In physical vapor deposition, the material is vaporized from a target and deposited on the areas of substratethat are in the line-of-sight of a vaporization source. To ensure uniform deposition, substrates are positionedon a turntable and rotated in a manner similar to the planetary rotation. In industrial deposition systems theturntable rotates around several axes andmoves substrates along different targets. The substrate rotation andthe target arrangement therefore determine the uniformity of the deposited material. When different targetmaterials are used coatings can be prepared in a layered structure; in such a case, the rotation and the targetarrangement also determine the layer structure of the coatings. In the present paper a computer simulationof coating growth in an industrial deposition system with a planetary type of rotation has been used to analyzethe influence of the rotation and the target arrangement on the uniformity and the periodicity of layered coat-ings. Results of simulations show that highly periodic modes of rotation, which are determined by the turntablegear ratio and the switch angle, cause large non-uniformities both in the thickness and the composition oflayered coatings. On the other hand, less periodic modes of rotation produce better coating uniformity althoughfor certain rotation parameters significant non-uniformities may also occur. Exact periodicity of layered coatingscan be calculated from the least common multiple of revolution times around individual axes. Calculations ofcoating thickness and composition on the perimeter of a round tool show that the uniformity also dependson the deposition time. Configuration with maximally separated targets produces better coating uniformitythan configuration with closely positioned targets.
Physical vapor deposition (PVD) techniques are widely usedmethods for the deposition of thin films. In the PVD, the material isvaporized from a target and deposited around the vacuum chamber.In contrast to chemical vapor deposition, the vapor flux in PVD is highlydirectional since the material originates from a spatially located area,i.e., from a target or more specifically from a small part of the target(e.g. racetrack or cathode spot). Due to directionality of the flux thematerial is deposited on those surfaces that are in the line-of-sight ofthe target. Line-of-sight deposition is disadvantageous for many appli-cationswhere a whole or most of the substrate with complex shape hasto be covered.
Several approaches are used to improve uniformity of the line-of-sight deposition. One way is to increase scattering of the vaporizedmaterial and thus reduce the directionality of the flux. This can beachieved by working at high pressures where mean free path of thevaporized species is significantly smaller than the target–substrate
rights reserved.
distance. Majority of PVD techniques operate at pressures wherethe mean free path of vaporized species is larger or comparable tothe target–substrate distance therefore only a small amount of materialreaches shaded areas of a substrate through scattering. High-pressurePVD techniques have been developed to improve uniformity of deposi-tion but such methods are rarely used in practice and acceptableuniformity can be realized only for specific deposition parametersand substrate geometries [1,2]. More importantly, high-pressure PVDtechniques are disadvantageous compared to low-pressure PVD tech-niques since the scattering reduces energy of the vaporized speciesand thus produces thin films with lower quality.
Uniformity of the deposited material can be also improved ifvaporized material is highly ionized and bias voltage is applied to thesubstrate. PVD techniques with high ionization degree include highpower impulse magnetron sputtering and cathodic arc [3]. In the high-ly ionized plasma, ions are influenced by electric and magnetic fields,thus, they can be guided to those areas that are not in the direct viewof the vaporization source. Yet, high ionization degree and biasing donot fully solve the problem of uniformity on large substrates. Ions areguided by electric fields that are mainly present close to cathode [4],biased surfaces and chamber walls (magnetic fields, in e.g. magnetron,are normally too weak to have significant influence on the path of
ions). Biased or unbiased surfaces immersed in plasma modify plasmapotential only in the close vicinity of the surface; the drop in the poten-tial occurs in the so called (pre)sheath area— typically few hundreds ofmicrometers to several millimeters large [5]. The effect of biasing onthe trajectory of ion is thus only felt when ion arrives in the closeproximity of the substrate. The advantage of highly ionized plasmaon the coating uniformity is evident when coating small features thatare in the line-of-sight of vaporization source, such as, trenches withhigh aspect ratio [6,7]. In cases where large surface areas haveto be coated and only part of the surface is in the view of the target(e.g. tools), biasing and the high degree of ionization alone will notenable satisfactory uniformity of the deposition.
The surface of a large substrate can be uniformly coated if severaltargets are positioned around the substrate. This approach is notpractical since it requires several target sources and is only usefulfor coating one substrate; coating several substrates would be lessuniform due to shading by other substrates. The only practical wayto improve the uniformity of the deposition in the line-of-sightprocesses is to rotate the substrates. Rotation ensures that substrates,which usually have complex geometry, are more or less uniformly ex-posed to the vaporized material. Practically all PVD systems employsome type of substrate rotation which is performed by a turntable.Typically turntables perform planetary type of rotation where turnta-ble rotates around its central axis, while the substrate holders andsubstrates rotate around their axes.
Industrial PVD systems can be classified into the designs wheresubstrates are always in the view of the vaporization sources anddesigns where they are not. In the first type of the design the targetsources are normally positioned on the bottom of the depositionchamber whereas the turntable is on the top of the chamber. In thisconfiguration, substrates are always in the view of the targets butonly one side of the substrate is coated. If the opposite side has tobe covered then substrates are turned around and deposition is re-peated. PVD systems with the top-bottom design are mainly utilizedfor the deposition of optical coatings (e.g. anti-reflective coatings onlenses) [8]. The second type of PVD systems employs a designwhere several vaporization sources (normally four) are positionedaround the vacuum chamber in vertical position while the turntableis in the center of the chamber [9,10]. Such design is used for coatingthe whole surface area of numerous substrates in a single depositionprocess. The rotation ensures that the surface gets uniformly exposedby all target sources; in this way, a relatively good coating uniformitycan be achieved. This type of PVD systems is mainly utilized for thedeposition of protective coatings on substrates with complex geome-try where small variations in coating uniformity can be tolerated(e.g. tools, machine components, consumer products, medical instru-ments). In this paper, the analysis of the deposition uniformity willaddress only this type of deposition systems.
If the rotation in the PVD system is not designed properly, it maycause substantial non-uniformities. In the top-bottom design wheresubstrates are always in the view of the vaporization sources,variations in the deposition rate are minor. On the contrary, in thedeposition systems where the turntable is positioned between thevaporization sources, variations in the deposition rate are large. Inthese systems the distance and the angle between the substrate and in-dividual source change significantly due to rotation. When the sub-strate travels toward a particular vaporization source, the depositionrate increases to the maximum and decreases when the substratemoves away from the source [11]. If the rotation is highly periodicand substrates return into the same position and orientation forevery rotational cycle of the turntable then a large coating non-uniformity will be produced.
The deposition rate variations are especially critical in the reactivedeposition mode because they can affect the composition of coatingsand, consequently, their properties. In the reactive deposition processthe partial pressure of the reactive gas is constant, while the flux of
the material on the surface of the substrate changes due to rotation.Since the ratio between the flux of the reactive gas and the vaporizedmaterial is not constant, variations in the stoichiometry can occur.Studies of nanolayered TiAlN/CrN coatings deposited in an industrialmagnetron sputtering system with planetary type of rotationrevealed that the rotation can produce large differences in the stoichi-ometry; in some cases, variations were so large that the growth ofhexagonal Cr2N phase was initiated within the cubic CrN phase [12].It was also shown that variations in the deposition flux caused bythe rotation have a strong influence on the microstructure andmechanical properties of coatings [13].
In the literature, the influence of the substrate rotation on thedeposition and properties of coatings has not been given much atten-tion; there are only a few studies on this topic [14–18]. Findings inRefs. [12,13] suggest that the rotation should be regarded as one ofthe deposition parameters since it can significantly influence proper-ties of coatings. Substrate rotation can have an important effectparticularly in the case of nanolayered coatings where the thicknessof individual layers determines the mechanical properties of coatings.As numerous studies have shown, a very high hardness is obtainedonly when the thickness of the individual layers is approximately2–10 nm [19–22]. Substrate rotation can cause significant variationsin the thickness of individual layers therefore required layer thicknesscannot be satisfied for all layers. The uniformity and the layerstructure of nanolayered coatings are also influenced by the targetarrangement. Targets can be arranged in different configurations,e.g., targets of the same type can be next or opposite to each other.In nanolayered coatings the sharpness of interfaces plays an impor-tant role on the mechanical properties [23] therefore targets shouldbe arranged in such a way that the intermixing between the materialsis minimized.
In the present paper, a previously developed computer simulationof coating growth in an industrial deposition system with planetarytype of substrate rotation has been utilized [11]. The computer simu-lation enables calculation of the deposition rate on the surface of arotating substrate and calculation of individual layer thicknesses thatcan be visually represented in the form of a layer structure. The accura-cy of the calculations was verified by comparing calculated layeredstructures with the deposited nanolayered coatings [24]. The aim ofthe present study is to perform computer simulation for different setsof parameters and give a detailed analysis of the substrate rotationand target arrangement on the periodicity and uniformity of layeredcoatings.
2. Calculation procedure
2.1. Geometry of the deposition system and simulation parameters
Simulations were performed for the magnetron sputtering systemCemeCon CC800/9 SinOx ML. A schematic diagram of the depositionsystems is shown in Fig. 1. The system has four unbalanced planarmagnetron sources with the turntable situated in the center of the de-position chamber. The turntable has six substrate towers and abilityof 3-fold planetary type of rotation. Geometry of the CC800/9 systemis summarized in Table 1. Rotation around the third axis is non-continuous and is realized by a switch (a metal strip). The switchturns the substrate holder for an angle α for every rotation of thetower around its axis. The revolution time of the turntable can beadjustable from 38 to 97 s, while the revolution time of the substratetower is determined by the gear ratio (g) between the turntable andthe substrate tower. The gear ratio for the CC800/9 system is 100:37 —
the turntable gear has 100 teeth while the substrate tower gear has37 teeth.
Calculations of a layer structure are based on previously developedmodel of coating growth [11]. The model considers the geometry of thedeposition system, rotation of substrates and particle flux from
1st rotation
2nd rotation
switch 3rd rotation
substrate
target 1
target 2target 3
target 4
turntable
substratetower
163 mm
472 mm
ϕ1
ϕ2
Fig. 1. Schematic diagram of the CemeCon CC800/9 SinOx ML deposition system.
individual targets. The deposition rate on a surface of the substrate is afunction of time, which is defined by the rotation. It depends on thedistance of a substrate to the target, the orientation between the sub-strate and the target, and angular distribution of the particle fluxfrom the target. The thickness of individual layers is calculated by inte-grating the deposition rate with respect to the time, which is thengraphically represented as a layer structure.
In the simulations, an angular distribution of particles from thetarget was considered to be cosine, while the flux intensity for allsources was set at the same value (i.e., the same power on all targets).Unless noted differently, in all simulations, the substratewas positionedand oriented toward targets 1 and 2 and the time of the turntable rota-tion was set to 60 s (i.e. 1 rpm). In all the cases, except in Section 3.4,targets 1 and 2 were taken as one type of material and targets 3 and 4as another type ofmaterial. The rotational parameters used in particularsimulations are noted in the text.
2.2. Example of simulation
An example of a simulation is illustrated in Fig. 2 showing resultsfor the 3-fold rotation. Normalized deposition rate as a function oftime is presented in Fig. 2(a). Calculation was made for arrangementof targets shown in Fig. 1 and for two rotational cycles of the turntable.The deposition rate strongly depends on the rotation of the substrate.In each rotational cycle, the material is deposited from all targets butthe amount of material from a particular target differs considerably.For example, in the first rotational cycle, the deposition rate from thetarget 2 is much higher than the deposition rate from other targets.On the other hand, in the second rotational cycle, the deposition ratefrom the target 2 is low, while the deposition rate from the target 3is high.
Table 1Geometry of CemeCon CC800/9 ML deposition system.
Parameter CC800/9 ML
Distance between the first and the second axis [mm] 137Distance between the second and the third axis [mm] 40–58Dimension of the targets (w × h) [mm2] 88 × 500Distance between targets and the center of the turntable [mm] 250Distance between the targets 1 and 2 [mm] 163Distance between the targets 1 and 3 [mm] 472Gear ratio between the turntable and towers 100/37Revolution time of the turntable [s] 38–97
The deposition rate as function of time can be better understood byfollowing the trajectory of the substrate. Fig. 2(b) depicts calculatedtrajectories for the first and the second rotational cycle. Black arrowsindicate the position and the orientation of the substrate in the first ro-tational cycle and gray ones in the second cycle. In the first rotationalcycle, the substrate travels close to the target 2 while facing its direc-tion, therefore, the deposition rate is high. In the second rotationalcycle, the substrate also travels close to the target but is facing awayfrom the target, therefore, the deposition rate is low. Note that afterthe first rotational cycle the substrate does not return in the sameinitial position and orientation. High deposition rate occurs when thesubstrate is close to the target and facing its direction, whereas thedeposition rate is low when the substrate is far from the target and/or faces the target under a high angle. In positions where the substrateis not in the direct view of the target the deposition rate is zero. Theanimation of the substrate rotation and corresponding deposition ratecan be viewed on-line (see Animation 1).
The thickness of the individual layers and the total thickness of thecoating can be calculated by integrating the deposition rate withrespect to the time. The calculated layer structure is presented inthe horizontal direction above Fig. 2(a). It can be seen that the indi-vidual layers have considerably different thicknesses. The first layerproduced by the target 2 is much thicker than the second layer pro-duced by the same target. In contrast, the first layer produced bythe target 3 is much thinner than the second layer produced by thesame target, whereas the first and the second layer from the target4 have approximately equal thickness. Such simulations wereperformed for different rotation parameters and will be discussed inthe following sections.
3. Results and discussion
3.1. The 1-, 2- and 3-fold rotations
The size of substrate (e.g. tool) determines the type of rotationused in a particular deposition process. The 1-fold rotation, i.e., rota-tion of a turntable around its axis, is used for coating large tools(e.g. large molds), while 2- and 3-fold rotations are used for coatingsmaller tools. Trajectory of the substrates is defined by the revolutiontimes around the individual axes and by the distance between theaxes — these are determined by the design of the system and cannotbe varied independently. The revolution time around the first and thesecond axis is defined by the gear ratio between the turntable and thesubstrate tower, thus, the coupling between these two gears deter-mines the revolution time around the second axis. The rotationaround the third axis, in most PVD systems, is realized by a switch;normally a steel strip fixed on the rod that turns the substrate for acertain angle for every rotation of the substrate tower around itsaxis. Switch angle is difficult to control because it depends on thefixture of the switch, as well as on the size and the weight of thetool. The only parameter, which can be adjusted, is the speed of theturntable. Despite practical limits of the rotation parameters the analy-sis will be performed for different revolution times, more specifically,for different turntable-to-tower gear ratios and switch angles.
First, a comparison of layer structures produced by different typesof rotations will be performed. The calculated deposition rate andlayer structures for 1-, 2- and 3-fold rotations are shown in Fig. 3.Calculations were made for the target arrangement with two types oftarget material on one side of the deposition chamber, and the othertype of material on the other side. The material from the targets 1and 2 (see Fig. 1) is represented by darker gray color (i.e. material A),while material from targets 3 and 4 is represented by brighter graycolor (i.e. material B).
Layer structures produced by 1-, 2-, and 3-fold rotations differconsiderably between each other. In the simple case of 1-fold rota-tion, a periodic layer structure is formed. Calculated deposition rate
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120
v/vmax
t [s]
target 1target 2target 3target 4
1
23
4
(a) (b)
first cycle second cycle
initialposition
switchedorientation
layer structure
Fig. 2. (a) Normalized deposition rate in dependence of time and corresponding layered structure for the 3-fold rotation. (b) Trajectory of the substrate for two rotational cycles.Larger (blue) arrows indicate positions where the switch turns the substrate for 165°. The animation of the substrate rotation and corresponding deposition rate can be viewedon-line (Animation 1).
(Fig. 3(a)) displays that the substrate is periodically exposed to thetargets, thus, the individual layers are equally thick. The 1-fold rota-tion is trivial as the substrate travels on the identical trajectory for
100 200 300 400 500 600
0.2
0.4
0.6
0.8
1
100 200 300 400 500 600
1-fold
2-fold
100 200 300 400 500 600t [s]
material Amaterial B
3-fold
200
400
600
200
400
200
400
600
800
1000
1200
1400
1600
v/ vmax
v/ vmax
v/ vmax d[nm ]
d[nm
d[nm
]
(a)
(b)
(c)
800
0
0
0
t [s]
t [s]
0.2
0.4
0.6
0.8
1
0.2
0.4
0.6
0.8
1
Fig. 3. Deposition rate and layer structure for (a) 1-, (b) 2-, and (c) 3-fold rotations.Calculations are made for 10 rotational cycles of the turntable with a revolution timeof 60 s.
each rotation of the turntable. On the other hand, the depositionrates and the layer structures for 2-fold and 3-fold rotations aremore complicated and will be discussed later.
The thickness of individual layers and the total thickness of thecoatings differ between the three types of the rotation. For the selectedcalculation parameters, the thickness of individual layers in 1-fold rota-tion is 90 nm, while in the 2- and 3-fold rotations the thickness variesbetween 35–40 nm and 7–32 nm, respectively. As a consequence, thetotal thicknesses of coatings are: ~1800 nm for 1-fold rotation,~700 nm for 2-fold rotation and ~400 nm for 3-fold rotation, and thecorresponding average deposition rates are ~3 nm/s, ~1.2 nm/s, and~0.7 nm/s. The rotation around the additional axis considerably lowersthe average deposition rate. In this case, the average deposition rate forthe 2-fold rotation is ~40% of the deposition rate for the 1-fold rotation,while for the 3-fold rotation the deposition rate is only ~20% of thedeposition rate for the 1-fold rotation. Note, that these values are thesame for different rotational speeds of the turntable and for differentinitial positions of substrates.
3.2. The 2-fold rotation
3.2.1. The periodicity of layer structuresIn the 2-fold rotation substrates rotate around the axis of the turn-
table and the axis of the substrate tower. In the industrial batchsystems, it is used for tools with the size comparable to the diameterof the substrate tower. The periodicity of layer structures producedby 2-fold rotation is determined by the gear ratio between the turnta-ble and the substrate tower. If the gear ratio is an integer number(e.g. 90:30 — turntable with 90 gears and the substrate tower with30 gears) then the substrate will return into the same position forevery rotation of the turntable. If the gear ratio is a non-integer number(e.g. 100:37 ~ 2.7) then the turntable will rotate several times beforethe substrate will return into the same position. In both cases thelayer structure is periodic, only the periodicity is different. Layerstructures produced by the turntable with an integer gear ratio havea period of one rotational cycle, whereas layer structures produced bynon-integer gear ratios have longer periods. In the literature, theterm “bilayer period” is used for the thickness of two deposited layers.This term is not adequate for the description of periodicity becauselayer structures can have longer periods. For this reason, a more gener-ic term “modulation period” will be used in the text, which denotesthickness of all layers in one period of the layer structure.
Fig. 4 shows a calculated deposition rate for different turntable-to-tower gear ratios. For the integer gear ratios (g = 60:30 andg = 90:30), the deposition rate changes periodically for every rota-tional cycle of the turntable, thus, these two layer structures have a
25 50 75 100 125 150 175
25 50 75 100 125 150 175 25 50 75 100 125 150 175
25 50 75 100 125 150 175
0.2
0.4
0.6
0.8
1
0.2
0.4
0.6
0.8
1
0.2
0.4
0.6
0.8
1
0.2
0.4
0.6
0.8
1
60:30 70:30
100:37 90:30
(a) (b)
(c) (d)
t [s] t [s]
t [s] t [s]
d [nm] d [nm]
d [nm] d [nm]
v/vmax
v/vmax
v/vmax
v/vmax
100
50
150
200
0
250
100
50
150
200
0
250
100
50
150
200
0
250
100
50
150
200
0
250
Fig. 4. Deposition rates and layer structures calculated for the 2-fold rotation and gear ratios: (a) 60:30, (b) 70:30, (c) 100:37 and (d) 90:30. The animation of the substrate rotationfor 70:30 gear ratio and corresponding deposition rate can be viewed on-line (Animation 2).
Table 2Number of the turntable rotations required for substrate to return into the same positionin the 2-fold rotation. Calculations are made for different numbers of turntable and sub-strate tower gears.
Number of turntable gears Number of substrate tower gears
periodicity of two layers (i.e. a bilayer period). In the case of a 70:30non-integer gear ratio, the turntable has to make three rotationsaround its axis before the substrate returns into the same position(see on-line Animation 2). This layer structure has periodicity of 6layers (2 × 3 = 6; in every rotational cycle 2 layers are deposited).In the case of g = 100:37 (Fig. 4(c)), the deposition rate is differentfor all the three calculated rotational cycles. The turntable has tomake 37 rotational cycles before the substrate returns into identicalposition — the layer structure has periodicity of 74 layers(2 × 37 = 74).
The periodicity of the layer structure, i.e., the number of layersthat repeat in the layer structure, is determined by the substrate rota-tion, whereas the substrate rotation is defined by the number of gearson the turntable and on the substrate tower. The exact periodicity ofthe layer structure can be found by calculating the least commonmultiple of revolution times around the first and the second axis.For example, if the gear ratio is 70:30 and revolution time aroundthe first axis is 60 s, then the rotation time around the second axisis (60 s) × 30/70 ~ 25.7 s. The least common multiple of 60 s and60 × 30/70 s is 180 s. This means that the turntable has to makethree rotations (180 s/60 s = 3) before the substrate returns intothe same position. Such calculations were made for different gearratios and are presented in Table 2. Turntables with integer gearratio (e.g. 60:30, 70:35, and 80:40) produce rotation where trajectoryrepeats in every rotational cycle; hence, the layer structures have a bi-layer period although the two constituting layers do not necessaryhave equal thickness. If the gear ratio is not an integer number,more rotations of the turntable are needed for the substrate to returninto the same position. The layer structures produced by the turnta-bles with non-integer gear ratio have longer modulation periods.The number of layers that repeat in the layer structure, i.e. periodicity
of layer structure, is not affected by the initial position of the sub-strate. The periodicity of layer structures is solely determined by thegear ratio of the turntable, whereas initial position affects the thick-ness of the individual layers.
3.2.2. The uniformity of coatingsIn industrial PVD systems numerous substrates are coated simulta-
neously. Substrates are fixed on various positions around the turntableand travel on different trajectories. Uniformity of the coating dependson the repeatability of the trajectory. A better uniformity can beexpected if the substrate travels on several different trajectories beforeit returns into the same position. Turntables with non-integer gearratio produce rotation where substrate travels on many different tra-jectories. This means that the substrate is more or less equally exposedto the vaporized material, thus, the total thickness and the composition
of the coating do not vary considerably around the perimeter of thetool. Still, not all non-integer gear ratios are appropriate for the sub-strate rotation; for example, g = 75:30 or 80:32 are non-integer gearratios but do not provide enough uniformity because the substratereturns into the same position after two rotations of the turntable(see Table 2). To achieve good uniformity, such gear ratio shouldbe chosen that the turntable makes several different trajectories(e.g. more than five) before the substrate returns into the same posi-tion. The uniformity of coating will be discussed more in detail for3-fold rotation.
3.3. The 3-fold rotation
A higher degree of rotation ensures better exposure of a substrateto the vaporized flux from the targets. In the industrial production,the 3-fold planetary rotation is the most commonly utilized, whereashigher degrees of rotation (e.g. 4-fold) are less common. Rotationaround three axes is applied only to small tools (e.g. drills, mills,and cutting inserts). The 3-fold rotation can be regarded as a superpo-sition of the 2-fold rotation and the rotation around the third axis. Therotation around the third axis, in majority of PVD systems, is non-continuous due to the switch, which turns substrate for specificangle after each rotation around substrate tower. In the following,we will analyze the influence of the turntable-to-tower gear ratioand the switch angle on the periodicity and the uniformity of layeredcoatings.
3.3.1. The periodicity of layer structures
3.3.1.1. Different gear ratios and selected switch angle. Fig. 5 shows thelayer structures calculated for different turntable-to-tower gear ratiosand for the switch angle of 120°. Calculations were made for 10 rota-tions of the turntable. If we compare these layer structures with theones produced by the 2-fold rotation (cf. Figs. 4 and 5), it can beseen that rotation around the third axis considerably changes theperiodicity of the layer structure. In the layer structure produced bythe 60:30 gear ratio, a periodicity of six layers is visible in thestructure (Fig. 5). The periodicity repeats for every three rotationsof the turntable. In the three rotational cycles, the substrate towermakes six rotations, while the substrate makes two rotations aroundits axis (see on-line Animation 3). In the first rotational cycle, the sub-strate returns into the same initial position but has different orienta-tions because it rotates for 240° around its axis. In the second rotationof the turntable the substrate rotates for 480°, while in the thirdrotation the substrate returns into the same position and orientation
100
50
150
200
250
300
350
400
450
0
90:30100:3760:30 70:30
thic
knes
s [n
m]
Fig. 5. Layer structures produced by 3-fold rotation calculated for different turntable-to-tower gear ratios and 120° switch angle. Animation of the substrate rotation and thedeposition rate for 60:30 gear ratio can be seen in the on-line Animation 3 and for90:30 gear ratio in the on-line Animation 4.
since it rotates around its axis for 720° (2 × 360°); the layer structurerepeats after three rotations of the turntable. Layer structuresproduced by turntables with 70:30 and 100:37 gear ratios havemore complex periodicity. We will discuss the periodicity of thesestructures in the next section.
In the layer structure produced by g = 90:30, the periodicity ismore distinct; it repeats every two deposited layers, meaning that thesubstrate returns into exactly the same position and orientation foreach rotation of the turntable (see Animation 4). In one rotationalcycle the tower makes three rotations around the second axis, whilethe substrate also makes three rotations around its axis (3 × 120° =360°); hence, the substrate returns into the same position andorientation.
3.3.1.2. Selected gear ratio and different switch angles. The switch turnsthe substrate for a certain angle when the substrate tower makes onerotation around its axis. The angle of the rotation depends on thelength of the steel strip, its fixture, and on the size of the tool. Sincetools, which rotate with 3-fold rotation, have small diameter com-pared to the diameter of the substrate tower and the turntable, theswitch does not influence the trajectory considerably; it mainlychanges orientation of a tool. Here, we will discuss the layer struc-tures calculated for different switch angles and for selected integerand non-integer gear ratios.
Fig. 6 shows the layer structures for several switch angles calculatedfor the turntable with 90:30 and 100:37 gear ratios. Clearly, the switchangle has a strong influence on the layer structure. In the case of g =90:30 different modulation periods are visible (Fig. 6(a)). For instance,the layer structures produced by α = 120° and α = 240° repeat aftertwo deposited layers, for α = 60° and 180° the structure repeatsafter four deposited layers, for α = 80°, 160° and 200° the structure re-peats after six deposited layers, etc. On the other hand the periodicityof layer structures produced by the turntable with 100:37 gear ratiois more irregular (Fig. 6(b)).
Similar to calculations made in Section 3.2.1, the periodicity oflayer structures can be found by analyzing the number of rotationalcycles needed for the substrate to return into the same position andorientation. The identical positions and orientations of the substratecan be found by calculating the least common multiple of the revolu-tion times around the three axes. Calculations of the least commonmultiples for the selected turntable-to-tower gear ratios and switchangles are shown in Table 3. For the turntables with an integer gearratio, only a few rotational cycles are needed for the substrate toreturn into the same position and orientation, whereas for non-integer gear ratios, the turntable has to make many rotations beforethe substrate returns in the same position and orientation. Forinstance, let us compare periodicity of layer structures produced bythe 90:30 and 100:37 gear ratios for the 120° switch angle. In thecase of the 90:30 gear ratio, the substrate returns into the same posi-tion and orientation after one revolution of the turntable, while forthe 100:37 gear ratio it takes 111 rotations. Hence, the periodicityin the layer structure for the 90:30 gear ratio repeats for each rotationof the turntable, while the periodicity for the 100:37 gear ratiorepeats after 111 rotations of the turntable. In the latter case, theperiodicity can be too long to be observed in the coatings. If duringthe deposition time the turntable does not perform enough rotations(e.g. for 1 rpm the turntable makes 60 rotations in 1 h) then suchlayer structure is aperiodic.
Turntables with non-integer gear ratio in general produce layerstructures with a very long periodicity, yet in some layer structuresa shorter periodicity than the one calculated in Table 3, appears toform. Shorter periodicity, for example, is visible in the layer structuresproduced by 100:37 gear ratio (Fig. 6(b)) and the switch angles of120°, 220° and 240°. However, the periodicity of these structures isnot perfect. In these cases, after certain rotational cycles, the substratearrives into the position and orientation, which is similar to the initial
Fig. 6. Layer structures calculated for different switch angles and for (a) 90:30 and (b) 100:37 gear ratios. Calculations are made for 20 rotational cycles and for the same initialposition and orientation of the substrate.
position and orientation, but not exactly the same. The exact period-icity of these structures is the one that is calculated in Table 3; 111 ro-tational cycles for α = 120° and α = 240°, and 333 rotational cyclesfor α = 220°.
3.3.2. The uniformity of coatingsThe uniformity of coatings prepared by the 3-fold rotation was an-
alyzed for a round tool with diameter of 10 mm. Fig. 7 shows a layerstructure on the perimeter of the tool calculated for the turntablewith short periodicity (g = 90:30 and α = 120° — structure repeatsafter one rotational cycle) and for the turntable with long periodicity(g = 100:37 and α = 120° — structure repeats after 111 rotationalcycles). It can be seen that the rotation with the short periodicity(Fig. 7(a)) causes large non-uniformities both in the total thicknessof the coating as well as in its composition. The total thickness variesconsiderably over the perimeter of the tool; one side of the tool hasalmost twice the coating thickness than the other side. The largestdifferences in the layer structure are between the opposite surfaces;i.e., azimuthal angles φa = 0° and φa = 180°. The layer structure
Table 3Number of the turntable rotations required for substrate to return into the same position aswitch angles.
for φa = 0° has thick dark layers, while the opposite surface hasvery thin dark layers; consequently, one side of the tool has largeramount of one type material while the other side has a larger amountof the opposite material. Such large non-uniformity occurs becausethe substrate travels on the identical trajectory, while the orientationfollows the same pattern.
A better uniformity is achieved if the tool travels on differenttrajectories for several rotational cycles. Such is the case for the turn-table with g = 100:37 and α = 120° (Fig. 7(b)) — the substratemakes 111 rotations around its axis before it returns into the sameposition and orientation. Since the tool is more equally exposed bythe vaporized material it has more uniform composition and totalthickness.
More data about the uniformity of coatings on a round tool ispresented in Fig. 8. The figure shows the thickness and compositionuniformity for various switch angles calculated for integer (90:30)and non-integer (100:37) gear ratios. Composition uniformity wascalculated from the ratio between the total thickness of the darklayers and the total thickness of the coating. Calculations were
nd orientation in the 3-fold rotation. Calculations are made for selected gear ratios and
Fig. 7. Layer structure on the perimeter of a round tool calculated in 30° steps of azimuthal angle. Calculations are made for turntables with (a) 90:30 and (b) 100:37 gear ratios and120° switch angle. Initial position of the tool was set at: φ10 = 0° and φ20 = 0° (i.e. positioned between the dark targets).
made for the same initial position and for 20 rotational cycles. Data isshown for selected switch angles, which produce different periodicity(the periodicity can be found in Table 3).
In the case of 90:30 gear ratio (Fig. 8(a) and (b)) the largest varia-tions occur for the α = 120° since the substrate follows on the sametrajectory and orientation. For the rotation which repeats every two ro-tational cycles (α = 60°), the thickness uniformity is much better al-though variations in the composition are still significant. Similarbehavior can be seen for other switch angles with longer periodicity;i.e., α = 90°, α = 100° and α = 130°, which produce layer structureswith periodicity of four, six and twelve layers, respectively. The data forthis set of parameters shows that uniformity does not considerably im-prove with longer periodicity. This is because the turntable with 90:30gear ratio rotates substrates on the same trajectory and the uniformitycannot be improved considerably; orientation alone cannot improvethe uniformity significantly if the substrate follows the same trajectory.
Calculations of thickness and composition uniformity for the100:37 gear ratio and the same switch angles are shown in Fig. 8(c)and (d). In general the uniformity is better than in the case of 90:30gear ratio, however, the switch angles of 90° and 130° produce largenon-uniformities. In the case of 90° switch angle, the total thicknessvaries more than 20%, while the composition does not deviate fromthe average. The opposite case is for 130° switch angle; the totalthickness does not vary considerably while the composition does —
the amount of dark material varies for 30% over the perimeter ofthe tool. The reasons for such non-uniformities will be discussed inthe next section.
3.3.2.1. Influence of the deposition time on the uniformity.Non-uniformitiesin the case of 100:37 gear ratio are partially result of the depositiontime. Calculations were made for 20 rotational cycles. If rotationspeed of the turntable is 1 rpm, then the deposition time is 20 min,during this time the tool does not make a whole periodic cycle. For
α = 90° the rotation repeats every 37 rotational cycles and for α =130° it repeats for 333 rotational cycles, meaning that calculationswere made for time shorter than is the period of the layer structures.Still, non-uniformities for these two switch angles are not only resultof the short calculation time because other switch angles, which alsohave long modulation periods, produce considerably more uniformcoatings (see Fig. 8(c) and (d)). It can be concluded that in the caseof 90° and 130° switch angles the substrate arrives in a similar positionand orientation after certain numbers of rotations causing significantnon-uniformity.
To analyze the influence of the deposition time on the uniformityof the coating, calculations were made for longer deposition times.Fig. 9(a) shows the thickness uniformity for the 90° switch angleand 100:37 gear ratio. It can be seen that with increasing number ofrotational cycles the thickness uniformity improves. Layer structurecalculated for 40 rotational cycles has better uniformity than theone for 30 or 20 rotational cycles, however, for 50 rotational cyclesthe uniformity declines again. Similar can be seen in Fig. 9(b),which shows the composition uniformity for 130° switch angle and100:37 gear ratio. Coating calculated for 20 rotational cycles has avery non-uniform composition, for 30 rotational cycles uniformityimproves, whereas for 40 rotational cycles the composition onlyslightly deviates from the average, however, for 50 and 60 rotationalcycles the non-uniformity increases again.
Clearly, the time of the deposition process influences the unifor-mity of the coating. If the deposition time does not coincide withthe periodicity of the layer structure, then large non-uniformitiesmay occur for certain rotation parameters.
3.3.2.2. Influence of substrate position on the uniformity. Tools in a batchare arranged on different positions around the turntable and as a resulteach tool travels on unique trajectory and has different layer structures.Previous analysis considered only one particular initial position — the
Fig. 8. Thickness uniformity on a round tool for different switch angles calculated for (a) 90:30 and (c) 100:37 gear ratios, and composition uniformity calculated for (b) 90:30 and(d) 100:37 gear ratios. Calculations are made for 20 rotational cycles.
tool was located on the substrate tower positioned between two tar-gets of the dark material (φ10 = 0° and φ20 = 0°). The same calcula-tions were performed for a round tool positioned on adjacentsubstrate tower (φ10 = 60°) at initial positions of φ20 = 0° and
0.95
1.00
0.75
0.80
0.85
0.90
# of cycles
t/tm
ax
20 30 40 5060(a)
0.70
azimuthal angle ϕa (°)
60
0 30 60 90 120 150 180 210 240 270 300 330 360
Fig. 9. (a) Thickness uniformity on the perimeter of a round tool calculated for 100:37 gear ra(b) Composition uniformity on the perimeter of a round tool calculated for 100:37 gear ratio a
φ20 = 180°. Fig. 10 shows the layer structures for these two initial po-sitions calculated for 90:30 gear ratio (only this gear ratio is consideredsince it creates the largest differences in the thickness and compositionuniformity). Comparing Figs. 10 and 7(a), it can be seen that the layer
60
65
70
35
40
45
50
55
# of cycles
t dark/t to
tal (
%)
20 30 40 5060(b)
0 30 60 90 120 150 180 210 240 270 300 330 36030
azimuthal angle (°)
60
tio and 90° switch angle; thickness distribution is normalized to the maximum thickness.nd for 130° switch angle. Calculations are made for different numbers of rotational cycles.
structures are different, however, the magnitude of the variations inthe thickness and composition is similar. In all cases, the thicknessvaries between 300 nm and 500 nm. Hence, the initial position of thetool does not considerably change the variations in the total thickness;it only shifts it. The periodicity of coatings is the same regardless of theinitial position of the tool — in all cases two layers are depositedalthough the thickness of layers is not the same. The periodicity is de-termined by the rotation, i.e. gear ratio and switch angle, whereas theposition of the tool on the turntable determines the thickness of indi-vidual layers within the modulation period.
The composition uniformity with respect to the positions of thetool on the turntable is more complex. In Fig. 7(a), where the toolwas positioned between two dark targets, the dark material is mostabounded at azimuthal angle of 0°, while the bright material com-poses most of the coating at azimuthal angles between 150° and210° (there the coating is the thinnest). The uniformity in composi-tion is different for tools on the adjacent substrate tower (Fig. 10).In Fig. 10(a) the bright material composes most of the coating at azi-muthal angles between 60° and 120° where the thickness is betweenminimum and maximum coating thicknesses. For the layer structurein Fig. 10(b), the bright material composes most of the coating atthe azimuthal angles between 210° and 270°, where the coating isthe thinnest; this layer structure is more similar to the one inFig. 7(a). Hence, initial position of the tool has significant influenceon the composition uniformity; part of the tool gets coated withhigher amount of one type of material than the other. Interpretationof the composition uniformity is more difficult because it has tobe considered together with the thickness uniformity, which alsodepends on the initial position of the tool.
400
450
500
50
100
150
200
250
300
350
0
thic
knes
s [n
m]
400
450
500
50
100
150
200
250
300
350
0
thic
knes
s [n
m]
1150°120°90°60°30°0°
1150°120°90°60°30°0°
azimuthal
azimuthal
Fig. 10. Layer structure on the perimeter of a round tool calculated for initial positions: (a) φ1
tool are schematically shown above the layer structures. Calculations are made for turntabl
Calculations were also performed for other initial positions of theround tool and show that magnitude of the variations in the thicknessand composition is approximately the same, i.e., 300 nm–500 nm forthe total thickness and approximately 30% deviations from theaverage composition. It can be concluded that the magnitudes of var-iations are approximately the same regardless of the initial positioneven though the initial position significantly alters the layer structureof coatings.
It should be emphasized that extreme variations in the thicknessand composition described above occur only in highly periodic sub-strate rotations. In the less periodic rotations, either for 100:37 or90:30 gear ratio, the uniformity in the thickness and composition isconsiderably better (see Figs. 7(b) and 8).
3.4. Target arrangement
The layer structure of coatings also depends on target position andtheir configuration. Industrial PVD systems are designed with severaltarget sources, which are positioned around the turntable. Usually aneven number of targets is used (typically four, sometimes six) due tothe symmetry of the systems and plasma confinement through theunbalanced magnetron configuration. Layered coatings are normallydeposited from two types of target materials (less commonly more)that can be arranged in different configurations: e.g., targets of thesame material may be next to each other or opposite to each other.In addition, some deposition systems also have the possibility ofadjusting the position of targets. Hence, both the target configurationand their position influence the layer structure of the coatings. In thefollowing these two parameters will be analyzed.
(a)
(b)
330° 360°300°270°240°210°80°
330° 360°300°270°240°210°80°
angle ϕa[°]
angle ϕa[°]
0 = 60° and φ20 = 0° and (b) φ10 = 60° and φ20 = 180°. Initial positions of the roundes with 90:30 gear ratio and 120° switch angle.
3.4.1. The periodicity of layered coatings for different target configurationsTargets in the CC800/9 deposition systems can be set into position
where two targets are either close to each other (Fig. 11(a)) ormaximally separated between each other (Fig. 11(b)). In the previoussections the layer structures were calculated for the adjacent-targetconfiguration with angular separation of 40°, which is normallyused in the CC800/9 ML deposition system. Here, we will considerconfiguration with maximally separated targets. Fig. 11(a) and(b) shows the comparison between layer structures calculated for tar-gets with close configuration (40°) and maximally separated targets(90°). The periodicity of the layer structures for both target positionsis the same only the thickness of individual layers, and consequently,the total thickness of coatings are different (cf. Fig. 11(a) and (b)).This is the most apparent for the layer structures produced by the60:30 and 90:30 gear ratios. Since the periodicity of coatings is deter-mined by the substrate rotation and not by the position of the targets,the same conclusions about the periodicity of coatings are valid asdescribed in Sections 3.3.1 and 3.3.1.2.
All simulations in previous sections considered target configurationwhere the same type of material was adjacent to each other. A targetconfiguration where the same type of target material is positioned op-posite to each other can be also used when depositing layered coatings.Such type of target arrangement and corresponding layer structures is
(a) 60:3
350
400
450
40°
100
50
150
200
250
300
350
0
(b)
90°
100
150
200
250
300
350
400
(c)
50
0
250
300
350
400
180°
100
50
150
200
250
0
Fig. 11. Target configuration with angular separation of (a) 40°, (b) 90°, and (c) 180°, and coCalculations are made for 120° switch angle and identical initial position.
shown in Fig. 11(c). The layer structures produced in this type of con-figuration appear very different from those produced in the adjacent-target configuration (Fig. 11(b)). In the opposite-target configurationfour layers are deposited in every rotational cycle, while in theadjacent-target configuration only two are deposited. Targets arrangedin the opposite-target configuration will therefore produce twice asmany layers than in the adjacent-target configuration and twice aslong periodicity. The periodicity of layer structures is determined bythe rotation whereas the target configuration influences the numberof layers that are produced in one rotational cycle and the thicknessof those layers. The modulation period of layered structures (i.e., thetotal thickness of layers in one period of the layer structure) in theopposite- and adjacent-target configuration is the same for the samegear ratios (cf. Fig. 11(b) and (c)).
3.4.2. The uniformity of coating in the opposite-target configurationIn the previous sections we analyzed uniformity of coatings on a
round tool in the configuration where targets of the same materialwere adjacent and close to each other (see Section 3.3.2). Here wewill examine the uniformity of coating on a round tool for theopposite-target configuration and maximally separated targets. Fig. 12shows the thickness and composition uniformity of the coating on theperimeter of a round tool with diameter of 10 mm (calculations were
70:30 100:37 90:300
rresponding layer structures calculated for the 3-fold rotation with different gear ratios.
Fig. 12. Thickness uniformity on a round tool for different switch angles calculated for (a) 90:30 and (c) 100:37 gear ratios, and composition uniformity calculated for (b) 90:30 and(d) 100:37 gear ratios. Calculations are made for the opposite-target configuration with maximally separated targets and 20 rotational cycles.
made in the same way as in Section 3.3.2). In general, the uniformity inthe opposite-target configurations is better than the uniformity of coat-ings prepared in the adjacent-target configuration (see Fig. 8). There is aconsiderable improvement in the uniformity of thickness, particularlyin the cases of 90:30 gear ratio and 120° switch angle, and 100:37gear ratio and 90° switch angle. In the case of g = 90:30 and α =120° where substrate travels on the identical trajectory there is asym-metry in the thickness uniformity. Onemight expect to have a symmet-rical coating thickness around the perimeter of the tool but this is notthe case (see Fig. 12a). The asymmetry is related to the fact that thesymmetry of the rotation does not coincide with the symmetry of thetarget arrangement. A substrate makes three rotations around its axisbefore it returns into identical position, while the symmetry oftargets is four-fold. If the substrate made four rotations around its axis(e.g. 80:20 gear ratio) then the thickness uniformity would besymmetrical.
Uniformity of coating composition (Fig. 12(b) and (d)) also variesconsiderably for some of the rotation parameters, particularly for120° switch angle and 90:30 gear ratio, and 90° switch angle and100:37 gear ratio. For the 100:37 gear ratio and 90° switch anglevariations in the composition are large, while for the same parametersin the adjacent-target configuration the composition does not deviateconsiderably from the average value of 50% (cf. Figs. 8(d) and 12(d)).In contrast, for the 100:37 gear ratio and 130° switch angle the varia-tions in the composition are small, while for the same parameters inthe adjacent-target configuration the composition variations are large
(cf. Figs. 8(d) and 12(d)). Hence, target arrangement also has consider-able influence on the composition uniformity of the coatings; it canimprove composition uniformity for certain rotation parameters butit can also degrade it for other rotation parameters.
In general, the opposite-target arrangement improves the uniformityof coatings because targets are separated further apart, and because suchconfiguration has a higher degree of symmetry than the adjacent-targetconfiguration. Nevertheless, considerable non-uniformities still occur fora particular set of rotation parameters.
4. Summary and conclusions
A computer simulation of the thin film growth in an industrial de-position system with a planetary type of rotation was used to analyzethe influence of the rotation and target arrangement on the periodic-ity and uniformity of layered coatings. Analysis shows that the bestuniformity in coating thickness and composition is achieved whensubstrates travel on several different trajectories before returninginto the same position and orientation. Very periodic rotations,where substrate travels on the identical trajectory, cause significantnon-uniformities.
The periodicity of the layer structure is determined by the turntable-to-tower gear ratio, and in the case of 3-fold rotation additionally by theswitch angle. Exact periodicity of layered coatings can be calculatedfrom the least commonmultiple of revolution times around the individ-ual axes. Turntables with an integer gear ratio perform rotation where
substrates travel on the identical trajectory therefore layer structureshave shorter periodicity. Overall, turntables with an integer gear ratio,both for 2-fold or 3-fold rotations, produce larger non-uniformities inthe thickness and composition of layered coatings. A better coatinguniformity is obtained by turntables with non-integer gear ratio, sincesubstrate travels on different trajectories causing a more equal expo-sure to the vaporized material from the targets. Such coatings havelayer structures with longer modulation periods. Turntables withnon-integer gear ratios, in the most cases, produce better coating uni-formity, however, in the case of 3-fold rotation, certain combinationsof gear ratio and switch angle can cause large non-uniformities.
The uniformity of the coating also depends on the deposition time;if the deposition time does not coincide with the periodicity of thelayer structure, then large non-uniformities may occur for particularrotation parameters. In principle, the planetary rotation always pro-duces periodic layer structures, however, for typical depositiontimes (e.g. 1 h) the layer structures can be periodic or aperiodic.The layer structure is periodic if the substrate returns into the sameposition and orientation during the deposition time, if not, the layerstructure is aperiodic.
Position of the targets in the deposition system does not changethe periodicity of layer structures, i.e., the number of layers thatrepeat in the layer structure, it only affects the thickness of the individ-ual layers. Configuration with maximally separated targets producesbetter coating uniformity than configuration with closely positionedtargets. Targets arranged in the opposite-target configuration producetwice as many layers than in the adjacent-target configuration; layerstructures appear different yet the modulation period, i.e., the totalthickness of layers in one period of the layer structure, is the same.
In conclusion, the thickness and composition uniformity of lay-ered coatings are a complex problem that is not easily predictedeven though the periodicity of layer structures can be exactly calcu-lated. The most practical way to understand the effects of the rotationand the target configuration on the uniformity of coatings is to per-form computer simulations. For this reason such simulations, aspresented in this paper, should be useful tool for future design ofPVD systems and for better optimization of the deposition processes,particularly for the deposition of nanolayered coatings.
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.surfcoat.2013.06.126.
Acknowledgments
The author would like to acknowledge the Slovenian ResearchAgency for the financial support. Asst. Prof. Miha Čekada and Dr.Peter Panjan are kindly acknowledged for introducing problem ofplanetary substrate rotation in industrial magnetron sputtering sys-tems and for reviewing the manuscript. Special credit goes to TomažPeterman for his contribution to the development of the computersimulation.
References
[1] D.D. Hass, Y. Marciano, H.N.G. Wadley, Surf. Coat. Technol. 185 (2004) 283–291.[2] S. Tang, U. Schulz, Surf. Coat. Technol. 204 (2009) 1087–1091.[3] U. Helmersson, M. Lattemann, J. Bohlmark, A.P. Ehiasarian, J.T. Gudmundsson,
Thin Solid Films 513 (2006) 1–24.[4] A. Rauch, R.J. Mendelsberg, J.M. Sanders, A. Anders, J. Appl. Phys. 111 (2012)
083302.[5] A. Anders, Handbook of Plasma Immersion Ion Implantation and Deposition,
Wiley, 2000.[6] J. Alami, P.O.A. Persson, D. Music, J.T. Gudmundsson, J. Bohlmark, U. Helmersson, J.
Vac, Sci. Technol. A 23 (2005) 278–280.[7] J. Weichart, S. Kadlec, Application of HiPIMS to through silicon via metallization
in three-dimensional wafer packaging, US 20090111216, Oerlikon Balzers AG,(April 30th 2009).
[10] C. Ducros, V. Benevent, F. Sanchette, Surf. Coat. Technol. 163–164 (2003) 681–688.[11] M. Panjan, M. Čekada, P. Panjan, T. Peterman, Plasma Process. Polym. 4 (2007)
S921–S926.[12] M. Panjan, S. Šturm, P. Panjan, M. Čekada, Surf. Coat. Technol. 203 (2008)
554–557.[13] M. Panjan, M. Čekada, P. Panjan, F. Zupanič, W. Kölker, Vacuum 86 (2012) 699–702.[14] H.A. Jehn, B. Rother, Surf. Coat. Technol. 112 (1999) 103–107.[15] B. Rother, G. Ebersbach, H.M. Gabriel, Surf. Coat. Technol. 116–119 (1999)
694–698.[16] Z. Zhou, W.M. Rainforth, B. Rother, A.P. Ehiasarian, P.E. Hovsepian, W.D. Münz,
591–596.[18] J. Haider, M.S.J. Hashmi, Adv. Mater. Res. 83–86 (2010) 977–984.[19] P.C. Yashar, W.D. Sproul, Vacuum 55 (1999) 179–190.[20] S. Veprek, J. Vac, Sci. Technol. A 17 (1999) 2401–2420.[21] M. Stueber, H. Holleck, H. Leiste, K. Seemann, S. Ulrich, C. Ziebert, J. Alloys Compd.
483 (2009) 321–333.[22] C. Ziebert, S. Ulrich, J. Vac. Sci. Technol. A, and Films, 24 (2006) 554–583.[23] X. Chu, S.A. Barnett, J. Appl. Phys. 77 (1995) 4403–4411.[24] M. Panjan, T. Peterman, M. Čekada, P. Panjan, Surf. Coat. Technol. 204 (2009)
