H2/CO ratio in the plasma DRM using
a DBD reactor (discharge power 50 W, total flow rate 50 ml/min). Reproduced from [93] with permission.

Higher CO2 content in the CO2/CH4 mixture leads to higher CO selectivity. In addition to direct CO2
dissociation to CO, more C2-C4 hydrocarbons generated by CH4 dissociation could be oxidised by O
atoms from CO2 dissociation, resulting in the enhanced CO selectivity and decreased selectivity to C2-
C4. For instance, the CO selectivity was increased from ~20% to over 80% when changing the CO2/CH4
molar ratio from 1:4 to 4:1 [93]. On the other hand, the lower content of CO2 in the feed gas leads to a
higher selectivity of C2-C4 hydrocarbons. Zhang et al. suggested that lower CO2/CH4 ratio decreased
the availability of O radicals in the reaction, which enhanced the possibility of recombination of CHx
(x=1-3) species to form C2-C4 hydrocarbons compared with that of direct CH4 oxidation to form CO[64].
This explanation is consistent with the decreasing trend in CO selectivity as a result of decreasing the
CO2 content in the feed gas. Wang et al. investigated the effect of CO2/CH4 molar ratio on the synthesis
of oxygenates via DRM using a water-cooled DBD system. The selectivity of acetic acid and methanol
increased initially and then decreased when changing the CO2/CH4 molar ratio from 3:1 to 1:2, with
the highest selectivity achieved at a CO2/CH4 molar ratio of 1:1. By contrast, the selectivity of ethanol
decreased continuously when decreasing the CO2/CH4 molar ratio[78]. Zhang et al. also reported that
there exists an optimum CH4/CO2 molar ratio for the maximum selectivity of target oxygenates[72].

Other process parameters also affect the performance of the plasma DRM process. Khoja et al.
evaluated the effect of discharge gap (1-4 mm) on the plasma DRM at a constant SIE in a DBD reactor.
The highest conversion of CH4 and CO2 and H2 selectivity was achieved at a discharge gap of 3 mm
using quartz as a dielectric material[94]. In most of the previous works, a discharge gap between 1 and
5 mm was used. The most appropriate discharge gap may be 2-3 mm for adequate residence time and
effective collision between electron-molecules, as studied in many cases. Enlarging the discharge gap
can increase the residence time of the reactants in the discharge zone, which can have a positive effect
on the conversion. However, increasing the gap distance at a constant input power decreases the
power density due to the increased discharge volume, which in turn negatively affects the conversion.
The balance between these opposite effects determines whether the change of the gap distance has
a positive or negative effect on the conversion[88]. A partial discharge is more likely to form at a larger
discharge gap, resulting in reduced conversion of the reactants. Li et al. found that a wider discharge
gap (1.8 mm) is more favourable for the formation of methanol and ethanol in the plasma reforming
of CO2 with CH4 at a lower CO2/CH4 feed ratio, while a smaller discharge gap (1.1 mm) produced more
acetic acid[76]. In the first reactor, the roughness of the inner electrodes was demonstrated to play an
important role on the conversion and efficiency levels of methane[95]. Zhu et al. investigated the effect
of pressure on plasma-based DRM using a kHz spark discharge plasma. Their results showed that
increasing the pressure from 1 to 2 bars enhanced the conversion of CO2 and CH4 by 7-14.8% and
reduced the energy costs by 7.7-15.2% for the conversion of the reactants[80].

Considerable effort has been devoted to further improving the performance of the dry reforming
process to maximise the conversion of CO2 and CH4 while reducing the energy consumption of the
plasma process through the development of new plasma reactor designs. Wu et al. designed a novel
rotating gliding arc co-driven by a magnetic field and tangential flow for the conversion of CO2 with
CH4. A total conversion of 39% with an energy cost of 1 eV per molecule was achieved in this process
[96]
. Very recently, Cleiren et al. applied a novel gliding arc plasmatron for the reforming of CH4 with
CO2 with syngas being the major product. The CO2 and CH4 conversions reached their highest values
of approximately 18 and 10 %, respectively, at 25 % CH4 in the gas mixture, which corresponded to an
energy efficiency of 66%. This value was above the required energy efficiency target (i.e. 60%)
reported in literature to be competitive with thermal catalytic DRM processes [22]. Modification of a
plasma reactor design has been carried out to manipulate the product distribution with enhanced
selectivity of target products. Wang et al. proposed a multi-stage ionisation design to enhance the
conversion of reactants and syngas production. It was found that the multi-stage ionisation process
favoured a higher conversion of CO2, but lowered the conversion of CH4. Meanwhile, the selectivity to
CO and H2 was increased, while the selectivity to the by-products (C2-C6) was decreased[92]. Ozkan et
al. developed a new geometry of a DBD reactor with multiple electrodes for the processing of high gas
flow rates in DRM. In their work, the main products were syngas, C2H4 and C2H6 when Ar or He was
used as the carrier gas[97]. Wang et al. developed a specially designed coaxial DBD reactor using water
as both the ground electrode and cooling for the direct synthesis of a range of oxygenates, with acetic
acid being the dominant liquid chemical via plasma DRM reaction [78].

3.2.2 Plasma-catalysis

The combination of non-thermal plasma with heterogeneous catalysis has been demonstrated as a
promising solution to further enhance the conversion and energy efficiency of the plasma process, as
well as the selectivity towards target products (e.g. syngas). Plasma-catalytic DRM has been carried
out using different plasma systems, including DBD, corona discharge, glow discharge and gliding arc
plasma. Those catalysts that have successfully demonstrated their activities in thermal catalytic dry
reforming are generally used as a starting point in the plasma-catalytic DRM reaction.

Catalysts can be coupled with non-thermal plasma in different ways, which in turn affects the
interaction between the plasma and catalyst and the formation of a plasma-catalytic synergy. Tu et al.
found that a fully packed Ni/Al2O3 catalyst into the entire discharge region of a DBD reactor decreased
the conversion of CH4 and CO2 in comparison to the plasma reaction without a catalyst[86]. Packing
catalyst pellets into the discharge area was found to shift the discharge mode from typical
microdischarges in the gas to a combination of surface discharge and weak microdischarges, which
could partly contribute to the negative effect of the plasma-catalyst coupling. A similar negative effect
from the integration of plasma and catalyst was also reported in previous studies[90, 98, 99]. These results
suggest that the generation of plasma-catalytic synergy at low temperatures (without extra heating)
in the DRM reaction is conditional and depends on the balance between the change in discharge
properties induced by the catalyst and the catalyst activity generated by the plasma[86]. The former
strongly depends on how the catalysts are packed into the discharge volume and the packing
geometry significantly affects the interactions between the plasma and the catalyst. Tu and
Whitehead compared the influence of three different catalyst packing methods on the plasma-catalyst
interactions and the resulting plasma-catalytic synergy in the DRM reaction (Figure 13)[65]. They found
that partially packing a Ni/γ-Al2O3 catalyst in flake form into the bottom of the discharge gap
significantly enhanced the reaction performance with a doubled CH4 conversion and hydrogen yield
in comparison to a fully packed reactor. This is because the discharge in the partially packed-bed
reactor retains the strong filamentary discharge, whereas the reduction in discharge volume in the
fully packed bed DBD reactor strongly supresses the formation of microdischarges and changes the
discharge mode to surface discharge and spatially limited microdischarges[65].

Fig. 13: Different catalyst packing methods in a DBD plasma reactor. Reproduced from [65]
with
permission.

Ray et al. also investigated the effect of catalyst packing volume (0%, 25%, 50% and 100%) on the
plasma dry reforming reaction in a DBD reactor[100]. They found that 25% catalyst packing showed the
highest conversion of CH4 and CO2, while fully (100%) packing the catalyst into the discharge zone
decreased the conversion of CH4 and CO2 compared to the plasma DRM reaction without packing.
Wang et al. studied the synergistic effect of catalyst and non-thermal plasma on DRM in fluidised bed
and packed-bed DBD reactors with a Ni/γ-Al2O3 catalyst. They concluded that both interaction modes
between the plasma and catalytic particles could promote the reaction at relatively low temperatures
(e.g. 673 K)[101].

The use of zeolites has been shown to be effective for enhancing selectivities towards higher
hydrocarbons, particularly C2-C4 species. Zeolites are known for their adsorbent properties, which are
beneficial to plasma reactions, because they allow species to be adsorbed onto the zeolite surface or
inside the pore structure, which can increase the residence time of the reactant species in the plasma
discharge. This can lead to an increased probability of successful collisions with active plasma species.
Eliasson et al. reported the direct formation of higher hydrocarbons with reduced carbon formation
in the plasma dry reforming of CH4 when zeolite NaX was used. They found that the presence of zeolite
NaX in the discharge zone reduced the overall conversion but increased the concentration of C2 to C4
hydrocarbons in the products[69]. Zhang et al. compared the effects of zeolite X, zeolite HY, and zeolite
NaY on the plasma DRM reaction at ambient conditions. Zeolite NaY was found to be the most
promising catalyst for the production of syngas and liquid hydrocarbons (C5+)[89], while Zeolite HY
showed the best performance in the generation of syngas and C4 hydrocarbons (C4H8, n-C4H10 and i-
C4H10) with high selectivity[102]. Jiang et al. reported that the coupling of a DBD with zeolite A inhibited
the formation of carbon black and polymers and resulted in a higher selectivity towards valuable
hydrocarbons (C2-C4) compared to the use of zeolite X in the DBD[103]. Li et al. reported the formation
of phenol in the plasma DRM combined with a HZSM-5 catalyst using a corona discharge[98].

Supported metal catalysts have been extensively used in thermal catalytic dry reforming, with
transition metals being prevalent due to their low cost and wide availability. Catalysts with high
activity in the thermal catalytic process have been used as the starting point for the plasma-catalytic
DRM process. Ni/γ-Al2O3 [64, 65, 68, 81, 86, 101, 104], Ag/Al2O3[99], Pd/Al2O3[99], Cu-Ni/Al2O3[102], Cu/Al2O3[82, 105],
Co/γ-Al2O3[82], Mn/γ-Al2O3[82], Fe/Al2O3[106], La2O3/γ-Al2O3[107] have been tested in the plasma DRM, with
Ni/Al2O3 catalysts being the most commonly used. Song et al. reported that the presence of Ni/γ-Al2O3
in a DBD reactor enhanced the CO selectivity by 22% but had a weak effect on the CO2 conversion. In
addition, they found that Ni loading (2-10 wt. %) had no effect on the conversion of CO2 and CH4, the
selectivity of gas products and H2/CO molar ratio[90]. Mahammadunnisa et al. also investigated the
effect of Ni loading (10, 20 and 30 wt. %) on the plasma DRM over partially packed Ni/Al2O3 catalysts
in a DBD reactor. The coupling of the DBD with 20 wt.% Ni/Al2O3 showed the highest reactant
conversion and syngas selectivity with doubled hydrogen yield and H2/CO molar ratio[108]. Zhu et al.
reported that increasing the Ni loading of Ni/Al2O3 from 6 wt. % to 10 wt. % enhanced the conversion
of CH4 with maximum CH4 conversion of 58.5% in the plasma DRM using a rotating gliding arc plasma,
which could be attributed to the increased catalytic effect due to the decreased Ni particle size and
enhanced Ni dispersion on the catalyst surface at a higher Ni loading (10 wt.%)[84]. Tu and Whitehead
evaluated the influence of calcination temperature (300-800 oC) of Ni/Al2O3 on the plasma-catalytic
DRM at low temperatures (~250 oC) in a DBD reactor. The results showed a synergistic effect from the
combination of the DBD with partially packed Ni/Al2O3 catalyst calcined at 300 oC, which almost
doubled the conversion of CH4 (56%) and hydrogen yield (17.5%) compared to the plasma reaction
without a catalyst[65]. Long et al. carried out the plasma DRM reaction using a cold plasma jet coupled
with a 12 wt.%Ni/Al2O3 catalyst placed in the downstream of the plasma jet. Compared to the reaction
using plasma only, the combination of the plasma jet and Ni/Al2O3 enhanced the conversion of CO2
and CH4 by 6-14% and the yield of hydrogen and CO by 11-18% at a discharge power of 700 W[109].
Zeng et al. carried out the plasma DRM reaction over different supported metal catalysts, i.e., M/γ-
Al2O3 (M = Ni, Co, Cu and Mn), in a DBD reactor. They found that the combination of the plasma with
Ni/γ-Al2O3 or Mn/γ-Al2O3 significantly enhanced the conversion of CH4 in comparison to the reaction
without catalyst. The presence of Ni/γ-Al2O3 in the plasma showed the highest activity for syngas
production[81]. However, the use of these catalysts did not improve the CO2 conversion[81]. Sentek et
al found that the presence of a Pd/Al2O3 catalyst in a DBD reactor slightly decreased the conversion of
CO2 and CH4 compared to the reaction without packing, but significantly changed the distribution of
C2-C4 hydrocarbons with the formation of more C2 and less C3-C4[99].

The catalyst support is of primary importance, as the support, along with its interactions with the
active metal, can affect the reaction performance. Mei et al. investigated the use of a Ni catalyst
supported on γ-Al2O3, TiO2, MgO and SiO2 in plasma-catalytic DRM[110]. The results of this experiment
concluded that the γ-Al2O3 support was most beneficial on the reaction performance, giving the
highest CO2 (26.2 %) and CH4 (44.1 %) conversions, as well as the maximum energy efficiency and
highest yields of CO and H2 (Figure 14). This was attributed to the increased reducibility of the Ni/γ-
Al2O3 catalyst and the number of stronger basic sites present on its surface (which facilitate CO2
chemisorption and activation), along with its higher specific surface area and greater dispersion of
smaller NiO particles[110]. Carbon deposition also occurred to a lower extent on this catalyst, as the
increase in CO2 chemisorption and activation may have resulted in adsorbed CO2 undergoing
gasification by surface adsorbed oxygen[110]. Weaker interactions between the catalyst and support
are favourable as this increases the reducibility of the catalyst, increasing its activity[65].

Fig. 14: Effect of catalyst support on the yield of H2 (a) and CO (b) as a function of discharge power in
the plasma-catalytic DRM reaction (total flow rate 50 ml/min, CO2/CH4 molar ratio 1:1). Reproduced
from [110] with permission.
The addition of dopants and use of bimetallic catalysts has also been studied in the plasma DRM
reaction. Zhang et al. investigated the effect of Cu/Ni ratio in Cu-Ni/γ-Al2O3 catalysts and found that
the 12 wt. %Cu-12 wt. % Ni/γ-Al2O3 catalyst gave the optimum results for both CH4 and CO2 conversion
and showed a synergistic effect of plasma-catalysis at 450 oC (Figure 15)[111]. This catalyst also achieved
the maximum CO selectivity of 75 %. However, this selectivity was also achieved when using the 5 wt.%
Ni-12 wt.% Cu/γ-Al2O3 catalyst, whereas the maximum selectivity to H2 was achieved with 16 wt.% Ni-
12 wt.% Cu/γ-Al2O3 and 20 wt.% Ni-12 wt.% Cu/γ-Al2O3 catalysts[111]. Ray et al. found that the addition
of Mn to a Ni/Al2O3 catalyst enhanced the conversion of CO2 and CH4, and the yield of H2 and CO in the
plasma DRM[100]. In addition, the coupling of the DBD with the Ni-Mn/Al2O3 bimetallic catalyst reduced
the carbon formation on the catalyst surface compared to Ni/Al2O3[100]. Zhang et al. investigated the
effect of La2O3/Al2O3 catalysts on the production of C2 hydrocarbons in the plasma DRM using a pulsed
corona discharge. They found that the La2O3/Al2O3 catalysts with different La loadings (5-12 wt. %)
gave a C2 hydrocarbon selectivity of more than 60% with C2H2 being the major C2 product, and
maintained the methane conversion of ~24%. Note the La2O3/Al2O3 catalysts with different La loadings
showed no change in the distribution of C2 products. Adding 0.01 wt.% Pd to 5 wt.% La2O3/Al2O3 still
gave a high C2 selectivity of 70% but significantly changed the distribution of C2 hydrocarbons with
C2H4 being the major C2 product (65 vol.%)[112]. Kado et al. reported that packing a Ni0.03Mg0.97O catalyst
into a flow-type tubular pulsed discharge reactor significantly changed the selectivity of gas products
compared to the plasma reaction without a catalyst: the selectivity of C2 drastically decreased from
33.6% to 1%, and the CO selectivity increased from 65.4% to 99%, at a CO2/CH4 molar ratio of 1:1[113].
More recently, K-, Mg- and Ce-promoted Ni/Al2O3 catalysts have also been evaluated in plasma-
catalytic DRM at 160 oC[85]. The addition of promoters (K, Mg and Ce) into the Ni/Al2O3 catalyst
enhanced the conversion of CH4, the yield of H2 and the energy efficiency of the plasma process. The
highest conversion of CO2 (22.8%) and CH4 (31.6%) was achieved by placing the K-promoted Ni/Al2O3
catalyst in the plasma reforming process. In addition, compared to the un-promoted Ni/Al2O3 catalyst,
although the use of the promoted catalysts increased the carbon deposition on the surface of the
spent catalysts by 22%-26%, the total amount of deposited carbon was still less than that reported in
high temperature catalytic dry reforming processes. More than 80% of the increased carbonaceous
species was in the form of reactive carbon species, which can be easily oxidised by CO2 and O atoms
and maintain the stability of the catalysts during the reforming reaction[85]. In this study, the behaviour
of K, Mg and Ce promoters in the low temperature plasma-catalytic DRM reforming was opposite to
that in high temperature thermal catalytic DRM process in terms of the conversion of CH4 and carbon
deposition, which could be ascribed to the temperature-dependent character of the promotors[85].
These results also suggest that those catalysts that have shown poor catalytic activity (e.g. conversion)
in thermal catalytic reactions might work well in low temperature plasma-catalytic processes, and vice
versa.

Fig. 15: DRM using plasma only, Cu-Ni/Al2O3 catalyst only and plasma-catalysis at 450 oC (total flow
rate 60 ml/min, argon flow rate 30 ml/min, CO2/CH4 molar ratio 1:1, discharge power 60 W, GHSV
1800 h-1, a 60 ml/min, 50%Ar in the feed, CO2/CH4 molar ratio 1:1, GHSV 1800 h-1, 0.1 g catalyst).
Reproduced from [111] with permission.

Core-shell structured catalysts have attracted significant interest in DRM as the metallic active sites
could be uniformly distributed within the shells. The strong interaction between the cores and shells
is ascertained to be highly capable of preventing metallic nanoparticles (NPs) from carbon deposition
and sintering even at high temperatures. Zheng et al. reported that the combination of a DBD plasma
with LaNiO3@SiO2 core-shell nanoparticle catalysts showed a better catalytic performance in plasma-
based DRM with higher reactant conversion, product selectivity and catalytic stability, compared to
the traditional Ni-based catalysts (Ni/SiO2, LaNiO3/SiO2 and LaNiO3)[114, 115]. The conversion of CH4 and
CO2 reached 88.3% and 77.8%, and the selectivity of CO and H2 was 92.4% and 83.7%, respectively.
Their results suggest that the SiO2 shell is capable of preventing Ni from sintering and mitigating
carbon deposition in the plasma-catalytic reaction (Figure 16)[114]. Compared to the supported Ni-
based catalysts (Ni/γ-Al2O3, NiFe/γ-Al2O3, NiFe/SiO2, and NiFe2O4), the use of spinel nickel ferrite
nanoparticles (NiFe2O4 NPs) embedded in silica (NiFe2O4#SiO2) also showed excellent catalytic
performance and high resistance to carbon formation in the plasma dry reforming under ambient
conditions without the involvement of extra heat. The results indicated that the special structure of
the as-synthesised NiFe2O4#SiO2 catalyst was capable of restraining the aggregation of NiFe alloy and
suppressing the carbon formation in the plasma reforming process[83].

Fig. 16: The conversion of CH4 and CO2 with time on stream over different Ni catalysts. Adopted from
with permission.
[114]

In addition, the catalytic effect of electrode materials on the plasma DRM reaction has also been
investigated. Li et al. evaluated the influence of different electrode materials (Ti, Al, Fe and Cu) on the
production of syngas and higher hydrocarbons in the plasma DRM using a DBD reactor. They found
that the Ti electrode showed the highest conversion of CH4 and CO2, while the other electrode
materials showed a similar performance[116]. Scapinello et al. reported that nickel and copper
electrodes are more efficient than stainless steel in producing carboxylic acids, in particular formic
acid in the plasma DRM using a DBD reactor[117]. However, no major catalytic effects of the metal
surface on the conversion of reactants (CO2 and CH4) and the production of H2 and CO were
observed[117].

The energy efficiency is higher in gliding arc discharges in comparison to other types of discharges,
and catalysts can increase this even more[81]. Placing a NiO/Al2O3 catalyst in the afterglow of the gliding
arc reactor was found to increase the energy efficiency by over 20 % in comparison to that achieved
using plasma only[81]. The H2 yield, along with the CO2 and CH4 conversions, was also increased. The
concentration of active metal was found to influence the reaction performance, as a 33 wt.%
NiO/Al2O3 catalyst resulted in a decrease in reaction performance in comparison to an 18 wt. %
NiO/Al2O3 catalyst; whilst a smaller catalyst diameter was found to be beneficial[81]. Goujard et al.
investigated the influence of the type of plasma power supply on the plasma-catalytic synergy for
DRM. Their experiments were performed in a DBD reactor packed with a cordierite honeycomb
monolith and excited by two different power supplies: a pulsed excitation and a sinusoidal excitation.
In the absence of a Ni catalyst, higher CO2 and CH4 conversion was achieved using the pulsed power
supply. However, when using a 2 wt.% Ni catalyst in the plasma, the reactive species generated by the
AC power supply promoted the activation of CO2 and CH4 on the Ni catalyst surface, leading to a
significant increase of CH4 and CO2 conversion[118].

3.3 Plasma CO2 Hydrogenation

CO2 hydrogenation for the synthesis of higher value fuels and chemicals has provided an attractive
route for CO2 conversion and utilisation, as this process has a lower thermodynamic limitation
compared to direct CO2 decomposition and dry reforming of methane. One of the key challenges
facing this process is the cost and source of hydrogen. In order for this process to be both economically
viable and sustainable, hydrogen must be generated using a low cost, environmentally friendly and
sustainable process, such as from water electrolysis using wind or solar power or from bioenergy. The
overall process should be CO2 neutral, which means that CO2 hydrogenation must convert a greater
amount of CO2 than renewable hydrogen production pathways generate. Although CO2 reduction with
H2 using heterogeneous catalysis has been extensively investigated in the past few years, there are still
significant challenges in developing active, selective and stable catalysts suitable for large-scale
commercialisation. In addition, it is key to lower the operating temperature of the CO2 hydrogenation
to minimise the energy consumption of the process.

3.3.1 Plasma conversion

The direct hydrogenation of CO2 mainly produces three types of C1 chemicals: CO via reverse water
gas shift reaction (RWGS, (7)), CH4 via CO2 methanation (8), and CH3OH via CO2 hydrogenation (9). Up
until now, very limited research has been concentrated on CO2 hydrogenation using non-thermal
plasmas[119-123]. The majority of this research reports CO as the dominant chemical, with CH4 formed
as a minor product and no or trace CH3OH detected[124-126].

CO2 (g) + H2 (g) ↔ CO (g) + H2O (g) ΔH° = +40.9 kJ mol-1 (7)

CO2 (g) + 4H2 (g) → CH4 (g) + 2H2O (g) ΔH° = -165.3 kJ mol-1 (8)

CO2 (g)+ 3H2 (g) → CH3OH (g) + H2O (g) ΔH° = -49.9 kJ mol-1 (9)

The reverse water-gas shift reaction converts CO2 and H2 to produce CO and H2O. CO is an important
chemical feedstock for Fischer-Tropsch synthesis (FTS) to produce higher hydrocarbons such as
liquefied petroleum gas (LPG), naphtha, gasoline and diesel; or for the synthesis of valorised chemicals
such as acetic acid, phosgene and formic acid. Recently, Porosoff et al. has proposed the combination
of the RWGS reaction with FTS for the synthesis of hydrocarbons[127].

Zeng and Tu investigated the influence of H2/CO2 ratio on the RWGS reaction using a DBD reactor[128].
They found that the conversion of CO2 increased almost linearly with the increase of the H2/CO2 ratio
from 1:1 to 4:1. Increasing the H2/CO2 ratio significantly enhanced the yield of CO, while the CO
selectivity was only slightly increased[128]. The dependence of CO selectivity on the gas flow rate at a
fixed H2/CO2 ratio of 4:1 was investigated using a low-pressure radio-frequency discharge[129]. The
selectivity of CO increased gradually when increasing the total flow rate. This phenomenon is most
likely due to the decreased residence time associated with the increase of the flow rate, resulting in
suppression of the recombination of CO and O. The effect of argon on the plasma RWGS was evaluated
in a DBD plasma reactor at 150 oC. In the absence of a catalyst, the CO2 conversion increased from
18.3% to 38% at a discharge power of 30 W and a fixed H2/CO2 ratio of 4:1 when increasing the Ar
content from 0 to 60% in the gas mixture[121]. The presence of metastable argon species in the DBD
creates new reaction pathways for the dissociation of CO2, resulting in enhanced CO2 conversion.

In the CO2 methanation reaction, CO2 reacts with hydrogen to produce methane and water. This
reaction was first discovered by Sabatier and Senderens in 1902. The CO produced during methanation
has been recognised as an important intermediate in the CO2 methanation pathways (10).

CO (g) + 3H2 (g) → CH4 (g) + H2O (g) ΔH° = -249.8 kJ mol-1 (10)

However, limited efforts have been devoted to the use of non-thermal plasmas for CO2 methanation,
especially in the plasma reaction without catalyst. CO is the major product with CH4 being the minor
one in CO2 methanation. Zeng and Tu showed that the selectivity of CH4 (2-5%) was significantly lower
than that of CO (>90%) in the plasma processing of CO2 with H2 at low temperature (150 oC)[121]. In the
plasma CO2 methanation process, a higher H2/CO2 ratio is desirable as this increases the conversion of
CO2 and the selectivity of CH4, which has been demonstrated both experimentally[128, 130] and through
the use of a 1D fluid model [131]. Optimizing the total flow rate can also maximize the CH4 selectivity
and CO2 conversion. A very low total flow rate can lead to reverse reactions occurring, reforming CO
from CH4 according to (11), due to the longer residence time, increasing the interactions between the
CO2 hydrogenation products and the reactive species in the plasma[129].

CH4 (g) + H2O (g)  CO (g) + 3H2 (g) ∆H° = 206 kJ mol-1 (11)

Zeng and Tu reported that adding argon up to 60% in the CO2/H2 mixture significantly enhanced the
CH4 selectivity by 85%, which suggests that the presence of metastable argon species in the discharge
creates new reaction routes for the production of methane. For DBD plasmas, the use of alumina as a
dielectric material instead of quartz is beneficial on CO2 methanation, which might be attributed to
the relatively higher dielectric constant of alumina[130]. Addition of a magnetic field to a plasma system
enhanced the CO2 conversion and CH4 selectivity by over 10 % at a discharge power of 30 W, whilst
the energy efficiency of the process was tripled[124]. This process however employed low pressure (200
Pa), reducing simplicity of design and requiring extra energy input. Increasing input power generally
results in a higher selectivity to CH4 due to the increased power density [124, 130]. However, it has been
found that at high power input (>160 W), energy is transferred to the electrodes through heating
rather than being used for plasma chemical reactions, resulting in the saturation of CH4 selectivity[124].
A smaller discharge gap is beneficial on the CO2 conversion and CH4 selectivity due to the rise in input
power density and enhanced electric field[124]. In fact, a smaller discharge gap can achieve the same
CH4 selectivity at a lower input power than when using a larger discharge gap[124]. A reduction in
discharge gap can also increase the production efficiency of the plasma process.

CO2 hydrogenation to liquid fuels (e.g. methanol and ethanol) is one of the most attractive routes for
CO2 conversion and utilisation. CH3OH is a valuable fuel substitute and additive, and is also a key
feedstock for the synthesis of other higher value chemicals. In addition, methanol is considered a
promising hydrogen carrier, suitable for storage and transportation[132]. In the late 1990s, Eliasson and
co-workers investigated CO2 hydrogenation to CH3OH using a DBD plasma reactor at pressures up to
10 bar[133]. However, the major products were CO and H2O with a CO selectivity of over 90%. Only
trace amounts of CH3OH were produced in the plasma CO2 hydrogenation without a catalyst, with a
maximum CH3OH yield of 0.06% (selectivity 0.4-0.5%) obtained at 8 bars, a relatively high plasma
power of 400 W, a total flow rate of 250 mL/min, and a H2/CO2 ratio of 3:1[133]. Increasing the wall
temperature from 100 to 220 oC had a limited effect on the selectivity and yield of methanol[133]. The
formation of trace amounts of CH3OH in the plasma CO2 hydrogenation was also reported using a radio
frequency (RF) impulse discharge at low pressures (1-10 Torr)[129]. Very recently, Wang et al. reported
that the methanol production via plasma-assisted CO2 hydrogenation was strongly dependent on the
structure of the DBD plasma reactor (Figure 17). The proposed single dielectric DBD reactor with a
special design using water as a ground electrode and cooling significantly enhanced the production of
methanol with the highest methanol selectivity of 54% achieved in the plasma hydrogenation of CO2
without a catalyst at atmospheric pressure (1 bar) and room temperature (30 oC). The concentration
and yield of CH3OH, as well as the conversion of CO2 were affected by the H2/CO2 molar ratio.
Increasing the H2/CO2 molar ratio from 1:1 to 3:1 increased the yield of CH3OH from 6.0% to 7.2%,
while the selectivity of CO decreased from 40.0% to 30.0%[123].

Fig. 17: Effect of DBD reactor structure (Reactor I, II and III) on plasma CO2 hydrogenation to
oxygenates without a catalyst at a fixed discharge power of 10 W and a H2/CO2 molar ratio 3:1 (a)
concentration of oxygenates; (b) selectivity of gas products and oxygenates. Reproduced from [123] with
permission.

3.3.2 Plasma-catalysis

Catalysts are the key to manipulate the selectivity of target products in the plasma hydrogenation of
CO2. Zeng and Tu investigated the influence of Al2O3 supported metal catalysts (Mn, Cu and Cu-Mn)
on the plasma RWGS reaction in a DBD reactor at atmospheric pressure (1 bar). Compared to the
reaction using plasma alone, the presence of these catalysts in the DBD enhanced the conversion of
CO2 by up to 36%, with the Mn/γ-Al2O3 catalyst showing the best activity for CO2 conversion at a
H2/CO2 molar ratio of 1:1. The coupling of the DBD with Mn/γ-Al2O3 also significantly enhanced the
yield of CO by 114%, followed by Cu-Mn/γ-Al2O3 (91%) and Cu/γ-Al2O3 (71%). As a result, the energy
efficiency for CO production was significantly enhanced by up to 116%[128]. However, only the Cu/γ-
Al2O3 catalyst was found to enhance the selectivity of CH4 compared to the reaction using plasma
alone [128]. The weaker activity of Cu/γ-Al2O3 in CO2 conversion in comparison to the Mn/γ-Al2O3
catalyst might be attributed to the increased prevalence of the water gas shift (WGS) reaction in the
presence of Cu/γ-Al2O3 as Cu catalysts are often used for catalysing the WGS reaction [128]. It is
therefore important to select a catalyst that will supress the WGS reaction and simultaneously
increase the CO2 conversion and the selectivity to CH4.

The combination of plasma and suitable catalysts enables the CO2 methanation reaction to occur at
much lower temperatures than those required in the thermal catalytic process[125]. Nizio et al.
evaluated the activity of ceria and zirconia-promoted Ni-containing hydrotalcite-derived catalysts in
the plasma-catalytic hydrogenation of CO2 to methane using a low temperature DBD reactor. Below
250 °C, negligible CO2 conversion occurs for the catalytic process using a Ce-Zr supported Ni catalyst.
However, when combined with a DBD plasma the CO2 conversion reached 80 %, with 90 % selectivity
to CH4 even at 110 oC[125]. This is due to the creation of excited species in the plasma, which generate
new pathways for CO2 dissociation; hence the reaction is not limited by the rate of CO2 dissociation at
the catalyst surface as it is in the thermal catalytic process[126]. The use of nickel containing hydrotalcite
catalysts also showed promising results in the plasma-catalytic CO2 methanation reaction, with a CO2
conversion of 80 % and a CH4 selectivity of nearly 100 % achieved[134].

Eliasson et al. investigated the plasma-catalytic hydrogenation of CO2 to methanol over a

Cu/ZnO/Al2O3 catalyst in a DBD reactor at a relatively high pressure (up to 10 bar). Compared to the
plasma reaction without catalyst (see previous section), the presence of the Cu/ZnO/Al2O3 catalyst in
the DBD increased the methanol yield (from 0.1 to 1.0%), methanol selectivity (from 0.4 to 10.0%),
and CO2 conversion (from 12.4% to 14.0%) at 8 bars, 100 oC and a H2/CO ratio of 3:1. However, the
methanol yield and selectivity were still significantly lower than those reported in thermal catalytic
CO2 hydrogenation processes [133]. Wang et al. has successfully demonstrated the synthesis of
methanol with a high selectivity via plasma-catalytic CO2 hydrogenation using a water-cooled DBD
reactor at room temperature and atmospheric pressure (Figure 18). Packing Cu/γ-Al2O3 or Pt/γ-Al2O3
into the DBD significantly enhanced the CO2 conversion and methanol yield compared to the plasma
hydrogenation of CO2 without a catalyst (Figure 19). The maximum methanol yield of 11.3% and
methanol selectivity of 53.7% were achieved using the Cu/γ-Al2O3 catalyst with a CO2 conversion of
21.2% in the plasma-catalytic process, while no reaction occurred at ambient conditions without using
plasma[123].

Fig. 18: Scheme of CO2 hydrogenation to methanol. Reproduced from [123] with permission.

Fig. 19: Effect of H2/CO2 ratio and catalysts on the plasma-catalytic CO2 hydrogenation to oxygenates
at a discharge power of 10 W. (a) methanol yield and CO2 conversion; (b) selectivity of gas products
and oxygenates. Reproduced from [123] with permission.

In addition, the production of dimethyl ether (DME) from plasma CO2 hydrogenation was reported
using an atmospheric pressure surface discharge, with a CO2 conversion of 15% and a H2/CO2 molar
ratio of 1:1[122]. Compared to thermal catalytic CO2 hydrogenation to value-added fuels and chemicals,
which has been carried out using a wide range of catalysts for a range of target products, very limited
catalysts that are active for the thermal catalytic process have been examined in plasma
hydrogenation of CO2 at low temperatures.

3.4. CO2 with water

3.4.1 Plasma conversion
Compared to the large amount of work performed for plasma-based DRM, only limited research has
been performed for the simultaneous conversion of CO2 and H2O, i.e., so-called artificial
photosynthesis.
CO2 (g) + H2O (g) ↔ CO (g) + H2 (g) + ½ O2 (g) ΔH° = 525 kJ mol-1 (12)

CO2 (g) + 2 H2O (g) ↔ CH3OH (g) + 3/2 O2 (g) ΔH° = 676 kJ mol-1 (13)

Futamura et al. [135] investigated a CO2/H2O mixture diluted to 0.5–2.5 % in N2 in a DBD reactor, and
reported a CO2 conversion of only 0.5 %, with product yields of 0.7 % for H2, 0.5 % for CO and no O2,
and also no mention on oxygenated products. Mahammadunnisa et al. [136] obtained a CO2 conversion
of 12–25 % in a DBD reactor, depending on the SEI, with a selectivity of 18–14 % for H2 and 97–99 %
for CO, yield a syngas ratio of 0.55–0.18.
Snoeckx et al. [137] performed a combined experimental and computational study for CO2/H2O
conversion in a DBD. Adding a few % of H2O to the CO2 plasma was found to cause a steep drop in the
CO2 conversion, and both the CO2 and H2O conversion were quite low. CO, H2, O2 and H2O2 (up to 2 %)
were the major products and no oxygenates were detected. The experimental data could be explained
by a chemical kinetics model (see Figure 20). The main reactive species created were OH, CO, O and
H, and the model reveals that the OH radicals quickly recombine with CO into CO2, which explains the
limited CO2 conversion upon H2O addition. In addition, the O and H atoms recombine in a few
subsequent reactions to form H2O again, explaining why also the H2O conversion was limited. Finally,
the fast reaction between O/OH and H atoms explains why no oxygenated products were formed.

Fig. 20: Reaction scheme to illustrate the main pathways for CO2 and H2O conversion and their
interactions. The arrow lines represent the formation rates of the species, with full green lines being
formation rates over 1017 cm-3∙s-1, orange dashed lines between 1017 and 1016 cm-3∙s-1 and red dotted
lines between 1016 and 1015 cm-3∙s-1. Adopted from [137] with permission.

Ihara et al. [138] investigated a 1:1 CO2/H2O mixture in a MW plasma, and detected low yields of oxalic
acid and H2O2 They also assumed that H2 and O2 are generated, but these products were not
measured. In their follow-up study [139] they varied the CO2/H2O gas mixing ratio from 1:4 to 1:1, and
detected methanol instead of H2O2 and oxalic acid, albeit again in very low concentrations < 0.01 %.
However, a rise in the methanol yield by a factor 3.5 was observed upon increasing the pressure from
240 to 400 Pa. They suggested two pathways for methanol formation, i.e., (i) direct formation from
CO2 and H2O in the plasma, and (ii) the reformation of deposited polymeric material on the walls
during the plasma reaction with H2O. Chen et al. [140] detected syngas (in a ratio close to 1, for a 1:1
CO2/H2O ratio) and O2 in a surface-wave MW plasma, but no hydrocarbons or oxygenates. In a follow-
up study [141], they reported that adding 10 % H2O to a CO2 MW plasma yielded a higher CO2 conversion
(i.e., 31 % vs. 23 %) and a lower energy cost (i.e., 22.4 eV/molec vs. 30.2 eV/molec), along with a lower
gas temperature, which was attributed to the higher heat capacity of water and the induced
endothermic reactions. This lower gas temperature can explain the higher conversion and lower
energy cost, due to less VT relaxation (causing vibrational depopulation; see section 2.2 above) and a
lower reaction rate of the recombination of CO back into CO2.
Indarto et al. [142] reported that H2O addition (in the range of 5 to 31 %) to a CO2 plasma in a classical
GA yielded a drop in CO2 conversion (around 7.1 – 3.0 %, compared to 13.4 % in pure CO2) and a higher
energy cost (around 89 – 189 eV/molec, compared to 53 eV/molec in pure CO2), which they attributed
to instabilities in the plasma upon adding H2O. Nunnally et al. [58] also found a higher energy cost when
adding 1 % H2O to a GAP, i.e., 14.8 eV/molec vs. 9.5 eV/molec in pure CO2, but they did not observe
arc instabilities. Instead, they attributed the higher energy cost to VT relaxation, causing depopulation
of the CO2 vibrational levels, which is much faster for collisions with H2O than with CO2 [2, 58].
Hayashi et al. [122] compared the conversion and product formation for a 1:1 mixture of CO2/H2O and
CO2/H2 in a surface discharge, and reported much lower values in the CO2/H2O mixture than in the
CO2/H2 mixture (i.e., 5 % vs. 15 %), but the same products were detected, i.e., CO, CH4, dimethyl ether
(DME) and formic acid. Finally, Guo et al. [143] studied the combined CO2/H2O conversion in a negative
DC corona discharge, for H2O contents between 10 and 43 % and pressures between 1 and 4 bar.
Again, a drop in CO2 conversion upon increasing H2O content was observed. The main products formed
were H2 and CO, as well as ethanol and methanol, in roughly a 3:1 ratio, with a total molar yield up to
4.7 %, increasing with pressure.
In general, it is clear from the above literature overview that adding even small amounts of H2O (1 – 2
%) yields a significant drop in CO2 conversion, followed by a further drop upon addition of even more
H2O. For a DBD, this can be explained by the chemical pathways presented in [137] (see Figure 20 above).
For a GA, the fast quenching of the CO2 vibrational levels by VT relaxation with H2O molecules is the
most probable explanation [2, 58]. This would be expected in a MW plasma as well, but Chen et al. [141]
reported a higher CO2 conversion upon H2O addition. However, this MW setup operates at low
pressures (30-60 Torr), where VT relaxation and thus quenching is less important, and cooling upon
H2O addition might be the dominant effect, which reduces VT relaxation, and thus enhances the CO2
conversion and corresponding energy efficiency.
The main products formed by the combined CO2/H2O conversion are H2 and CO, like for DRM (see
Section 3.2), as well as O2, but some papers also report the production of hydrogen peroxide (H2O2)
[137, 138]
, oxalic acid (C2H2O4) [138], formic acid (CH2O2) [122], methane (CH4) [122, 136], dimethyl ether (C2H6O,
DME) [122]
, methanol (CH3OH) [136, 139, 143], ethanol (C2H5OH) al. [143], acetylene (C2H2) [136], propadiene
(C3H4) [136] and even carbon nanofibres (CNFs) [136]. However, most of these data are only qualitative
and mainly incomplete, so we cannot deduce a general trend on product yields or selectivities.
Nevertheless, the formation of oxygenates, and other value-added compounds, in a one-step process
seems limited, in the absence of a catalyst. The reason is that too many steps are involved in creating
these oxygenates, and all of them involve H atoms, which will rather recombine quickly with OH into
H2O or with O2 into HO2, which also reacts further with O into OH—and hence H2O. In other words,
the interactions of H atoms with oxygen species (i.e., OH, O3, O2 or O atoms) are too fast and their
tendency to form H2O is too strong.
On the other hand, the H2/CO ratio produced in CO2/H2O plasmas can be very high (even up to 8.5, for
sufficient amounts of H2O addition), and they can be easily controlled, as revealed by computer
modelling [137], which might be useful for the production of value-added chemicals in a two-step
process.

3.4.1. Plasma catalysis

It is clear from above that the direct production of value-added compounds in CO2/H2O plasmas
requires suitable catalysts. Futamura et al. [135] investigated the potential of a ferroelectric packed bed
reactor with BaTiO3 pellets for the diluted CO2/H2O mixture mentioned above. They obtained a CO2
conversion of 12.3 %, with product yields of 12.4 % for H2, 11.8 % for CO and 2.8 % for O2, hence much
higher than in the non-packed DBD reactor (cf. above), but nothing was mentioned on the formation
of oxygenates. Likewise, adding a Ni/γ-Al2O3 catalyst, in both unreduced and partially reduced form,
Mahammadunnisa et al. [136] obtained a higher conversion and syngas ratio than without catalyst (see
above), i.e., 18–28 % and 0.95–0.45 for the unreduced catalyst (NiO/γ-Al2O3), and 24–36 % and 0.66–
0.35 for the partially reduced catalyst (Ni/γ-Al2O3). The NiO catalyst was found to yield a reduction of
the produced CO to CH4, CH3OH, C2H2, propadiene, while the Ni catalyst also gave rise to carbon
nanofibers.
Chen et al. [141] studied the effect of NiO/TiO2 catalysts (treated with an Ar plasma) in a MW plasma.
Compared to their results without catalyst (see above), the CO2 conversion was further enhanced to
48 %, with an energy cost of 14.5 eV/molec, but no oxygenated products were detected. The authors
suggested that CO2 is adsorbed at oxygen vacancies on the catalyst surface, reducing the threshold for
dissociative electron attachment into CO, adsorbed O atoms at the vacancies and electrons. The
adsorbed O atoms may subsequently recombine with gas phase O atoms or OH radicals into O2 (and
H atoms). Hence, the catalyst seems to have a beneficial effect in tuning O/OH from the plasma into
O2, by means of adsorbed O atoms at the vacancies, before they recombine again into CO2 and H2O,
which seems the limiting step in CO2/H2O conversion, at least in a DBD, as revealed by computer
modelling [137] (see above).
Although the above papers use catalysts, and report beneficial effects, the production of oxygenates
or other value-added compounds is very limited, if reported at all. Hence, clearly more research is
needed for tailored catalyst design. These catalysts should allow the plasma-generated CO and H2 to
selectively react into oxygenates, such as methanol, and subsequently they should separate the
methanol from the mixture. As mentioned in previous section, Eliasson et al. [133] applied a
CuO/ZnO/Al2O3 catalyst in a CO2/H2 DBD, which gave a 10 times higher methanol yield and selectivity.
Other possible candidates could be Ni-zeolite catalysts, for which methanation is reported [126], a
Rh10/Se catalyst, yielding an ethanol selectivity up to 83 % [144], and a Ni-Ga catalyst for the conversion
into methanol [145].
4. Summary and steps to be taken for further improvement
It is clear from the above sections that plasma-based CO2 conversion is promising, but more research
is needed before it can be implemented in industry. Although plasma creates a very reactive
environment, and is thus chemically not selective in producing value-added compounds, plasma is
selective in another way, in the sense that it can selectively populate the vibrational levels of CO2,
without activating the other degrees of freedom, i.e., without the need to heat the gas. This selectivity
induces thermal non-equilibrium, and explains the good energy efficiency compared to thermal
conversion, at least for some plasma types, like MW and GA plasmas (and maybe APGDs).
To compete with classical and other emerging technologies, Snoeckx and Bogaerts [12] stated that an
energy efficiency of 60 % (or an energy cost below 4.27 eV/molec for DRM) would be required (see
also Figures 6 and 9 in sections 3.1). MW and GA plasmas already reach energy efficiencies close to,
or above, this defined efficiency target. This shows their great potential, attributed indeed to the
important role of vibrational excitation for energy-efficient CO2 dissociation. This is especially true for
MW plasmas at reduced pressure for CO2 conversion, which clearly exhibit a thermal non-equilibrium
between the vibrational and gas temperature. However, for DRM, very limited results have been
reported in MW plasmas, while many promising results are published for GA discharges, and also ns-
pulsed discharges, spark discharges and APGDs show potential.
DBDs are most frequently studied, both for CO2 conversion and DRM, but we believe that even with
further improvements the energy efficiency will remain too low for industrial implementation. Indeed,
they have typical energy efficiencies of 5-10 %, with some exceptions up to 20 %, so significant
improvements, by a factor 3-5, would still be needed to reach the defined efficiency target of 60 %, in
order to make them competitive with other emerging technologies.
It should be realized, however, that the efficiency target of 60 % was defined for the production of
syngas by DRM, which is indeed the major product in plasma-based DRM. Nevertheless, also other
higher value products are formed in the plasma, although not selectively, but when suitable catalysts
can be found, plasma catalysis can produce these value-added compounds, such as higher
hydrocarbons or oxygenates, in a one-step process at low temperatures and ambient pressure. This
would significantly reduce the energy efficiency target to be competitive with other technologies, if
the latter would need a two-step process, because indeed, the subsequent Fischer-Tropsch or
methanol synthesis from syngas is quite energy intensive.
In this sense, even DBD could become suitable, especially because of their simple design, allowing
easy upscaling and straightforward implementation of catalysts. However, much more research is
needed to design catalysts, tailored to the plasma environment, to directly produce such value-added
chemicals with high selectivity. The latter does not only apply to DRM, but certainly also to CO2
hydrogenation (CO2/H2) and artificial photosynthesis (CO2/H2O mixtures), where the formation of
value-added compounds without catalyst seems even less straightforward.
To improve the capabilities of plasma-based CO2 conversion, first of all the energy efficiency should
be further improved. We believe this should be realized by a low enough reduced electric field (order
of 5-100 Td), in combination with high enough plasma power for sufficient vibrational excitation,
which is the most energy-efficient CO2 dissociation pathway, and with a low gas temperature, to
minimise vibrational losses upon collision with other gas molecules (so-called VT relaxation), i.e.,
strong thermal non-equilibrium. MW and GA plasmas already make use of this most energy-efficient
dissociation pathway, but at atmospheric pressure, the VT relaxation is quite significant, and hence,
the vibrational distribution function is too close to thermal, so for further improvement, the non-
equilibrium should be further exploited.
In addition, the conversion should be further improved, along with the product yield/selectivity.
Indeed, the major disadvantage of plasma-based CO2 conversion is in our opinion the need for a post-
reaction separation step, as the gas conversion is typically (far) below 100 %, and many different
reaction products can be formed. In case of CO2 splitting, the separation of CO and O2 was calculated
to yield the largest energy cost [146]. When a H-source is added (either CH4, H2 or H2O), mainly syngas
is formed, but some minor side-products are observed as well. The syngas mixture does not really
pose a problem, when it is subsequently used for Fischer-Tropsch or methanol synthesis. Moreover,
plasma technology can deliver a wide variety of syngas ratios, depending on the initial feed gas mixing
ratio. However, when higher hydrocarbons or oxygenates could be directly produced with sufficient
selectivity and yields, the post-reaction separation step would not be so critical, as these (liquid)
compounds can more easily be separated. This brings us back, however, to the crucial need for tailored
catalysts, specifically designed to the plasma environment.
In spite of the further improvement of energy efficiency or selectivity being required, we believe that
plasma-based CO2 conversion is very promising, especially because of its overall flexibility, in terms of
(i) feed gas (i.e., pure CO2 splitting, but also mixtures with any H-source are possible), (ii) energy source
(solar, wind, hydro, wave and tidal power, as well as nuclear power), and (iii) fast on/off switching and
modular upscaling capabilities. This flexibility makes plasma very useful as a so-called “turnkey”
process, which might be able to use renewable electricity in a flexible way, following its intermittency,
and convert it into fuels or chemicals. An example of how a (microwave) plasma based process for
pure CO production could be implemented at industrial scale has been evaluated in terms of its
economics[146]. It indicates that CO production cost price translates into 0.22 $/kWhr stored in CO at
an electricity price of 0.05 $/kWhr based on present day technology and performance.

References (EndNote)

[1] Bogaerts, A., et al., Gas discharge plasmas and their applications. Spectrochimica Acta Part B:
Atomic Spectroscopy, 2002. 57(4): p. 609-658.
[2] Fridman, A., Plasma chemistry. 2008: Cambridge university press.
[3] Itikawa, Y., Cross sections for electron collisions with carbon dioxide. Journal of Physical and
Chemical Reference Data, 2002. 31(3): p. 749-767.
[4] Semiokhin, I., Y.P. Andreev, and G. Panchenkov, Dissociation of Carbon Dioxide Circulating in
a Silent Electric Discharge. Russian Journal of Physical Chemistry, 1964. 38(8): p. 1126.
[5] Azizov, R., et al. The nonequilibrium plasmachemical process of decomposition of CO2 in a
supersonic SHF discharge. in Soviet Physics Doklady. 1983.
[6] Maltsev, A., E. Eremin, and V. IVANTER, DISSOCIATION OF CARBON DIOXIDE IN GLOW
DISCHARGE. 1967, INTERPERIODICA PO BOX 1831, BIRMINGHAM, AL 35201-1831. p. 633-&.
[7] Andreev, Y., et al., DISSOCIATION OF CARBON-DIOXIDE IN A PULSE DISCHARGE. RUSSIAN
JOURNAL OF PHYSICAL CHEMISTRY, USSR, 1971. 45(11): p. 1587-&.
[8] Rusanov, V., A. Fridman, and G. Sholin, The physics of a chemically active plasma with
nonequilibrium vibrational excitation of molecules. Physics-Uspekhi, 1981. 24(6): p. 447-474.
[9] Asisov, R.F., AA; Givotov, VK; Krasheninnikov, EG; Petrushev, BI; Potapkin, BV; Rusanov, VD;
Krotov, MF and Kurchatov, IV, Carbon dioxide dissociation in non-equilibrium plasma, in 5th
Int. Symp. on Plasma Chemistry 1981: Edinburgh, Scotland. p. 774.
[10] De Bie, C., J. van Dijk, and A. Bogaerts, The Dominant Pathways for the Conversion of Methane
into Oxygenates and Syngas in an Atmospheric Pressure Dielectric Barrier Discharge. Journal
of Physical Chemistry C, 2015. 119(39): p. 22331-22350.
[11] Kogelschatz, U., Dielectric-barrier discharges: their history, discharge physics, and industrial
applications. Plasma Chemistry and Plasma Processing, 2003. 23(1): p. 1-46.
[12] Snoeckx, R. and A. Bogaerts, Plasma technology–a novel solution for CO2 conversion?
Chemical Society Reviews, 2017. 46(19): p. 5805-5863.
[13] Van Laer, K. and A. Bogaerts, Fluid modelling of a packed bed dielectric barrier discharge
plasma reactor. Plasma Sources Science and Technology, 2015. 25(1): p. 015002.
[14] Michielsen, I., et al., CO2 dissociation in a packed bed DBD reactor: First steps towards a better
understanding of plasma catalysis. Chemical Engineering Journal, 2017. 326: p. 477-488.
[15] Uytdenhouwen, Y., et al., A packed-bed DBD micro plasma reactor for CO2 dissociation: Does
size matter? Chemical Engineering Journal, 2018. 348: p. 557-568.
[16] Bogaerts, A. and E.C. Neyts, Plasma Technology: An Emerging Technology for Energy Storage.
ACS Energy Letters, 2018. 3(4): p. 1013-1027.
[17] Azizov, R., et al. Nonequilibrium plasmachemical process of CO2 decomposition in a supersonic
microwave discharge. in Akademiia Nauk SSSR Doklady. 1983.
[18] Kozák, T. and A. Bogaerts, Splitting of CO2 by vibrational excitation in non-equilibrium plasmas:
a reaction kinetics model. Plasma Sources Science and Technology, 2014. 23(4): p. 045004.
[19] Berthelot, A. and A. Bogaerts, Modeling of CO2 splitting in a microwave plasma: how to
improve the conversion and energy efficiency. The Journal of Physical Chemistry C, 2017.
121(15): p. 8236-8251.
[20] Sun, S., et al., CO2 conversion in a gliding arc plasma: Performance improvement based on
chemical reaction modeling. Journal of CO2 Utilization, 2017. 17: p. 220-234.
[21] Ramakers, M., et al., Gliding Arc Plasmatron: providing an alternative method for carbon
dioxide conversion. ChemSusChem, 2017.
[22] Cleiren, E., et al., Dry Reforming of Methane in a Gliding Arc Plasmatron: Towards a Better
Understanding of the Plasma Chemistry. ChemSusChem, 2017. 10(20): p. 4025-4036.
[23] Scapinello, M., et al., Conversion of CH4/CO2 by a nanosecond repetitively pulsed discharge.
Journal of Physics D: Applied Physics, 2016. 49(7): p. 075602.
[24] Zhu, B., et al., Kinetics study on carbon dioxide reforming of methane in kilohertz spark-
discharge plasma. Chemical Engineering Journal, 2015. 264: p. 445-452.
[25] Indarto, A., et al., Effect of additive gases on methane conversion using gliding arc discharge.
Energy, 2006. 31(14): p. 2986-2995.
[26] Li, D., et al., CO2 reforming of CH4 by atmospheric pressure glow discharge plasma: a high
conversion ability. International Journal of Hydrogen Energy, 2009. 34(1): p. 308-313.
[27] Neyts, E.C., et al., Plasma catalysis: synergistic effects at the nanoscale. Chemical Reviews,
2015. 115(24): p. 13408-13446.
[28] Whitehead, J.C., Plasma–catalysis: the known knowns, the known unknowns and the unknown
unknowns. Journal of Physics D: Applied Physics, 2016. 49(24): p. 243001.
[29] Tu, X., J.C. Whitehead, and T. Nozaki, Plasma-catalysis: Fundamentals and applications. 2018:
Springer.
[30] Blin‐Simiand, N., et al., Removal of 2 ‐ Heptanone by Dielectric Barrier Discharges – The
Effect of a Catalyst Support. Plasma Processes and Polymers, 2005. 2(3): p. 256-262.
[31] Hong, J., et al., Cobalt species and cobalt-support interaction in glow discharge plasma-
assisted Fischer–Tropsch catalysts. Journal of Catalysis, 2010. 273(1): p. 9-17.
[32] Zou, J.-J., Y.-p. Zhang, and C.-J. Liu, Reduction of Supported Noble-Metal Ions Using Glow
Discharge Plasma. Langmuir, 2006. 22(26): p. 11388-11394.
[33] Demidyuk, V. and J.C. Whitehead, Influence of temperature on gas-phase toluene
decomposition in plasma-catalytic system. Plasma Chemistry and Plasma Processing, 2007.
27(1): p. 85-94.
[34] Shang, S., et al., Research on Ni/γ-Al2O3 catalyst for CO2 reforming of CH4 prepared by
atmospheric pressure glow discharge plasma jet. Catalysis Today, 2009. 148(3-4): p. 268-274.
[35] Liu, C.-j., R. Mallinson, and L. Lobban, Nonoxidative methane conversion to acetylene over
zeolite in a low temperature plasma. Journal of Catalysis, 1998. 179(1): p. 326-334.
[36] Holzer, F., F. Kopinke, and U. Roland, Influence of ferroelectric materials and catalysts on the
performance of non-thermal plasma (NTP) for the removal of air pollutants. Plasma Chemistry
and Plasma Processing, 2005. 25(6): p. 595-611.
[37] Guaitella, O., et al., C2H2 oxidation by plasma/TiO2 combination: influence of the porosity,
and photocatalytic mechanisms under plasma exposure. Applied Catalysis B: Environmental,
2008. 80(3-4): p. 296-305.
[38] Liu, C.-j., et al., Floating double probe characteristics of non-thermal plasmas in the presence
of zeolite. Journal of Electrostatics, 2002. 54(2): p. 149-158.
[39] Kim, H.-H., J.-H. Kim, and A. Ogata, Microscopic observation of discharge plasma on the surface
of zeolites supported metal nanoparticles. Journal of Physics D: Applied Physics, 2009. 42(13):
p. 135210.
[40] Tu, X., et al., Dry reforming of methane over a Ni/Al2O3 catalyst in a coaxial dielectric barrier
discharge reactor. Journal of Physics D: Applied Physics, 2011. 44(27): p. 274007.
[41] Nozaki, T., et al., Dissociation of vibrationally excited methane on Ni catalyst: Part 2. Process
diagnostics by emission spectroscopy. Catalysis Today, 2004. 89(1-2): p. 67-74.
[42] Mizuno, A., Generation of non-thermal plasma combined with catalysts and their application
in environmental technology. Catalysis Today, 2013. 211: p. 2-8.
[43] Hensel, K., Microdischarges in ceramic foams and honeycombs. The European Physical Journal
D, 2009. 54(2): p. 141.
[44] Rousseau, A., et al., Combination of a pulsed microwave plasma with a catalyst for acetylene
oxidation. Applied Physics Letters, 2004. 85(12): p. 2199-2201.
[45] Neyts, E. and A. Bogaerts, Understanding plasma catalysis through modelling and
simulation—a review. Journal of Physics D: Applied Physics, 2014. 47(22): p. 224010.
[46] Bongers, W., et al., Plasma‐driven dissociation of CO2 for fuel synthesis. Plasma processes
and polymers, 2017. 14(6): p. 1600126.
[47] den Harder, N., et al., Homogeneous CO2 conversion by microwave plasma: Wave propagation
and diagnostics. Plasma Processes and Polymers, 2017. 14(6): p. 1600120.
[48] Berthelot, A. and A. Bogaerts, Pinpointing energy losses in CO 2 plasmas–Effect on CO 2
conversion. Journal of CO2 Utilization, 2018. 24: p. 479-499.
[49] Kozák, T. and A. Bogaerts, Evaluation of the energy efficiency of CO2 conversion in microwave
discharges using a reaction kinetics model. Plasma Sources Science and Technology, 2014.
24(1): p. 015024.
[50] Snoeckx, R. and A. Bogaerts, Plasma technology–a novel solution for CO 2 conversion?
Chemical Society Reviews, 2017. 46(19): p. 5805-5863.
[51] Ozkan, A., A. Bogaerts, and F. Reniers, Routes to increase the conversion and the energy
efficiency in the splitting of CO2 by a dielectric barrier discharge. Journal of Physics D: Applied
Physics, 2017. 50(8): p. 084004.
[52] Tsuji, M., et al., Decomposition of CO2 into CO and O in a microwave-excited discharge flow of
CO2/He or CO2/Ar mixtures. Chemistry Letters, 2001. 30(1): p. 22-23.
[53] Vesel, A., et al., Dissociation of CO2 molecules in microwave plasma. Chemical Physics, 2011.
382(1-3): p. 127-131.
[54] Spencer, L. and A. Gallimore, CO2 dissociation in an atmospheric pressure plasma/catalyst
system: a study of efficiency. Plasma Sources Science and Technology, 2012. 22(1): p. 015019.
[55] Goede, A.P.H., et al., Production of solar fuels by CO2 plasmolysis. EPJ Web of Conferences,
2014. 79: p. 01005.
[56] Van Rooij, G., et al., Taming microwave plasma to beat thermodynamics in CO2 dissociation.
Faraday discussions, 2015. 183: p. 233-248.
[57] Chen, G., et al., Plasma assisted catalytic decomposition of CO2. Applied Catalysis B:
Environmental, 2016. 190: p. 115-124.
[58] Nunnally, T., et al., Dissociation of CO2 in a low current gliding arc plasmatron. Journal of
Physics D: Applied Physics, 2011. 44(27): p. 274009.
[59] Kim, S.C., M.S. Lim, and Y.N. Chun, Reduction characteristics of carbon dioxide using a
plasmatron. Plasma Chemistry and Plasma Processing, 2014. 34(1): p. 125-143.
[60] Wang, W., et al., CO2 conversion in a gliding arc plasma: 1D cylindrical discharge model.
Plasma Sources Science and Technology, 2016. 25(6): p. 065012.
[61] Wang, W., et al., Gliding arc plasma for CO2 conversion: better insights by a combined
experimental and modelling approach. Chemical Engineering Journal, 2017. 330: p. 11-25.
[62] Heijkers, S. and A. Bogaerts, CO2 Conversion in a Gliding Arc Plasmatron: Elucidating the
Chemistry through Kinetic Modeling. The Journal of Physical Chemistry C, 2017. 121(41): p.
22644-22655.
[63] Trenchev, G., et al., CO2 Conversion in a Gliding Arc Plasmatron: Multidimensional Modeling
for Improved Efficiency. The Journal of Physical Chemistry C, 2017. 121(44): p. 24470-24479.
[64] Zhang, A.-J., et al., Conversion of greenhouse gases into syngas via combined effects of
discharge activation and catalysis. Chemical Engineering Journal, 2010. 156(3): p. 601-606.
[65] Tu, X. and J.C. Whitehead, Plasma-catalytic dry reforming of methane in an atmospheric
dielectric barrier discharge: Understanding the synergistic effect at low temperature. Applied
Catalysis B-Environmental, 2012. 125: p. 439-448.
[66] Tu, X. and J.C. Whitehead, Plasma dry reforming of methane in an atmospheric pressure AC
gliding arc discharge: Co-generation of syngas and carbon nanomaterials. International
Journal of Hydrogen Energy, 2014. 39(18): p. 9658-9669.
[67] Snoeckx, R., et al., Plasma-based dry reforming: improving the conversion and energy
efficiency in a dielectric barrier discharge. RSC Adv., 2015. 5(38): p. 29799-29808.
[68] Mei, D.H., S.Y. Liu, and X. Tu, CO 2 reforming with methane for syngas production using a
dielectric barrier discharge plasma coupled with Ni/γ-Al 2 O 3 catalysts: Process optimization
through response surface methodology. Journal of CO2 Utilization, 2017. 21: p. 314-326.
[69] Eliasson, B., C.J. Liu, and U. Kogelschatz, Direct conversion of methane and carbon dioxide to
higher hydrocarbons using catalytic dielectric-barrier discharges with zeolites. Industrial &
Engineering Chemistry Research, 2000. 39(5): p. 1221-1227.
[70] Snoeckx, R., et al., Plasma-based dry reforming: a computational study ranging from the
nanoseconds to seconds time scale. The Journal of Physical Chemistry C, 2013. 117(10): p.
4957-4970.
[71] Yao, S.L., et al., Plasma reforming and coupling of methane with carbon dioxide. Energy &
Fuels, 2001. 15(5): p. 1295-1299.
[72] Zhang, Y.P., et al., Plasma methane conversion in the presence of carbon dioxide using
dielectric-barrier discharges. Fuel Processing Technology, 2003. 83(1-3): p. 101-109.
[73] Kozlov, K.V., P. Michel, and H.E. Wagner, Synthesis of organic compounds from mixtures of
methane with carbon dioxide in dielectric-barrier discharges at atmospheric pressure. Plasmas
and Polymers, 2001. 5(3-4): p. 129-150.
[74] Martini, L.M., et al., Oxidation of CH4 by CO2 in a dielectric barrier discharge. Chemical Physics
Letters, 2014. 593: p. 55-60.
[75] Zou, J.-J., et al., Starch-Enhanced Synthesis of Oxygenates from Methane and Carbon Dioxide
Using Dielectric-Barrier Discharges. Plasma Chemistry and Plasma Processing, 2003. 23(1): p.
69-82.
[76] Li, Y., et al., Synthesis of oxygenates and higher hydrocarbons directly from methane and
carbon dioxide using dielectric-barrier discharges: Product distribution. Energy & Fuels, 2002.
16(4): p. 864-870.
[77] Bogaerts, A., et al., Plasma based CO2and CH4conversion: A modeling perspective. Plasma
Processes and Polymers, 2016.
[78] Wang, L., et al., One-Step Reforming of CO2 and CH4 into High-Value Liquid Chemicals and
Fuels at Room Temperature by Plasma-Driven Catalysis. Angew Chem Int Ed Engl, 2017.
56(44): p. 13679-13683.
[79] Chung, W.-C. and M.-B. Chang, Simultaneous Generation of Syngas and Multiwalled Carbon
Nanotube via CH4/CO2 Reforming with Spark Discharge. ACS Sustainable Chemistry &
Engineering, 2017. 5(1): p. 206-212.
[80] Zhu, B., et al., Pressurization effect on dry reforming of biogas in kilohertz spark-discharge
plasma. International Journal of Hydrogen Energy, 2012. 37(6): p. 4945-4954.
[81] Abd Allah, Z. and J.C. Whitehead, Plasma-catalytic dry reforming of methane in an
atmospheric pressure AC gliding arc discharge. Catalysis Today, 2015. 256: p. 76-79.
[82] Zeng, Y.X., et al., Plasma-catalytic dry reforming of methane over gamma-Al2O3 supported
metal catalysts. Catalysis Today, 2015. 256: p. 80-87.
[83] Zheng, X.G., et al., Plasma-assisted catalytic dry reforming of methane: Highly catalytic
performance of nickel ferrite nanoparticles embedded in silica. Journal of Power Sources,
2015. 274: p. 286-294.
[84] Zhu, F., et al., Plasma-catalytic reforming of CO 2 -rich biogas over Ni/γ-Al 2 O 3 catalysts in a
rotating gliding arc reactor. Fuel, 2017. 199: p. 430-437.
[85] Zeng, Y.X., et al., Low temperature reforming of biogas over K-, Mg- and Ce-promoted Ni/Al 2
O 3 catalysts for the production of hydrogen rich syngas: Understanding the plasma-catalytic
synergy. Applied Catalysis B: Environmental, 2018. 224: p. 469-478.
[86] Tu, X., et al., Dry reforming of methane over a Ni/Al2O3 catalyst in a coaxial dielectric barrier
discharge reactor. Journal of Physics D-Applied Physics, 2011. 44(27).
[87] Snoeckx, R. and A. Bogaerts, Plasma technology-a novel solution for CO2conversion? Chemical
Society Reviews, 2017. 46(19): p. 5805-5863.
[88] Mei, D. and X. Tu, Conversion of CO2in a cylindrical dielectric barrier discharge reactor: Effects
of plasma processing parameters and reactor design. Journal of CO2 Utilization, 2017. 19: p.
68-78.
[89] Zhang, K., U. Kogelschatz, and B. Eliasson, Conversion of greenhouse gases to synthesis gas
and higher hydrocarbons. Energy & Fuels, 2001. 15(2): p. 395-402.
[90] Song, H.K., et al., Synthesis gas production via dielectric barrier discharge over Ni/gamma-
Al2O3 catalyst. Catalysis Today, 2004. 89(1-2): p. 27-33.
[91] Aziznia, A., et al., Comparison of dry reforming of methane in low temperature hybrid plasma-
catalytic corona with thermal catalytic reactor over Ni/gamma-Al2O3. Journal of Natural Gas
Chemistry, 2012. 21(4): p. 466-475.
[92] Wang, Q., et al., Investigation of Dry Reforming of Methane in a Dielectric Barrier Discharge
Reactor. Plasma Chemistry and Plasma Processing, 2009. 29(3): p. 217-228.
[93] Mei, D., Plasma-catalytic conversion of greenhouse gas into value-added fuels and chemicals,
in Department of Electrical Engineering and Electronics. 2016, University of Liverpool:
Liverpool.
[94] Khoja, A.H., M. Tahir, and N.A.S. Amin, Dry reforming of methane using different dielectric
materials and DBD plasma reactor configurations. Energy Conversion and Management, 2017.
144: p. 262-274.
[95] Rico, V.J., et al., Evaluation of Different Dielectric Barrier Discharge Plasma Configurations As
an Alternative Technology for Green C-1 Chemistry in the Carbon Dioxide Reforming of
Methane and the Direct Decomposition of Methanol. Journal of Physical Chemistry A, 2010.
114(11): p. 4009-4016.
[96] Wu, A., et al., Study of the dry methane reforming process using a rotating gliding arc reactor.
International Journal of Hydrogen Energy, 2014. 39(31): p. 17656-17670.
[97] Ozkan, A., et al., CO2-CH4 conversion and syngas formation at atmospheric pressure using a
multi-electrode dielectric barrier discharge. Journal of Co2 Utilization, 2015. 9: p. 74-81.
[98] Li, M.W., et al., Effects of catalysts in carbon dioxide reforming of methane via corona plasma
reactions. Energy & Fuels, 2006. 20(3): p. 1033-1038.
[99] Sentek, J., et al., Plasma-catalytic methane conversion with carbon dioxide in dielectric barrier
discharges. Applied Catalysis B-Environmental, 2010. 94(1-2): p. 19-26.
[100] Ray, D., P.M.K. Reddy, and C. Subrahmanyam, Ni-Mn/γ-Al 2 O 3 assisted plasma dry reforming
of methane. Catalysis Today, 2018. 309: p. 212-218.
[101] Wang, Q., Y. Cheng, and Y. Jin, Dry reforming of methane in an atmospheric pressure plasma
fluidized bed with Ni/gamma-Al2O3 catalyst. Catalysis Today, 2009. 148(3-4): p. 275-282.
[102] Zhang, K., B. Eliasson, and U. Kogelschatz, Direct conversion of greenhouse gases to synthesis
gas and C-4 hydrocarbons over zeolite HY promoted by a dielectric-barrier discharge. Industrial
& Engineering Chemistry Research, 2002. 41(6): p. 1462-1468.
[103] Jiang, T., et al., Plasma methane conversion using dielectric-barrier discharges with zeolite A.
Catalysis Today, 2002. 72(3-4): p. 229-235.
[104] Bo, Z., et al., Plasma assisted dry methane reforming using gliding arc gas discharge: Effect of
feed gases proportion. International Journal of Hydrogen Energy, 2008. 33(20): p. 5545-5553.
[105] Kroker, T., et al., Catalytic Conversion of Simulated Biogas Mixtures to Synthesis Gas in a
Fluidized Bed Reactor Supported by a DBD. Plasma Chemistry and Plasma Processing, 2012.
32(3): p. 565-582.
[106] Krawczyk, K., et al., Methane conversion with carbon dioxide in plasma-catalytic system. Fuel,
2014. 117: p. 608-617.
[107] Pham, M.H., et al., Activation of methane and carbon dioxide in a dielectric-barrier discharge-
plasma reactor to produce hydrocarbons-Influence of La2O3/gamma-Al2O3 catalyst. Catalysis
Today, 2011. 171(1): p. 67-71.
[108] Mahammadunnisa, S., et al., Catalytic Nonthermal Plasma Reactor for Dry Reforming of
Methane. Energy & Fuels, 2013. 27(8): p. 4441-4447.
[109] Long, H., et al., CO2 reforming of CH4 by combination of cold plasma jet and Ni/gamma-Al2O3
catalyst. International Journal of Hydrogen Energy, 2008. 33(20): p. 5510-5515.
[110] Liu, C.J., et al., Methane conversion to higher hydrocarbons in the presence of carbon dioxide
using dielectric-barrier discharge plasmas. Plasma Chemistry and Plasma Processing, 2001.
21(3): p. 301-310.
[111] Hoang Hai, N., et al., Analysis on CO2 reforming of CH4 by corona discharge process for various
process variables. Journal of Industrial and Engineering Chemistry, 2015. 32: p. 58-62.
[112] Zhang, X.L., et al., The simultaneous activation of methane and carbon dioxide to C-2
hydrocarbons under pulse corona plasma over La2O3/gamma-Al2O3 catalyst. Catalysis Today,
2002. 72(3-4): p. 223-227.
[113] Kado, S., et al., Low temperature reforming of methane to synthesis gas with direct current
pulse discharge method. Chemical Communications, 2001(5): p. 415-416.
[114] Zheng, X., et al., LaNiO3@SiO2 core-shell nano-particles for the dry reforming of CH4 in the
dielectric barrier discharge plasma. International Journal of Hydrogen Energy, 2014. 39(22): p.
11360-11367.
[115] Zheng, X.G., et al., Silica-coated LaNiO3 nanoparticles for non-thermal plasma assisted dry
reforming of methane: Experimental and kinetic studies. Chemical Engineering Journal, 2015.
265: p. 147-156.
[116] Li, Y., et al., Co-generation of syngas and higher hydrocarbons from CO2 and CH4 using
dielectric-barrier discharge: Effect of electrode materials. Energy & Fuels, 2001. 15(2): p. 299-
302.
[117] Scapinello, M., L.M. Martini, and P. Tosi, CO2 Hydrogenation by CH4 in a Dielectric Barrier
Discharge: Catalytic Effects of Nickel and Copper. Plasma Processes and Polymers, 2014. 11(7):
p. 624-628.
[118] Goujard, V., J.M. Tatibouet, and C. Batiot-Dupeyrat, Influence of the Plasma Power Supply
Nature on the Plasma-Catalyst Synergism for the Carbon Dioxide Reforming of Methane. Ieee
Transactions on Plasma Science, 2009. 37(12): p. 2342-2346.
[119] Zou, J.-J. and C.-J. Liu, Utilization of carbon dioxide through nonthermal plasma approaches.
Carbon dioxide as chemical feedstock, 2010: p. 267-290.
[120] Amouroux, J., S. Cavadias, and A. Doubla. Carbon dioxide reduction by non-equilibrium
electrocatalysis plasma reactor. in IOP conference series: materials science and engineering.
2011. IOP Publishing.
[121] Zeng, Y. and X. Tu, Plasma-catalytic hydrogenation of CO2 for the cogeneration of CO and CH4
in a dielectric barrier discharge reactor: effect of argon addition. Journal of Physics D: Applied
Physics, 2017. 50(18): p. 184004.
[122] Hayashi, N., T. Yamakawa, and S. Baba, Effect of additive gases on synthesis of organic
compounds from carbon dioxide using non-thermal plasma produced by atmospheric surface
discharges. Vacuum, 2006. 80(11-12): p. 1299-1304.
[123] Wang, L., et al., Atmospheric Pressure and Room Temperature Synthesis of Methanol through
Plasma-Catalytic Hydrogenation of CO2. ACS Catalysis, 2018. 8(1): p. 90-100.
[124] Arita, K. and S. Iizuka, Production of CH4 in a low-pressure CO2/H2 discharge with magnetic
field. Journal of Materials Science and Chemical Engineering, 2015. 3(12): p. 69.
[125] Nizio, M., et al., Hybrid plasma-catalytic methanation of CO2 at low temperature over ceria
zirconia supported Ni catalysts. International Journal of Hydrogen Energy, 2016. 41(27): p.
11584-11592.
[126] Jwa, E., et al., Plasma-assisted catalytic methanation of CO and CO2 over Ni–zeolite catalysts.
Fuel processing technology, 2013. 108: p. 89-93.
[127] Porosoff, M.D., B. Yan, and J.G. Chen, Catalytic reduction of CO 2 by H 2 for synthesis of CO,
methanol and hydrocarbons: challenges and opportunities. Energy & Environmental Science,
2016. 9(1): p. 62-73.
[128] Zeng, Y. and X. Tu, Plasma-catalytic CO2 hydrogenation at low temperatures. IEEE
Transactions on Plasma Science, 2015. 44(4): p. 405-411.
[129] Kano, M., G. Satoh, and S. Iizuka, Reforming of Carbon Dioxide to Methane and Methanol by
Electric Impulse Low-Pressure Discharge with Hydrogen. Plasma Chemistry and Plasma
Processing, 2012. 32(2): p. 177-185.
[130] Mora, E.Y., A. Sarmiento, and E. Vera, Alumina and quartz as dielectrics in a dielectric barrier
discharges DBD system for CO 2 hydrogenation. Journal of Physics: Conference Series, 2016.
687(1): p. 012020.
[131] De Bie, C., J. van Dijk, and A. Bogaerts, CO ₂ Hydrogenation in a Dielectric Barrier
Discharge Plasma Revealed. The Journal of Physical Chemistry C, 2016. 120(44): p. 25210-
25224.
[132] Wang, W., et al., Recent advances in catalytic hydrogenation of carbon dioxide. Chemical
Society Reviews, 2011. 40(7): p. 3703-3727.
[133] Eliasson, B., et al., Hydrogenation of Carbon Dioxide to Methanol with a Discharge-Activated
Catalyst. Industrial & Engineering Chemistry Research, 1998. 37(8): p. 3350-3357.
[134] Nizio, M., et al., Low temperature hybrid plasma-catalytic methanation over Ni-Ce-Zr
hydrotalcite-derived catalysts. Catalysis Communications, 2016. 83: p. 14-17.
[135] Futamura, S. and H. Kabashima, Synthesis gas production from CO2 and H2O with nonthermal
plasma, in Studies in Surface Science and Catalysis. 2004, Elsevier. p. 119-124.
[136] Mahammadunnisa, S., et al., CO2 reduction to syngas and carbon nanofibres by plasma-
assisted in situ decomposition of water. International Journal of Greenhouse Gas Control,
2013. 16: p. 361-363.
[137] Snoeckx, R., et al., The Quest for Value‐Added Products from Carbon Dioxide and Water in a
Dielectric Barrier Discharge: A Chemical Kinetics Study. ChemSusChem, 2017. 10(2): p. 409-
424.
[138] Ihara, T., M. Kiboku, and Y. Iriyama, Plasma reduction of CO2 with H2O for the formation of
organic compounds. Bulletin of the Chemical Society of Japan, 1994. 67(1): p. 312-314.
[139] Ihara, T., et al., Formation of methanol by microwave-plasma reduction of CO2 with H2O.
Bulletin of the Chemical Society of Japan, 1996. 69(1): p. 241-244.
[140] Chen, G., et al., Simultaneous dissociation of CO2 and H2O to syngas in a surface-wave
microwave discharge. International Journal of Hydrogen Energy, 2015. 40(9): p. 3789-3796.
[141] Chen, G., et al., An overview of CO2 conversion in a microwave discharge: the role of plasma-
catalysis. Journal of Physics D: Applied Physics, 2017. 50(8): p. 084001.
[142] Indarto, A., et al., Gliding arc plasma processing of CO2 conversion. Journal of Hazardous
Materials, 2007. 146(1-2): p. 309-315.
[143] Guo, L., et al., A novel method of production of ethanol by carbon dioxide with steam. Fuel,
2015. 158: p. 843-847.
[144] Centi, G. and S. Perathoner, Opportunities and prospects in the chemical recycling of carbon
dioxide to fuels. Catalysis Today, 2009. 148(3-4): p. 191-205.
[145] Studt, F., et al., Discovery of a Ni-Ga catalyst for carbon dioxide reduction to methanol. Nature
Chemistry, 2014. 6(4): p. 320.
[146] Van Rooij, G., et al., Plasma for electrification of chemical industry: a case study on CO2
reduction. Plasma Physics and Controlled Fusion, 2017. 60(1): p. 014019.

Fig. 1

Fig. 2
Fig. 3

Fig. 4
Fig. 5

Fig. 6

Fig. 7
Fig. 8

Fig 9

Fig. 10
 (a)

(b)

Fig. 11

Fig. 12
 Fig. 13

(a) (b)
Fig. 14

Fig. 15
Fig. 16

(a)

(b)
Fig. 17

Fig. 18

(a)

(b)

Fig. 19
Fig. 20