Microwave Plasmas as a Processing Tool for Tailoring the Surface Properties of Ceramic Coatings

1. Introduction

Microwaves are electromagnetic waves with wavelengths from 1 mm to 1 m and corresponding frequencies between 0.3 MHz and 300 GHz [1, 2]. The use of microwave treatments (non-plasma), as an energy source for the heat treatment of metals, ceramics and composites has been reported to be more effective than conventional furnace heating methods, due to the improved microstructure and properties, reduction in processing time, etc. [1, 2, 3]. These microwave treatments are usually carried out at 0.915 GHz, 2.45 GHz and 20–30 GHz frequencies in agreement with the industrial, scientific and medical (ISM) radio bands set aside for non-communication purposes [2, 4]. During microwave processing, energy is supplied by an electromagnetic field directly to the material by the interaction of the molecules of the material with electromagnetic field. In contrast, conventional thermal processing involves the transfer of energy by conduction, convection and radiation [5].

Clark et al. [1] reviewed the effectiveness of microwave heating and found it to be substrate material dependant. Materials were, thus, grouped into categories (depending on the electromagnetic field-material interaction) as: transparent (low dielectric loss materials); opaque (conductors); absorbing (high dielectric loss materials); and mixed absorbing where there is selective absorption due to differences in dielectric loss in the materials.

Several researchers have demonstrated the capability of using microwave to process metallic and ceramic materials [6, 7, 8, 9, 10, 11]. It has been reported that a distinction in the microstructure and porosity distribution can be made between conventional and microwave sintered materials with respect to the pore shape. Microwave sintered samples exhibit pores with more rounded edges than the conventionally sintered materials. The sphericity of pore is reported to be critical in the ductility and strength of the sintered material [11].

In addition to the use of microwaves directly, a plasma discharge can also be formed using the microwave energy, if the appropriate conditions (gas type, pressure, resonator, etc.) are in place. Briefly by way of introduction to this technique, a plasma is described as a collection of unbound charged particles, photons and neutral atomic or molecular species which are electrically neutral on average [12, 13]. Plasma generation arises from the excitation and ionization of gases by energy transfer from various sources to the gases to form excited, charged and neutral species [14]. There are a range of plasma types including high frequency, DC and low frequency discharges [15, 16]. The latter include systems such as glow discharge, corona discharge, hollow cathode discharge [16], while the high frequency consists of radio frequency (RF) and microwave plasma discharges [17].

Microwave plasma systems are usually electrodeless systems and thus, differ from DC and RF systems which mostly use electrodes to generate plasma. In these systems, microwaves are usually guided using waveguide into the chamber where energy is impacted to the gas to form plasma by partially ionizing the gas [18, 19]. A typical example of microwave plasma system is the circumferential antenna plasma (CAP) microwave reactor, which operates at 2.45 GHz (Figure 1). In this CAP system, a cylindrical quartz ring window, 345 mm in diameter, embedded in the wall of the plasma chamber guides the microwaves into the plasma chamber. The microwaves are directed to the plasma chamber through a coaxial waveguide, and expands radially as it moves towards the cylindrical quartz ring window. A perfect rotational symmetry is obtained when the microwave which has uniform amplitude and phase distribution passes through the quartz window [20]. The cylindrical geometry of the configuration ensures that the formation of microwave power density at the window is lower than in the center of the chamber; thus, no plasma is formed adjacent to the window [21].

The use of microwave plasma source for surface and bulk treatment of metal and ceramics has the potential to be more effective than DC and RF plasmas, because of the higher density of active atomic species that can be generated using this type of discharge [22]. Microwave discharges generally have higher electron kinetic temperature and number density because the power is usually coupled through radiation and thus, bypasses sheath losses [17, 23]. A further advantage of these plasmas is that they can be operated in a wide gas pressure range [24, 25], which combine with the higher electron temperature generally obtained, makes them capable of providing a higher fraction of ionization and dissociation [24].

Metal oxide layers can be obtained using a range of techniques including sol-gel, hydrothermal synthesis, electrophoretic deposition, sputtering, electrochemical treatments (anodic oxidation), chemical vapor deposition (CVD), physical vapor deposition and ion implantation, etc. [26, 27, 28, 29, 30, 31, 32]. Most of the techniques (except sputtering and direct metal oxide growth on metal surfaces) require sintering to enhance the packing density, porosity and the adhesion of the deposited coatings. A range of characterisation techniques usually applied includes a FEI Quanta 3D FEG DualBeam scanning electron microscope (SEM) (Morphology); a focused ion beam (FIB)(Thickness); Energy-dispersive X-ray spectroscopy (EDX) (Elemental data); Veeco NT1100 optical profilometer (Surface roughness). Siemens D500 XRD system (Phase identification), a LASCON QP003 and LPC03 ratio pyrometers from Dr. Mergenthaler GmbH & Co. KG. (Temperature measurements) and a J.A. Woollam ellipsometer with a Tauc-Lorentz fitting via completeEASE® software (Band gap and coating thickness).

Having provided an introduction to microwave plasma treatments along with the associated characterization techniques, the following sections firstly provide an overview of the use of microwave plasmas for the sintering of metal oxide powders. The use of microwave plasma treatments for sintering, has previously been described as rapid discharge sintering (RDS). The specific focus is on the sintering of nanoparticles of nickel oxide (NiO) and titanium dioxide (TiO₂), used as electrodes in dye sensitized solar cells (DSSCs). The effect of the plasma treatments on thermal sensitive crystalline phase changes, is also discussed. A further application reported is that of the use of microwave plasma treatments for the oxidation of metal surfaces, in order to produce porous oxide ceramic (TiO₂) layers on the metal surface.

2. Microwave plasma sintering of nanoparticles

Considerable research has been carried out to understand the interaction of microwave radiation with materials [33, 34]. The ability of microwaves to penetrate work pieces enables volumetric heating at a rapid rate, thus avoiding the need of external heat sources [33]; this makes it an attractive technique in material processing. However, poor coupling in non-plasma microwave processing may result in a non-uniform heating of the substrates [35]. One of the advantages of microwave plasma treatment is the combined advantage inherent in microwave and plasma heating in terms of volumetric, homogeneous and rapid heating [2, 36, 37]. For example, the work of Twomey et al. [37] demonstrates that homogenous heating of substrates could be achieved using microwave plasma treatments including considerably lower cycle times [38] compared to non-plasma microwave and furnace heating. In this section, the use of microwave plasma processing for the sintering of Nickel oxide (NiO) and Titanium dioxide (TiO₂) particles is presented.

2.1. Sintering of nickel oxide (NiO) nanoparticles

NiO coatings are used extensively as a photocathode in the construction of p-type DSSCs [39, 40, 41, 42] due to their p-type semiconductivity [43], and well defined electrical and optical properties [44]. Furthermore, its bandgap energy which range from 3.6 to 4.0 eV helps to make it a model semiconductor substrate [43]. Also, it is considered a good electron donor for numerous photo sensitizers due to its valance bond potential; thus, the NiO coatings are readily quenched with many dye sensitizers [45]. In this section, the use of microwave plasma processing for the sintering of NiO particles to produce coatings for use in DSSC electrodes is discussed [42, 46].

2.1.1. Rapid discharge sintering vs. conventional furnace sintering

In order to process NiO using the RDS technique, a spray technique is used to deposit solvent slurry of NiO nanoparticles onto conductive glass substrates (Figure 2). The spray technique involves suspending the NiO nanoparticles in alcohol; the mixture is then applied via spraying onto fluorine-doped tin oxide (FTO) coated glass substrates to form the coating. A subsequent step is carried out after spraying in order to enhance the interconnectivity between the NiO particles by sintering, as well as an increased level of adhesion to the conductive glass substrate. This sintering step was investigated using both microwave plasma and furnace treatments [42, 46]. The latter was carried out using a CWF 1200 Carbolite tube air furnace. Typical furnace sintering temperatures in the range of 300–450°C have been reported previously [47, 48, 49, 50, 51]. In this study, a treatment temperature of approximately 450°C was used for a 5-min period for both the furnace and RDS treatments. This treatment time was chosen because it yielded the best photovoltaic performance for RDS treatment. Overall, a cycle time of ~120 min was required for the furnace sintering including heat up, maintenance at the max 450°C temperature and cool down. In contrast, the RDS processing required only 15 min, included all stages such as pump down, plasma treatment, and cooling.

In this Section, the properties of the furnace sintered coatings, such as morphology, crystal structure, dye adsorption, chemical composition and photovoltaic performance are compared with the NiO coatings treated for 5 min, using the RDS technique.

2.1.1.1. Morphological analyses of porous NiO coatings

A typical SEM micrograph of the morphology of RDS and furnace treated coatings is given in Figure 3. A higher level of interconnectivity with reduced grain growth for the RDS treated coatings when compared with the furnace sintered coatings can be observed. The increased grain growth in the furnace sintered coatings could be due to a slower cooling rate (10°C/min) used in the furnace process. In contrast, the RDS technique requires only a 5-min cooling time following the plasma treatment.

A further examination of NiO coatings using an FIB/SEM cross section micrograph (Figure 4) indicated that the RDS treated coating displays an improved level of porosity as well as interfacial connection between the NiO coating and FTO substrate. This could be due to the difference in the type of heating mechanisms in both sintering techniques. RDS treatment encompasses volumetric heating, which delivers more efficient heating inside the NiO coating matrix than that obtained with furnace sintering where a conductive type of heating is present. Also, furnace sintering could have resulted in inhomogeneous heating of coatings’ surface to yield a heat affected zone [53]; thus, a more open structure is observed for the RDS treated coatings when compared with the furnace sintered samples.

2.1.1.2. Crystal structure, crystallite size and dye adsorption evaluations of NiO coatings

As shown in Figure 3, the NiO grain size obtained using the RDS process was smaller than that obtained after treatment using the furnace. A measurement of the crystallite size of NiO coatings obtained using similar sintering techniques to Figure 3 indicated that the sizes were approximately 6.5 and 14.0 nm, respectively [42]. Thus, the RDS technique results in a finer grain size, in addition to a more uniform heating/sintering of the NiO nanoparticles.

The effect of a smaller grain size and a more open structure (Figures 3 and 4) obtained using the RDS technique may be observed in the UV-vis absorption spectra of the NiO coatings obtained using the two sintering techniques (Figure 5). The level of dye adsorption increased up to 44% for the case of RDS treated NiO coatings when compared with the furnace sintered coatings of the same thickness.

2.1.1.3. Comparison of photovoltaic performance of porous NiO coatings

Another parameter which can be used to portray the performance of the porous NiO structure is its photovoltaic performance (I-V characteristic) when evaluated as part of a p-type DSSCs. A comparison of the photovoltaic performance of NiO coatings prepared by various researchers is given in Table 1. From this Table, it is clear that the light-to-current conversion efficiency increase of almost 10-fold for the RDS treated NiO coatings (1–2 μm and 2.5 μm thick), when compared with the furnace sintered coatings. The conversion efficiency of the furnace treated coatings (1–2 μm thick) is comparable to the values reported in the literature [42, 46]. A notable observation is that a 2.5 μm thick NiO coatings, prepared using the RDS treatment, exhibited the best performance; it is likely due to improved dye adsorption with enhanced active sites, which results in better photochemical reaction.

Awais et al. [42, 46]/Reported data (NiO thickness)	Treatment time (min)	VOC (mV)	JSC (mAcm−2)	FF	Efficiency (η)
RDS treated (~2.5 μm thick)	5	120.00	1.05	36	0.0450
Furnace sintered (~2.5 μm thick)	5	84.00	0.22	25	0.0050
Furnace sintered (~2.5 μm thick)	30	35.29	0.21	26	0.0023
RDS treated (~1–2 μm thick)	5	72.14	0.53	28	0.0118
Furnace sintered (~1–2 μm thick)	5	50.30	0.24	28	0.0037
He et al. [54] (~1 μm thick)	60	83.00	0.20	27	0.0070
Nattestad et al. [48] (~1.6 μm thick)	20	120.00	0.36	26	0.0110

Table 1.

Photovoltaic performance of RDS and furnace NiO coatings used as photocathode in construction of p-type DSSC (compared with the literature). AM 1.5 solar simulator (I: 870 W m⁻²), 0.5 M LiI and 0.05 M I₂ in propylene carbonate as an electrolyte [52]. Values in italics represent reported literature data.

2.2. Sintered titanium dioxide (TiO₂) nanoparticles

Since the use of TiO₂ as a photoanode in the fabrication of a DSSC in 1991 by O’Regan and Grätzel [55], many researchers employ TiO₂ for DSSC and other related applications. The phase transformation of TiO₂ coatings is generally carried out using furnace sintering, however, as discussed earlier, it is time consuming. In this section, we will present the use of RDS as viable unconventional sintering technique to convert amorphous TiO₂ to crystalline phase(s). The effect of carbon doping on the resultant coating is also discussed. In this review, the TiO₂ coating of interest with respect to sintering will be that obtained by Dang et al. [56], for example. The following type of coatings are analyzed for their performance: furnace sintered C-doped, RDS treated undoped and C-doped RDS treated.

2.2.1. Comparison of RDS and air furnace treatments of porous TiO₂

TiO₂ coatings were also used to fabricate DSSC electrodes. TiO₂ and carbon-doped TiO₂ were deposited as coatings onto unheated titanium and silicon wafer substrates using a DC closed-field magnetron sputtering system [56]. The C-doped TiO₂ coatings were obtained by introducing low concentrations of carbon dioxide into the argon/oxygen plasma during the sputtering of the metal. The resultant coatings had an amorphous structure and a post-deposition heat treatment is required to convert this amorphous structure into the photoactive crystalline phase(s) of TiO₂. This was achieved using the CAP microwave plasma heat treatment using a nitrogen plasma. During the plasma treatment, the substrate temperature was about 550°C. At this temperature and for treatment times as short as 1 min, 0.25 μm thick coatings converted into the anatase crystalline phase of TiO₂. Further treatments of the coatings at higher temperatures resulted in anatase-to-rutile crystalline phase transformation [56]. As reported earlier for the NiO layers the use of microwave plasma heat treatments facilitated a much more rapid processing compared with furnace heat treatments. It was also observed that the plasma treated TiO₂ coatings also exhibited higher photocurrent density. This is possibly as a result of higher level of surface roughness and consequently a higher available surface area observed for these coatings [56].

As for the plasma treated TiO₂ coatings, C-doped coatings also exhibited a higher roughness when plasma treated compared to furnace treated (Figure 6). To investigate the phase transformation efficiency of the two treatment methods as well as the effect of carbon doping, an XRD profile of undoped and C-doped TiO₂ coatings was obtained. As shown in Figure 7(a), anatase phase peaks were observed for coatings treated between 550 and 850°C. As the temperature is increased to 875°C, peaks indicative of the rutile phase could be observed. Interestingly, for the C-doped TiO₂ coatings, the rutile phase can be observed at 750°C (Figure 7(b)); thus, it is possible that carbon doping lowered the transitional phase change temperature of anatase to rutile.

In order to evaluate the effect of doping on the efficiency of the TiO₂ coated electrodes, photocurrent density (I_Ph) measurements were obtained, as presented in Table 2. The highest value of photocurrent density was obtained for the case of carbon doped plasma treated TiO₂ coatings. There was a 19% increase in the I_Ph value in comparison to those treated in the furnace. Possible reasons for the superior performance could be due to the enhanced porosity level of plasma treated coatings, which may have provided more active sites for charge production [56]. A further factor could be that the retention of more doped carbon in the RDS-treated TiO₂ coatings further reduced the band gap of RDS-treated compared to the furnace-treated TiO₂ coatings (See Figure 8).

TiO2 coatings	Photocurrent density values, IPh (μA/cm2)
As-deposited (C-doped)	108
Furnace sintered (C-doped)	181
RDS treated (un-doped)	167
RDS treated (C-doped)	216

Table 2.

Photocurrent density values for different TiO₂ coatings.

3. Direct porous ceramics growth from metal substrates

In this study, the performance of microwave plasma treatments as a technique for the oxidation of metallic surfaces is investigated. Oxides of titanium formed either on the metal surface or alternatively as TiO₂ coatings (discussed in Section 2.2) have found applications in areas ranging from medical devices (cell attachment), solar cells (light capture), air and water purification, gas sensing, wear protective coatings, etc. [57, 58, 59, 60, 61, 62]. This is due to the oxide’s photocatalytic, biocompatibility properties, as well as its physical and chemical stability. These properties would depend on the morphology, surface roughness and porosity of the oxides.

Compared to furnace heat treatments, microwave plasma-treated TiO₂ coatings discussed in Section 2.2 possess higher level of surface roughness and hence photoactivity. However, the enhanced surface roughness is still only in the order of a few nanometers and thus, the photoactivity of the resulting coatings is relatively limited [63]. An alternative fabrication method which has been investigated to address this shortcoming is presented in this section.

3.1. RDS oxidation of titanium metal substrates

The plasma treatments were carried out as before using the CAP microwave reactor, in this case using an oxygen discharge [63, 64]. Prior to plasma oxidation, the titanium disc test substrates were polished to a mirror finish and then solvent cleaned. The oxidized substrates exhibited a white appearance, as shown in Figure 9; this is in contrast to the metallic appearance of the non-oxidized titanium metal.

A typical FIB/metallography SEM cross-section image of a RDS treated TiO₂ structure obtained using a focused ion beam (FIB) and metallographic technique for comparison is shown in Figure 10. It can be observed that the two sample preparation methods (FIB and metallography) did not alter the porous structure of the ceramics.

3.2. Influence of substrate temperature on the pore structure morphology in a RDS grown oxide-layer

The porosity of the oxide-layer structure was found to generally increase with increase in thickness and treatment temperature (Figure 11). Examination of the porous oxide structures indicates that oxide-layers grown at temperature below 880°C exhibits a relatively porous structure, in the upper oxide layer compared to that closer to the metal substrate. Oxides fabricated above 880°C exhibited a further increase in porosity levels throughout the oxide layer. It is well known that α to β phase transformation of titanium occurs at approximately 882°C; this may have influenced porosity distribution observed in the oxide-layers grown on titanium substrates above 880°C [64].

3.3. Comparison of the porosity of sintered TiO₂, air furnace and RDS grown oxide-layer

Figure 12 shows the oxide layer morphology of a RDS and air furnace grown ceramic structures. The structure in Figure 12(a) was obtained after 5-hours furnace treatment in air and its thickness was found to be 4.17 μm. In contrast, after treatment in an oxygen microwave plasma for 10 min, the thickness obtained was 6.96 μm (Figure 12(b)). It can be observed that the oxide-layer obtained using the microwave plasma oxidation exhibited a relatively rough morphology, with large grains. In contrast, the slower growing oxide formed using the furnace oxidation exhibited a much denser morphology with little or no porosity observed. A possible reason for the increased porosity obtained for the RDS treatments is a result of a preferential grain growth in specific directions due to a van der Drift type of competition with increasing thickness of oxide layer [64].

To conclude it is interesting to compare the morphology of two TiO₂ oxide layers obtained as part of this study as shown in Figure 13. Figure 13(a) was obtained by oxidation of metal oxide nanoparticles and Figure 13(b) by the oxidation of the titanium metal. Both were obtained using the microwave plasma treatments. It is clear from Figure 13 that the sintering of the TiO₂ nanoparticles yields a much more homogeneous oxide pore structure. Its fabrication, however, involves two steps (spray deposition and sintering), compared with the single microwave plasma oxidation step. Both treatments, however, demonstrate the flexibility of the microwave plasma treatments.

4. Conclusions

The use of microwave plasma treatments has been reviewed both for sintering nanoparticles (referred to as rapid discharge sintering (RDS), as a thermal processing route for crystalline phase changes, as well as for the oxidation of metal surfaces. For the fabrication of porous ceramics, the plasma treatments have the advantage of being rapid and flexible allowing for the tailoring and control of pore structures with processing conditions. Compared to the conventional methods such as air furnace treatment, the plasma treatments were demonstrated to yield improved results in terms of shorter treatment time, lower energy requirement and enhanced performance. The latter was demonstrated in the case of metal oxide layers (NiO and TiO₂), when evaluated for use in solar energy applications.