Features of copper coatings growth at high-rate deposition using magnetron sputtering systems with a liquid metal target

Surface & Coatings Technology 324 (2017) 111–120

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Features of copper coatings growth at high-rate deposition using magnetron sputtering systems with a liquid metal target G.A. Bleykher a,⁎, A.O. Borduleva a, A.V. Yuryeva a, V.P. Krivobokov a, J. Lančok b, J. Bulíř b, J. Drahokoupil b, L. Klimša b, J. Kopeček b, L. Fekete b, R. Čtvrtlìk c, J. Tomaštik c a

Institute of Physics and Technology, Tomsk Polytechnic University, Lenin Avenue, 2a, Tomsk 634028, Russia Institute of Physics of Czech Academy of Sciences, Na Slovance 2, CZ-18221 Prague, Czech Republic Institute of Physics of Czech Academy of Sciences, Joint Laboratory of Optics at Palacky University and Institute of Physics of Czech Academy of Sciences, 17.listopadu 12, 772 07 Olomouc, Czech Republic

b c

a r t i c l e

i n f o

Article history: Received 17 February 2017 Revised 20 April 2017 Accepted in revised form 23 May 2017 Available online 24 May 2017 Keywords: Magnetron sputtering Evaporation High-rate coating deposition Coating properties Cu coatings

a b s t r a c t The article focuses on the study of growth conditions for metal coatings during the operation of magnetron sputtering systems with a liquid-phase target. It also discusses the trends of coating properties formation (in terms of copper example) depending on the growth conditions. The data of the experiments and calculations confirm that the appearance of intensive evaporation on the target surface allows increasing more than 10 times the deposition rate and deposited particles flux density in comparison to conventional sputtering with a cooled solid target at the same magnetron power. In the case of the magnetron sputtering system with a liquid-phase target, energy fluxes and particles ones towards the substrate during the coating growth and the substrate heating rate are much greater than at conventional magnetron deposition with a solid target. The combination of calculations and experiments have made it possible to reveal the structure of energy and particles fluxes towards the substrate during the operation of the magnetron sputtering system with a liquid-phase target. The contribution of these fluxes formed by different mechanisms in a wide range of magnetron power has also been found out. The analysis of structural and mechanical properties of copper films being deposited at different energy and particles fluxes ratio towards the substrate has been carried out. It has been found that at intensive evaporation the coating surface is smooth and uniform, and the size of grains decreases. The mechanical properties of coatings (adhesion and microhardness) have got higher values compared to the cases related to only sputtering. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Magnetron sputtering is a commonly used technique for deposition of modifying coatings on the surface of workpieces and materials. However, despite the development level of this technology, it has a number of disadvantages such as a low coating deposition rate and low deposited particles flux density [1–3]. Regardless the power supply, the deposition rate of metal coatings is usually not more than 10 nm/s [2,4,5]. It happens due to the fact that the main mechanism of atoms emission from the target surface is the sputtering with the ions of working gas. The sputtering rate is proportional to the ion current power density applied to the target from the magnetron plasma region. Due to the fact that there are technical limitations on the magnetron power supply, the sputtering rate cannot be increased significantly.

⁎ Corresponding author. E-mail addresses: [email protected] (G.A. Bleykher), [email protected] (V.P. Krivobokov).

http://dx.doi.org/10.1016/j.surfcoat.2017.05.065 0257-8972/© 2017 Elsevier B.V. All rights reserved.

The investigations of various research teams show that the evaporation initialization in addition to sputtering on the target surface of magnetron sputtering systems (MSS) can significantly increase the coatings deposition rate [6–11]. The evaporation rate increases almost exponentially with the rise in target surface temperature, which has a non-linear dependence on ion current power density of the magnetron discharge [12]. Moreover, the increase in the atoms emission rate from the target surface leads to the instantaneous growth of deposited particles flux density. A substantial increase in the temperature of a target can be achieved by its thermal isolation using special inserts between the target assembly and the magnetron body with a cooled magnet system (Fig.1). As a result, the energy which comes on the target from plasma will not flow away due to heat conductivity, and the target substance can melt. In this case the crucible made from higher-melting-point material (e.g. Mo, Graphite) should be used to preserve the target shape. The coatings deposition rate can be increased by several tens of times at the same power density in comparison to sputtering cooled solid targets [9,12].

112

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

trends in properties change depending on the operating parameters of MSS with liquid metal targets. 2. Calculation method of energy and particles fluxes Since the energy and particles fluxes are difficult to be measured in experiments, therefore, their density has been calculated depending on the power of MSS. The atoms flux from the highly heated target surface consists of two independent components: sputtered and evaporated particles. Total rate of surface erosion can be represented as: V ¼ V sput þ V ev ;

Fig. 1. Layout of the cathode assembly and substrate: 1 – magnetron body, 2 – ceramic insert, 3 – magnetic core, 4 – target, 5 – crucible, 6 – magnets cooled by water, 7 – ceramic insulator, 8 – substrate, 9 – substrate holder, 10 – the place of thermocouple mounting.

There is a demand for magnetron technologies of high performance such as deposition of protective, conductive and decorative layers on the surface of materials and workpieces. Therefore, it is necessary to make coating properties meet the operating requirements in addition to high performance. It should be noted that in this case we deal with totally different coatings growth conditions compared to the conventional magnetron sputtering of cooled solid targets because the total energy and particles fluxes towards the substrate are much higher. That is why the structural and functional properties of coatings when using magnetrons with liquid phase targets can also be different. It is known that the substrate heating influences the following processes such as adsorption, desorption, mobility of adatoms, chemical reactions etc., which further define the coating properties. The models of coatings structure formation at sputtering cooled solid targets have already been developed [13,14]. These models use normalized temperature T n as one of the important criterion (Tn = Tsub/Tmelt, where Tsub is substrate temperature, Tmelt is melting temperature of the target material). The impact of the ion bombardment on the coatings structure has led to the introduction of another criterion – the amount of energy that comes to the substrate fallen within one deposited atom (Ea) [15–17].This criterion is used for an analysis of coatings structure and properties obtained by sputtering cooled solid targets. However, it is not clear whether this criterion is suitable in our case. There are a lot of works devoted to revealing dependencies between the deposition parameters and coatings properties formed by different types of MSS with solid targets, such as [18–21]. However, there is no any research that focuses on connecting deposition parameters with properties of coatings obtained by MSS with liquid phase targets. To predict what properties the coatings might have depending on the operating parameters of MSS with liquid phase targets and to be aware of the methods of a parametric control, it is necessary to reveal the relation between the characteristics of energy and particles fluxes coming to the substrate and the peculiarity of the growth structure and functional properties of formed metal coating. This issue has not found all-around consideration in the scientific community therefore this article is to present the results of the research related to this area. Making calculations of energy and particles fluxes on the substrate, running experiments related to coatings deposition, and doing subsequent analysis of coatings properties, it has been attempted to reveal the

ð1Þ

where Vsput and Vev are the rates of surface erosion due to sputtering and evaporation. The models of thermal and erosion processes based on solution of heat conductivity equation at considering heat consumption for phase transitions and sputtering have been used to calculate the target temperature and evaporation rate; the boundary conditions are determined by the energy balance in the “target in the crucible” system under the influence of the ion current from magnetron plasma. The models are described in [10,22]. The sputtering rate is calculated according to the Sigmund formula for the primary knock-out mode [23]. The growth rate of the deposited coating Vdep in any element on the substrate with coordinates (X, Y) is influenced by the contribution of sputtered and evaporated particles and calculated by the Lambert-Knudsen law [24]:
V dep ðX; Y Þ ¼ F dep =n0 ¼ F dep;sput þ F dep;ev =n0      V sput xtarg ; ytarg þ V ev xtarg ; ytarg dxtarg dytarg ð2Þ L2 ¼ ∬  2 2 π Starg
2  L2 þ X−xtarg þ Y−ytarg

Here Fdep is a total deposited particles flux density; Fdep,sput and Fdep,ev are the deposited flux density of sputtered and evaporated atoms respectively; n0 is nuclear density of the coating, which is assumed to be equal to the target nuclear density; L is a distance between target and substrate placed in parallel; Starg is target surface area. The rates of sputtering (Vsput) and evaporation (Vev) on the element of the target surface with coordinates (xtarg, ytarg) should be expressed in m/c. The application of this method for calculating the deposition rate is appropriate in our case since the working pressure does not exceed 0.2 Pa, and the mean free path of emitted atoms is commensurate with the distance between the target and substrate. The energy flux density on the substrate Qtotal is equal to the following: Q total ¼ Q rad þ Q cond;sput þ Q cond;ev þ Q kin;sput þ Q kin;ev ;

ð3Þ

where Qrad is the flux density of heat radiation from the target; Qcond,sput and Qcond,ev are the energy fluxes densities due to condensation of sputtered and evaporated atoms; Qkin,sput and Qkin,ev are the fluxes densities due to kinetic energy of sputtered and evaporated atoms being deposited on the substrate. The heat radiation flux density per an element of substrate surface with coordinates (X, Y) is calculated on the basis of the laws of Lambert and Stephen-Boltzmann for a gray body, taking into account the uneven distribution of the temperature on the target surface:     εp σ SB T 4targ xtarg ; ytarg −T 4sub dxtarg dytarg L2 ∫ : Q rad ðX; Y Þ ¼  2 2 π Starg
2  L2 þ X−xtarg þ Y−ytarg

ð4Þ

Here εp is specific emissivity (εp = 1/(1/εtarg + 1/εsub − 1); εtarg and εsub are emissivity of the target and substrate surface respectively; σSB

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

is Stefan-Boltzmann constant; Ttarg is temperature on the target surface in an element with coordinates (xtarg, ytarg); Tsub is substrate temperature. The flux density due to heat released during the condensation of evaporated or sputtered particles being deposited on the substrate is equal to: Q cond;sput=ev ðX; Y Þ ¼ F dep;sput=ev ðX; Y Þ  U S ;

ð5Þ

where Fdep,sput/ev is the flux density of the sputtered or evaporated atoms being deposited on the substrate; Us is surface binding energy. The kinetic energy of the evaporated particles being deposited on the substrate is estimated at an average rate of the Maxwell distribution:       4kB  T targ xtarg ; ytarg =π  V ev xtarg ; ytarg dxtarg dytarg L2 n0 ∫ Q kin;ev ðX; Y Þ ¼ :  2 2 π Starg
2  L2 þ X−xtarg þ Y−ytarg

ð6Þ Here kB is Boltzmann constant. The most probable energy in Thompson spectrum [25] is used to estimate the kinetic energy flux of the sputtered atoms being deposited on the substrate: Q kin;sput ðX; Y Þ ¼

US F ðX; Y Þ: 2 dep;sput

ð7Þ

The substrate temperature is calculated by the heat conductivity equation with boundary condition (3) on the surface, at which the coating is growing. On the back surface of the substrate, the heat exchange with the environment is due to heat radiation. The energy per a depositing atom is calculated as: Ea ¼

Q total : F dep

ð8Þ

In Table 1, there is an overview of equations and formulas parameters used throughout the text. 3. Experiments details and coatings diagnostic methods To determine the influence of the target temperature-phase condition and evaporation on coatings properties, the experiments were

113

Table 2 Geometric parameters of the MSS targets in the experiments and calculations. Target type Solid cooled target Heat isolated Cu target in the crucible of

a) Graphite b) Mo

Shape Disc Ring

Sizes Starg = 63.5 cm2 Starg = 62 cm2, hc = 1.5 cm, wc = 0.3 cm, Starg+c = 76.9 cm2

Note: Starg – target surface area; hc and wc – crucible height and wall thickness; Starg+c – total area of the target surface and crosscut walls of the crucible.

carried out with three copper targets having the same sputtering area but different heat exchange with the environment. For the first one, the complete cooling was provided, in other words conventional sputtering of a solid phase target. The second target was placed in a heat insulated Graphite crucible that had high emissivity (ε = 0.9). The third target was located in a Mo crucible that had low emissivity (ε = 0.1). The layout of the cathode assembly for the last two cases is shown in Fig. 1. Geometric parameters of the MSS targets are in Table 2. In all the cases, Cu target purity was 99.99%, the base pressure was 9∙ 10−3 Pa, and sputtering process was performed in the argon atmosphere at the working pressure of 0.18 Pa. The discharge power was 3 kW in DC mode. The distance between the substrate and target was 14 cm. The coatings were deposited on glass and Si (100) substrates with the dimensions of 2.54 × 7.62 × 0.1 cm3 and Ø3x0.04 cm, respectively. The substrates ion pre-treatment was being carried out for 20 min at the discharge current of 38 mA and voltage of 2.8 kV. The substrate temperature was measured by thermocouple M422 Pt 100. The scheme of the thermocouple fixing is shown in Fig. 1. The surface morphology was studied by scanning electron microscope (SEM) TESCAN FERA3 GM. The elemental compositions of the coatings were determined using a built-in energy-dispersive X-ray analyzer of the SEM. The root-mean-square roughness of the Cu samples was obtained by ambient Bruker Dimension Icon atomic force microscope (AFM). X-ray structure analysis (XRD) was carried out with the diffractometer X'Pert Pro from PANalytical Company equipped with Co tube. The lattice parameters were determined by Pawley mode in the Rietveld analysis [26] using the Topas software tool [27]. The films thickness was measured by a rapid analysis using spherical metallographic microsection (Calotest CAT-S-0000, CSEM).

Table 1 Overview of the parameters in the equations and formulas. Parameters

Descriptions

Pmag V Vev, Vsput Vdep Fdep Fdep,ev, Fdep,sput Ea

Magnetron power density Total rate of a target surface erosion Rates of a target surface erosion due to evaporation and sputtering Total rate of coating deposition Total deposited particles flux density Flux densities of evaporated and sputtered atoms being deposited on the substrate Amount of energy coming to the substrate and been fallen within one deposited atom Averaged temperature of a target surface Total energy flux density on the substrate The density of energy flux towards the substrate due to heat radiation from the target The densities of energy fluxes on the substrate due to condensation of evaporated and sputtered atoms respectively The densities of energy fluxes on the substrate due to kinetic energy of sputtered or evaporated atoms being deposited on the substrate; this energy releases on the substrate at thermalization of deposited particles Emissivity of a crucible surface

Ttarg Qtotal Qrad Qcond,ev, Qcond,sput Qkin,ev, Qkin,sput

ε

Fig. 2. Dependence of the coating growth rate Vdep on the magnetron power density Pmag in the case of MSS with a heat isolated Cu target in Mo (1), Graphite (2) crucible and with a cooled solid target (3). Solid lines – calculations, dots – experiment data.

114

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

The hardness and reduced modulus were obtained by nanoindentation with a diamond pyramidal Berkovich indenter at the indentation load of 1 mN. The adhesive properties of Cu coatings were explored by the Scratch test performed up to maximum normal load of 500 mN with a diamond Rockwell indenter (nominal radius of 10 μm). Standard procedures were used to analyze the experiment results in both cases. 4. Results and discussion 4.1. Particles and energy fluxes on the substrate In calculations, the ion current power density was equal to the magnetron power density, all geometric parameters and target sizes were the same as in the experiments (see Table 2). Fig. 2 presents calculations (solid lines) and experimental results (dots) of a deposition rate (Vdep) of Cu coatings at different modes depending on the magnetron power density (Pmag). Curve 1 presents the mode of deposition by sputtering of a liquid target placed into a Mo crucible; curve 2 is deposition using liquid target in a Graphite crucible; and curve 3 is deposition using solid target. The correctness of calculations is confirmed by the experimental results (dots in Fig. 2). In the case of MSS with a solid target, the deposition rate linearly depends on the power density of the magnetron because only sputtering takes place (curve 3). In the case of MSS with a liquid target, the nonlinear growth of an erosion rate is observed with a magnetron power density increase (curves 1 and 2). At low power density, the erosion of the targets surface is mainly due to the sputtering. The target temperature increases with a magnetron power rise, hence an evaporation rate also increases and at some instant, in the case of the Mo crucible at Pmag ≥ 20 W/cm2 and Graphite crucible at Pmag ≥ 60 W/cm2, it becomes predominant. In Fig.2 it can be seen that the evaporation increases the deposition rate 10 times more in comparison with conventional solid target MSS. The quantitative difference of dependencies Vdep(Pmag) presented by curves 1 and 2 is connected with varying intensity of heat radiation coming from the crucibles surfaces that have different values of emissivity (ɛ). The emissivity for Mo is equal to 0.1 and for Graphite is 0.9. The higher deposition rates can be achieved (almost an order of magnitude higher) by using a crucible with lower emissivity at the same magnetron power density. For example, at Pmag of 39 W/cm2 the deposited particles flux density for a Mo crucible is the value of about 1018 atoms/ cm2s, and for a Graphite one is the value of about 1017 atoms/cm2s. In Fig. 3, the calculated energy fluxes towards the substrate versus magnetron power density are shown. In the picture, the calculations only for the Mo crucible are shown. In the case of a liquid target, the

contribution of heat radiation from the surface target is rather significant. However, when Pmag N 50 W/cm2, the energy released because of condensation of evaporated particles being deposited on the substrate becomes predominant. The contribution of kinetic energy of deposited particles, both sputtered and evaporated, is negligible. For the Graphite crucible, a similar situation was observed. In a qualitative sense, the fluxes ratio was the same but the numerical values were different. The amount of energy coming to the substrate and been fallen within one deposited atom (Ea) in the case of using solid as well as liquid targets is presented in Fig. 4. For a solid target, this value is equal to about 5 eV/atom, and the value doesn't depend on the magnetron power density if there is no any bias on the substrate. In the case of a liquid target in a Mo crucible, Ea reaches maximum of about 78 eV/atom. It is almost 15 times more than at conventional sputtering. The maximum of Ea occurs when the target is highly heated but Vsput N Vev. This value decreases with an increase in Pmag because the flux density of deposited particles rises due to evaporation. At Pmag equal to 70 W/cm2, Ea in the case of a Mo crucible decreases to the level of a conventional magnetron, i.e. when the energy due to condensation is predominant, and mostly evaporated particles come to the substrate. It is known that evaporated particles have just the thermal energy kBT, which is lower than the kinetic one of sputtered

Fig. 3. Energy fluxes on the substrate versus magnetron power density in the case of MSS with a Cu target in a Mo crucible (1 – Qcond,ev; 2 – Qrad; 3 – Qkin,ev; 4 – Qcond,sput; 5 – Qkin,sput).

Fig. 5. The substrate heating evolution during the coating deposition using a liquid Cu target in a Mo (1) and Graphite (2) crucible at magnetron power of 3 kW (Pmag = 39 W/cm2). Solid lines – calculations, dots – experiment data.

Fig. 4. The energy coming to a substrate per one deposited atom versus magnetron power density in the case of MSS with a heat isolated Cu target in a Mo (1), Graphite (2) crucible and with a cooled solid target (3).

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120 Table 3 The calculated data on substrate average temperature and density of energy and particles fluxes towards the substrate formed in different modes of the target temperature-phase condition at the same magnetron power of 3 kW. Fdep, at./cm2s

Fev/Fdep Q, W/cm2

Mode

Pmag, W/cm2

Ttarg, K

Solid Cu target Liquid Cu target in Graphite crucible Liquid Cu target in Mo crucible

47 39

~300 1.98∙1017 0 1380 1.68∙1017 0.02

0.123 0 5.3 0.72 0.58 26.9

39

1760 1.48∙1018 0.89

2.46

Qtotal

Qrad

E a, eV/at.

1.55 10.4

particles. Thus during coatings deposition by MSS with a liquid target in a Mo crucible at Pmag N 70 W/cm2 mostly particles with the thermal energy come to the surface of a substrate. The energy fluxes on a substrate at the operation of MSS with liquid metal targets are much higher than the energy fluxes at conventional magnetron sputtering. Therefore, the substrate is heated rapidly if there is no its special intensive cooling. To reveal the substrate heating rate depending on the deposition conditions, we have made calculations for temperature evolution in the substrates in the cases of both crucibles at the same magnetron power of 3 kW (or power density 39 W/cm2). The calculations are shown in Fig. 5. Here the results of the experiments on the substrate temperature measurement during coating deposition are also presented. The substrate heating rate is about 2 K/s when intensive evaporation occurs, i.e. in the case of a molten Cu target in the Mo crucible, whereas it is about 0.3 K/s in the case of sputtering of a molten Cu target in the Graphite crucible. The calculations show that a thin substrate (1 mm glass) is uniformly heated. A good agreement of the calculations results and experiments indicates the reliability of the energy balance in the substrate, which is described by Eq. (3).

4.2. The coatings properties In the experiments with a magnetron power of 3 kW, the Cu target with full cooling remained solid. The Cu target in a Graphite crucible, which had high emissivity, reached the melting point, but the evaporation rate was low. The third target, which was located in a Mo crucible with low emissivity, was not only melted but also heated up to the temperature at which the evaporation rate was greater than the sputtering rate. The calculated values of the target temperature, the energy and particles fluxes, which come to the substrate, are presented in Table 3.

115

The structural and mechanical properties of Cu coatings deposited in the modes with a liquid target were studied. These results were compared with Cu coatings obtained by sputtering of a cooled solid target. The main difference of deposition using a liquid target and the usual MSS with a cooled target is a high flux of heating radiation towards the substrate and high deposited particles flux density in the case with a Mo crucible (Table 3). In Fig. 6, the results of the morphology study of Cu coatings deposited at different modes are presented. It can be seen that in the case of the Graphite crucible the coating has a coarse-grained and more regular structure than in case of deposition using a solid target. Here apparently the heat radiation from the target plays an important role. The coating in the case of using a Mo crucible has much more regular structure in comparison with the deposition from a cooled solid target and much more fine grains than in the case presented in Fig. 6b. If we look at the data in Table 3, we can see that the coating in Fig. 6c is obtained by involving a significant portion of evaporated atoms. Perhaps, the combination of the high particles flux with high energy flux, which come to the substrate during growth of the coating, has led to the formation of a regular and fine-grained structure. There is a noticeable difference in a growth structure of the coating deposited by a liquid target in a Mo crucible from the coating deposited by a cooled solid target and a liquid one in a Graphite crucible. In the case of the deposition using a cooled solid target and a liquid one in the Graphite crucible, the coatings have a columnar structure. In the case of a liquid target in the Graphite crucible, there are a lot of voids that can be seen in the Fig. 7b. If the substrate temperature is quite low and there is no additional impact (bias towards the substrate or its rotation), the porous coatings are often formed [14,20]. It is quite possible that a similar situation is observed for coatings formed by sputtering the liquid phase target placed in the Graphite crucible. At such target temperature, the substrate is probably still insufficiently heated to ensure the intensive diffusion and the formation of dense coating. The deposition conditions using the target placed in the Mo crucible differ due to the presence of a significant amount of evaporated particles in the deposited flow. The evaporated particles increase the value of Fdep by several times. Another difference is a high flux of the heat radiation (Table 3). All of this together has led to strong heating of the substrate (Fig. 5). A substrate temperature rise leads to an enhancement of atoms mobility on its surface layers. Therefore, the formed coating is denser than in other cases.

Fig. 6. SEM surface morphology of Cu coatings: (a) – solid target, (b) – liquid target in a Graphite crucible, (c) – liquid target in a Mo crucible.

116

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

Fig. 7. Cross-section of Cu coatings deposited on glass substrate: (a) – solid target, (b) – liquid target in the Graphite crucible, (c) – liquid target in the Mo crucible, (d) – liquid target in the Mo crucible (lower resolution).

The research group [28] has described that the grain size increases with a substrate temperature rise. It is confirmed by the data obtained by SEM for coatings deposited by sputtering a liquid target in the Graphite crucible (Fig. 6b). Films deposited in the Mo crucible by MSS with liquid targets have a morphology that can't be explained by the dependence mentioned above. Such films have a small grained structure even at a high

Table 4 Proportions of different elements in Cu coatings formed in different modes of the target temperature-phase condition at the same magnetron power of 3 kW. Element

C O Cu

Solid cooled Cu target

Liquid Cu target in Graphite crucible

Liquid Cu target in Mo crucible

wt%

at.%

wt%

at.%

wt%

at.%

2.22 1.20 96.57

10.39 4.23 85.38

1.77 0.83 97.41

8.50 2.98 88.52

1.78 1.12 97.10

8.49 4.02 87.49

temperature of the substrate. We suppose that in the latter case a significant increase in the deposited particles flux density due to evaporated atoms has a critical impact on the structure of the coating growth. For the samples deposited using a liquid phase target the energy dispersive spectroscopy on SEM has been carried out. The proportions of different elements are given in Table 4. According to the results, it can be concluded that the presence of a crucible during the deposition process does not change the chemical composition of formed coatings. Figs. 8 and 9 contain the XRD results of the Cu coatings deposited at the different modes. These measurements have shown that all the films have a crystalline structure. The high intensity of peaks in XRD patterns of the Cu coating deposited using the Graphite crucible (see Fig. 8) and Pole figures (Fig. 9a) indicate a strong texture of this coating, i.e. strong preferred orientation of (111) direction along the surface normal. The Cu coating deposited using a Mo crucible is almost without a texture, and a weak texture is observed in (111) direction and it has maximum along the surface normal. The data on lattice parameters, a crystallite size and microstrain from XRD measurement are presented in Table 5.

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

Fig. 8. The XRD patterns of the Cu coatings deposited on a Si substrate.

The lattice parameters of all coatings are close to bulk copper but the crystallite size is very different. In the mode of films deposition using the liquid phase target in the Graphite crucible, the crystallite size of formed coating is three times larger than in the case of sputtering a solid target, and the crystallite size is four times larger in the case of sputtering a liquid phase in the Mo crucible. The crystallite size correlates with microstrain. The films with a larger crystallite size have greater microstrain therefore they contain more defects. The breadth of the peaks that can be simply taken as a measure of microstructure is similar to all three films. According to the data obtained, it can be concluded that the presence of high temperature leads to an increase in the crystallite size of formed coating (deposition using liquid phase in the Graphite crucible), whereas the combination of high deposited particles flux density and temperature (Mo crucible) reduces the crystallite size. The data from AFM has shown that the coating deposited using the solid target has the highest roughness Ra, and the lowest Ra is in the case of the Мо crucible (Table 6). The obtained data correlate with the results of SEM (Fig. 6). The combination of high substrate temperature

117

and high deposited particles flux density creates conditions for the roughness reduction of the growing coating. The Cu coating deposited using the MSS with a liquid target in the Mo crucible has the highest hardness value. The coatings obtained by sputtering a solid target and liquid target in the Graphite crucible have almost the same hardness values. Thus, in this case a substrate temperature increase doesn't significantly influence the hardness values. The obvious explanation for the hardness change in the case of the Mo crucible is a high deposition rate due to evaporation, because of which the deposited particles flux density increases several times. The reduced modulus increases in all cases in comparison with the bulk Cu (100 GPa). Copper is a high-ductile metal. It can be seen from the obtained results that the coating deposited by the MSS with the liquid target in the Mo crucible has good ductility, and the coating in case of using the Graphite crucible is less ductile. These results are in the good agreement with the adhesion data, which will be presented further. In general, the mechanical properties depend on the structure of coatings. According to the Hall–Petch strengthening (hardness is proportional to 1/d1/2, where d is the grain size) [29,30] the film hardness increases with decreasing d, which is confirmed in our case. Coatings obtained by the MSS with the liquid target in the Mo crucible have a small-grained structure, which can be seen from morphology and XRD results (Fig. 6, Table 5). Researchers [31] showed that increasing the temperature of the substrate leads to a decrease in the hardness of the coatings. It does not correlate with the data obtained for the MSS with the liquid target in the Mo crucible, where the substrate temperature is the highest. Therefore, in this case it seems that the factor of high density of the deposited particle flux is very important for the coating hardness. Figs. 10–12 represent the adhesion/cohesion properties of Cu coatings. The top part of the figures provides information on the applied load and scan distance. The bottom part corresponds to the scan distance and visualizes the moment of coating failure or its indentation into the substrate. The red lines in the bottom part help to identify the indenter traces.

Fig. 9. The pole figures plotted in XRD analysis of the coating, deposited using the molten Cu target in the Graphite (a) and Mo (b) crucible for a different orientation (111), (200), (220).

118

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

Table 5 XRD data of Cu coatings deposited on the surface of Si substrate in different modes of the target temperature-phase condition at the same magnetron power of 3 kW. Mode

Lattice constant a (Å) (bulk Cu a = 3.615 Å)

Crystallite size [nm]

Microstrain [%]

Solid Cu target Liquid Cu target in Graphite crucible Liquid Cu target in Mo crucible

3.617 3.611 3.607

25 81 18

0.14 0.18 0.13

Table 6 Mechanical properties of Cu coatings formed in different modes of the target temperature-phase condition at the same magnetron power of 3 kW. Mode

Thickness, μm

Roughness (Ra), nm (5 × 5 μm)

Hardness, GPa

Reduced Modulus, GPa

Solid Cu target Liquid Cu target in Graphite crucible Liquid Cu target in Mo crucible

0.78 1.18 1.50

15.0 7.5 6.1

3.8 3.9 5.0

144 152 131

The character of the residual groove for the coatings obtained during MSS with the liquid target in the Graphite crucible clearly displays the brittle character of the film and poor adhesion. It can be noticed from the unstable behavior of the residual scratch grooves on the top part of Fig. 10 and the large spalled area on its bottom part. On the other hand, the coatings deposited by sputtering the solid target and the liquid one in the Mo crucible, exhibit ductile behavior with ploughing at the edges of the scratch track. These two films get thinner as the normal load is increased under the indenter and are finally eliminated (see Figs. 11, 12). Comparison of obtained results shows that the factor of the substrate high temperature without any significant increase of the deposited atoms flux density can contribute to coating adhesion degradation.

5. Conclusion Evaporation on the surface of a highly heated magnetron target results in an increase of the metal coatings deposition rate up to 100– 1000 nm/s. These values exceed by 10–100 times the deposition rates typical for conventional magnetrons that perform only sputtering. Heat processes in the target and its heat exchange with environment are the key factors in the increase of atoms emission intensity and as consequence coating deposition rate. The crucible surface emissivity, in which a melting target should be placed, is one of the most important factors. The less crucible surface emissivity is, the less magnetron power density is required to create a noticeable evaporation.

Fig. 10. Scratch test evaluation: (a) – depth-time records, b, c, d – residual scratch grooves. Cu coatings deposited on the surface of glass substrate using liquid target placed in the Graphite crucible.

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

119

Fig. 11. Scratch test evaluation: (a) – depth-time records, b, c, d – residual scratch grooves. Cu coatings deposited on the surface of glass substrate using liquid target placed in the Mo crucible.

Fig. 12. Scratch test evaluation: (a) – depth-time records, b, c, d – residual scratch grooves. Cu coatings deposited on the surface of glass substrate using solid target.

120

G.A. Bleykher et al. / Surface & Coatings Technology 324 (2017) 111–120

The conditions of the coatings growth on a substrate surface at operation of MSS with a strong heated target are characterized by high flux density of heat radiation from the target surface. The contribution of condensation heat of evaporated atoms into the energy flux coming to the substrate is also significant. As the power of the magnetron increases, the contribution becomes dominant. The rate of substrate heating under similar conditions of deposition is almost an order of magnitude higher than that of conventional magnetron sputtering. Peculiarities of deposition conditions at MSS with evaporation of targets regarded with much higher heating of the substrate by energy flows from the target and the multiply increased density of deposited particles have a significant effect on the structural and mechanical properties of the coatings. In the case of copper, it has been found that the crystallites sizes, coatings structure, their mechanical and adhesion properties nontrivially depend on the combination of energy and particles flux densities. Thus, the tendency for decreasing the crystallite size, increasing the smoothness and uniformity of the coating surface, compacting its structure, increasing adhesion and microhardness in the presence of a predominant fraction of evaporated particles in the deposited flow is observed.

Acknowledgements This research has been supported by Russian Science Foundation (project №15-19-00026). The authors acknowledge the financial support from the Ministry of Education, Youth and Sport of the Czech Republic, research infrastructures SAFMAT LM2015088 and LO1409. References [1] J. Musil, Recent advances in magnetron sputtering technology, Surf. Coat. Technol. 100–101 (1998) 280–286. [2] G. Bräuer, B. Szyszka, M. Vergöhl, R. Bandorf, Magnetron sputtering – milestones of 30 years, Vacuum 84 (2010) 1354–1359. [3] A. Anders, Deposition rates of high power impulse magnetron sputtering: Physics and economics, J. Vac. Sci. Technol. A 28 (4) (2010) 783–790. [4] Y. Jin, W. Wu, L. Li, J. Chen, J. Zhang, Y. Zuo, J. Fu, Effect of sputtering power on surface topography of dc magnetron sputtered Ti thin films observed by AFM, Appl. Surf. Sci. 255 (2009) 4673–4679. [5] A.A. Solovyev, V.O. Oskirko, V.A. Semenov, K.V. Oskomov, S.V. Rabotkin, Comparative study of Cu films prepared by DC, high-power pulsed and burst magnetron sputtering, J. Electron. Mater. 45 (8) (2016) 4052–4060. [6] M. Laurikaitis, J. Čyvienė, J. Dudonis, Deposition of Zr–ZrOX and Y–YXOY films by reactive magnetron sputtering, Vacuum 78 (2005) 395–399. [7] J. Vlcek, B. Zustin, J. Rezek, K. Burcalova, J. Tesar, Pulsed Magnetron Sputtering of Metallic Films Using a Hot Target, 52nd Annual Technical Conference Proceedings, Santa Clara, CA, May 9–14 2009, pp. 219–223. [8] D. Mercs, F. Ferry, A. Billard, Hot target sputtering: a new way for high-rate deposition of stoichiometric ceramic films, Surf. Coat. Technol. 201 (2006) 2276–2281.

[9] G.A. Bleykher, V.P. Krivobokov, A.V. Yurjeva, I. Sadykova, Energy and substance transfer in magnetron sputtering systems with liquid-phase target, Vacuum 124 (2016) 11–17. [10] G.A. Bleykher, A.O. Borduleva, V.P. Krivobokov, D.V. Sidelev, Evaporation factor in productivity increase of hot target magnetron sputtering systems, Vacuum 132 (2016) 62–69. [11] D.V. Sidelev, G.A. Bleykher, V.P. Krivobokov, Z. Koishybayeva, High-rate magnetron sputtering with hot target, Surf. Coat. Technol. 308 (2016) 168–173. [12] G.A. Bleykher, V.P. Krivobokov, The Erosion of the Surface of a Solid under Influence of Powerful Beams of Charged Particles, Nauka, Novosibirsk, 2014. [13] J.A. Thornton, Influence of apparatus geometry and deposition conditions on the structure and topography of thick sputtered coating, J. Vac. Sci. Technol. A 11 (4) (1974) 666–670. [14] S. Mahieu, P. Ghekiere, D. Depla, R. De Gryse, Biaxial alignment in sputter deposited thin films, Thin Solid Films 515 (4) (2006) 1229–1249. [15] I. Petrov, F. Adibi, J.E. Greene, L. Hultman, J.-E. Sundgren, Average energy deposited per atom: a universal parameter for describing ion-assisted film growth, Appl. Phys. Lett. 63 (1) (1993) 36–38. [16] S.D. Ekpe, S.K. Dew, Theoretical and experimental determination of the energy flux during magnetron sputter deposition onto an unbiased substrate, J. Vac. Sci. Technol. A 21 (2) (2003) 476–483. [17] T.P. Drusedau, T. Bock, T.-M. John, W. Eckstein, Energy transfer into the growing film during sputter deposition: an investigation by calorimetric measurements and Monte Carlo simulations, J. Vac. Sci. Technol. A 17 (5) (1999) 2896–2905. [18] R.D. Arnell, P.J. Kelly, Recent advances in magnetron sputtering, Surf. Coat. Technol. 112 (1999) 170–176. [19] G. Abadias, Y.Y. Tse, Ph. Guérin, V. Pelosin, Interdependence between stress, preferred orientation, and surface morphology of nanocrystallineTiN thin films deposited by dual ion beam sputtering, J. Appl. Phys. 99 (2006) 113519-1–113519-13. [20] I. Petrov, P.B. Barna, L. Hultman, J.E. Greene, Microstructural evolution during film growth, J. Vac. Sci. Technol. A 21 (5) (2003) 117–128. [21] S. Mahieu, D. Depla, R. De Gryse, Characterization of the hardness and the substrate fluxes during reactive magnetron sputtering of TiN, Surf. Coat. Technol. 202 (2008) 2314–2318. [22] G.A. Bleykher, V.P. Krivobokov, A.V. Yuryeva, Magnetron deposition of coatings with evaporation of the target, Tech. Phys. 12 (2015) 1790–1795. [23] P. Sigmund, Sputtering by ion bombardment Theoretical Concepts, in: R. Behrisch (Ed.), Sputtering by Particle Bombardment I: Physical Sputtering of Single-Element Solids, Topics in Applied Physics, Vol. V. 47, Springer Verlag, Berlin, Heidelberg, New York 1981, pp. 9–71. [24] R. Glang, Vacuum evaporation, in: L.I. Maissel, R. Glang (Eds.), Handbook of Thin Film Technology, V. 1, McGraw-Hill Book Company, 1970. [25] W.O. Hofer, Angular, energy and mass distribution of sputtered particles, in: R. Behrisch, K. Wittmaack (Eds.), Sputtering by Particle Bombardment III: Characteristics of Sputtered Particles, Technical Applications, Topics in Applied Physics, 64, Springer Verlag, Berlin, Heidelberg 1991, pp. 15–90. [26] H.M. Rietveld, A profile refinement method for nuclear and magnetic structures, J. Appl. Crystallogr. 2 (1969) 65–71. [27] R.W. Cheary, A.A. Coelho, J.P. Cline, Fundamental parameters line profile fitting in laboratory diffractometers, J. Res. Natl. Inst. Stand. Technol. 109 (2004) 1–25. [28] C.R.M. Grovenor, H.T.G. Hentzell, D.A. Smith, The development of grain structure during growth of metallic films, Acta Metall. 32 (5) (1984) 773–781. [29] J. Musil, Hard nanocomposite coatings: thermal stability, oxidation resistance and toughness, Surf. Coat. Technol. 207 (2012) 50–65. [30] S. Firstov, V. Kulikovsky, T. Rogul, R. Ctvrtlik, Effect of small concentrations of oxygen and nitrogen on the structure and mechanical properties of sputtered titanium films, Surf. Coat. Technol. 206 (2012) 3580–3585. [31] P.J. Kelly, R.D. Arnell, The influence of substrate temperature on the properties of Al, Zr and W coatings deposited by closed-field unbalanced magnetron sputtering, Vacuum 49 (1998) 43–47.