INTRODUCTION

Diamond-like carbons (DLC) possess a series of qualities which make them attractive for nanotechnology applications (such as chemical resistance, hardness, and other mechanical properties comparable to those in crystalline diamond [1, 2, 3]). In this context, one can mention methods for surface nanostructuring based on scanning probe microscopy (SPM) with the application of an electric bias between the SPM probe (tip) and the conductive surface of the sample; the probe serves as a local moving electrode. Attention is paid to electrochemical interactions in the aqueous medium by analogy with works on Si and Ti [4]. The local modification of films induced by the probe is caused by the structural transformation, oxidation [5, 6, 7, 8, 9 ], or even graphitization [10, 11,12 ] of the surface layers. The last circumstance predetermines the interest of researchers to the implementation of DLС-based superhydrophobic structures [13], which are implied owing to the combination of hydrophobic properties of the surface (i.e., graphite) with a developed microrelief.

The anodic oxidation of carbons can induce the removal of material and the formation of grooves [7, 14, 15, 16, 17], as well as the growth of protrusions, whose emergence is interpreted as a result of the incomplete oxidation and the formation of solid carbon oxides [8, 9, 14, 15, 17, 18]. The height of protrusions is usually below 10 nm.

The object of the present study is titanium-doped ultrathin diamond-like a-C:H,Si:O nanocomposite films. The doping of DLC films by a metal (titanium) is usually applied to increase the thermal stability of electroconductive coatings [19]. While exposed to SPM nanostructuring, a-C:H,Si:O films exhibit unique properties [8, 18, 20]. The metal-free films reveal the radial ridges with branches that look like a dendrite upon electric nanolithography with the tip. These surface modifications arise in the film at only high environmental humidity (RH > 50%). The lateral sizes of dendrites achieve several hundred nanometers, and their height enriches tens of nanometers [8, 18], which is comparable to the film thickness, testifying to the bulk processes.

There are various assumptions about the mechanisms of dendrite formations [20]. As is known, the growth of a dendrite structure is observed not only in crystalline but also in noncrystalline media, or a substrate. What they both have in common is a violation of the equilibrium growth conditions due to a lack or excess of the substrate. One example is the emergence of a dendritelike porous structure when the impurity–defect silicon subsystem is decomposed [21]. It is noticeable that the dendrite formation in C:H,Si:O also yields a porous structure [22]. In this work, the electric field around the charged SPM probe induces the process, and the two-component substrate is ionized water and carbon, which interplay during the electrochemical reaction.

In this work, special focus is given to studies of the influence of titanium on the nanostructuring of C:H,Si:O film, initiated by a SPM probe, upon the result of nanolithography. Furthermore, the friction properties of films were analyzed via in situ SPM methods that allow monitoring any changes in the surface subjected to lithography.

EXPERIMENTAL

Both pristine and Ti-doped diamond-like nanocomposite (DLN) C:H,Si:O films were grown onto the Si substrates via chemical vapor deposition (CVD) from the poly phenyl methyl siloxane gas phase [23, 24]. During the process, a metal (Ti) was embedded in the film via the magnetron sputtering of the metal target [24]. The deposition conditions are available in work [25]. The SPM nanostructuring was performed on ultrathin DLN films (with a thickness of ~100 nm) with different Ti contents (0, 15, 22, and 33 at %). Here and hereinafter, DLN films with 15%, 22% and 33% Ti are designated as 15%Ti-DLN, 22%Ti-DLN and 33%Ti-DLN, respectively. In accordance with the electron diffraction, the metal (Тi) is injected in the film in the form of a fine (~5 nm) individual phase, being nonstoichiometric titanium carbide TiC [26].

A choice of the ~100-nm film thickness is due to the ability to compare the results of the present work with those earlier gathered via the SPM nanolithography on the metal-free DLN films with identical structure [8, 20]. The studies were carried out with a SPM NTEGRA Spectra M (NT-MDT) using the contact silicon probes with a conductive Pt coating. The radius of tip curvature was Rtip ~ 30 nm and the force constant was k ~0.6 N/m. The nanolithography was performed in the SPM contact mode by applying a series of N rectangular voltage pulses with an amplitude U = 10 V between the sample and the grounded probe at the chosen surface points of the sample. The total duration, or exposure time t, was varied in a range of 0.1–1000 s. The load on the tip of the probe was F = 120 nN. The surface relief of a structured film was mapped at the same load on the probe in the contact mode. The friction properties of samples were explored via the SPM lateral-force method. The technique for the determination of nanofriction forces in the specified SPM operation mode is thoroughly described in work [20]. The nanostructuring dynamics and changes in electric conductivity of the film were inspected during nanolithography by recording the relief height and current amplitude as the functions of time at the contact point of the probe with a surface.

Furthermore, the state of the surface of the modified area and regions outside it were characterized by evaluating the pull-off (adhesive) force Fpull-off upon removing the SPM cantilever tip from the sample surface. The Fpull-off force was found from the formula Fpull-off = kδZ, where k is a force constant of the probe and δZ is the change in displacement of the probe normal to the surface (along the axis Z) between the moment it touches the surface to the its detachment from the water adsorbate.

During the measurements, the relative air humidity was kept at RH = 40–70%.

RESULTS AND DISCUSSION

The surface reliefs of nanostructures formed by the point electric nanolithography on the pristine and Ti-doped DLN films are plotted in Fig. 1.

Fig. 1.
figure 1

(Color online) Images of nanoobjects formed by applying the point voltage pulses (U = 10 V) on various ultrathin DLN films: (a) undoped DLN, t = 16 s, RH = 60%; (b) undoped DLN, t = 250 s, RH = 60%; (c) 22%Ti–DLN, t = 16 s, RH = 56%; and (d) 15%Ti–DLN, t = 1000 s, RH = 56%.

Full size image

As seen in Figs. 1a and 1b, high humidity (RH > 50%) causes the growth of dendrite nanostructures on the undoped films independently of the exposure time. In the case of Ti-DLN films, the formation of hill-like nanostructures occurs (Figs. 1c, 1d). Long-term electric probe exposure leads to the uniform lateral growth of the nanostructure, but with no dendrites manifested (Fig. 1d). At the same time, the modified regions of Ti-doped films are more localized in comparison with a pristine film for all exposure times. Moreover, the Ti-doped films exhibit the altered friction properties (an increase in friction force) in the affected area, which is also characteristic of metal-free DLN films [20], and the friction is enhanced in the nanostructured area at all exposure times. Figure 2 displays the friction force distributed across the a-C:H,Si:O film with 22 at % Ti after several series of the probe exposure at the specified points (three points per series) with duration t of the voltage pulse applied to a sample. The appropriate exposure times are given inside the image in Fig. 2 alongside with the corresponding series.

Fig. 2.
figure 2

(Color online) Friction forces distributed over the 22%Ti–DLN film surfaces after a series of nanolithographic exposures at various voltage pulses of U = 10 V applied to a surface.

Full size image

Electric current Ip passing through a point probe–sample contact exposed to SPM, measured versus exposure time t during the lithography, exhibits the same behavior in terms of conductivity for both the pristine and Ti-doped DLN films (Fig. 3). The inspection of dependences in Fig. 3 reveals their similar feature, namely, the abrupt decrease (by an order of magnitude) in conductivity (the constant decay time was τз ~ 0.1 s) after the start of the SPM exposure.

Fig. 3.
figure 3

(Color online) Characteristic dependences of the electric current Ip through the probe-surface contact on the probe exposure time t for undoped DLN film (black line) and for Ti DLN film (red line). In the plot: Ip(0) is the onset current of nanolithography.

Full size image

Figure 4 shows the F(z) dependences characteristic to the Ti-DLN films, when the tip is above the surface, which allow one to evaluate the adhesive forces Fpull-off of the SPM probe on the initial and modified surfaces.

Fig. 4.
figure 4

(Color online) Force–distance curves F(z) during the unloading of a SPM tip for modified and initial surface. The Ti content in this film is 15%.

Full size image

The average adhesion forces (averaged over 6 measurements for each case) were found to be Fpull-off = 16.2 ± 3 nN for the modified area and Fpull-off = 21.1 ± 3.4 nN for the initial surface. For a sample with aqueous adsorbate, the adhesive force depends on the wettability of the surface; i.e., it decreases together with wettability. In turn, lower wettability means higher hydrophobicity of the probed surface zone [27]. The dependences in Fig. 4 behave similarly with those F(z) characteristic of pure films [20]. Surface properties shifting towards hydrophobic in the modified area is a common peculiarity of the DLN films used in this study. By analogy with pure films [20], it can be assumed that the changes observed in hydrophobicity do not impact the nanofriction properties of the modified areas for titanium-doped films, but the feasible reason for the increase in friction forces in a zone that underwent nanolithography can be the local decrease in their hardness [22] due to increased porosity of the modified film structure.

As follows from an analysis of results and probe nanolithography processes for nanocomposites, neither the conductivity change dynamics nor the transformation from the surface properties to the hydrophobic in the modified areas and the altered friction properties are specific to Ti-DLN films relative to unalloyed films. In this respect, these mechanisms cannot be referred to the main cause of the missing ability of Ti-containing films to form the dendrite nanostructures.

Let us now interpret the experimental data on the geometry of protrusions formed. For this aim, we calculated the volume of nanoobjects. As is seen in Fig. 5, volume Vs of the modified material rises monotonically with increasing exposure time t for all cases.

Fig. 5.
figure 5

(Color online) Volume of nanoobjects as a function of exposure time for films with various Ti contents in DLN films. The plots are given in the double logarithmic scale. The solid lines correspond to the appropriate linear approximations.

Full size image

The Vs(t) dependence can be expressed as Vs ~ t  n, where parameter n is determined by a slope of the Vs(t) dependence in the double logarithmic scale. The n value for undoped DLN tends to n = 1, meaning the bulk process. Below we consider that each dimension of the object takes n = 1/3; for a two-dimensional case (plane), n = 2/3. For Ti-DLN, the exponent n goes down from n = 0.8–0.9 to n ~ 0.5 with increasing Ti content, exhibiting the trend to “flattening” of the nanoobject. The average modification rate dVs/dt was evaluated to vary from ~3 × 105 nm3/s for 0%Ti-DLN to ~3 × 103 nm3/s for 33%Ti–DLN. As was shown in work [26], the metal (Тi) is injected into the film in the form of ~5 nm clusters of individual phase, being a nonstoichiometric titanium carbide TiC. The TiC compound is a well-studied material with high chemical and thermal resistances (above those of Ti). According to data on the formation of undulating objects with a height of ~10 nm onto a pure titanium film exposed to SPM at times of ~1–10 s [28], one can assume that the electrolytic reaction rate in SPM for a TiC phase is Vs \( \ll \) 5 × 102 nm3/s.

As was mentioned above, a two-component substrate upon the formation of the modified area is presented by (i) water on the surface and in the bulk of the a-C:H,Si:O film and (ii) DLNs themselves that interact during the electrochemical reaction [20].

We suggest that the rate of this electrochemical reaction increases so much that its specific value (per unit length of the reaction-zone perimeter) exceeds the equilibrium one. This circumstance is followed by the enlargement in the electrochemical reaction zone perimeter via the formation of radial beams that maintains its equilibrium course.

Based on experimental data, the growth rate of the modified volume of ~3 × 105 nm3/s is enough for the manifestation of dendrite-like structures, whereas the rate of ~1.5 × 105 nm3/s (for a 15% Ti-DLN film) seems too low. Hence, one can assume that the optimal threshold rate of the dendrite formation is between these values, i.e. ~2–2.5 × 105 nm3/s.

The experimentally measured electric conductivity allows one to imply the oxidative scenario for studied DLN samples. According to the simplest model, one suggests the anodic oxidation of atomic groups containing the basic DLN components (C, Si), localized at the aqueous meniscus around the SPM probe tip. The oxidation of atomic groups with Ti impurity (TiC) occurs at a much lower rate and can therefore be neglected. The anodic oxidation of carbon (by the example of DLN) and silicon is well known and usually causes the formation of highly ohmic layers near the film surface [5, 6, 7, 8, 9, 10 ], which is likely the case in the present work. Meanwhile, attention has to be paid to the fact that this oxidation process should reduce the electrochemical reaction rate proportionally to the Ti content in the film. However, the dVs/dt rate exhibits abrupt drop, falling by an order of magnitude with the Ti concentration increasing from 0 to 22% and thus ensuring the oxidation of the C/Si component across the large surface area. This enables us to assume another factor favoring a decrease in the reaction rate and consequently the inhibition of the dendrite formation in Ti-DLN films. In our opinion, his factor is change of the atomic-structure in the a-C:H,Si:O film caused by Ti.

We will proceed from the analogy with pure DLN a-C:H,Si films, as well as those containing the metal Mo dopant (a-C:H,Si:Mo), grown via a similar way [18]. It is worth noting that a-C:H,Si films, by analogy with DLN a-C:H,Si:O structures, are prone to form dendrites, but adding Mo elements reduces this ability to nothing. In accordance with XRD data reported in work [18], the a-C:H,Si films have a nanocrystalline structure including various atomic groups (methyl and methylene) that are characteristic of hydrocarbon polymers, such as polyethylene and polypropylene. The embedding of Mo leads to a gain in the amount of structurally ordered atomic groups against the present items in a-C:H,Si films. In connection with this, the influence of the atomic structure of films on the nanostructuring processes without a doubt requires further complementary study.

CONCLUSIONS

Summarizing the results obtained in this work, we conclude that one important factor of inhibition of the dendrite structure formation in a-C:H:Si films reinforced with titanium is the atomic structure realignment of the film. Its external manifestation is a decrease in the rate of the electrochemical oxidation induced by a charged SPM tip below the threshold value for the unbalanced process. The SPM characterization of the tribological characteristics of nanostructures formed by the electric point nanolithography reveals no peculiarities associated with the titanium doping of DLN films.