- Email: [email protected]
A comparative study on the properties of chromium coatings deposited by magnetron sputtering with hot and cooled target
Vacuum 143 (2017) 479e485
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
A comparative study on the properties of chromium coatings deposited by magnetron sputtering with hot and cooled target D.V. Sidelev a, *, G.A. Bleykher a, M. Bestetti a, b, **, V.P. Krivobokov a, A. Vicenzo b, S. Franz b, M.F. Brunella b a b
Tomsk Polytechnic University, 30, Lenin Avenue, Tomsk, 634050, Russian Federation Politecnico di Milano, Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Milano Italy
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 December 2016 Received in revised form 11 March 2017 Accepted 14 March 2017 Available online 18 March 2017
This article reports on the analysis of energy flux density to the substrate and results of the detailed study of properties of Cr coatings deposited by magnetron sputtering with hot and cooled Cr targets. It was demonstrated that deposition rates of Cr films increase and particle flow density on the substrate changes by target sublimation. Heat radiation of the hot target leads to more rapid heating of the substrate and this energy flux has a main contribution in enhancement of total energy flux density on the substrate (from 0.06 to 0.43 W/cm2). As consequence, energy per deposited atom increases in 5.8 times for the hot Cr target sputtering even taking into account higher deposition rates. Cr films deposited by the cooled target sputtering have a (110) crystal texture, columnar microstructure, low surface roughness (~2.66 nm) and hardness from 13.8 to 14.2 GPa. In the case of the hot target sputtering, the competitive crystal growth of (110) and (200) directions is observed, microstructure of the Cr films is denser and homogeneous, grain size increases up to 200e300 nm and film surface becomes coarse (Ra ~11.75 nm), hardness of the Cr films drops by a factor of about 2. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Hot target sputtering Energy balance Chromium film High deposition rate Film properties
1. Introduction Magnetron sputtering systems are extensively used techniques for conventional deposition of nanostructured coatings due to high quantity of adjusting work parameters (pressure, discharge power, ion current density on substrate, etc.). Nowadays, the extra interest in magnetron sputtering is directed to extend the range of operation parameters, which are required to meet new challenges in coatings deposition. One of these directions is the increase of deposition rates without degradation of functional properties of deposited films. Hot target magnetron sputtering is a high-rate method of surface modification in comparison with conventional sputtering [1e4]. In the case of the hot target sputtering, the target can be heated up to elevated temperatures by the energy of bombarding ions, when the target has a particular heat insulation from a water-
* Corresponding author. ** Corresponding author. Tomsk Polytechnic University, 30, Lenin Avenue, Tomsk, 634050, Russian Federation. E-mail addresses: [email protected] (D.V. Sidelev), massimiliano.bestetti@ polimi.it (M. Bestetti). http://dx.doi.org/10.1016/j.vacuum.2017.03.020 0042-207X/© 2017 Elsevier Ltd. All rights reserved.
cooled magnetron body. In this way, the target material additionally sublimates coupled with sputtering. The combination of these erosion processes results in the increase of deposition rates by several times, or more significantly, dependent on material properties and target temperature. For instance, the hot Ti target sputtering revealed the enhancement of deposition rates by up to 1.9 times [1]. The mathematical simulation of the hot target sputtering showed the technological possibilities to increase deposition rates by 5 and 20 times for Ti and Cr film deposition, respectively [5]. Besides, this sputtering technique has another strengths, it is effective to reduce a hysteresis effect in a reactive sputtering processes, even in an oxygen atmosphere [2,3]. In Ref. [6], the hot target sputtering was used to improve adhesion of metal films to the substrate, but the main feature of the hot target sputtering is that the heated (hot) target is a source of additional heat flux and flow of sublimated particles onto the substrate [5,7,8]. This causes significant changes of energy flux and particles flow to substrate during a pulse period, which are critical to structural and functional properties of deposited coatings. In this study chromium films were deposited using cooled and hot target magnetron sputtering at typical power density of the conventional sputtering (20 … 40 W/cm2). The interest in the
480
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
chromium film deposition is based on the satisfactory sublimation properties of Cr [9] and high demand of metallic Cr coatings in decorative, mechanical and corrosion resistance applications [10,11]. The influence of the state of the sputtered target on the structure, morphology, topography and mechanical properties was investigated under the equal deposition conditions for both cases (hot and cooled target sputtering). Furthermore, the mathematical modelling was used to calculate the energy density and particle flow on the substrate over a wide range of deposition conditions. The analysis of these results by comparison with microstructure and mechanical properties of the Cr films reveals key factors of the hot target magnetron sputtering. 2. Experimental 2.1. Film deposition Cr films were deposited in an Ar atmosphere (at 0.2 Pa) by sputtering of cooled and hot Cr targets (99.95%, 90 mm in dia, 8 mm thick) using a pulsed DC power supply APEL-M Series (operation frequency - 15 kHz, duty cycle - 0.25) with increased capacitance up to 2 mF. The schematic diagram of the deposition system is shown in Fig. 1a. The base pressure was 103 Pa. The magnetron sputtering system has an indirect cooling of the Cr target by dint of a copper diaphragm (2 mm). In the case of the hot target sputtering, another construction of the Cr target with a radial cavity was used. The schematic diagram of the hot Cr target magnetron is presented in Fig. 1b. Additional description and construction details of the hot Cr target magnetron are reported in Ref. [4]. AISI 321 stainless steel plates and Si (100) wafers were used as substrates. The target to substrate distance was 80 mm. Before the Cr film deposition, the substrates were set to face a wide-aperture ion source and treated by Arþ for 20 min (2.5 kV, 50 mA). Then, the samples were rotated to the magnetron sputtering system and the Cr films were deposited at constant discharge power of 1.75 kW (target power density over pulse period Wav ¼ 27.5 W/cm2). Deposition conditions are presented in Table 1. In the case of the hot target sputtering, the Cr target was previously sputtered for 8 min to heat and to stabilize the target temperature which was determined by the attainment of the discharge current maximum. Three different positions of substrates were fixed on the substrate holder in one deposition process (Fig. 1c). The substrates were thermally and electrically insulated from the substrate holder by using alumina ceramic interlayers. The substrates were separately grounded from the vacuum chamber. The scheme to measure the substrate current during films deposition is shown in Fig. 1a. The data were recorded using a digital oscilloscope Tektronix TDS 2022B. The substrate temperature was measured by using a chromel-copel thermocouple, which is connected to the backside of the central substrate. 2.2. Calculation of energy balance on the substrate Previously, we used the mathematical model of heat and erosion processes for pulsed magnetron sputtering of a solid-state target taking into account sublimation and local evaporation. The previously obtained data of [5] is the basis of the current calculations. For the high-rate magnetron sputtering with a hot target, the energy density on the substrate (FSUM) comprises heat flux density by target radiation (FRAD), heat flux density of condensation of deposited particles (FCS - sputtered particles; FCE e sublimated particles) and energy density by kinetic energy of sputtered (FKS) and sublimated (FKE) particles deposited on the substrate:
Fig. 1. (a) Schematic diagram of the deposition system, (b) the magnetron sputtering system with the Cr hot target and (c) the arrangement of samples on the substrate holder.
FSUM ¼ FRAD þ FCS þ FCE þ FKS þ FKE :
(1)
At the stabilized deposition mode, FSUM is a superposition of continuous and pulsed components of the energy density on the substrate. The energy flux density due to target radiation, condensation and kinetic energy of sublimated particles is continuous as is the target temperature (Ttag) and these components do not change during a pulse period [5]. However, the temperature field of the target surface is non-uniform. So, FRAD on the elemental area (1 cm2) with (X,Y) coordinates was calculated based on the Lambert and Stefan-Boltzmann laws:
4 4 Z εp s SB Ttag xtag ; ytag Tsub dxtag dytag Frad ðX; YÞ ¼ 2 2 ; p 2 Stag L2 þ X xtag þ Y ytag L2
(2)
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
481
Table 1 Deposition conditions and film thickness. Sample
Target type
Pos.
Wav, W/cm2
U, V
Iav, A
t, min
d, mm
Cr-1-1 Cr-1-2 Cr-1-3 Cr-2-1 Cr-2-2 Cr-2-3
cooled
1 2 3 1 2 3
27.5
640
2.71
13
605
3.05
12
3.42 3.31 3.43 5.96 6.82 6.39
hot
± ± ± ± ± ±
0.02 0.05 0.06 0.29 0.10 0.08
Isub, mA/cm2
Tsub, K
2.8
375
8.4
612
Note: Pos. e substrate position; Wav e target power density over pulse period; Iav e averaged target current over pulse period; t e deposition time; d e film thickness; Isub e current density on the substrate averaged over pulse; Tsub e substrate temperature.
where L is the distance between target and substrate surfaces, εp is the emissivity of the target surface, sSB is Stefan-Boltzmann constant, Ttag (xtag, ytag) is the target temperature on the elemental area with (xtag, ytag) coordinates, Tsub is the substrate temperature, Stag is the target surface square. The heat flux density by condensation of sublimated particles on the substrate was calculated as:
FCE ðX; YÞ ¼ vdep;ev ðX; YÞ,US ¼
US L 2
p
∬
Stag
Vev xtag ; ytag dxtag dytag
2 ; 2 L2 þ X xtag þ ðY yc Þ2
(3)
where VEV(X,Y) is the sublimation rate of the target surface with (xtag, ytag) coordinates (atom/(m2∙s)), US is the binding energy of the target surface. The kinetic energy of sublimated particles was evaluated by average speed of Maxwell distribution:
FKE ðX; YÞ ¼
L2
p
Z Stag
2.3. Film characterization The film thickness was measured using a Fischerscope XRAY XAN (number of measurements - 5; time per measurement - 30 s; voltage 50 kV, anode current - 1000 mA, measurement distance - 0.06 mm). The crystal structure was investigated by the grazing incident X-ray diffraction (XRD) technique with Cu Ka radiation in Bragg-Brentano configuration with the scanning angle of 3 . The surface morphology and cross-section microstructure were determined using a Zeiss EVO 50 EP at high vacuum of 103 Pa (accelerating voltage 17.5 kV). The film topography was investigated by atomic force microscopy (AFM) in the contact mode by sampling an area of the desired size with a NT-MDT Solver SPM instrument (Solver PRO-M). The Ar content in the Cr films was analyzed by glow discharge optical emission spectrometry (Spectruma GDA750 analyzer). The indentation tests were performed using a Fischerscope HV100VP XY (load 100 mN; time - 10 s; number of measurements - 5).
. 4kB ,Ttag xtag ; ytag p ,Vev xtag ; ytag dxtag dytag ; 2 2 2 L2 þ X xtag þ Y ytag
where kB is the Boltzmann constant. Due to the impulse behavior of magnetron sputtering, the energy flux density by condensation and kinetic energy of sputtered particles also has a repetitively pulsed nature. Also, the deposition rate by sputtering is unevenly distributed along the target surface. Thus:
FCS ðX; Y; tÞ ¼ vdep;sput ðX; Y; tÞ,US ;
(5)
FKS ðX; Y; tÞ ¼ Vdep;sput ðX; Y; tÞUS =2;
(6)
0 Vsput xtag ; ytag ; t ti dxtag dytag ∬ vdep;sput ðX; Y; tÞ ¼ 2 2 ; p Stag 2 L2 þ X xtag þ Y ytag
(7)
L2
energy per single deposited atom (Ea).
where t’ is the transit time of particles from elemental area (xi,yi) of the target surface to the substrate. The kinetic energy of sputtered particles is taken into account in (6) as the most probable energy of the Thompson spectrum (0.5$US [12]). It should be noted that the additional energy fluxes (by target radiation, condensation and kinetic energy of sublimated particles) all have to be taken into the account for the calculations of the
(4)
3. Results and discussion 3.1. Energy balance on substrate The thickness measurements revealed that the average deposition rate is higher for the hot target sputtering (Table 1) due to sublimation of the heated target and this is confirmed by our previous experiments and calculations [4]. So, the energy flux density on the substrate averaged over pulse period (FSUM) significantly increases and the distribution of the particles flow density changes in time in comparison with the cooled Cr target sputtering (Table 2). It is should be noted that the Cr target during sputtering is heated (hot) and in a solid state for the given value of Wav and the present configuration of a partial thermal insulation of the Cr target from the AISI 321 magnetron body. The phase transition of the Cr target to a molten state occurs only at 145 W/cm2. The process of the hot target sputtering is described schematically in Fig. 2. Due to the pulse behavior of magnetron discharge, sputtered particles deposit by portions independent of the target temperature [13,14]. For the hot target sputtering, the film deposition also occurred between sputtering pulses by the presence of the sublimated particle flow onto the substrate, when the Cr target is heated (hot). This results in the enhancement of the deposition
482
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
Table 2 Particles and energy flux density to the substrate by cooled and hot target magnetron sputtering. Target type
cooled hot
v, nm/s exp./calc. 4.38/6.91 8.88/8.36
Q1, atom/(cm2$s) (calc.)
Q2, atom/(cm2$s) (calc.)
by sput.
by subl.
by sput.
by subl.
2.4$1017 2.5$1017
e 9.5$1015
e e
e 9.5$1015
FSUM, W/cm2 (calc.)
Ea, eV/atom (calc.)
Tsub, K exp./calc.
0.06 0.43
6.2 35.7
375/352 612/616
Note: v e deposition rate; FSUM e total energy density on the substrate averaged on pulse period; Q1 e particle flux density during deposition of sputtered portion (averaged over pulse-on); Q2 e particle flux density between deposition of sputtered portion (averaged over pulse-off); Ea e energy per deposited atom on the substrate.
Fig. 2. Schematic diagram of a steady-state deposition process by the hot target sputtering: tpulse e pulse-on time; T e pulse period.
rates for our experimental conditions but, the ratio of sublimated to sputtered particles is not high for the given Wa (Table 2).
Fig. 3a presents the total energy flux density (FSUM) on the substrate averaged over a pulse period for cooled and hot Cr target sputtering. Fig. 3b shows the decomposition of the energy flux density on the substrate averaged over a pulse period (according to eq. (1)) for the hot Cr target magnetron sputtering. It was revealed that heat radiation flux from the elevated temperature target is predominant up to 65 W/cm2. At Wa > 65 W/cm2, the energy by condensation of sublimated particles has a predominant role. Condensation of sputtered particles and kinetic energy of deposited particles has a minor impact in the total energy flux density on the substrate. Similar ratios of components of the energy flux density are observed, when one pulse period of the stabilized deposition mode is considered (Fig. 3c). The heat flux by target radiation is also dominant during pulse period. To evaluate the contribution of the deposited energy by plasma ions in FSUM, we additionally measured the substrate current density averaged over the pulse (Table 1). It was 2.8 and 8.4 mA/cm2 to cooled and hot target sputtering, respectively. The previous data of optical emission spectroscopy in the cases of cooled and hot Cr target sputtering showed a noticeable difference in the quantity of plasma ions [4]. The increase of the current density on the substrate with the rise of the target temperature was also observed for the Ti films deposition [6]. However, we did not take into account this energy in the calculations of the total energy flux density due to the
Fig. 3. (a) The total energy density on the substrate over the pulse period for cooled and hot target sputtering as a function of Wa. (b) The energy density on the substrate averaged over pulse period for the hot target magnetron sputtering as a function of Wa. (c) The energy density on the substrate during one pulse period for the hot target sputtering (FSUM e total energy density, FCE e by condensation of sublimated particles, FCS - by condensation of sputtered particles, FRAD e by target radiation, FKS e by kinetic energy of sputtered particles, FKE - by kinetic energy of sublimated particles).
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
483
the distributions of the particle flux density on the substrate over pulse period are different for cooled and hot target sputtering. 3.2. Microstructure
Fig. 4. XRD patterns of the Cr samples.
low accelerating voltage of ions to the substrate. According to previous Langmuir-probe measurements of the cooled Cr target sputtering, the plasma potential is equal to þ3 V and substrates were grounded (Usub ¼ 0). The high heat flux from the heated (hot) Cr target limits the measurements of plasma characteristics for the hot target sputtering. Therefore, aspects of the hot target sputtering must be taken into consideration for future analysis of the Cr films: -the heat target radiation leads to significant increase of the heat flux on the substrate [15]. In our experiments, FSUM is 0.06 and 0.43 W/cm2 for cooled and hot Cr target sputtering, respectively (Table 2); -heat target radiation and energy by condensation of sublimated particles and their kinetic energy are additional sources of the energy on the substrate in comparison with the cooled target sputtering. This resulted in the increase of the energy per single deposited atom (Ea) from 6.2 to 35.7 eV/atom; -the flow density of deposited particles on the substrate comprises sputtered atoms or ions and sublimation particles, but the second component is low under our experimental conditions (in ratio of 1e7); -the obvious target sublimation process starts at elevated temperatures and the sublimation particle flux is constant, while the temperature of the Cr target surface is stable. The pulse nature of sputtering process and low duty-cycle of the power source of the magnetron causes a pulse character of the sputtered particle flux on the substrate. So,
Fig. 4 shows the XRD survey of the Cr films that have a body center cubic (bcc) structure. In order to show the XRD lines in one graph, the intensities have been normalized to the maximum intensity of the XRD pattern. The Cr samples deposited by the cooled target sputtering have a strong Cr (200) peak, while Cr (110) and Cr (200) reflections are detected in the case of the hot Cr target sputtering. According to [16,17], the formation of Cr (110) texture is inherent to the deposition of thick Cr films on a room temperature substrate that is caused by the lowest surface energy of Cr (110) intensity in bcc Cr films. For the hot target sputtering, the enhancement of deposition rates leads to a decrease in time to reach equilibrium states of deposited particles in the film structure. Additionally to this, the presence of the continuous particle flow on the substrate influences this time and decreases it. In this case, the Cr (110) microstructure should be also predominant. However, due to the increased adatom mobility, the surface diffusion can occur more intensively at the higher substrate temperature and it results in the formation of the Cr films with the competitive growth of Cr (110) and Cr (200) crystal directions. So, the observed changes in the Cr film microstructure is an aggregated result of the changes of energy and particle flows on the substrate. Also, the shift of Cr (200) peak to lower values of 2q is observed for both cases (for the cooled target sputteringe64.28 ; for the hot target sputteringe64.41 ) relative to the value of the unstrained bcc Cr crystal (64.58 ). It is indicative on the formation of compressive stresses in the Cr films. The magnitude of these stresses increases with decrease of 2q position of X-ray reflection. Lower compressive stresses in the Cr films deposited by the hot target sputtering are caused by higher substrate temperature and increased adatom mobility. Fig. 5 presents the cross-sectional microstructure of the Cr films on Si substrates. The Cr samples deposited by the cooled target sputtering exhibited a tapered-shaped columnar grain structure. The columnar width is approximately 200e250 nm. Based on Thornton's structure model zone [18], the observed microstructure in a good correlation with zone I (Tsub/Tm ¼ 0.176, where Tm is the melting point). For the hot target sputtering, the Cr film microstructure is denser and more homogenous. In this case, the Tsub/Tm ratio is 0.287, where the similar microstructure to typical zone I should be expected. However, the SEM micrograph (Fig. 6) revealed another microstructure type, which is characteristic to zone T and is
Fig. 5. SEM images of cross-sectional microstructure of the Cr films: a - cooled target sputtering; b - hot target sputtering.
484
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
Fig. 6. Surface morphology of the Cr films deposited by (a) cooled and (b) hot target sputtering.
Fig. 7. AFM images of the Cr films: a - cooled target sputtering; b - hot target sputtering. The average surface roughness is presented.
observed between zone I and zone II at low Ar pressures [19]. 3.3. Surface morphology and topography Figs. 6 and 7 show SEM and AFM images that revealed the influence of the target temperature on surface morphology and topography of the Cr samples. In the case of the cooled target sputtering, the Cr film surface is more homogeneous and smoother. There are few surface spikes and substrate defects. Average surface roughness and peak-to-peak distance are equal to 2.66 nm and 147.74 nm, respectively. On the contrary, the Cr films deposited by the hot target sputtering have rougher surface with a high amount of spikes and irregularities. The hot target sputtering of the Cr films
is characterized by the presence of spherical micro-droplets with diameter of 1.5e3.5 mm and high center-to-center distance (10e13 mm). Nevertheless, average surface roughness and peak-topeak distance are close in a magnitude and equal to 11.75 nm and 248.18 nm, respectively. The grain sizes of the Cr films deposited by cooled and hot target sputtering strongly differ. It is evident that the increase of the substrate temperature results in significant growth of crystallites up to 200e300 nm. The grain size of the Cr films in the cooled target sputtering case is significant lower. Due to higher substrate temperature, the adatom mobility enhances and it results in the rise of the size of stable film nucleus. But, the observed changes in the film morphology are also caused
Fig. 8. GDOES profiles of Ar and Fe in the Cr samples deposited by (a) cooled and (b) hot target sputtering.
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485 Table 3 Indentation results. Sample
dmax, mm
Substrate Cr-1-1 Cr-1-2 Cr-1-3 Cr-2-1 Cr-2-2 Cr-2-3
0.873 0.431 0.423 0.428 0.542 0.532 0.559
± ± ± ± ± ± ±
0.013 0.013 0.008 0.009 0.004 0.017 0.007
H, GPa
E, GPa
3.5 ± 0.1 13.8 ± 0.9 14.2 ± 0.6 13.9 ± 0.6 8.3 ± 0.1 8.5 ± 0.5 7.7 ± 0.2
195 261 274 271 250 257 253
± ± ± ± ± ± ±
H/E ratio 15 4 8 5 6 8 6
e 0.053 0.052 0.051 0.033 0.033 0.030
485
- the crystal structure of the Cr films changes from (110) texture to a structure with competitive growth of (110) and (200), when the hot Cr target is sputtered; - grain sizes increased up to 200e300 nm and the average surface roughness changed from 2.66 to 11.75 nm caused by the higher substrate temperature and film thickness for the hot target sputtering; - a reduction of the Cr film hardness by a factor of about 2 and a decrease in the film toughness. It should be noted that such crucial changes are caused by a significant increase of the substrate temperature and grain growth.
by higher deposition rates that was discussed in Section 3.2. 3.4. Elemental composition and indentation Fig. 8 shows GDOES results of argon and iron content in the Cr films deposited by cooled and hot target sputtering. It is only a qualitative measure to determine the influence of the Cr target state on the Ar content in the Cr samples. There is no differences of the Ar content in the Cr samples that can be expected for the hot target sputtering case due to higher (3 times) current density on the substrate. The independence of the Ar content from the substrate temperature is caused by the high deposition rates of chromium. The indentation results of the Cr films deposited on AISI 321 substrates are presented in Table 3. The double decrease of the Cr films hardness (H) was observed in the case of the hot target sputtering. It is widely accepted that coatings have maximum hardness at the critical grain size and H significantly decreases with the increase of grain size (Hall-Petch effect [20]). According to the SEM and AFM results, the noticeable growth of crystal grains (up to 200e300 nm) is observed in the case of the hot target sputtering. Moreover, our previous study of the Cr films deposited by cooled and hot target sputtering showed that the drop of the film hardness was less, when mechanical properties of the Cr films with the equal thickness (2 mm) were compared [4]. Enhanced substrate temperature and increased film thickness influence on the hardness changes and are main effects of hot target sputtering on the mechanical properties of the deposited films. The analysis of the Cr film resistance to mechanical degradation, which is described by H/E ratio [21], showed that the Cr samples deposited by the cooled target sputtering have higher values (0.051e0.053) in comparison with those of the Cr films obtained by hot target sputtering (0.033). This is indicated by the lower toughness of the Cr films deposited by the hot target magnetron sputtering. 4. Conclusion This study has demonstrated differences of the energy flux density on the substrate and changes in microstructure, film morphology and functional properties of the Cr films deposited by cooled and hot target sputtering for equal power density for a pulse period (27.5 W/cm2). It was found that: - the additional target sublimation results in higher deposition rates of the Cr films and a change in the distribution of particle flow density on the substrate during the pulse period for the hot target sputtering; - there is a significant increase of the heat flux on the substrate averaged over the pulse period (0.06e0.43 W/cm2) and enhancement of the energy per deposited atom from 6.2 to 35.7 eV/atom, these effects primary determined by target heat radiation;
Acknowledgements This study was supported by Russian Science Foundation (project No 15-19-00026).
References [1] J. Vlcek, B. Zustin, J. Rezek, K. Burcalova, J. Tesar, Pulsed magnetron sputtering of metallic films using a hot target, in: 52nd Annual Technical Conference Proceedings of the Society of Vacuum Coaters, 2009, pp. 219e223. Santa Clara. [2] 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) 2276e2281. [3] M. Laurkaitis, J. Cyviene, J. Dudonis, Deposition of Zr-ZrOX and Y-YXOY films by reactive magnetron sputtering, Vacuum 78 (2005) 395e399. [4] D.V. Sidelev, G.A. Bleykher, V.P. Krivobokov, Z. Koishybayeva, High-rate magnetron sputtering with hot target, Surf. Coat. Technol. 308 (2016) 168e173. [5] 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) 62e69. [6] Y. Chao, J. Bailing, L. Zheng, F. Lin, H. Juan, Nanocrystalline titanium films deposited via thermal-emission-enhanced magnetron sputtering, Thin Solid Films 597 (2015) 117e124. [7] J. Tesar, J. Martan, J. Rezek, On surface temperatures during high power pulsed magnetron sputtering using a hot target, Surf. Coat. Technol. 206 (2011) 1155e1159. [8] 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) 11e17. [9] R.E. Honig, Vapor pressure data for the more common elements, RCA Rev. 18 (1957) 195e204. [10] J. Stockemer, R. Winand, P. Vanden Brande, Comparison of wear and corrosion behaviors of Cr and CrN sputtered coatings, Surf. Coat. Technol. 115 (1999) 230e233. [11] J. Lin, I. Dahan, Nanostructured chromium coatings with enhanced mechanical properties and corrosion resistance, Surf. Coat. Technol. 265 (2015) 154e159. [12] 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, vol. 64, Springer Verlag, Berlin, Heidelberg, 1991, pp. 15e90. [13] A. Kuzmichev, I. Goncharuk, Simulation of the sputtered atom transport during a pulse deposition process in single- and dual-magnetron systems, IEEE Trans. Plasma Sci. 31 (2003) 994e1000. [14] L.B. Jonsson, Т. Nyberg, I. Katardjiev, S. Berg, Frequency response in pulsed DC reactive sputtering process, Thin Solid Films 365 (2000) 43e48. [15] A. Caillard, M. El’Mokh, N. Semmar, R. Dussart, T. Lecas, A.-L. Thomann, Energy transferred from a hot nickel target during magnetron sputtering, IEEE Trans. Plasma Sci. 42 (2014) 2802e2803. [16] Y.C. Feng, D.E. Laughlin, D.N. Lambeth, Formation of crystallographic texture in rf sputter-deposited Cr thin films, J. Appl. Phys. 76 (11) (1994) 7311e7316. [17] S.L. Duan, J.O. Artman, B. Wong, D.E. Laughlin, Study of the growth characteristics of sputtered Cr thin films, J. Appl. Phys. 67 (1990) 4913e4915. [18] J.A. Thornton, Influence of apparatus geometry and deposition conditions on the structure and topography of thick sputtered coatings, J. Vac. Sci. Technol. 11 (1974) 666e670. [19] J.A. Thornton, Influence of substrate temperature and deposition rate on structure of thick sputtered Cu coatings, J. Vac. Sci. Technol. 12 (1975) 830e835. [20] T.G. Nieh, J. Wadsworth, Hall-Petch relation on nanocrystalline solids, Scripta Metallurgica Materiala 25 (1991) 955e958. [21] J. Musil, M. Jirout, Toughness of hard nanostructured ceramic thin films, Surf. Coat. Technol. 201 (2007) 5148e5152.
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
A comparative study on the properties of chromium coatings deposited by magnetron sputtering with hot and cooled target D.V. Sidelev a, *, G.A. Bleykher a, M. Bestetti a, b, **, V.P. Krivobokov a, A. Vicenzo b, S. Franz b, M.F. Brunella b a b
Tomsk Polytechnic University, 30, Lenin Avenue, Tomsk, 634050, Russian Federation Politecnico di Milano, Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Milano Italy
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 December 2016 Received in revised form 11 March 2017 Accepted 14 March 2017 Available online 18 March 2017
This article reports on the analysis of energy flux density to the substrate and results of the detailed study of properties of Cr coatings deposited by magnetron sputtering with hot and cooled Cr targets. It was demonstrated that deposition rates of Cr films increase and particle flow density on the substrate changes by target sublimation. Heat radiation of the hot target leads to more rapid heating of the substrate and this energy flux has a main contribution in enhancement of total energy flux density on the substrate (from 0.06 to 0.43 W/cm2). As consequence, energy per deposited atom increases in 5.8 times for the hot Cr target sputtering even taking into account higher deposition rates. Cr films deposited by the cooled target sputtering have a (110) crystal texture, columnar microstructure, low surface roughness (~2.66 nm) and hardness from 13.8 to 14.2 GPa. In the case of the hot target sputtering, the competitive crystal growth of (110) and (200) directions is observed, microstructure of the Cr films is denser and homogeneous, grain size increases up to 200e300 nm and film surface becomes coarse (Ra ~11.75 nm), hardness of the Cr films drops by a factor of about 2. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Hot target sputtering Energy balance Chromium film High deposition rate Film properties
1. Introduction Magnetron sputtering systems are extensively used techniques for conventional deposition of nanostructured coatings due to high quantity of adjusting work parameters (pressure, discharge power, ion current density on substrate, etc.). Nowadays, the extra interest in magnetron sputtering is directed to extend the range of operation parameters, which are required to meet new challenges in coatings deposition. One of these directions is the increase of deposition rates without degradation of functional properties of deposited films. Hot target magnetron sputtering is a high-rate method of surface modification in comparison with conventional sputtering [1e4]. In the case of the hot target sputtering, the target can be heated up to elevated temperatures by the energy of bombarding ions, when the target has a particular heat insulation from a water-
* Corresponding author. ** Corresponding author. Tomsk Polytechnic University, 30, Lenin Avenue, Tomsk, 634050, Russian Federation. E-mail addresses: [email protected] (D.V. Sidelev), massimiliano.bestetti@ polimi.it (M. Bestetti). http://dx.doi.org/10.1016/j.vacuum.2017.03.020 0042-207X/© 2017 Elsevier Ltd. All rights reserved.
cooled magnetron body. In this way, the target material additionally sublimates coupled with sputtering. The combination of these erosion processes results in the increase of deposition rates by several times, or more significantly, dependent on material properties and target temperature. For instance, the hot Ti target sputtering revealed the enhancement of deposition rates by up to 1.9 times [1]. The mathematical simulation of the hot target sputtering showed the technological possibilities to increase deposition rates by 5 and 20 times for Ti and Cr film deposition, respectively [5]. Besides, this sputtering technique has another strengths, it is effective to reduce a hysteresis effect in a reactive sputtering processes, even in an oxygen atmosphere [2,3]. In Ref. [6], the hot target sputtering was used to improve adhesion of metal films to the substrate, but the main feature of the hot target sputtering is that the heated (hot) target is a source of additional heat flux and flow of sublimated particles onto the substrate [5,7,8]. This causes significant changes of energy flux and particles flow to substrate during a pulse period, which are critical to structural and functional properties of deposited coatings. In this study chromium films were deposited using cooled and hot target magnetron sputtering at typical power density of the conventional sputtering (20 … 40 W/cm2). The interest in the
480
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
chromium film deposition is based on the satisfactory sublimation properties of Cr [9] and high demand of metallic Cr coatings in decorative, mechanical and corrosion resistance applications [10,11]. The influence of the state of the sputtered target on the structure, morphology, topography and mechanical properties was investigated under the equal deposition conditions for both cases (hot and cooled target sputtering). Furthermore, the mathematical modelling was used to calculate the energy density and particle flow on the substrate over a wide range of deposition conditions. The analysis of these results by comparison with microstructure and mechanical properties of the Cr films reveals key factors of the hot target magnetron sputtering. 2. Experimental 2.1. Film deposition Cr films were deposited in an Ar atmosphere (at 0.2 Pa) by sputtering of cooled and hot Cr targets (99.95%, 90 mm in dia, 8 mm thick) using a pulsed DC power supply APEL-M Series (operation frequency - 15 kHz, duty cycle - 0.25) with increased capacitance up to 2 mF. The schematic diagram of the deposition system is shown in Fig. 1a. The base pressure was 103 Pa. The magnetron sputtering system has an indirect cooling of the Cr target by dint of a copper diaphragm (2 mm). In the case of the hot target sputtering, another construction of the Cr target with a radial cavity was used. The schematic diagram of the hot Cr target magnetron is presented in Fig. 1b. Additional description and construction details of the hot Cr target magnetron are reported in Ref. [4]. AISI 321 stainless steel plates and Si (100) wafers were used as substrates. The target to substrate distance was 80 mm. Before the Cr film deposition, the substrates were set to face a wide-aperture ion source and treated by Arþ for 20 min (2.5 kV, 50 mA). Then, the samples were rotated to the magnetron sputtering system and the Cr films were deposited at constant discharge power of 1.75 kW (target power density over pulse period Wav ¼ 27.5 W/cm2). Deposition conditions are presented in Table 1. In the case of the hot target sputtering, the Cr target was previously sputtered for 8 min to heat and to stabilize the target temperature which was determined by the attainment of the discharge current maximum. Three different positions of substrates were fixed on the substrate holder in one deposition process (Fig. 1c). The substrates were thermally and electrically insulated from the substrate holder by using alumina ceramic interlayers. The substrates were separately grounded from the vacuum chamber. The scheme to measure the substrate current during films deposition is shown in Fig. 1a. The data were recorded using a digital oscilloscope Tektronix TDS 2022B. The substrate temperature was measured by using a chromel-copel thermocouple, which is connected to the backside of the central substrate. 2.2. Calculation of energy balance on the substrate Previously, we used the mathematical model of heat and erosion processes for pulsed magnetron sputtering of a solid-state target taking into account sublimation and local evaporation. The previously obtained data of [5] is the basis of the current calculations. For the high-rate magnetron sputtering with a hot target, the energy density on the substrate (FSUM) comprises heat flux density by target radiation (FRAD), heat flux density of condensation of deposited particles (FCS - sputtered particles; FCE e sublimated particles) and energy density by kinetic energy of sputtered (FKS) and sublimated (FKE) particles deposited on the substrate:
Fig. 1. (a) Schematic diagram of the deposition system, (b) the magnetron sputtering system with the Cr hot target and (c) the arrangement of samples on the substrate holder.
FSUM ¼ FRAD þ FCS þ FCE þ FKS þ FKE :
(1)
At the stabilized deposition mode, FSUM is a superposition of continuous and pulsed components of the energy density on the substrate. The energy flux density due to target radiation, condensation and kinetic energy of sublimated particles is continuous as is the target temperature (Ttag) and these components do not change during a pulse period [5]. However, the temperature field of the target surface is non-uniform. So, FRAD on the elemental area (1 cm2) with (X,Y) coordinates was calculated based on the Lambert and Stefan-Boltzmann laws:
4 4 Z εp s SB Ttag xtag ; ytag Tsub dxtag dytag Frad ðX; YÞ ¼ 2 2 ; p 2 Stag L2 þ X xtag þ Y ytag L2
(2)
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
481
Table 1 Deposition conditions and film thickness. Sample
Target type
Pos.
Wav, W/cm2
U, V
Iav, A
t, min
d, mm
Cr-1-1 Cr-1-2 Cr-1-3 Cr-2-1 Cr-2-2 Cr-2-3
cooled
1 2 3 1 2 3
27.5
640
2.71
13
605
3.05
12
3.42 3.31 3.43 5.96 6.82 6.39
hot
± ± ± ± ± ±
0.02 0.05 0.06 0.29 0.10 0.08
Isub, mA/cm2
Tsub, K
2.8
375
8.4
612
Note: Pos. e substrate position; Wav e target power density over pulse period; Iav e averaged target current over pulse period; t e deposition time; d e film thickness; Isub e current density on the substrate averaged over pulse; Tsub e substrate temperature.
where L is the distance between target and substrate surfaces, εp is the emissivity of the target surface, sSB is Stefan-Boltzmann constant, Ttag (xtag, ytag) is the target temperature on the elemental area with (xtag, ytag) coordinates, Tsub is the substrate temperature, Stag is the target surface square. The heat flux density by condensation of sublimated particles on the substrate was calculated as:
FCE ðX; YÞ ¼ vdep;ev ðX; YÞ,US ¼
US L 2
p
∬
Stag
Vev xtag ; ytag dxtag dytag
2 ; 2 L2 þ X xtag þ ðY yc Þ2
(3)
where VEV(X,Y) is the sublimation rate of the target surface with (xtag, ytag) coordinates (atom/(m2∙s)), US is the binding energy of the target surface. The kinetic energy of sublimated particles was evaluated by average speed of Maxwell distribution:
FKE ðX; YÞ ¼
L2
p
Z Stag
2.3. Film characterization The film thickness was measured using a Fischerscope XRAY XAN (number of measurements - 5; time per measurement - 30 s; voltage 50 kV, anode current - 1000 mA, measurement distance - 0.06 mm). The crystal structure was investigated by the grazing incident X-ray diffraction (XRD) technique with Cu Ka radiation in Bragg-Brentano configuration with the scanning angle of 3 . The surface morphology and cross-section microstructure were determined using a Zeiss EVO 50 EP at high vacuum of 103 Pa (accelerating voltage 17.5 kV). The film topography was investigated by atomic force microscopy (AFM) in the contact mode by sampling an area of the desired size with a NT-MDT Solver SPM instrument (Solver PRO-M). The Ar content in the Cr films was analyzed by glow discharge optical emission spectrometry (Spectruma GDA750 analyzer). The indentation tests were performed using a Fischerscope HV100VP XY (load 100 mN; time - 10 s; number of measurements - 5).
. 4kB ,Ttag xtag ; ytag p ,Vev xtag ; ytag dxtag dytag ; 2 2 2 L2 þ X xtag þ Y ytag
where kB is the Boltzmann constant. Due to the impulse behavior of magnetron sputtering, the energy flux density by condensation and kinetic energy of sputtered particles also has a repetitively pulsed nature. Also, the deposition rate by sputtering is unevenly distributed along the target surface. Thus:
FCS ðX; Y; tÞ ¼ vdep;sput ðX; Y; tÞ,US ;
(5)
FKS ðX; Y; tÞ ¼ Vdep;sput ðX; Y; tÞUS =2;
(6)
0 Vsput xtag ; ytag ; t ti dxtag dytag ∬ vdep;sput ðX; Y; tÞ ¼ 2 2 ; p Stag 2 L2 þ X xtag þ Y ytag
(7)
L2
energy per single deposited atom (Ea).
where t’ is the transit time of particles from elemental area (xi,yi) of the target surface to the substrate. The kinetic energy of sputtered particles is taken into account in (6) as the most probable energy of the Thompson spectrum (0.5$US [12]). It should be noted that the additional energy fluxes (by target radiation, condensation and kinetic energy of sublimated particles) all have to be taken into the account for the calculations of the
(4)
3. Results and discussion 3.1. Energy balance on substrate The thickness measurements revealed that the average deposition rate is higher for the hot target sputtering (Table 1) due to sublimation of the heated target and this is confirmed by our previous experiments and calculations [4]. So, the energy flux density on the substrate averaged over pulse period (FSUM) significantly increases and the distribution of the particles flow density changes in time in comparison with the cooled Cr target sputtering (Table 2). It is should be noted that the Cr target during sputtering is heated (hot) and in a solid state for the given value of Wav and the present configuration of a partial thermal insulation of the Cr target from the AISI 321 magnetron body. The phase transition of the Cr target to a molten state occurs only at 145 W/cm2. The process of the hot target sputtering is described schematically in Fig. 2. Due to the pulse behavior of magnetron discharge, sputtered particles deposit by portions independent of the target temperature [13,14]. For the hot target sputtering, the film deposition also occurred between sputtering pulses by the presence of the sublimated particle flow onto the substrate, when the Cr target is heated (hot). This results in the enhancement of the deposition
482
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
Table 2 Particles and energy flux density to the substrate by cooled and hot target magnetron sputtering. Target type
cooled hot
v, nm/s exp./calc. 4.38/6.91 8.88/8.36
Q1, atom/(cm2$s) (calc.)
Q2, atom/(cm2$s) (calc.)
by sput.
by subl.
by sput.
by subl.
2.4$1017 2.5$1017
e 9.5$1015
e e
e 9.5$1015
FSUM, W/cm2 (calc.)
Ea, eV/atom (calc.)
Tsub, K exp./calc.
0.06 0.43
6.2 35.7
375/352 612/616
Note: v e deposition rate; FSUM e total energy density on the substrate averaged on pulse period; Q1 e particle flux density during deposition of sputtered portion (averaged over pulse-on); Q2 e particle flux density between deposition of sputtered portion (averaged over pulse-off); Ea e energy per deposited atom on the substrate.
Fig. 2. Schematic diagram of a steady-state deposition process by the hot target sputtering: tpulse e pulse-on time; T e pulse period.
rates for our experimental conditions but, the ratio of sublimated to sputtered particles is not high for the given Wa (Table 2).
Fig. 3a presents the total energy flux density (FSUM) on the substrate averaged over a pulse period for cooled and hot Cr target sputtering. Fig. 3b shows the decomposition of the energy flux density on the substrate averaged over a pulse period (according to eq. (1)) for the hot Cr target magnetron sputtering. It was revealed that heat radiation flux from the elevated temperature target is predominant up to 65 W/cm2. At Wa > 65 W/cm2, the energy by condensation of sublimated particles has a predominant role. Condensation of sputtered particles and kinetic energy of deposited particles has a minor impact in the total energy flux density on the substrate. Similar ratios of components of the energy flux density are observed, when one pulse period of the stabilized deposition mode is considered (Fig. 3c). The heat flux by target radiation is also dominant during pulse period. To evaluate the contribution of the deposited energy by plasma ions in FSUM, we additionally measured the substrate current density averaged over the pulse (Table 1). It was 2.8 and 8.4 mA/cm2 to cooled and hot target sputtering, respectively. The previous data of optical emission spectroscopy in the cases of cooled and hot Cr target sputtering showed a noticeable difference in the quantity of plasma ions [4]. The increase of the current density on the substrate with the rise of the target temperature was also observed for the Ti films deposition [6]. However, we did not take into account this energy in the calculations of the total energy flux density due to the
Fig. 3. (a) The total energy density on the substrate over the pulse period for cooled and hot target sputtering as a function of Wa. (b) The energy density on the substrate averaged over pulse period for the hot target magnetron sputtering as a function of Wa. (c) The energy density on the substrate during one pulse period for the hot target sputtering (FSUM e total energy density, FCE e by condensation of sublimated particles, FCS - by condensation of sputtered particles, FRAD e by target radiation, FKS e by kinetic energy of sputtered particles, FKE - by kinetic energy of sublimated particles).
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
483
the distributions of the particle flux density on the substrate over pulse period are different for cooled and hot target sputtering. 3.2. Microstructure
Fig. 4. XRD patterns of the Cr samples.
low accelerating voltage of ions to the substrate. According to previous Langmuir-probe measurements of the cooled Cr target sputtering, the plasma potential is equal to þ3 V and substrates were grounded (Usub ¼ 0). The high heat flux from the heated (hot) Cr target limits the measurements of plasma characteristics for the hot target sputtering. Therefore, aspects of the hot target sputtering must be taken into consideration for future analysis of the Cr films: -the heat target radiation leads to significant increase of the heat flux on the substrate [15]. In our experiments, FSUM is 0.06 and 0.43 W/cm2 for cooled and hot Cr target sputtering, respectively (Table 2); -heat target radiation and energy by condensation of sublimated particles and their kinetic energy are additional sources of the energy on the substrate in comparison with the cooled target sputtering. This resulted in the increase of the energy per single deposited atom (Ea) from 6.2 to 35.7 eV/atom; -the flow density of deposited particles on the substrate comprises sputtered atoms or ions and sublimation particles, but the second component is low under our experimental conditions (in ratio of 1e7); -the obvious target sublimation process starts at elevated temperatures and the sublimation particle flux is constant, while the temperature of the Cr target surface is stable. The pulse nature of sputtering process and low duty-cycle of the power source of the magnetron causes a pulse character of the sputtered particle flux on the substrate. So,
Fig. 4 shows the XRD survey of the Cr films that have a body center cubic (bcc) structure. In order to show the XRD lines in one graph, the intensities have been normalized to the maximum intensity of the XRD pattern. The Cr samples deposited by the cooled target sputtering have a strong Cr (200) peak, while Cr (110) and Cr (200) reflections are detected in the case of the hot Cr target sputtering. According to [16,17], the formation of Cr (110) texture is inherent to the deposition of thick Cr films on a room temperature substrate that is caused by the lowest surface energy of Cr (110) intensity in bcc Cr films. For the hot target sputtering, the enhancement of deposition rates leads to a decrease in time to reach equilibrium states of deposited particles in the film structure. Additionally to this, the presence of the continuous particle flow on the substrate influences this time and decreases it. In this case, the Cr (110) microstructure should be also predominant. However, due to the increased adatom mobility, the surface diffusion can occur more intensively at the higher substrate temperature and it results in the formation of the Cr films with the competitive growth of Cr (110) and Cr (200) crystal directions. So, the observed changes in the Cr film microstructure is an aggregated result of the changes of energy and particle flows on the substrate. Also, the shift of Cr (200) peak to lower values of 2q is observed for both cases (for the cooled target sputteringe64.28 ; for the hot target sputteringe64.41 ) relative to the value of the unstrained bcc Cr crystal (64.58 ). It is indicative on the formation of compressive stresses in the Cr films. The magnitude of these stresses increases with decrease of 2q position of X-ray reflection. Lower compressive stresses in the Cr films deposited by the hot target sputtering are caused by higher substrate temperature and increased adatom mobility. Fig. 5 presents the cross-sectional microstructure of the Cr films on Si substrates. The Cr samples deposited by the cooled target sputtering exhibited a tapered-shaped columnar grain structure. The columnar width is approximately 200e250 nm. Based on Thornton's structure model zone [18], the observed microstructure in a good correlation with zone I (Tsub/Tm ¼ 0.176, where Tm is the melting point). For the hot target sputtering, the Cr film microstructure is denser and more homogenous. In this case, the Tsub/Tm ratio is 0.287, where the similar microstructure to typical zone I should be expected. However, the SEM micrograph (Fig. 6) revealed another microstructure type, which is characteristic to zone T and is
Fig. 5. SEM images of cross-sectional microstructure of the Cr films: a - cooled target sputtering; b - hot target sputtering.
484
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485
Fig. 6. Surface morphology of the Cr films deposited by (a) cooled and (b) hot target sputtering.
Fig. 7. AFM images of the Cr films: a - cooled target sputtering; b - hot target sputtering. The average surface roughness is presented.
observed between zone I and zone II at low Ar pressures [19]. 3.3. Surface morphology and topography Figs. 6 and 7 show SEM and AFM images that revealed the influence of the target temperature on surface morphology and topography of the Cr samples. In the case of the cooled target sputtering, the Cr film surface is more homogeneous and smoother. There are few surface spikes and substrate defects. Average surface roughness and peak-to-peak distance are equal to 2.66 nm and 147.74 nm, respectively. On the contrary, the Cr films deposited by the hot target sputtering have rougher surface with a high amount of spikes and irregularities. The hot target sputtering of the Cr films
is characterized by the presence of spherical micro-droplets with diameter of 1.5e3.5 mm and high center-to-center distance (10e13 mm). Nevertheless, average surface roughness and peak-topeak distance are close in a magnitude and equal to 11.75 nm and 248.18 nm, respectively. The grain sizes of the Cr films deposited by cooled and hot target sputtering strongly differ. It is evident that the increase of the substrate temperature results in significant growth of crystallites up to 200e300 nm. The grain size of the Cr films in the cooled target sputtering case is significant lower. Due to higher substrate temperature, the adatom mobility enhances and it results in the rise of the size of stable film nucleus. But, the observed changes in the film morphology are also caused
Fig. 8. GDOES profiles of Ar and Fe in the Cr samples deposited by (a) cooled and (b) hot target sputtering.
D.V. Sidelev et al. / Vacuum 143 (2017) 479e485 Table 3 Indentation results. Sample
dmax, mm
Substrate Cr-1-1 Cr-1-2 Cr-1-3 Cr-2-1 Cr-2-2 Cr-2-3
0.873 0.431 0.423 0.428 0.542 0.532 0.559
± ± ± ± ± ± ±
0.013 0.013 0.008 0.009 0.004 0.017 0.007
H, GPa
E, GPa
3.5 ± 0.1 13.8 ± 0.9 14.2 ± 0.6 13.9 ± 0.6 8.3 ± 0.1 8.5 ± 0.5 7.7 ± 0.2
195 261 274 271 250 257 253
± ± ± ± ± ± ±
H/E ratio 15 4 8 5 6 8 6
e 0.053 0.052 0.051 0.033 0.033 0.030
485
- the crystal structure of the Cr films changes from (110) texture to a structure with competitive growth of (110) and (200), when the hot Cr target is sputtered; - grain sizes increased up to 200e300 nm and the average surface roughness changed from 2.66 to 11.75 nm caused by the higher substrate temperature and film thickness for the hot target sputtering; - a reduction of the Cr film hardness by a factor of about 2 and a decrease in the film toughness. It should be noted that such crucial changes are caused by a significant increase of the substrate temperature and grain growth.
by higher deposition rates that was discussed in Section 3.2. 3.4. Elemental composition and indentation Fig. 8 shows GDOES results of argon and iron content in the Cr films deposited by cooled and hot target sputtering. It is only a qualitative measure to determine the influence of the Cr target state on the Ar content in the Cr samples. There is no differences of the Ar content in the Cr samples that can be expected for the hot target sputtering case due to higher (3 times) current density on the substrate. The independence of the Ar content from the substrate temperature is caused by the high deposition rates of chromium. The indentation results of the Cr films deposited on AISI 321 substrates are presented in Table 3. The double decrease of the Cr films hardness (H) was observed in the case of the hot target sputtering. It is widely accepted that coatings have maximum hardness at the critical grain size and H significantly decreases with the increase of grain size (Hall-Petch effect [20]). According to the SEM and AFM results, the noticeable growth of crystal grains (up to 200e300 nm) is observed in the case of the hot target sputtering. Moreover, our previous study of the Cr films deposited by cooled and hot target sputtering showed that the drop of the film hardness was less, when mechanical properties of the Cr films with the equal thickness (2 mm) were compared [4]. Enhanced substrate temperature and increased film thickness influence on the hardness changes and are main effects of hot target sputtering on the mechanical properties of the deposited films. The analysis of the Cr film resistance to mechanical degradation, which is described by H/E ratio [21], showed that the Cr samples deposited by the cooled target sputtering have higher values (0.051e0.053) in comparison with those of the Cr films obtained by hot target sputtering (0.033). This is indicated by the lower toughness of the Cr films deposited by the hot target magnetron sputtering. 4. Conclusion This study has demonstrated differences of the energy flux density on the substrate and changes in microstructure, film morphology and functional properties of the Cr films deposited by cooled and hot target sputtering for equal power density for a pulse period (27.5 W/cm2). It was found that: - the additional target sublimation results in higher deposition rates of the Cr films and a change in the distribution of particle flow density on the substrate during the pulse period for the hot target sputtering; - there is a significant increase of the heat flux on the substrate averaged over the pulse period (0.06e0.43 W/cm2) and enhancement of the energy per deposited atom from 6.2 to 35.7 eV/atom, these effects primary determined by target heat radiation;
Acknowledgements This study was supported by Russian Science Foundation (project No 15-19-00026).
References [1] J. Vlcek, B. Zustin, J. Rezek, K. Burcalova, J. Tesar, Pulsed magnetron sputtering of metallic films using a hot target, in: 52nd Annual Technical Conference Proceedings of the Society of Vacuum Coaters, 2009, pp. 219e223. Santa Clara. [2] 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) 2276e2281. [3] M. Laurkaitis, J. Cyviene, J. Dudonis, Deposition of Zr-ZrOX and Y-YXOY films by reactive magnetron sputtering, Vacuum 78 (2005) 395e399. [4] D.V. Sidelev, G.A. Bleykher, V.P. Krivobokov, Z. Koishybayeva, High-rate magnetron sputtering with hot target, Surf. Coat. Technol. 308 (2016) 168e173. [5] 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) 62e69. [6] Y. Chao, J. Bailing, L. Zheng, F. Lin, H. Juan, Nanocrystalline titanium films deposited via thermal-emission-enhanced magnetron sputtering, Thin Solid Films 597 (2015) 117e124. [7] J. Tesar, J. Martan, J. Rezek, On surface temperatures during high power pulsed magnetron sputtering using a hot target, Surf. Coat. Technol. 206 (2011) 1155e1159. [8] 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) 11e17. [9] R.E. Honig, Vapor pressure data for the more common elements, RCA Rev. 18 (1957) 195e204. [10] J. Stockemer, R. Winand, P. Vanden Brande, Comparison of wear and corrosion behaviors of Cr and CrN sputtered coatings, Surf. Coat. Technol. 115 (1999) 230e233. [11] J. Lin, I. Dahan, Nanostructured chromium coatings with enhanced mechanical properties and corrosion resistance, Surf. Coat. Technol. 265 (2015) 154e159. [12] 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, vol. 64, Springer Verlag, Berlin, Heidelberg, 1991, pp. 15e90. [13] A. Kuzmichev, I. Goncharuk, Simulation of the sputtered atom transport during a pulse deposition process in single- and dual-magnetron systems, IEEE Trans. Plasma Sci. 31 (2003) 994e1000. [14] L.B. Jonsson, Т. Nyberg, I. Katardjiev, S. Berg, Frequency response in pulsed DC reactive sputtering process, Thin Solid Films 365 (2000) 43e48. [15] A. Caillard, M. El’Mokh, N. Semmar, R. Dussart, T. Lecas, A.-L. Thomann, Energy transferred from a hot nickel target during magnetron sputtering, IEEE Trans. Plasma Sci. 42 (2014) 2802e2803. [16] Y.C. Feng, D.E. Laughlin, D.N. Lambeth, Formation of crystallographic texture in rf sputter-deposited Cr thin films, J. Appl. Phys. 76 (11) (1994) 7311e7316. [17] S.L. Duan, J.O. Artman, B. Wong, D.E. Laughlin, Study of the growth characteristics of sputtered Cr thin films, J. Appl. Phys. 67 (1990) 4913e4915. [18] J.A. Thornton, Influence of apparatus geometry and deposition conditions on the structure and topography of thick sputtered coatings, J. Vac. Sci. Technol. 11 (1974) 666e670. [19] J.A. Thornton, Influence of substrate temperature and deposition rate on structure of thick sputtered Cu coatings, J. Vac. Sci. Technol. 12 (1975) 830e835. [20] T.G. Nieh, J. Wadsworth, Hall-Petch relation on nanocrystalline solids, Scripta Metallurgica Materiala 25 (1991) 955e958. [21] J. Musil, M. Jirout, Toughness of hard nanostructured ceramic thin films, Surf. Coat. Technol. 201 (2007) 5148e5152.