Introduction

In recent years thin film technology has become pervasive in many applications, including microelectronics [1], optics [2, 3], hard and corrosion resistant coatings [4], micro-mechanics [5, 6], etc. Nowadays there is vast number of deposition methods available [7–10].

The selection of the best deposition technique is difficult. In addition, because of ecological demands, the industry is forced to search for alternative environmentally benign methods. One of them might be the application of cold plasmas. The research of cold plasma techniques has generally experienced rapid growth. Unfortunately, for many industrial applications it is not practicable to use vacuum technology with load locks for large scale processes, and even if it were practicable the high capital and running costs of vacuum equipment can become prohibitive [11].

Thus plasma technologies operating at atmospheric pressure have been attracting more and more attention [11–14]. Atmospheric plasma methods are already widely used e.g. for surface modification [15], etching [16] and sterilization of medical instruments [17].

The feasibility of atmospheric plasma methods for thin films deposition has been investigated over the years, but no industrial applications have been known so far, and research in this area is still restricted only to the laboratory stage. Due to practical reasons the films that are most tested are SiOx and TiOx coatings. For deposition, atmospheric plasma jets or plasma torches and atmospheric dielectric barrier discharges (ADBD) were mostly used (For details about ADBD see [11, 18, 19]). ADBD has proven applicability to various purposes and also scalability to very large systems. ADBD arrangement is well adaptable with regard to its geometrical design, operating parameters and process parameters, i.e. to requirements of industrial processes like roll to roll production, deposition in large volumes and in-line continuous manufacturing.

For deposition of organic and organo-inorganic thin films on different types of substrates, including thermal sensitive polymer substrates, such as polymeric ones, thin film chemical vapour deposition (CVD) with ADBD plasma might be ideally suited [20, 21]. The goal of this paper is applicability study of a ADBD plasma reactor for TiOx film deposition on glass substrates with PECVD method and search for the best deposition conditions in the case when TTIP, less toxic than TiCl4 used in [22–24], was applied as a precursor. We investigated deposited TiOx films’ characteristics—chemical composition, surface morphology, wettability and aging. ADBD was operated in the filamentary mode, in an atmosphere containing TTIP and oxygen as reactants and argon as a working gas.

Experimental Part

ADBD Reactor and Experimental Conditions

Experiments were carried out in a (90 × 79 × 41) mm Plexiglas reactor. The CVD system consisted of the reactor, gas input and a AC power supply (Fig. 1). Plasma sustained between two brass electrodes [(45 × 8 × 18) mm and (40 × 17 × 18) mm], which were placed within the reactor. The grounded electrode (with larger cross-section (17 × 18) mm) was covered with a glass plate ((70 × 46 × 1) mm). The inter-electrode distance between electrodes was fixed at 4 mm. The plasma reactor was an open flow-through type. Before the initialisation of any test, deposited film had to be removed from the reactor and surface-related chemical reactions/contamination by ambient air species already started during removal. Reactions proceeded also during stocking of films. Samples were stocked in darkness at room temperature (20 ÷ 23) °C, relative humidity (30 ÷ 40) %, in plastic boxes in air at atmospheric pressure.

Fig. 1
figure 1

Schematic view of the experimental setup. 1 ground electrode, 2 dielectric barrier, 3 substrate, 4 HV electrode, 5 evaporator, 6 mass flow controller, 7 HV supply

Full size image

Electrical parameters of the discharge were measured with the LeCroy LT264 oscilloscope equipped with HV probe Tektronix P6015A used for sampling of supply voltage between electrodes (for characteristic time behaviour of ADBD electric current i(t) and voltage u(t) see Fig. 2). Discharge current i(t) was determined by a voltage drop on 1Ω resistor. A single measurement contained 250,000 samples taken in the 60 ms time span.

Fig. 2
figure 2

Characteristic time behaviour of current i and voltage U of ADBD sustained in air

Full size image

Film deposition was performed with discharge power of about 350 mW [14 kV, 50 Hz], the highest value that was at our disposal. For supply voltages above 17 kV, the danger of dielectric barrier discharge transition to spark or arc discharge was imminent.

ADBD sustained in the filamentary mode. Ambient atmosphere relative humidity and temperature were (36 ÷ 47) % and (20 ÷ 22) °C, respectively.

TTIP (Ti[OCH(CH3)2]4, 97 % purity) in the evaporator volatilized at a temperature of approx. 30 °C. TTIP vapour was mixed with argon and transported into the HV electrode cavity, where oxygen (O2) was added into the mixture. The TTIP/Ar/O2 mixture was fed through the hole (diameter 3 mm) in the HV electrode into discharge space. TTIP vapour concentration in the discharge region was controlled by argon flow changes.

We investigated the influence of TTIP/Ar flow ((0.5 ÷ 5) slm) and of oxygen flow ((2.5 ÷ 10) slm on thin film composition and surface quality. All films were deposited on glass substrates.

Film Characterization and Analysis

Surface chemical composition of films was investigated by X-ray photoelectron spectroscopy (XPS). A multi-channel hemispherical electrostatic analyser (Phobios 100, Specs) was used. The Al Kα line (1,486.6 eV) was used with the incidence angle of X-ray 45 ° to the surface plane. The analyser was operated in the retarding-field mode, applying the pass energy of 40 eV for survey scans and 10 eV for all core level data. XPS peak positions were referenced at aliphatic carbon component at 285.0 eV and were determined with an accuracy of 0.1 eV.

The thickness of deposited films was determined using a surfometer (Planer Industrial, SF 200). To analyse hydrophylicity a contact angle test (CA) was applied. CA was measured by a sessile drop technique when the constant time (30 s) passed after the dropping of distilled water with a volume of about 0.5 μl. CA of each sample was performed at 5 different positions at room temperature.

The surface morphology and surface roughness of TiOx films was examined using the atomic force microscopy (AFM) technique in noncontact mode performed under ambient conditions on a FRT (Fries Research and Technology) AFM Scanning Probe Microscope at the Department of Material Science, Technical University of Liberec. All scans (2 × 2) μm were processed by Gwyddion software for SPM (scanning probe microscopy) data visualization and analysis.

Film surface topography was studied by means of scanning electron microscopy (SEM).

Results and Discussion

Deposition of films proceeds in substrate surface-limited reactions.

Initial deposition tests were performed with TTIP/Ar flows (0.5 ÷ 4) slm (inter-electrode distance 4 mm). Film deposition by surface-limited reactions at TTIP/Ar flows higher than 4 slm is accompanied by volume-limited reactions generating dust. Soft, low quality films were produced at TTIP/Ar flows under 0.5 slm. With increasing oxygen concentration in TTIP/Ar/O2 mixture the deposition rate decreased. Further deposition tests were performed only with TTIP/Ar flows 1 and 2 slm, respectively (Table 1).

Table 1 Deposition parameters after initial tests
Full size table

Effective deposition rate, and discharge power were interdependent. Due to low disposable discharge power and related deposition times and with regard to supply voltage source characteristics and risk of ADBD transition to spark or arc discharge, ADBD had been sustained at maximum permissible voltage value (14 kV, 50 Hz) during experiments.

Film thickness varied in the range of (80–210) nm. The connection between TiOx film thickness and oxygen flow rate is shown in Fig. 3. The highest deposition rate 180 nm/10 min was detected at oxygen flow of 2.5 slm. This value is comparable with rates described in [25, 26]. The films with the best surface characteristics and hardness were deposited with (1 ÷ 2) slm TTIP/Ar flows (0.05 ÷ 0.1 %) TTIP content) and 2.5 slm oxygen flow rate. Mostly films deposited at these conditions are discussed further.

Fig. 3
figure 3

TiOx film thickness dependence on oxygen flow rate (QAr/TTIP = 1 slm (X series); QAr/TTIP = 2 slm (Y series))

Full size image

Chemical Composition

The chemical composition of the films deposited with (1 ÷ 2) slm TTIP/Ar flows (0.05 ÷ 0.1 %) TTIP content) and oxygen flow rate (1 ÷ 4) slm was almost constant. High-resolution spectra for the main elements in films are shown in Fig. 4. It can be seen that the extent of hydrocarbon contamination on film surface is relatively high. This contamination might have originated during the deposition process, but it was most probably caused by post-discharge reactions and adsorption of various species from ambient air atmosphere after film removal from reactor.

Fig. 4
figure 4

XPS spectra of the a C 1 s peaks, b O 1 s peaks and c Ti 2p of TiOx films [TTIP/Ar flow rate 2 slm, O2 flow rate 2.5 slm]

Full size image

Carbon oxygen is an abundant element in TiOx films. It can react both with titanium and with hydrocarbon contaminants. The carbon (contamination) is partially bonded with titanium (C–Ti bond is about 283 eV). Carbon forms mostly C–C (285 eV) backbone chains, some of which are partly oxidized (Fig. 4a).

Oxygen is bonded with titanium forming titanium dioxide (main peak at 530.5 eV). O 1 s spectra are shown in Fig. 4b. Note, that oxygen-titanium peak width is more than 2 eV. Such a broadening is evidently caused by substoichiometric titanium oxides [27]. Furthermore, the second peak at 532.8 eV can be considered as the contribution from single and double oxygen–carbon bonds [28].

Titanium is reactive element and easily forms oxides and carbides, they can be seen in the curve of Ti 2p. (Fig. 4c). Titanium is partially presented in three states: Ti(IV), Ti(III), Ti(II). The Ti 2p3/2 and Ti 2p1/2 spin orbit peaks are located at 458.8 and 465 eV, respectively [29]. Obviously, the location of the strongest peaks at 458.8 and 465 eV indicate that the main titanium species is TiO2. Small peaks at 457 and 462.6 eV indicate a mixed presence of substoichiometric titanium oxides TixOy and titanium carbides TiC. XPS spectra demonstrate that titanium in the near surface area is strongly oxidized, the dominating species being TiO2 and substoichiometric titanium oxides.

The changes of TiOx films elemental composition in dependence with oxygen flow rate are presented in Fig. 5. A measurement of 14.2 % content of titanium is the maximal concentration achieved in films deposition experiments performed with Ar/TTIP and O2 flow rates 1 and 10 slm, respectively (the most diverse experimental conditions). All samples exhibited contamination with carbon. The purest sample contained approximately 32.7 at. % of carbon.

Fig. 5
figure 5

Chemical composition of TiOx/glass thin films as a function of O2 flow rate, a QAr/TTIP = 2 slm, b QAr/TTIP = 1 slm

Full size image

The deposition process resulted more likely in production of TiOxCy films instead of primarily desired TiOx films.

XPS results exhibit insignificant differences between samples deposited with various oxygen flow rates. Initially it was expected that growth of oxidant flow must lead in films to carbon content reduction and titanium content increase. However, as shown in the graphs, titanium content in the film had not significantly grown with increasing oxygen flow rate. Oxygen flow rate rise led to slightly decreasing carbon and slightly increasing titanium content in the chemical composition of films.

Formation of more carbon–oxygen single and double bonds could be a result of post discharge reactions after finalization of the deposition process, or a result of atmospheric oxygen and carbon dioxide reactions with precursor radicals in film surface between the moment of film removal from reactor and the beginning of tests, as already mentioned above. We suppose contamination was probably only superficial. Unfortunately, not being equipped and able to clean the film profile by sputtering and perform XPS profile tests, our research had to focus on film surface tests only.

Surface Morphology

The surface structure was interconnected with precursor concentration, with the TTIP/Ar ratio. Typical topography of films deposited under the most diverse experimental conditions—TTIP/Ar flow rate 1 slm and O2 flow rates 2.5 and 10.0 slm respectively are on Fig. 6 (AFM scans) and Fig. 7 (SEM images). AFM scans and SEM images were not taken from identical samples, but from samples deposited under the same conditions.

Fig. 6
figure 6

Topography of the TiOx films: QAr/TTIP = 1 slm. a O2 flow rate 2.5 slm, b O2 flow rate 10.0 slm

Full size image
Fig. 7
figure 7

SEM images: QAr/TTIP = 1 slm. a O2 flow rate 2.5 slm, b O2 flow rate 10.0 slm

Full size image

The films were deposited in ADBD sustaining in the filamentary mode. The surface topography was influenced with unstable, non-equilibrium character of discharge and random time and space distribution of filaments, weakly ionized, tiny breakdown channels. The deposited surface is irregularly corrugated, hummocky, with growing thickness in spots, where filaments bridged electrodes. Growing/diminishing thickness of the film area reflects local electric field inhomogeneities associated with existence of filaments. The surface topography is similar to results described in [23].

The surface in AFM scan images (Fig. 6) slightly differs in dependence on TTIP/Ar/O2 concentration (O2 flow rates 2.5 and 10.0 slm, respectively, TTIP/Ar flow rate 1 slm).

Figure 6a represents surface deposited at optimal conditions. It is characterised with higher hummocks in comparison with Fig. 6b, when film deposition in surface- limited reactions was accompanied by volume-limited reactions generating dust. Losses in volume-limited chemical reduction were probably interconnected with deposition rate reduction.

RMS values of roughness ranged from 15 to 23 nm in dependence on TTIP/Ar/O2 concentration. There was no distinct relationship between film roughness and gas flow rate values. We suppose that roughness of the surface was most likely related with etching of the film surface by oxygen atoms.

SEM images (Fig. 7) did not reveal any relevant essential difference of the film surface due to oxygen concentration inside the reactor. Both samples (Fig. 7) exhibit similar microstructure with small spherical clusters. Their dimensions reduced with an increase of the oxygen flow rate.

Contact Angle Measurement

The dependence of static CA of water on time passed after deposition determined for various TTIP/Ar/O2 concentrations is shown in Fig. 8. CA values were measured immediately after deposition and in following 28 days. Measurements were performed with a standard error of about 10 %.

Fig. 8
figure 8

Thin film aging: variation of contact angle with time

Full size image

All samples were hydrophilic immediately after deposition. For Ar/TTIP flow rate 1 slm CA value attained 5° immediately after deposition. Hydrophilicity of films remained almost invariable (Fig. 8) in first 7 days after deposition. Later on the films wettability had worsened and in 28 days after deposition the CA value of all samples exceeded 40 °. Films deposited with QAr/TTIP = 2 slm changed from hydrophilic to hydrophobic faster. Most hydrophilic film was prepared at flow rates QAr/TTIP = 1 slm; QO2 = 2.5 slm (Fig. 8).

Chemical reactions between film surface and contaminants from ambient air atmosphere (e.g. absorption of hydrophobic impurities and fall-out contamination) resulted in differences between CA values estimated in first days after deposition and later on. These changes are in agreement with results obtained in [30].

Conclusion

Film deposition on glass substrates was performed by the PECVD method in ADBD plasma reactor. The plasma reactor was an open flow-through type. ADBD sustained in filamentary mode.

The reactor atmosphere consisted of TTIP/O2/Ar mixture. The mixture composition was controlled by TTIP/Ar flow and oxygen flow changes. Mixture composition influenced the quality of films, hence the best deposition conditions had to be recognised and then used.

Film thickness varied in the range of (80–210) nm. The highest detected deposition rate 180 nm/10 min corresponds to deposition rates values described in the literature [25, 26].

XPS spectra demonstrate that titanium in the films near the surface area is strongly oxidized, dominating species being TiO2 and substoichiometric titanium oxides. Titanium maximal content in films was 14.2 %. All samples exhibit contamination with carbon. Contamination might have originated during the deposition process, but it was probably caused by post-discharge reactions and adsorption of various species from ambient air atmosphere after film removal from the reactor. Unfortunately, not being equipped and able to clean the film profile by sputtering and perform XPS profile tests, we can only assume that contamination was only superficial. The purest sample contained approximately 32.7 at. % of carbon.

XPS tests indicate that the deposition process resulted more likely in production of TiOxCy films instead of primarily desired TiOx films.

SEM images did not reveal any relevant difference of the film surface in dependence on oxygen concentration inside the reactor. It seems, that the reactor atmosphere composition affected more the wettability of films than their surface morphology.

The surface topography was influenced with unstable, non-equilibrium character of discharge and random time and space distribution of filaments, weakly ionized, tiny breakdown channels. The deposited surface is irregularly corrugated, hummocky, with growing thickness in spots, where filaments bridged electrodes. Growing/diminishing thickness of the film area reflects local electric field inhomogeneities associated with the existence of filaments. The surface topography is similar to results obtained in deposition with precursor TiCl4 [23].

All films were hydrophilic immediately after deposition. Later their wettability diminished and the contact angle value of all samples exceeded 40 ° after 28 days; changes were probably related with chemical reactions between film surface and chemical groups involved in air atmosphere.

PECVD process in ADBD reactor consists of two different mechanisms: surface-related chemical processes resulting in conventional PECVD film deposition, and undesired volume-related chemical processes resulting in dust production. Under certain working conditions (e.g. higher O2 flow rate values) a second mechanism prevails. Dust particles, when created, remain in the discharge region and their layer(s) at the substrate hinder effective creation of more homogeneous film and influence film quality.

Another problem of PECVD thin film deposition with ADBD seems to be with the filamentary character of the ADBD leading to generation of hummocks forming rough surface of films.