**4.1 Thin-film solar cells**

Among the plasma deposited amorphous semiconductors, which have so far been studied in greatest detail and also appear to be of greatest applied interest, are hydrogenated amorphous silicon films (a-Si:H). Initial studies of this material were done in the sixties of the 20th century, when the RF glow discharge deposition of a-Si:H from silane was demonstrated. This was followed by the very important works that reported on successful *n*- and *p*-doping a-Si:H. It was really surprising, because previous attempts to dope thermally evaporated a-Si had failed. As it turned out, hydrogen was responsible for this effect. In 1975 and 1976, research confirmed that plasma-deposited films from silane contained hydrogen. Hydrogen serves primarily to passivate the dangling-bond defects and

they should be produced from the same precursor (monomer), have practically nothing in common. The fundamental difference is that mer units cannot be defined in the case of plasma polymers. A large variety of chemical species created in the plasma, statistical combination them into high molecular structures and generally a high degree of their crosslinking cause that the structure of such a material is very often much closer to that of

Frequently plasma polymers are not classified in respect of the type of monomer but from a point of view of their chemical composition and morphology. For example, amorphous (a-) covalent material obtained by plasma polymerization of silane (SiH4), which is composed of silicon and hydrogen, can be termed as a-Si:H. In turn, amorphous plasma polymer deposited from acrylonitrile (C3H3N) can be called as plasma-polymerized (pp-) acrylonitrile or a-CXNY:H. It is usually met, but it is not a rule, that if the plasma polymer structure is close to a covalent glass structure, the latter notation is used. If plasma polymer reveals nano- or microcrystalline structures, prefixes nc- or µc- are put in the place of a-.

The structure and properties of plasma polymers are closely connected with a thin-film form, in which they are produced. In general, the thickness of the films is between a few nanometers and a few micrometers. Appropriate choice of precursors and plasma process parameters allow for the preparation of such thin films with a huge variety of structure and properties. Hence, there is a wide and diverse range of their current and anticipated applications, such as electronic and photoelectronic materials, insulating coatings, catalytic films, semi-permeable and electrolyte membranes, protecting layers, and many others. Some

Optical and electrical properties of plasma deposited films, sometimes unique indeed, as well as the easy of their deposition, at low temperature and low cost, on inexpensive substrates of almost any size and shape, render these materials very attractive for optoelectronic applications. The possibility to tailor optical parameters, such as refractive index and extinction coefficient, and what is particularly important - the ability to adjust parameters of the electronic structure, such as transport gap, optical gap, density of localized states, etc., recommend these plasma films as active photoelectric elements, e.g. for

Among the plasma deposited amorphous semiconductors, which have so far been studied in greatest detail and also appear to be of greatest applied interest, are hydrogenated amorphous silicon films (a-Si:H). Initial studies of this material were done in the sixties of the 20th century, when the RF glow discharge deposition of a-Si:H from silane was demonstrated. This was followed by the very important works that reported on successful *n*- and *p*-doping a-Si:H. It was really surprising, because previous attempts to dope thermally evaporated a-Si had failed. As it turned out, hydrogen was responsible for this effect. In 1975 and 1976, research confirmed that plasma-deposited films from silane contained hydrogen. Hydrogen serves primarily to passivate the dangling-bond defects and

covalent glasses than that of conventional polymers.

of these uses are also related to the electrochemical systems.

**4. Optoelectronically active materials** 

solar cells and water splitting cells.

**4.1 Thin-film solar cells** 

thus to decrease the density of localized states in the mid-gap.2 Thermally evaporated a-Si has about 1026 eV−1m−3 states in the mid-gap, whereas the density of states for typical a-Si:H films is about 1021 eV−1m−3, i.e. five orders of magnitude lower. This just explains why one can control the electronic properties of a-Si:H by doping with donor and acceptor centers, contrary to a-Si (LeComber & Spear, 1979).

The most frequent method used to prepare doped films is plasma copolymerization from a mixture of the film precursor and the dopant agent. Diborane (B2H6) and phosphine (PH3) are often used as sources of acceptor centers (boron atoms) and donor centers (phosphorus atoms), respectively. Recently, liquid compounds instead of these gases, such as triethylboron (B(C2H5)3) and trimethylphosphine (P(CH3)3) have become more and more popular dopant agents. They are less toxic, more stable and their low vapor pressure offers facilities for precise controlling of the doping process. In turn, as a-Si:H film precursor, one can use not only SiH4, but also, for example, a mixture of SiCl4 and H2, disilane (Si2H6), trisilane (Si3H8), cyclohexasilane (Si6H12), etc. (Pokhodnya et al., 2009; Searle, 1998; Tyczkowski, 2004). So, as one can see, the possibilities of designing and controlling the molecular and electronic structure of plasma deposited films are indeed enormous.

As an example of designing the electronic structure of a-Si:H films, the electrical conductivity of these films doped with acceptors (boron) and donors (phosphorus) is shown in Fig. 5. The room-temperature conductivity *σ* of the films is plotted against the ratio of the number of dopant agent molecules to the number of silane molecules in the gaseous mixture. In the center of the graph, the conductivity around 10−6 S/m is representative of undoped a-Si:H films, which typically are *n*-type material. Thus, even a small quantity of P atoms (donors) increases *σ* rapidly. In the case of B atoms (acceptors), however, we see that initially *σ* decreases to about 10−10 S/m. This is connected with the transition from *n*-type to *p-*type material (LeComber & Spear, 1979; Tyczkowski, 2004).

Fig. 5. Room temperature conductivity *σ* of *n*- and *p*-type a-Si:H, plotted as a function of the gaseous precursor ratio (Tyczkowski, 2004).

<sup>2</sup> For more detailed description of the electronic structure models for plasma deposited amorphous films, see, for example, (Tyczkowski, 2004).

Cold Plasma – A Promising Tool for the Development of Electrochemical Cells 113

Among the promising materials in this respect are thin films (e.g. a-GeXCY:H, a-SiXCY:H, a-SnXCY:H, a-PbXCY:H), which can exist in two forms with totally different electronic structures, namely, the semiconducting (a-S) and insulating (a-I) form. In general, a typical amorphous semiconductor is characterized by localized states that form only short tiles in the mid-gap above and below the valence and conduction bands of extended states whereas in amorphous insulators all states are localized. In consequence, drastic differences in electrical and optical properties of a-S and a-I films are observed. It turned out that a particularly useful system for the production of these two qualitatively different materials is the three-electrode reactor presented in *Sec. 2*. Both a-S and a-I can be fabricated in this reactor from a single precursor (organometallic or organosemimetallic compound, e.g. tetramethylsilane, tetramethyltin, etc.) in the same deposition process, only by changing the impact energy of ions bombarding the growing film. The ion energy is controlled by the sheath voltage of the small electrode (*V*(-)), which in turn is governed by the coupling capacitance (Fig. 3). It is especially striking that very small variation of the sheath voltage in a defined range of its values is sufficient to create the step change in the electronic structure of deposited film. This "transition" between two forms of the film (a-I and a-S) has been called the a-I–a-S transition. As an example, changes in electrical conductivity *σ* of selected films deposited in the three-electrode reactor are shown in Fig. 6 (Tyczkowski, 2004, 2006).

Fig. 6. Electrical conductivity *σ* of films deposited in the three-electrode AF reactor from organic derivatives of the carbon family as a function of V(−): ( ) a-SiXCY:H, ( ) a-GeXCY:H,

a-GeXCY:H films for water splitting are being undertaken now in our laboratory.

The existence of two qualitatively different, from the electronic structure point of view, types of plasma deposited films offers new possibilities for material technology consisting in preparation of a novel class nanocomposites formed from insulating and semiconducting fractions (in the form of clusters or layers) deposited in the same plasma process from the same single-source precursor. Without a doubt, such nanocomposites are not only interesting for solar cells, but also for other electrochemical systems. An attempt to use

( ) a-SnXCY:H, ( ) a-PbXCY:H (Tyczkowski, 2006).

Thin-film solar cells are without doubt one of the most spectacular application (except thinfilm transistors) of a-Si:H films obtained by plasma polymerization processes. Since the demonstration of the first a-Si:H photovoltaic devices at RCA laboratories in 1976 (Carlson & Wronski, 1976) there has been remarkable progress in the development of a-Si:H solar cells, spurred, first of all, by the wide demand for a low cost, clean and safe energy (Mueller, 2009).

The most extensively studied a-Si:H solar cells, due to their highest conversion efficiencies, are those fabricated in the form of *p-i-n* devices. Typically, the *p*-a-Si:H film is less than 10 nm, the undoped *i*-a-Si:H is between 200 and 700 nm and *n*-a-Si:H layer is approx. 30 50 nm. The layers are deposited on each other in successive plasma reactor chambers connected by vacuum locks. A metallic electrode is used as a substrate. An opposite optically transparent and electrically conducting electrode (e.g. ITO film) is deposited from the top. Today, the commercial large-area solar cell modules based on a-Si:H are fabricated with a stabilized conversion efficiency (the ratio of the maximum power output to the solar energy input) in the 4–6 % range (Green, 2007).

To improve the efficiency and stability of a-Si:H solar cells a lot of various experimental investigations coupled with theoretical device modeling and design analysis have been carried out for the past two decades. Among the most important ideas are multiple-junction solar cell structures fabricated from different amorphous semiconducting films, in addition to a-Si:H, such as a-SiXCY:H, a-SiXGeY:H, and other related films. These films, classified as silicon-based alloys, can be plasma deposited from single precursors, e.g. tetramethylsilane (Si(CH3)4) or mixtures of precursors such as SiH4 and CH4, SiH4 and GeH4, etc. The films can be also doped with donor or acceptor atoms introduced to their structure during the plasma deposition process by addition of an appropriate dopant agent. An example of such a multiple-junction solar cell could be a triple-junction system composed of three *p-i-n* structures plasma-deposited on top of one another, prepared consecutively with the following amorphous semiconductors: a-Si:H, a-SiX1GeY1:H and a-SiX2GeY2:H. The system gives the stable efficiency of 10.4 % for a 900 cm2 module (Green et al., 2011).

Apart from the thin films of amorphous semiconductors, nanocrystalline form of these materials has also attracted much attention due to its higher efficiency and stability. While an amorphous semiconductor, for example, a-Si:H is a single phase material, its nanocrystalline form (nc-Si:H, also very often, but less correctly, called µc-Si:H) can be described as a bi-phasic material consisting of a dispersion of silicon nanocrystals embedded in silicon or other silicon-based hydrogenated amorphous matrices, whose volume fraction could be varied by selecting the proper plasma deposition conditions. The nc-Si:H films have been recently demonstrated to be an interesting alternative to a-Si:H films (Conibeer et al., 2006).

Although the efficiency of a-(or nc-)Si:H based solar cells is considerably lower than, for example, that for Si crystalline cells (25.0±0.5 %) or GaAs crystalline cells (26.1±0.8 %) (Green et al., 2011), much lower cost and easy of production are the major arguments in favor of the plasma-deposited solar cells. It should be noted that the last word has not been said yet in this regard and further significant progress is expected, provided that the substantial evolution in the field of new materials will be achieved.

Thin-film solar cells are without doubt one of the most spectacular application (except thinfilm transistors) of a-Si:H films obtained by plasma polymerization processes. Since the demonstration of the first a-Si:H photovoltaic devices at RCA laboratories in 1976 (Carlson & Wronski, 1976) there has been remarkable progress in the development of a-Si:H solar cells, spurred, first of all, by the wide demand for a low cost, clean and safe energy (Mueller,

The most extensively studied a-Si:H solar cells, due to their highest conversion efficiencies, are those fabricated in the form of *p-i-n* devices. Typically, the *p*-a-Si:H film is less than 10 nm, the undoped *i*-a-Si:H is between 200 and 700 nm and *n*-a-Si:H layer is approx. 30 50 nm. The layers are deposited on each other in successive plasma reactor chambers connected by vacuum locks. A metallic electrode is used as a substrate. An opposite optically transparent and electrically conducting electrode (e.g. ITO film) is deposited from the top. Today, the commercial large-area solar cell modules based on a-Si:H are fabricated with a stabilized conversion efficiency (the ratio of the maximum power output to the solar

To improve the efficiency and stability of a-Si:H solar cells a lot of various experimental investigations coupled with theoretical device modeling and design analysis have been carried out for the past two decades. Among the most important ideas are multiple-junction solar cell structures fabricated from different amorphous semiconducting films, in addition to a-Si:H, such as a-SiXCY:H, a-SiXGeY:H, and other related films. These films, classified as silicon-based alloys, can be plasma deposited from single precursors, e.g. tetramethylsilane (Si(CH3)4) or mixtures of precursors such as SiH4 and CH4, SiH4 and GeH4, etc. The films can be also doped with donor or acceptor atoms introduced to their structure during the plasma deposition process by addition of an appropriate dopant agent. An example of such a multiple-junction solar cell could be a triple-junction system composed of three *p-i-n* structures plasma-deposited on top of one another, prepared consecutively with the following amorphous semiconductors: a-Si:H, a-SiX1GeY1:H and a-SiX2GeY2:H. The system

Apart from the thin films of amorphous semiconductors, nanocrystalline form of these materials has also attracted much attention due to its higher efficiency and stability. While an amorphous semiconductor, for example, a-Si:H is a single phase material, its nanocrystalline form (nc-Si:H, also very often, but less correctly, called µc-Si:H) can be described as a bi-phasic material consisting of a dispersion of silicon nanocrystals embedded in silicon or other silicon-based hydrogenated amorphous matrices, whose volume fraction could be varied by selecting the proper plasma deposition conditions. The nc-Si:H films have been recently demonstrated to be an interesting alternative to a-Si:H films (Conibeer et

Although the efficiency of a-(or nc-)Si:H based solar cells is considerably lower than, for example, that for Si crystalline cells (25.0±0.5 %) or GaAs crystalline cells (26.1±0.8 %) (Green et al., 2011), much lower cost and easy of production are the major arguments in favor of the plasma-deposited solar cells. It should be noted that the last word has not been said yet in this regard and further significant progress is expected, provided that the substantial

gives the stable efficiency of 10.4 % for a 900 cm2 module (Green et al., 2011).

evolution in the field of new materials will be achieved.

2009).

al., 2006).

energy input) in the 4–6 % range (Green, 2007).

Among the promising materials in this respect are thin films (e.g. a-GeXCY:H, a-SiXCY:H, a-SnXCY:H, a-PbXCY:H), which can exist in two forms with totally different electronic structures, namely, the semiconducting (a-S) and insulating (a-I) form. In general, a typical amorphous semiconductor is characterized by localized states that form only short tiles in the mid-gap above and below the valence and conduction bands of extended states whereas in amorphous insulators all states are localized. In consequence, drastic differences in electrical and optical properties of a-S and a-I films are observed. It turned out that a particularly useful system for the production of these two qualitatively different materials is the three-electrode reactor presented in *Sec. 2*. Both a-S and a-I can be fabricated in this reactor from a single precursor (organometallic or organosemimetallic compound, e.g. tetramethylsilane, tetramethyltin, etc.) in the same deposition process, only by changing the impact energy of ions bombarding the growing film. The ion energy is controlled by the sheath voltage of the small electrode (*V*(-)), which in turn is governed by the coupling capacitance (Fig. 3). It is especially striking that very small variation of the sheath voltage in a defined range of its values is sufficient to create the step change in the electronic structure of deposited film. This "transition" between two forms of the film (a-I and a-S) has been called the a-I–a-S transition. As an example, changes in electrical conductivity *σ* of selected films deposited in the three-electrode reactor are shown in Fig. 6 (Tyczkowski, 2004, 2006).

Fig. 6. Electrical conductivity *σ* of films deposited in the three-electrode AF reactor from organic derivatives of the carbon family as a function of V(−): ( ) a-SiXCY:H, ( ) a-GeXCY:H, ( ) a-SnXCY:H, ( ) a-PbXCY:H (Tyczkowski, 2006).

The existence of two qualitatively different, from the electronic structure point of view, types of plasma deposited films offers new possibilities for material technology consisting in preparation of a novel class nanocomposites formed from insulating and semiconducting fractions (in the form of clusters or layers) deposited in the same plasma process from the same single-source precursor. Without a doubt, such nanocomposites are not only interesting for solar cells, but also for other electrochemical systems. An attempt to use a-GeXCY:H films for water splitting are being undertaken now in our laboratory.

Cold Plasma – A Promising Tool for the Development of Electrochemical Cells 115

very recently also by creation of new materials. Invaluable in this respect seems to be the cold plasma technology, which allows to design the structure of fabricated materials in a

Fig. 7. Schematic representation of charge transfer within a photoelectrochemical cell involving a semiconducting photoanode and metal cathode (Nowotny et al., 2006).

daCruz et al., 2000; Maeda & Watanabe, 2005; Nakamura et al., 2001).

films (Brudnik et al., 2007; Dang et al., 2011; Huang et al., 2011).

The most attractive oxide to date, namely TiO2, has already been repeatedly produced by the cold plasma technology. In most reported works, either TiCl4 or Ti alkoxides (mainly titanium tetraisopropoxide, Ti(OC3H7)4) are used as the Ti-containing precursors of the plasma polymerization (PECVD) process, resulting in amorphous or crystalline films, with the nonstoichiometric (TiOX) or stoichiometric (TiO2) structure. For all of these films, their physicochemical properties are strongly dependent on the film structure, which can be effectively controlled by the deposition conditions (Battiston et al., 2000; Borrás et al., 2009;

 **PHOTO-ANODE AQUEOUS CATHODE (SEMICONDUCTOR) ELECTROLYTE (METAL)**

**H+**

**O2/H2O**

**1.23 eV**

**EF(C)**

**H+ /H2**

**hν**

**R**

TiO2 films are also obtained in a wide range by the reactive sputtering, usually using pure titanium as a target, and O2 as reactive gas. Similarly, as in the PECVD technique, also in this case, the sputtering process conditions control the structure of the deposited films, which in turn affects to a large extend the optical and photoelectrochemical properties of the

A particularly useful feature of the cold plasma technology is the possibility of co-deposition either by copolymerization of a mixture of precursors or by co-sputtering using more than one target or a mixture of several reactive gases. In this way we can get doped films as well as films with alloy-type structures. Numerous studies have been already done on the introduction into the TiO2 structure other atoms (e.g. C, N, S). For example, PECVD with the DC discharge carried out using mixtures of Ti(OC3H7)4 and nitrogen led to a Ti(OCN) film structure (Randeniya et al., 2007; Wierzchoń et al., 1993). Instead of nitrogen, ammonia can be introduced to plasma reactors (Weber et al., 1995). Precursors containing nitrogen in their chemical structure, e.g. tetrakis(dimethylamido) or (diethylamido)-titanium (Ti(N(CH3)2)4 or

very wide range (Walsh et al., 2009).

**EF(C)**

**e−**

**Eg**

**EC**

**e−**

**EV**

#### **4.2 Water splitting systems**

The most desirable method for production of hydrogen, which represents a sustainable fuel of the future, is photoelectrochemical (PEC) splitting of water by visible light. Theoretically, the PEC production of hydrogen has the capacity to provide global energy security at potentially low cost (James et al., 2009).

The most critical issue in PEC hydrogen generation is the development of a highperformance photoelectrode that exhibits high efficiency in the conversion of solar energy into chemical energy, resistance to corrosion in aqueous environment, and low processing costs. Metal oxides are most promising in this regard (Walter et al., 2010). After four decades of intensive research, however, no material has been found to simultaneously satisfy all the criteria required for widespread PEC application. No wonder that a broad search for new materials for photoelectrodes is still ongoing. Cold plasma technology is also involved in this activity (Randeniya et al., 2007; Slavcheva et al., 2007; Walsh et al., 2009; Zhu, F. et al., 2009).

A typical simple PEC cell with schematic representation of charge transfer is shown in Fig. 7. The cell is constructed from a semiconducting photoanode and metal cathode. The basic reaction steps involved in the PEC process in the cell are as follows (Nowotny et al., 2006): *1*. Photoionization over the band gap of the semiconductor:

$$h\nu \to e^{\cdot \cdot} + h^{\ast} \,, \tag{1}$$

where *h* is the Planck constant and *ν* is the light frequency; *2*. Charge separation:

$$e^{-} + h^{\*} \rightarrow e^{-} \\ \text{bulk} + h^{\*} \\ \text{surface} \\ \vdots \tag{2}$$

*3*. Reaction between water molecules and holes at the surface of the photoanode:

$$\rm H\_2O + 2lr^\* \to \rm 2H^+ + \rm 4\cdot O\_2 \tag{3}$$

*4*. Transport of hydrogen ions from the photoanode to the cathode through the liquid electrolyte; *5*. Transport of electrons to the cathode thorough the external circuit; and *6*. Reaction between electrons and hydrogen ions at the cathode:

$$2\mathbf{r}^- + 2\mathbf{H}^\* \to \mathbf{H}\_2 \,. \tag{4}$$

The first PEC cell for water splitting, with a rutile TiO2 photoanode and Pt counter cathode, was reported in 1972 (Fujishima & Honda, 1972). Following this discovery, intensive studies aiming at increasing the energy conversion efficiency of solar energy into chemical energy have been carried out, mainly on the analogous PEC cells, using TiO2 as the photoanode. Then other oxides, e.g. Fe2O3 and WO3, have been also tested. Despite the good catalytic activity of oxides such as TiO2, they are generally limited by too large band gaps (approx. 3 eV), which fail to absorb a significant fraction of visible light, resulting in poor solar to hydrogen conversion efficiencies under terrestrial conditions. This value should be reduced to 1.7–2.0 eV. Besides, there is a range of other problems, like incorrect alignment of band edges with respect to the water redox potentials, energy losses due to charge recombination, low density of surface active sites reacting with water molecules, low corrosion resistance, etc. Thus, the majority of PEC oxide research has focused on trying to solve these problems by the modification of known photoactive oxides (through their doping or alloying), and

The most desirable method for production of hydrogen, which represents a sustainable fuel of the future, is photoelectrochemical (PEC) splitting of water by visible light. Theoretically, the PEC production of hydrogen has the capacity to provide global energy security at

The most critical issue in PEC hydrogen generation is the development of a highperformance photoelectrode that exhibits high efficiency in the conversion of solar energy into chemical energy, resistance to corrosion in aqueous environment, and low processing costs. Metal oxides are most promising in this regard (Walter et al., 2010). After four decades of intensive research, however, no material has been found to simultaneously satisfy all the criteria required for widespread PEC application. No wonder that a broad search for new materials for photoelectrodes is still ongoing. Cold plasma technology is also involved in this activity (Randeniya et al., 2007; Slavcheva et al., 2007; Walsh et al., 2009; Zhu, F. et al.,

A typical simple PEC cell with schematic representation of charge transfer is shown in Fig. 7. The cell is constructed from a semiconducting photoanode and metal cathode. The basic reaction steps involved in the PEC process in the cell are as follows (Nowotny et al.,

 *hν* → *e*− + *h*+ , (1)

 *e*− + *h*+ → *e*−bulk + *h*+surface ; (2)

 H2O + 2*h*+ → 2H+ + ½ O2; (3) *4*. Transport of hydrogen ions from the photoanode to the cathode through the liquid electrolyte; *5*. Transport of electrons to the cathode thorough the external circuit; and *6*.

 2*e*− + 2H+ → H2 . (4) The first PEC cell for water splitting, with a rutile TiO2 photoanode and Pt counter cathode, was reported in 1972 (Fujishima & Honda, 1972). Following this discovery, intensive studies aiming at increasing the energy conversion efficiency of solar energy into chemical energy have been carried out, mainly on the analogous PEC cells, using TiO2 as the photoanode. Then other oxides, e.g. Fe2O3 and WO3, have been also tested. Despite the good catalytic activity of oxides such as TiO2, they are generally limited by too large band gaps (approx. 3 eV), which fail to absorb a significant fraction of visible light, resulting in poor solar to hydrogen conversion efficiencies under terrestrial conditions. This value should be reduced to 1.7–2.0 eV. Besides, there is a range of other problems, like incorrect alignment of band edges with respect to the water redox potentials, energy losses due to charge recombination, low density of surface active sites reacting with water molecules, low corrosion resistance, etc. Thus, the majority of PEC oxide research has focused on trying to solve these problems by the modification of known photoactive oxides (through their doping or alloying), and

where *h* is the Planck constant and *ν* is the light frequency; *2*. Charge separation:

*3*. Reaction between water molecules and holes at the surface of the photoanode:

2006): *1*. Photoionization over the band gap of the semiconductor:

Reaction between electrons and hydrogen ions at the cathode:

**4.2 Water splitting systems** 

2009).

potentially low cost (James et al., 2009).

very recently also by creation of new materials. Invaluable in this respect seems to be the cold plasma technology, which allows to design the structure of fabricated materials in a very wide range (Walsh et al., 2009).

Fig. 7. Schematic representation of charge transfer within a photoelectrochemical cell involving a semiconducting photoanode and metal cathode (Nowotny et al., 2006).

The most attractive oxide to date, namely TiO2, has already been repeatedly produced by the cold plasma technology. In most reported works, either TiCl4 or Ti alkoxides (mainly titanium tetraisopropoxide, Ti(OC3H7)4) are used as the Ti-containing precursors of the plasma polymerization (PECVD) process, resulting in amorphous or crystalline films, with the nonstoichiometric (TiOX) or stoichiometric (TiO2) structure. For all of these films, their physicochemical properties are strongly dependent on the film structure, which can be effectively controlled by the deposition conditions (Battiston et al., 2000; Borrás et al., 2009; daCruz et al., 2000; Maeda & Watanabe, 2005; Nakamura et al., 2001).

TiO2 films are also obtained in a wide range by the reactive sputtering, usually using pure titanium as a target, and O2 as reactive gas. Similarly, as in the PECVD technique, also in this case, the sputtering process conditions control the structure of the deposited films, which in turn affects to a large extend the optical and photoelectrochemical properties of the films (Brudnik et al., 2007; Dang et al., 2011; Huang et al., 2011).

A particularly useful feature of the cold plasma technology is the possibility of co-deposition either by copolymerization of a mixture of precursors or by co-sputtering using more than one target or a mixture of several reactive gases. In this way we can get doped films as well as films with alloy-type structures. Numerous studies have been already done on the introduction into the TiO2 structure other atoms (e.g. C, N, S). For example, PECVD with the DC discharge carried out using mixtures of Ti(OC3H7)4 and nitrogen led to a Ti(OCN) film structure (Randeniya et al., 2007; Wierzchoń et al., 1993). Instead of nitrogen, ammonia can be introduced to plasma reactors (Weber et al., 1995). Precursors containing nitrogen in their chemical structure, e.g. tetrakis(dimethylamido) or (diethylamido)-titanium (Ti(N(CH3)2)4 or

Cold Plasma – A Promising Tool for the Development of Electrochemical Cells 117

used to produce ternary cobalt spinel oxides of the type CoX2O4 (X = Al, Ga, In), taking the targets from Co3O4 and Al, Ga2O3 or In2O3. Preliminary research showed that although these materials combine excellent stability in solution and good visible light absorption properties, their performance as photoelectrochemical catalysts for water splitting is limited by the poor electrical transport properties. It is hoped, however, that the broad capabilities

Particularly significant is the finding that nanoclusters of Co3O4 are much more efficient in the PEC process than larger objects (e.g. micrometer-sized particles) of this oxide. Thus, it is very important to develop methods for producing cobalt oxides films containing Co3O4 nanoclusters. Recently, mesoporous silica has been used as a scaffold for growing Co3O4 nanocrystals within its naturally parallel nanoscale channels via a wet impregnation technique. It has been found that rod-shaped crystals measuring 8 nm in diameter and 50 nm in length are interconnected by short bridges to form bundled clusters. The bundles are shaped like a sphere with a diameter of 35 nm (Jiao & Frei, 2009). This report aroused great scientific interest. However, it should be noted that films composed of nanoclusters can be easily deposited on a flat surface without any special mesoporous structure, only involving the plasma deposition technique for this purpose. Using this method, nanocrystalline films of cobalt oxides have been already obtained. For example, small particles of CoOX in the range of 2–10 nm in diameter were deposited in this way on TiO2 support (Dittmar et al., 2004). If CoOX films were fabricated on a substrate at elevated temperature (150–400°C), then columnar grains with average diameter size at the film surface of 35–60 nm were formed (Fujii et al., 1995). Research conducted recently in our laboratory has led to the cobalt oxide films containing 4–8 nm sized Co3O4 crystals, whose size can be controlled by the plasma deposition process (Tyczkowski, 2011; Tyczkowski et al., 2007). These films will be discussed in more detail in *Sec. 5.1*. It should be noted, however, that we are also now

Although research on water splitting, conducted using the cold plasma technology to produce thin-film coatings on photoelectrodes, is only beginning, the obtained results give cause for great hope. Plasma deposited films by both PECVD and reactive sputtering can reveal very high incident photon conversion efficiency (Randeniya et al., 2007). These films also appear to be better as photoelectrodes than the corresponding materials produced by

The PEC process can be realized not only in cells with two electrodes separated from each other, which was discussed above, but also when the electrodes are in direct close contact. Much attention in this regard has been paid for systems where both electrodes are located within a single grain. Such a structure may be considered as a microsized PEC cell. The best analogue of the PEC cell shown in Fig. 7 is a microsized cell formed of a small semiconductor grain (e.g. TiO2) and noble metal islets (e.g. Pt) deposited on its surface. Then, the surface of the semiconductor grain and the metal act as photoelectrode and counter electrode, respectively. More sophisticated systems with a bifunctional catalyst have been also proposed. In this case, anodic and cathodic photoactive islets (several nm in size) are deposited onto the same semiconducting nanoparticle (tens of nanometers in size) (James et al., 2009). These microsized PEC cells produce, however, a mixture of oxygen and hydrogen. To receive these gases separately, reactors composed of two chambers connected by a diffusion bridge are used. In one chamber there is a suspension of nanoparticles only

of plasma technology will help overcome this problem (Walsh et al., 2009).

starting work on their application to the water splitting.

other method (Naseri et al., 2011).

Ti(N(C2H5)2)4) were also utilized (Raaijmakers, 1994). By properly adjusting the composition of the reaction mixture and the conditions of the plasma, a film structure similar to stoichiometric TiN can be obtained (Weber et al., 1995). In turn, the films composed mainly of titanium and carbon (TiCX) were fabricated leading PECVD process in a mixture of TiCl4 and hydrocarbons (Täschner et al., 1991). TiO2 films with N and C atoms were also obtained by reactive sputtering of titanium in an appropriate gas mixture. This method also proved to be useful for the production of nanocomposite thin films for photoanodes, e.g. Au:TiO2 films sputtered (using an RF discharge) from Ti and Au targets in O2 as the reactive gas (Naseri et al., 2011).

The cold plasma deposition method has been used to produce, in addition to films based on TiO2, other films that constitute an interesting material for the photoelectrodes. For example, iridium oxide (IrO2) (Slavcheva et al., 2007), tantalum nitride (Ta3N5) (Yokoyama et al., 2011), ruthenium sulfide (RuS2) (Licht et al., 2002), and tungsten trioxide (WO3) (Garg et al., 2005) films were prepared by reactive sputtering. WO3 was also deposited by PECVD technique, feeding the RF plasma reactor with a gas mixture of tungsten hexafluoride (WF6) and oxygen (Garg et al., 2005).

Recently, a proposal to employ a-SiXCY:H films as photoelectrodes for PEC cells has been presented (Zhu, F. et al., 2009). The films were fabricated by PECVD using a SiH4, H2 and CH4 gas mixture. It was found that the a-SiXCY:H photoelectrode behaves as a photocathode, where the photo-generated electrons are injected into the electrolyte and reduce H+ ions for hydrogen evolution. The use of this photoelectrode led to a solar-to-hydrogen conversion efficiency higher than 10%. It should be noted that a-SiXCY:H films deposited by PECVD technique have been extensively investigated for a long time. A lot of gas mixtures (e.g. SiH4 and hydrocarbons) or single compounds (e.g. tetramethylsilane (Si(CH3)4)) are used as precursors of the deposition process. In some cases, dopant agents are also added. In fact, the films can be accurately produced according to the designed electronic structure and photoelectronic properties, which can change over a very wide range. For instance, their electrical conductivity changes from 10−18 to 0.1 S/m and the optical gap shifts between 1.8 and 3.2 eV (Tyczkowski, 2004).

In the past five years, cobalt has emerged as the most versatile non-noble metal for the development of synthetic H2- and O2-evolving catalysts. Among the various structures containing cobalt atoms, cobalt oxides appear to be particularly promising materials. The possibility of using such oxides to catalyze water oxidation in neutral aqueous solutions has recently experienced a burst of interest (Artero et al., 2011). There have been many reports in the literature concerning the use of cobalt oxides, mainly Co3O4, as electrode coatings that catalyze water oxidation. Many different methods have been applied to prepare these coatings, among others, also the cold plasma deposition technique – both PECVD and reactive sputtering. In PECVD, cobalt oxide films were obtained from volatile precursors such as bis(acetylacetonate)cobalt(II) (Fujii et al., 1995), bis(2,2,6,6-tetramethylheptan-3,5 dionato)cobalt(II) (Barreca et al., 2011), bis(cyklopentadienyl)cobalt(II) (Donders et al., 2011) or cyclopentadienyl(dicarbonyl)cobalt(I) (Tyczkowski et al., 2007). Other volatile cobalt complexes (e.g. amidinates and cyclodexrtins) are now also proposed as the precursors (Li et al., 2008; Papadopoulos et al., 2010). The sputtering process, in turn, was conducted in the presence of pure Co or Co3O4 as targets and plasma generated in a gas mixture containing O2 (Ingler Jr et al., 2006; Schumacher et al., 1990). The reactive sputtering technique was also

Ti(N(C2H5)2)4) were also utilized (Raaijmakers, 1994). By properly adjusting the composition of the reaction mixture and the conditions of the plasma, a film structure similar to stoichiometric TiN can be obtained (Weber et al., 1995). In turn, the films composed mainly of titanium and carbon (TiCX) were fabricated leading PECVD process in a mixture of TiCl4 and hydrocarbons (Täschner et al., 1991). TiO2 films with N and C atoms were also obtained by reactive sputtering of titanium in an appropriate gas mixture. This method also proved to be useful for the production of nanocomposite thin films for photoanodes, e.g. Au:TiO2 films sputtered (using an RF discharge) from Ti and Au targets in O2 as the reactive gas (Naseri et

The cold plasma deposition method has been used to produce, in addition to films based on TiO2, other films that constitute an interesting material for the photoelectrodes. For example, iridium oxide (IrO2) (Slavcheva et al., 2007), tantalum nitride (Ta3N5) (Yokoyama et al., 2011), ruthenium sulfide (RuS2) (Licht et al., 2002), and tungsten trioxide (WO3) (Garg et al., 2005) films were prepared by reactive sputtering. WO3 was also deposited by PECVD technique, feeding the RF plasma reactor with a gas mixture of tungsten hexafluoride (WF6)

Recently, a proposal to employ a-SiXCY:H films as photoelectrodes for PEC cells has been presented (Zhu, F. et al., 2009). The films were fabricated by PECVD using a SiH4, H2 and CH4 gas mixture. It was found that the a-SiXCY:H photoelectrode behaves as a photocathode, where the photo-generated electrons are injected into the electrolyte and reduce H+ ions for hydrogen evolution. The use of this photoelectrode led to a solar-to-hydrogen conversion efficiency higher than 10%. It should be noted that a-SiXCY:H films deposited by PECVD technique have been extensively investigated for a long time. A lot of gas mixtures (e.g. SiH4 and hydrocarbons) or single compounds (e.g. tetramethylsilane (Si(CH3)4)) are used as precursors of the deposition process. In some cases, dopant agents are also added. In fact, the films can be accurately produced according to the designed electronic structure and photoelectronic properties, which can change over a very wide range. For instance, their electrical conductivity changes from 10−18 to 0.1 S/m and the optical gap shifts between 1.8

In the past five years, cobalt has emerged as the most versatile non-noble metal for the development of synthetic H2- and O2-evolving catalysts. Among the various structures containing cobalt atoms, cobalt oxides appear to be particularly promising materials. The possibility of using such oxides to catalyze water oxidation in neutral aqueous solutions has recently experienced a burst of interest (Artero et al., 2011). There have been many reports in the literature concerning the use of cobalt oxides, mainly Co3O4, as electrode coatings that catalyze water oxidation. Many different methods have been applied to prepare these coatings, among others, also the cold plasma deposition technique – both PECVD and reactive sputtering. In PECVD, cobalt oxide films were obtained from volatile precursors such as bis(acetylacetonate)cobalt(II) (Fujii et al., 1995), bis(2,2,6,6-tetramethylheptan-3,5 dionato)cobalt(II) (Barreca et al., 2011), bis(cyklopentadienyl)cobalt(II) (Donders et al., 2011) or cyclopentadienyl(dicarbonyl)cobalt(I) (Tyczkowski et al., 2007). Other volatile cobalt complexes (e.g. amidinates and cyclodexrtins) are now also proposed as the precursors (Li et al., 2008; Papadopoulos et al., 2010). The sputtering process, in turn, was conducted in the presence of pure Co or Co3O4 as targets and plasma generated in a gas mixture containing O2 (Ingler Jr et al., 2006; Schumacher et al., 1990). The reactive sputtering technique was also

al., 2011).

and oxygen (Garg et al., 2005).

and 3.2 eV (Tyczkowski, 2004).

used to produce ternary cobalt spinel oxides of the type CoX2O4 (X = Al, Ga, In), taking the targets from Co3O4 and Al, Ga2O3 or In2O3. Preliminary research showed that although these materials combine excellent stability in solution and good visible light absorption properties, their performance as photoelectrochemical catalysts for water splitting is limited by the poor electrical transport properties. It is hoped, however, that the broad capabilities of plasma technology will help overcome this problem (Walsh et al., 2009).

Particularly significant is the finding that nanoclusters of Co3O4 are much more efficient in the PEC process than larger objects (e.g. micrometer-sized particles) of this oxide. Thus, it is very important to develop methods for producing cobalt oxides films containing Co3O4 nanoclusters. Recently, mesoporous silica has been used as a scaffold for growing Co3O4 nanocrystals within its naturally parallel nanoscale channels via a wet impregnation technique. It has been found that rod-shaped crystals measuring 8 nm in diameter and 50 nm in length are interconnected by short bridges to form bundled clusters. The bundles are shaped like a sphere with a diameter of 35 nm (Jiao & Frei, 2009). This report aroused great scientific interest. However, it should be noted that films composed of nanoclusters can be easily deposited on a flat surface without any special mesoporous structure, only involving the plasma deposition technique for this purpose. Using this method, nanocrystalline films of cobalt oxides have been already obtained. For example, small particles of CoOX in the range of 2–10 nm in diameter were deposited in this way on TiO2 support (Dittmar et al., 2004). If CoOX films were fabricated on a substrate at elevated temperature (150–400°C), then columnar grains with average diameter size at the film surface of 35–60 nm were formed (Fujii et al., 1995). Research conducted recently in our laboratory has led to the cobalt oxide films containing 4–8 nm sized Co3O4 crystals, whose size can be controlled by the plasma deposition process (Tyczkowski, 2011; Tyczkowski et al., 2007). These films will be discussed in more detail in *Sec. 5.1*. It should be noted, however, that we are also now starting work on their application to the water splitting.

Although research on water splitting, conducted using the cold plasma technology to produce thin-film coatings on photoelectrodes, is only beginning, the obtained results give cause for great hope. Plasma deposited films by both PECVD and reactive sputtering can reveal very high incident photon conversion efficiency (Randeniya et al., 2007). These films also appear to be better as photoelectrodes than the corresponding materials produced by other method (Naseri et al., 2011).

The PEC process can be realized not only in cells with two electrodes separated from each other, which was discussed above, but also when the electrodes are in direct close contact. Much attention in this regard has been paid for systems where both electrodes are located within a single grain. Such a structure may be considered as a microsized PEC cell. The best analogue of the PEC cell shown in Fig. 7 is a microsized cell formed of a small semiconductor grain (e.g. TiO2) and noble metal islets (e.g. Pt) deposited on its surface. Then, the surface of the semiconductor grain and the metal act as photoelectrode and counter electrode, respectively. More sophisticated systems with a bifunctional catalyst have been also proposed. In this case, anodic and cathodic photoactive islets (several nm in size) are deposited onto the same semiconducting nanoparticle (tens of nanometers in size) (James et al., 2009). These microsized PEC cells produce, however, a mixture of oxygen and hydrogen. To receive these gases separately, reactors composed of two chambers connected by a diffusion bridge are used. In one chamber there is a suspension of nanoparticles only

Cold Plasma – A Promising Tool for the Development of Electrochemical Cells 119

Fig. 8. Schematic representation of a photovoltaic–electrolysis system based on a simply

The potential applications of plasma polymers as materials for electrochemistry are primarily associated with the possibility of designing their electronic structure and thereby with the designing of their electrical properties. Plasma deposition from diverse precursors and their mixtures, performed under various conditions of the process, leads to a huge variety of thin films characterized by a broad spectrum of electrical properties. In Fig. 9 a diagram of the conductivity typical for different plasma polymer types is presented. As one can see, the whole range from 10−18 S/m to 106 S/m is covered by plasma deposited films. Insulating, semiconducting (of different types of conductivity) and metallic films can be obtained in this way (Tyczkowski, 2004). Some of these films can reveal a significant activity in electrochemical processes. Photocatalytic activity of such films was already discussed in

Taking into account the electrical properties mentioned above as well as another important feature of plasma polymers, namely the membrane nature, which is very often revealed by these films, a new fascinating electrocatalyst structure has been proposed. An effective electrocatalyst must satisfy many requirements, such as high activity, high electrical conductivity, and long-term stability, which may be in conflict with each other. One possible way to solve these conflicts is the use of composite materials, where the matrix and the dispersed phase are independently selected. A successful approach is that of associating a highly conducting (though catalytically inert) matrix with an active (though less conducting) dispersed phase. When the matrix is permeable, one has a three-dimensional (3D) catalyst: all catalytic particles are active, irrespective of their position in the composite. Fig. 10

a-Si:H solar cell (Kelly & Gibson, 2006).

the previous section (*Sec. 4.2.*).

**5. New types of electrodes for electrochemical cells** 

schematically represents the operation of such a system.

with an anodic catalyst while in the second one – only with a cathodic catalyst (James et al., 2009).

In the field of microsized PEC cells, the plasma techniques have also proved to be very useful, both in the synthesis of semiconducting particles and in the deposition of active catalysts of them. For the production of powders, in a very wide range of grain sizes (from single nanometers to tens of micrometers) mainly the thermal plasma (see: Fig. 1) is utilized (Ctibor & Hrabovský, 2010; Karthikeyan et al., 1997). The PECVD technique is, however, also engaged for this purpose. By this technique, TiO2 nanocrystalline powder was prepared, using an AF glow discharge (40 kHz) and titanium tetraisopropoxide (Ti(OC3H7)4) with oxygen as a reactive mixture. The obtained nanocrystalline particles, with mean size of about 25–55 nm, revealed good photocatalytic activity (Ayllón et al., 1999). Another example is the synthesis of carbon-supported ultrafine metal particles by MW plasma from metal carbonyls (e.g. Fe(CO)5, Co2(CO)8, Mo(CO)6) as precursors (Brenner et al., 1997). In turn, noncrystalline organosilicon powder was produced by plasma polymerization of tetramethylsilane (Si(CH3)4). The ratio of elements (Si/C) as well as the chemical structure of the grains was highly dependent on the plasma process conditions (Fonseca et al., 1993).

As already mentioned earlier, an important feature of the PECVD method is the possibility of copolymerization. This route was used to prepare TiO2 nanoparticles doped with Sn4+ ions. The plasma process was performed in a mixture of TiCl4 and SnCl4 with an appropriate molar ratio. It was found that photocatalytic activity of TiO2–Sn4+ nanoparticles was much higher than those of the pure TiO2 (Cao et al., 2004). Wide possibilities of the plasma technique also allow to produce nanoparticles of doped semiconducting organic polymers, e.g. polypyrrole plasma doped with iodine (Cruz et al., 2010). This is only a matter of time when such organic nanoparticles with anchored a molecular water oxidation catalyst will be produced by plasma polymerization.

The deposition of catalytically active coatings onto the surface of already prepared particles has also been performed. In this case, however, to achieve efficient coating of the nanoparticles, a special construction of the plasma reactor chamber is needed, for example, with 360° continuous rotation. In such a reactor, TiO2 nanoparticles were coated with thin film produced by plasma polymerization of tetramethyltin (Sn(CH3)4). Subsequently, the coated particles were heated in air to remove the carbonaceous material while, simultaneously, oxidizing tin atoms to tin oxide. To obtain partially fluorinated tin oxide, hexafluoropropylene oxide (C3F6O) was added to Sn(CH3)4. It should be noted that significantly increased photocatalytic activity of TiO2 nanoparticles was achieved using the PECVD approach (Cho, J. et al., 2006).

Finally, one more type of water splitting cells should be mentioned, namely integrated photovoltaic–electrolysis (PV-PEC) cells. In this type of devices, the photovoltaic cell and the electrolyser are combined into a single system, in which the light-harvesting solar cell is one of the electrodes. Very often, thin-film solar cells fabricated by the cold plasma deposition method are employed in the PV-PEC devices (Kelly & Gibson, 2006). A diagram of such a system with a simply a-Si:H solar cell is shown in Fig. 8. There is no doubt that the role played by the cold plasma deposition technique in the creation of such systems is unquestionable (see: *Sec. 4.1.*).

with an anodic catalyst while in the second one – only with a cathodic catalyst (James et al.,

In the field of microsized PEC cells, the plasma techniques have also proved to be very useful, both in the synthesis of semiconducting particles and in the deposition of active catalysts of them. For the production of powders, in a very wide range of grain sizes (from single nanometers to tens of micrometers) mainly the thermal plasma (see: Fig. 1) is utilized (Ctibor & Hrabovský, 2010; Karthikeyan et al., 1997). The PECVD technique is, however, also engaged for this purpose. By this technique, TiO2 nanocrystalline powder was prepared, using an AF glow discharge (40 kHz) and titanium tetraisopropoxide (Ti(OC3H7)4) with oxygen as a reactive mixture. The obtained nanocrystalline particles, with mean size of about 25–55 nm, revealed good photocatalytic activity (Ayllón et al., 1999). Another example is the synthesis of carbon-supported ultrafine metal particles by MW plasma from metal carbonyls (e.g. Fe(CO)5, Co2(CO)8, Mo(CO)6) as precursors (Brenner et al., 1997). In turn, noncrystalline organosilicon powder was produced by plasma polymerization of tetramethylsilane (Si(CH3)4). The ratio of elements (Si/C) as well as the chemical structure of the grains was highly dependent on the plasma process

As already mentioned earlier, an important feature of the PECVD method is the possibility of copolymerization. This route was used to prepare TiO2 nanoparticles doped with Sn4+ ions. The plasma process was performed in a mixture of TiCl4 and SnCl4 with an appropriate molar ratio. It was found that photocatalytic activity of TiO2–Sn4+ nanoparticles was much higher than those of the pure TiO2 (Cao et al., 2004). Wide possibilities of the plasma technique also allow to produce nanoparticles of doped semiconducting organic polymers, e.g. polypyrrole plasma doped with iodine (Cruz et al., 2010). This is only a matter of time when such organic nanoparticles with anchored a molecular water oxidation

The deposition of catalytically active coatings onto the surface of already prepared particles has also been performed. In this case, however, to achieve efficient coating of the nanoparticles, a special construction of the plasma reactor chamber is needed, for example, with 360° continuous rotation. In such a reactor, TiO2 nanoparticles were coated with thin film produced by plasma polymerization of tetramethyltin (Sn(CH3)4). Subsequently, the coated particles were heated in air to remove the carbonaceous material while, simultaneously, oxidizing tin atoms to tin oxide. To obtain partially fluorinated tin oxide, hexafluoropropylene oxide (C3F6O) was added to Sn(CH3)4. It should be noted that significantly increased photocatalytic activity of TiO2 nanoparticles was achieved using the

Finally, one more type of water splitting cells should be mentioned, namely integrated photovoltaic–electrolysis (PV-PEC) cells. In this type of devices, the photovoltaic cell and the electrolyser are combined into a single system, in which the light-harvesting solar cell is one of the electrodes. Very often, thin-film solar cells fabricated by the cold plasma deposition method are employed in the PV-PEC devices (Kelly & Gibson, 2006). A diagram of such a system with a simply a-Si:H solar cell is shown in Fig. 8. There is no doubt that the role played by the cold plasma deposition technique in the creation of such systems is

2009).

conditions (Fonseca et al., 1993).

catalyst will be produced by plasma polymerization.

PECVD approach (Cho, J. et al., 2006).

unquestionable (see: *Sec. 4.1.*).

Fig. 8. Schematic representation of a photovoltaic–electrolysis system based on a simply a-Si:H solar cell (Kelly & Gibson, 2006).
