- Email: [email protected]
Evaluation of residual stress and adhesion of Ti and TiN PVD films by laser spallation technique
Optics and Laser Technology 104 (2018) 140–147
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Evaluation of residual stress and adhesion of Ti and TiN PVD films by laser spallation technique J. Radziejewska a,⇑, A. Sarzyn´ski b, M. Strzelec b, R. Diduszko c, J. Hoffman d a
Warsaw University of Technology, Faculty of Production Engineering, Narbutta Str.85, 02-524 Warsaw, Poland Institute of Optoelectronics, MUT, gen. Sylwestra Kaliskiego Str. 2, 01-476 Warsaw, Poland c Tele & Radio Research Institute, Ratuszowa Str.11, 03-450 Warsaw, Poland d ´ skiego Str. 5b, 02-106 Warsaw, Poland Institute of Fundamental Technological Research Polish Academy of Sciences, Pawin b
a r t i c l e
i n f o
Article history: Received 30 August 2017 Received in revised form 12 January 2018 Accepted 8 February 2018
Keywords: Laser spallation technique Residual stress Adhesion Thin layer PVD VISAR system
a b s t r a c t The laser spallation technique was applied for measurement of residual stress and adhesion of thin films. Two films of different properties, ductile and soft Ti, and hard and brittle TiN, were studied. The films were produced on 304 steel substrate by PVD method. The residual stress value obtained by laser spallation technique LST were compared with stress value from X-ray diffraction method. Good agreement of stress values measured by both methods was attained. Additionally, the interface strength of the films was tested by laser adhesion spallation technique LASAT with use of VISAR system. It was shown that shock wave induced by a nanosecond laser pulse adequately determines properties of PVD thin films on metal substrate. Ó 2018 Elsevier Ltd. All rights reserved.
1. Introduction The shock waves create many unique possibilities in materials engineering although evaluation of properties of materials and layers undergoing high speed deformation is a serious experimental issue [1,2]. There are several methods for material testing at high speed deformation, such as the Split Hopkinson Pressure Bar (SHPB) method, miniaturized direct impact test, shock and explosive methods, however all these tests are complex and destructive [2]. Mechanical properties at high strain rate may be tested also by means of dynamic hardness testers. Strain rate in these devices is about 103 s1. Dynamic-plastic tests are applied the most commonly by pressing the indenter with the Poldi’s hammer while in dynamic-elastic test (Shore method) a bounce of the indenter or spring is measured [3]. Recently, the reports on development of new methods for measuring the dynamic hardness using a high velocity gas gun have been elaborated. The strain rate in these methods is about 1500–2200 s1 [3,4]. In 60-ties the high-energy laser pulse was applied as pressure load [5]. The short laser pulse interaction with material generates plasma and a pressure wave. The shock waves with high amplitudes, causing compression stress exceeding yield point of metals ⇑ Corresponding author. E-mail address: [email protected] (J. Radziejewska). https://doi.org/10.1016/j.optlastec.2018.02.014 0030-3992/Ó 2018 Elsevier Ltd. All rights reserved.
can be obtained [6] for surface covered with inertial layer, transparent for the laser beam (confined regime) [7,8]. Nanosecond high power pulses and carefully selected absorption layer as well as inertial layers, allow generating a pressure wave from a few to several GPa [9]. The use of short laser pulses in order to generate high-pressure shock waves create many unique possibilities in materials testing. Contrary to collision systems, wider range of pressures, speed and deformation settings may be achieved as a result of changing a shape and time duration of the laser pulse. On the basis of the laser shock waves the new diagnostic methods of dynamic behaviour of material and layer [10], as well as adhesion of thin films could be developed [11]. Thin films are an important component of many microelectronic, optical and micromechanical systems as well as cutting tool coatings. During their manufacture a large amount of a residual stress is induced, that has significant influence on their mechanical properties and overall efficiency. In certain conditions, the residual stress may cause layer delamination from substrate or its cracking. The most well-known practical techniques of measuring the adhesion of thin layers are scratch, peel, pull, blister or indentation test. Laser Spallation Technique LST was first introduced by Vossen [12]. In this method a layer is loaded with the stress wave created by short laser pulses. Accompanied phenomena were wider described in [13–19], and the technique was called LASAT (Laser Shock
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Adhesion Test). Adhesion of films was examined by the LASAT in many systems, such as components of electronic devices [13], PVD and CVD coatings in tool applications [20], plasma layers [21], joined materials or composites [22]. Mostly the interface strength of thin films few micrometers thick, on ceramic substrate, was analyzed [11–16]. In the 1990s, attempts to use the pressure wave generated by a laser pulse for testing the adhesion of thin layers, obtained by PVD and CVD methods, were made. Tensile stress at the interface of the material/layer phases, which caused separation of the layer from the surface, have been studied [18]. In these methods adhesion of layers is calculated on the basis of a speed of the layer. The accuracy of the method depends mainly on the accuracy of measuring of velocity of the back sample’s surface. Due to a very short time of the process, from several to tens of nanoseconds, highly advanced measurement techniques are required [18,19]. A number of review articles on the LASAT measurement methods are published for specific applications and layer materials. Among others they refer to TiN layers [15,20], hydroxyapatite [22], thermal barrier coatings EB-PVD TBC with the use of shock wave propagation in two dimensions (LASAT 2D) [23] or carbon fibre reinforced composite CFRP [24]. Also the surface shapes of the substrate as well as configuration of samples are profoundly analysed [24,25]. Based on laser spallation technique new method of residual stress measurement was proposed by Ikeda at el. [26]. During delamination of thin films, a protuberance occurs in films that have compressive residual stresses [27,28]. In many thin films made by PVD and CVD methods compressive residual stress generate. The film and substrate are made of materials that has different thermal expansion coefficient thus during a cooling the film gets residual stress. To release this compression the film tends to fit off the substrate and fracture of interface can take place. Residual stress can be determined from a shape of protuberance without knowing elastic properties of film. Analytical solution for a pressurized membrane was applied taking into account the pressure and further neglecting residual stress release by the elastic deformation [26]. The advantage of this method compared to the X-ray diffraction is possibility of stress measuring of the film that has strong texture or poor crystallinity. In such case a conventional X-ray method cannot be used. The paper presents experimental results of residual stress measuring by a laser spallation technique. The technological thin films TiN and Ti, produced by PVD method on steel substrate, were studied. The residual stress value determined by the laser spallation technique were compared with stress value obtained by X-ray diffraction methods. The adhesion strength was also determined by LASAT method with use VISAR system. It was shown that shock wave induced by nanoseconds laser pulse can be suitable tool for determination of properties of technological thin films.
141
The PVD process consists of the following steps: vacuum generation, heating up to 450 K, ion etching, coating deposition and cooling. Nd:YAG Quantel YG 981E laser with a wavelength of 1.064 mm and pulse duration of 10 ns was applied for testing. The beam diameter was 2 mm. A diagram of the measurement system is shown in Fig. 1. The laser pulse (1) is directed through a glass (2) to the absorption layer (3) causing its evaporation and plasma generation. A pressure wave (4) is formed as a result of rapid expansion of a plasma plume and propagates into material (5). Graphite 5 mm thick was used as the absorption layer, while glass 1 mm thick was the inertial layer. At very short, nanosecond laser pulses, and a suitably selected type and thickness of the absorption layer, the thermal effects associated with the interaction between the beam and material are negligible [10,30]. This allows to examine such a case as pure mechanical interaction of the pressure wave with tested material. Different pulse energy levels and thicknesses of substrate were applied to obtain proper conditions for delamination of technological layers. The values of pulse energy were: 0.5, 0.7, 1.0 and 1.25 J while substrate thickness had three values: 1, 0.8 and 0.5 mm. Table 1 shows test parameters and presence of films delamination. In order to determine a strength of the interface a surface velocity was measured by a VISAR system. The studies were conducted at pulse energy 1.2 J for both films. Three thickness values of the steel substrate were applied. The parameters applied in the LASAT are denoted by the ‘‘⁄” mark in Table 1. The VISAR (Velocity Interferometer System for Any Reflector) determines the velocity of moving surface by measuring the Doppler shift of laser light reflected from the surface. It is sensitive to wavelength; therefore, it transforms changes in the wavelength to changes in the intensity of four output signals. Afterwards these intensities are converted by fast photodiodes to electrical signals recorded by an oscilloscope. Velocities can be determined in the range from m/s up to km/s and with sub-nanosecond time resolution with accuracy ±1%. The observed surface does not need to be mirror polished. Changes in its reflectivity or in the background light have no effect on derivation of velocity. The properties of the films were controlled before testing in order to confirm the producer declaration. The following properties were measured: roughness, thickness and hardness. A surface geometrical structure of the films was studied on scanning profilometer. The roughness parameters Ra, Rz were determined according to ISO 4287:1997. For selected samples the metallographic crosssections were made perpendicularly to the surface and thickness of the films was measured on a Scanning Electron Microscope SEM. A microhardness test of the films was carried out using the Vickers method at a load of 0.2 N (20 gf). The hardness values of
2. Experimental method The study of residual stress and adhesion of PVD thin films by laser spallation technique was carried out for typical commercial metals and thin films. As a substrate a stainless steel, EN X5CrNi18-10 1.4301 (304), was used. Two kinds of commercial thin films, TiN and Ti, deposited by PVD method in Surftech manufacture [29], were tested. Samples were prepared before the deposition process. Round samples of a diameter of 10 mm were cut by WEDM method out of sheets 0.5 mm thick. Surface of material was electrolytic polished and cleaned before the deposition. Multi-step degreasing was applied: first in tetra-chloroethyl activated by ultrasound, then in aqueous solution of a detergent with a rinse in water, afterwards cleaning in alcohol and acetone activated by ultrasound, finally in vapor of tetra-chloroethyl.
Fig. 1. Experimental scheme for testing of adhesion of thin films to the substrate: 1 – laser pulse; 2 – inert layer (glass); 3 – absorption layer (graphite); 4 – stress wave front; 5 – substrate; 6 – thin film; 7 – VISAR interferometer.
142
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Table 1 Delamination of tested films at different pulse energy and substrate thickness. ‘‘*” denotes - VISAR tests. Film
E [J]
0.5
0.7
1.0
Ti
0.5 0.7 1.0 1.2
No Yes Yes Yes*
No No Yes Yes*
No No No No*
1.0 1.2
Yes Yes*
No Yes*
No No*
TiN
Substrate thickness [mm]
The microhardness measurement were conducted on film surface, based on knowledge of films thickness and hardness of stainless steel substrate 220 HV. Hardness of the Ti film was estimated as 640 HV and 2500 HV for the TiN films. The measured values of hardness correspond to hardness given by the manufacturer [29] and confirm the proper realisation of the PVD process. 3.2. Residual stress measurement by LST
the films were determined based on knowledge of thickness of the films and hardness of the substrate according the method [31]. After delamination of the films by laser pulse the surface deformations were measured on a laser confocal microscope Keyence VK-X100. The diameter, d, and height, h, of the protuberance were determined based on a VK Analyzer program. The 3D views, maps and profiles of surface were generated. The height and diameter of protuberances were determined on profiles passing through the highest point of bulged area. To verify correctness of the stress measurement by the LST method the residual stress in the films were determined also by X-ray diffraction method (XRD). The studies were carried out on diffractometer Rigaku SmartLab 3 kW with use of the Cu tube and Ka radiations. The stress was calculated based on sin2w method. 3. Results and discussion 3.1. Properties of films Surface topography of the steel plate before deposition process was studied. The roughness of polished steel plate prior to the deposition was Ra = 0.018 mm, Rz = 0.096 mm. After the deposition roughness increased for both tested films. The parameters were Ra = 0.39 mm, Rz = 3.14 mm for Ti film, while for TiN: Ra = 0.26 mm, Rz = 2.91 mm The main reason of high roughness is a presence of particles, visible in Fig. 2a, created during the PVD process. Fig. 2a shows the 3D view after Ti film deposition, whereas Fig. 2b presents the representative surface profile. The thickness of the films was examined on the profilometer by step method and additionally using SEM on samples for which metallographic cross section were made. It was stated that thickness of the Ti film was 3.4 ± 0.6 mm whereas 1.8 ± 0.3 mm for the TiN film. Fig. 3a shows the cross section of stainless steel sample with the Ti film while Fig. 3b presents surface of the film. Due to relatively high roughness the film thickness is not uniform.
3.2.1. Ti film The surfaces of thin films processed by shock wave were examined on scanning profilometer. The presence of delamination was verified. Delamination of the films was observed for laser pulse energy 0.7 J and higher, and substrate thickness 0.8 mm and 0.5 mm (Table 2). For lower pulse energy and thicker substrate the shock wave was too weak to cause film delamination. Fig. 4 shows the 3D view of surface and profile of the Ti film after film debonding caused by the shock wave at energy level 0.7 J. Small protuberance is visible on the 3D view; its altitude is comparable with the height of surface roughness peak. The profile (Fig. 4b) shows that delamination of the film occurred and its maximum height is 1.4 mm. At higher pulse energy the protuberances were larger, the maximum height was 13.2 mm, diameter was 2.72 mm for laser pulse energy 1 J and substrate thickness 0.5 mm. Fig. 5 shows the map and the profile of surface after film debonding at the highest pulse energy 1.25 J. The process was repeated 3 times for each level of energy. The heights were in the range 1–13 mm, that is small when compared to diameter 2–3 mm of protuberance. For small film thickness and ratio of height to diameter of protuberance the value of residual stress can be calculated from a dimension according to relation [26]:
r¼
pr 2 ; 6th
ð1Þ
where p - atmospheric pressure, t - thickness of film; r, h – radius and height of protuberance, respectively. Residual stress can be released by elastic deformations of the protuberance. The value of released stress can be estimated assuming the plane state of stress from Eq. (2):
rd ¼
0
E l l ; 1m l
ð2Þ
where E – Young’s modulus m – Poisson’s ratio of the film and l diameter of protuberance, l0 - total length of protuberance. In case of Ti film the parameters are: E = 110 GPa, m ¼ 0:36 [32], thus a value of released stress rd = 4.9 MPa, while for TiN the film parameters are: E = 250 MPa, m ¼ 0:2, rd = 3.4 MPa. In both cases the values were small and they were neglected.
Length = 3.4 mm Pt = 4.26 µm Scale = 5 µm
µm 4 3.5 3 2.5 2 1.5 1 0.5 0 -0.5 0
a)
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
b)
Fig. 2. The 3D-view – a; and the roughness profile - b of the Ti film deposited by PVD method.
2.4
2.6
2.8
3
3.2
3.4 mm
143
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Fig. 3. Microstructure of the Ti film on 304 steel substrate - a (cross-section); b – surface of the Ti film. Confocal microscope.
compressive residual stress determined according to formula 1, for mean value, 3.4 mm, of film thickness was 1.01 ± 0.64 GPa.
Table 2 Laser pulse energy, height h and radius r of the delamination and determined residual stress for the Ti film calculated according to formula 1. E [J]
h [mm]
r [mm]
Residual stress [GPa]
0.7 0.7 0.7 1 1 1 1.2 1.2 1.2 1.25 1.25 1.25
0.0026 0.0032 0.0026 0.0026 0.0132 0.0127 0.0013 0.0037 0.0049 0.0062 0.0125 0.006
0.78 0.84 0.76 0.82 1.36 1.34 1.99 1.11 0.96 1.1 1.5 1.03
1.15 1.07 1.09 1.17 0.69 0.7 1.36 1.51 0.93 0.95 0.92 0.86
Mean value
3.2.2. TiN films The same study was conducted for TiN film. Delamination was observed at the laser pulse energy level 1.2 J. The size values of the protuberance for tested samples are: h = 0.0031 mm and r = 1.3 mm. Shapes of the bulges were different then compered to Ti films. The bulges were more flat. The mean value of compressive residual stress determined according to formula 1 was 6.3 ± 0.6 GPa (Table 3). Fig. 6 shows the map and photograph of a defect on film protuberance. Cracking was not observed, only a small collapse with irregular shape was detected. Localized instabilities were observed by Jin et al. [33] in systems consisting of a thin stiff film on an elastomeric substrate. The ridge formations were observed when bilayer system was compressed. In our case the film was stiff and had high compressive residual stress but instability was limited to a small area and was detected only in one case. No collapse was observed for plastic Ti films that have much lower value of compressive residual stress. Therefore, it suggests that instability can be connected with variations of the TiN film properties, adhesion or residual stress.
1.01
Size of the protuberances was different for the same energy level but calculated value of the residual stress was very close. Table 2 shows the dimensions of the protuberances and the calculated value of the residual stress for the Ti film. During the shape analysis of protuberance the problems occurred in exact determination of a radius due to absence of sharp boundary between delaminated and not delaminated parts of the film as well as relatively high surface roughness. This is probably the main reason of differences in the values of determined residual stress. Another reason is heterogeneity of the film thickness that varies from 2.8 to 4 lm. The thickness of the film affects residual stress thus the value can vary in different areas of the film. The mean value of
Alpha = 59°
3.3. Residual stress analysis by XRD method The recorded diffraction patterns for the Ti thin film is shown in Fig. 7. The film was polycrystalline exhibiting reflections related to hexagonal structure. The diffraction patterns indicated strong tex-
µm
Beta = 22°
8 7.5 7 6.5 6 5.5 5
8.06 µm
4.5 4 3.5 3 2.5 2
3.4 mm 3.94 mm
µm
Length = 4.24 mm Pt = 3.58 µm Scale = 5 µm
5 4 3 2
1.5 1
1
0.5 0
0 0
a)
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
3
b)
Fig. 4. The 3D view – a, and profile after debonding – b; the Ti film, pulse energy 0.7 J. Scanning profilometer.
3.2
3.4
3.6
3.8
4
4.2 mm
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
100 101
10000
Fig. 5. The map with a profile of surface after debonding of the Ti film at laser pulse energy 1.2 J, 10 ns. Confocal microscope.
E [J]
h [mm]
r [mm]
Residual stress [GPa]
1.2 1.2 1.2 1.25 1.25 1.25
0.0032 0.0033 0.0029 0.0062 0.0051 0.0015
1.21 1.25 1.38 1.85 1.84 1.15
4.9 8.6 6.4 5.0 6.2 6.7
60
80
100
300
212 120
140
2θ [deg] Fig. 7. The X-ray diffraction pattern of deposited Ti film and austenite crystalline phases. The star symbol marks the reflection from the austenite substrate, while hkl indexes mark the hexagonal Ti phase (red color indicates the reflection used to determine residual stress). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
100000
222
10000
*
*
**
* *
220
* 200
ture. A preferred orientation was a crystallographic plane 100 parallel to the surface. The residual stress was determined by sin2w method, the analysis was performed based on 300 reflection. For calculations the following data was assumed: Young’s modulus E = 110 GPa, Poisson’s ratio m = 0.36 [32]. The value of residual stress was r = 0.86 GPa, Dr ±0.07 GPa. Fig. 8 shows the recorded XRD patterns for TiN film on steel substrate. The TiN film was polycrystalline and the reflections indicate a cubic structure. A strong texture was stated. Almost all crystallites have preferential direction and they are orientated according to 111 crystallographic plane that is parallel to surface. Such anisotropy of microstructure of film limits the ability to perform measurements and calculations of stress because it gives large error, especially for the TiN layer. Nevertheless, an attempt was made and for calculating the following data: E = 250 GPa,
3 3 3/ 5 1 1
6.3
Intensity [cps]
Mean value
40
**
111
Table 3 Laser pulse energy, height h and radius r of the protuberance, and residual stress values determined according to formula 1.
20
211
* 1000
* 202
*
110 103 200 112 201
002
Intensity [cps]
* 102
144
1000
20
40
60
80
100
120
140
2θ [deg] Fig. 8. The X-ray diffraction pattern for deposited TiN film and austenite crystalline phases. The star symbol marks reflections from the austenite substrate, while the hkl indexes denote cubic TiN phase (red color indicates the reflection used to determine residual stress). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 6. The map with profile of the protuberance - a, photograph of defect - b, TiN film after debonding with laser pulse energy 1.25 J, 10 ns. Confocal microscope.
145
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
m = 0.20 were assumed. Value of the residual stress was calculated based on reflection 222. Compressive residual stress was estimated: r = 7.2 GPa, Dr ±1.7 GPa. Due to strong texture the value is approximate. Both applied methods, LSP and XRD, give similar values of residual stress. The value of stress for the Ti film obtained by LST, is several times smaller than for the TiN film. A difference in the stress value is similar to that obtained by the XRD method. 3.4. Interface strength by LASAT The adhesion of the films to the substrate was determined based on measurement of film velocity during delamination. Fig. 9a shows typical recorded signals from VISAR system made by shock waves. The signals were measured on back surface of steel samples, then the velocity profile were estimated based on them (Fig. 9b) [30]. The second peak of velocity corresponds to reflected pressure wave. The velocity of sound inside steel plate, estimated from experiment, is about 5.76 km/s, thus it is equal to the velocity of sound in steel AISI 304 with approximation of 3 %.
In Fig. 10 the signals from VISAR system for samples covered by TiN films are visible. The surface speed is lower and strong attenuation of wave pressure can be observed. This effect is caused by differences in acoustic impedance of substrate and films. Fig. 11a shows the signals from VISAR system for samples 0.7 mm thick covered by the Ti film for which delamination was observed, while the calculated velocity profile is shown in Fig. 11b. In Fig. 12 the data for non-delaminated film are presented. The shock wave that reaches the interface is divided into two waves: a reflected wave and a transmitted one. In case of weak shock waves (when shock stress is much lower than Young’s modulus of material) linear approximation may be used [34,35]. In our case shock wave pressure is in the order of 1 GPa and E = 110 GPa, E = 250 GPa, respectively for the Ti film and for the TiN film. Therefore, energy division between reflected and transmitted waves depends on a value of ratio of acoustic impedances of contacting materials. Factors of transformation are described by equations [34]:
P¼
2A ; 1þA
R¼
A1 ; Aþ1
A¼
q2 c 2 ; q1 c 1
Fig. 9. a - The VISAR interferometer signals recorded during deformation of steel plate 1 mm thick, diameter of 10 mm, b-velocity profile. Pulse energy pulse 1.2 J.
Fig. 10. The VISAR interferometer signals recorded during delamination of the TiN film, steel plate 0.5 mm thick, 10 mm of diameter. Pulse energy 1.2 J.
ð3Þ
146
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Fig. 11. The VISAR interferometer signals recorder during delamination of the Ti film; steel plate 0.7 mm thick, 10 mm diameter. Pulse energy 1.2 J.
Fig. 12. The VISAR interferometer signals recorded during deformation of steel plate (1 mm thick, 10 mm diameter) with the Ti film. Pulse energy 1.2 J.
where A – ratio of the acoustic impedances; P – relative amplitude of the wave transmitted from medium 1 to medium 2; R – relative amplitude of the wave reflected from material boundaries. . – film and substrate density, c – sound speed. In case of the Ti film on steel substrate the ratio of acoustic impedances is 4.5 (for steel 47 (g/cm3km/s). Relative amplitude of the wave reflected from material boundaries is R = 0.43 while the relative amplitude of wave transmitted from steel to film is R = 0.57. For the TiN film we obtained A = 2.4 (TiN 5.22 g/cm3, c = 3.8 km/s) [11], R = 0.41 and P = 0.59. In both cases acoustic impedances of films were less than substrate thus less than 60% of shock wave energy was transmitted to the films. In Fig. 10 weaker reflected wave from surface of substrate is visible. The delamination of film takes place during impacts of the first peak of pressure. Sharp changes in velocity profile are not visible. Relatively smooth course of velocity profile indicates on viscous failure of interface. In a model case when incident and reflected waves meets at interface the strength of interface can be estimated by simple relation [14,16,36]:
1 2
r ¼ qcDu
ð4Þ
where q – film density, c – sound speed, Du – surface velocity changes of the films during delamination.
The Du values for the Ti film were: 25, 50, 40 and 50 m/s. For the TiN films only two proper signals from ViSAR system were registered and the average value of velocity changes during delamination was Du = 58 m/s. The spall strength of interface, estimated from relation 4 for the TiN film, was 550 MPa and for the Ti film was 450 ± 53 MPa. The higher adhesion of the TiN film compared to the Ti film agrees with level of pulse energy that caused delamination. The delamination was observed at 0.7 J for Ti film while for the TiN film it occurred at 1.2 J. 4. Conclusions The study shows that laser spallation technique can be successfully used for estimating the value of residual stress of technological thin films deposited by PVD method. Until now there are no suitable methods to measure residual stress in thin films with poor crystallization or strong texture. This method is simple and quick. It allows obtain indicative value of compressive residual stress even for films with high roughness. The value of residual stress (6.3 GPa) for TiN films measured by the proposed method is similar to that obtained for films of comparable thickness deposited by PVD method on steel substrate. In the work [37,38] residual stress measured by X-ray diffraction was 8 GPa while the value obtained by Carvalho et al. [39] was 5 GPa for 5 lm thick film. On hard substrate a lower film stress, 3.58 GPa, was stated [40]. For Ti film the value of residual stress obtained by the LST was 1.01 GPa, whereas by the XRD it was 0.86 GPa. The results are comparable to stress value got by Leoni et al. [38] by the XRD, 1 GPa for 0.2 lm thick film on the same
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
substrate. Therefore, the LST method gives reliable value of compressive residual stress in tested thin films. The accuracy of the LST depends on precision of measurement of size of the bulge. Several techniques can be adopted, for example scanning profilometry or confocal microscope. For technological thin films with strong adhesion to metal substrate the high level of laser pulse energy is required. The thickness of substrate is limited and should be matched to laser pulse energy. The study shows that in case of technological films with good adhesion to steel substrate the LASAT can be successfully applied. The interface strength for the Ti films was 450 MPa whereas for the TiN film it was 550 MPa that is similar to that found by Zhou et al. [17] (371 MPa) on steel substrate. For more precise determination of the interface strength the future study should include analysis of reflected waves from the surface of film. The presented experimental results show that shock wave induced by nanoseconds laser pulse can be suitable new tool for determination of properties of technological thin films. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.optlastec.2018.02. 014. References [1] J.Z. Malinowski, J.R. Klepaczko, A unified analytic and numerical approach to specimen behavior in the SHPB, Int. J. Mech. Sci. 28 (1986) 381–391. [2] W. Moc´ko, Z.L. Kowalewski, Application of FEM in assessments of phenomena associated with dynamic investigations on miniaturised DICT testing stand, Kovove Mater. 51 (2013) 71–82. [3] B.J. Koeppel, G. Subhash, Characteristic of residual plastic zone under static and dynamic Vickers indentations, Wear 224 (1999) 56–67. [4] W. Kohlhofer, R.K. Penny, Dynamic hardness testing of metals, Int. J. Press. Vessel. Pip. 61 (1995) 65–75. [5] G.A. Askaryon, E.M. Moroz, Pressure on evaporation of matter in a radiation beam, JETP Lett. 16 (1963) 1638–1644. [6] N.C. Anderholm, Laser generated stress waves, Appl. Phys. Lett 16 (1970) 113– 118. [7] D.W. Gregg, S.J. Thomas, Momentum transfer produced by focused laser giant pulses, J. Appl. Phys. 27 (1966) 2787–2789. [8] C.H. Skeen, C.M. York, Laser-Induced ‘‘blow-off” phenomenon, Appl. Phys. Lett. 12 (1968) 369–371. [9] L. Berthe, R. Fabbro, P. Peyre, L. Tollier, E. Bartnicki, Shock waves from a water confined laser-generated plasma, J. Appl. Phys. 82 (1997) 2826–2832. [10] J. Radziejewska, Application of a nanosecond laser pulse to evaluate dynamic hardness under ultra-high strain rate, Optics Laser Technol. 78 (2016) 125– 133. [11] J. Wang, R.L. Weaver, N.R. Sottos, A parametric study of laser induced thin film spallation, Exp. Mech. 42 (1) (2002) 74–83. [12] J.L. Vossen, Adhesion Measurement of Thin Films, Thick Films and Bulk Coatings, ASTM STP640, 1978, 122–133. [13] I.G. Epishin, V.V. Suslov, V.A. Yanushkevich, Determination of adhesion strength of film structures of components in electronic devices using laser shock waves, Fizika i Khimiya Obrabotki Materialov 22 (5) (1988) 80–84. [14] V. Gupta, A.S. Argon, D.M. Parks, J.A. Cornie, Measurement of interface strength by a laser spallation technique, J. Mech. Phys., Solids 40 (1992) 141–147. [15] C. Tang, J. Zhu, The measurement of interface strength of TiN coating/substrate by laser spallation, Int. J. Refract. Hard Metals 14 (1996) 203–206.
147
[16] A.G. Youtsos, M. Kiriakopoulos, T. Timke, Experimental and theoretical/ numerical investigations of thin films bonding strength, Theor. Appl. Fract. Mech. 31 (1999) 47–59. [17] M. Zhou, Y.K. Zhang, L. Cai, Adhesion measurement of thin films by a modified laser spallation technique: theoretical analysis and experimental investigation, Appl. Phys. A 74 (2002) 475–480. [18] J. Yuan, V. Gupta, Measurement of interface strength by the modified laser spallation technique. I. Experiment and simulation of the spallation process, J. Appl. Phys. 74 (1993) 2388–2396. [19] Berthe, R. Fabbro, P. Peyre, E. Bartnicki, Wavelength dependent of laser shockwave generation in the water-confinement regime, J. Appl. Phys. 85 (11) (1999) 7552–7555. [20] M.F. Othman, A.R. Bushroa, Wan Normimi, Roslini Abdullah, Evaluation techniques and improvements of adhesion strength for TiN coating in tool applications: a review, J. Adhesion Sci. Technol. 29 (7) (2015) 569–591. [21] D. Courapied, R. Kromer, L. Berthe, P. Peyre, S. Costil, J. Cormier, M. Boustie, X. Milhet, Laser adhesion test for thermal sprayed coatings on textured surface by laser, J. Laser Appl. 28 (2) (2016), https://doi.org/10.2351/1.4944451. [22] V. Guipont, M. Jeandin, S. Bansard, K.A. Khor, M. Nivard, L.t. Berthe, J.-P. CuqLelandais, M. Boustie, Bond strength determination of hydroxyapatite coatings on Ti-6Al-4V substrates using the Laser Shock Adhesion Test (LASAT), J. Biomed. Mater. Res. A 95A (4) (2010) 1096–1104. [23] G. Bégué, G. Fabre, V. Guipont, M. Jeandin, P. Bilhe, J.Y. Guédou, F. Lepoutre, Laser Shock Adhesion Test (LASAT) of EB-PVD TBCs: Towards an industrial application, Surf. Coat. Technol. 237 (2013) 305–312. [24] B. Ehrhart, R. Ecault, F. Touchard, M. Boustie, L. Berthe, et al., Development of a laser shock adhesion test for the assessment of weak adhesive bonded CFRP structures, Int. J. Adhesion Adhesives 52 (2014) 57–65. [25] M. Arrigoni, J.-P. Cuq-Lelandais, M. Boustie, E. Gay, L. Berthe, An industrial challenge based on the wave propagation: the shock adhesion test, in: Wave propagation, Academy Publish Ed., WY (USA), 2012. [26] R. Ikeda, T. Uchiyama, H. Cho, T. Ogawa, M. Takemoto, An advanced method for measuring the residual stress of deposited film utilizing laser spallation technique, Sci. Technol. Adv. Mater. 7 (2006) 90–96. [27] B. Audoly, Mode-dependent toughness and the delamination of compressed thin films, J. Mech. Phys. Solids 48 (2000) 2315–2332. [28] Yu. Hong-Hui, John W. Hutchinson, Influence of substrate compliance on buckling delamination of thin films, Int. J. Fracture 113 (2002) 39–55. [29] http://www.surftec.pl/firma/powlekanie-pvd/ [30] W. Mocko, J. Radziejewska, A. Sarzyn´ski, M. Strzelec, J. Marczak, A. Rycyk, Analysis of the plastic deformation of AISI 304 steel induced by nanosecond laser pulse, Optics Laser Technol. 90 (2017) 165–173. [31] S.R.J. Saunders, H.R. Vetters, Standardisation of test methods for the mechanical properties of thin coatings, Thin Solid Films 299 (1997) 82–87. [32] http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=mq304a. [33] L. Jin, A. Takei, J.W. Hutchinson, Mechanics of wrinkle/ridge transitions in thin film/substrate systems, J. Mech. Phys. Solids 81 (2015) 22–40. [34] H. Kolsky, Stress waves in solids, (tłum. ros.), U.Rokmcrbq, Bokys yagpa;eybq d ndeplsx nekax, Bplanek]cndo Byocnpayyoq Kbnepanyps, Mocrda, 1955. [35] T. de Rességuier, On the shock-based determination of the adhesive strength at a substrate-coating interface, J. Adhesion Sci. Technol. (2017), https://doi.org/ 10.1080/01694243.2017.1332502. [36] F. Bauer, PVDF shock compression sensors in shock wave physics, CP706, Shock Compression of Condensed Matter, in: M.D. Furnish, Y.M. Gupta, J.W. Forbes (Ed.), 2003, pp. 1121–1126. [37] L. Zhang, H. Yang, X. Pang, K. Gao, A.A. Volinsky, Microstructure, residual stress, and fracture of sputtered TiN films, Surf. Coat. Technol. 224 (2013) 120– 125. [38] M. Leoni, P. Scardi, S. Rossi, L. Fedrizzi, Y. Massiani, (Ti, Cr)N and Ti/TiN PVD coatings on 304 stainless steel substrates: Texture and residual stress, Thin Solid Films 345 (2) (1999) 263–269. [39] N.J.M. Carvalho, E. Zoestbergen, B.J. Kooi, J.Th.M. De Hosson, Stress analysis and microstructure of PVD monolayer TiN and multilayer TiNy(Ti, Al)N coatings, Thin Solid Films 429 (2003) 179–189. [40] P.C. Jindal, A.T. Santhanam, U. Schleinkofer, A.F. Shuster, Performance of PVD TiN TiCN, and TiAlN coated cemented carbide tools in turning, Int. J. Refractory Metals Hard Mater. 17 (1999) 163–170.
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/optlastec
Full length article
Evaluation of residual stress and adhesion of Ti and TiN PVD films by laser spallation technique J. Radziejewska a,⇑, A. Sarzyn´ski b, M. Strzelec b, R. Diduszko c, J. Hoffman d a
Warsaw University of Technology, Faculty of Production Engineering, Narbutta Str.85, 02-524 Warsaw, Poland Institute of Optoelectronics, MUT, gen. Sylwestra Kaliskiego Str. 2, 01-476 Warsaw, Poland c Tele & Radio Research Institute, Ratuszowa Str.11, 03-450 Warsaw, Poland d ´ skiego Str. 5b, 02-106 Warsaw, Poland Institute of Fundamental Technological Research Polish Academy of Sciences, Pawin b
a r t i c l e
i n f o
Article history: Received 30 August 2017 Received in revised form 12 January 2018 Accepted 8 February 2018
Keywords: Laser spallation technique Residual stress Adhesion Thin layer PVD VISAR system
a b s t r a c t The laser spallation technique was applied for measurement of residual stress and adhesion of thin films. Two films of different properties, ductile and soft Ti, and hard and brittle TiN, were studied. The films were produced on 304 steel substrate by PVD method. The residual stress value obtained by laser spallation technique LST were compared with stress value from X-ray diffraction method. Good agreement of stress values measured by both methods was attained. Additionally, the interface strength of the films was tested by laser adhesion spallation technique LASAT with use of VISAR system. It was shown that shock wave induced by a nanosecond laser pulse adequately determines properties of PVD thin films on metal substrate. Ó 2018 Elsevier Ltd. All rights reserved.
1. Introduction The shock waves create many unique possibilities in materials engineering although evaluation of properties of materials and layers undergoing high speed deformation is a serious experimental issue [1,2]. There are several methods for material testing at high speed deformation, such as the Split Hopkinson Pressure Bar (SHPB) method, miniaturized direct impact test, shock and explosive methods, however all these tests are complex and destructive [2]. Mechanical properties at high strain rate may be tested also by means of dynamic hardness testers. Strain rate in these devices is about 103 s1. Dynamic-plastic tests are applied the most commonly by pressing the indenter with the Poldi’s hammer while in dynamic-elastic test (Shore method) a bounce of the indenter or spring is measured [3]. Recently, the reports on development of new methods for measuring the dynamic hardness using a high velocity gas gun have been elaborated. The strain rate in these methods is about 1500–2200 s1 [3,4]. In 60-ties the high-energy laser pulse was applied as pressure load [5]. The short laser pulse interaction with material generates plasma and a pressure wave. The shock waves with high amplitudes, causing compression stress exceeding yield point of metals ⇑ Corresponding author. E-mail address: [email protected] (J. Radziejewska). https://doi.org/10.1016/j.optlastec.2018.02.014 0030-3992/Ó 2018 Elsevier Ltd. All rights reserved.
can be obtained [6] for surface covered with inertial layer, transparent for the laser beam (confined regime) [7,8]. Nanosecond high power pulses and carefully selected absorption layer as well as inertial layers, allow generating a pressure wave from a few to several GPa [9]. The use of short laser pulses in order to generate high-pressure shock waves create many unique possibilities in materials testing. Contrary to collision systems, wider range of pressures, speed and deformation settings may be achieved as a result of changing a shape and time duration of the laser pulse. On the basis of the laser shock waves the new diagnostic methods of dynamic behaviour of material and layer [10], as well as adhesion of thin films could be developed [11]. Thin films are an important component of many microelectronic, optical and micromechanical systems as well as cutting tool coatings. During their manufacture a large amount of a residual stress is induced, that has significant influence on their mechanical properties and overall efficiency. In certain conditions, the residual stress may cause layer delamination from substrate or its cracking. The most well-known practical techniques of measuring the adhesion of thin layers are scratch, peel, pull, blister or indentation test. Laser Spallation Technique LST was first introduced by Vossen [12]. In this method a layer is loaded with the stress wave created by short laser pulses. Accompanied phenomena were wider described in [13–19], and the technique was called LASAT (Laser Shock
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Adhesion Test). Adhesion of films was examined by the LASAT in many systems, such as components of electronic devices [13], PVD and CVD coatings in tool applications [20], plasma layers [21], joined materials or composites [22]. Mostly the interface strength of thin films few micrometers thick, on ceramic substrate, was analyzed [11–16]. In the 1990s, attempts to use the pressure wave generated by a laser pulse for testing the adhesion of thin layers, obtained by PVD and CVD methods, were made. Tensile stress at the interface of the material/layer phases, which caused separation of the layer from the surface, have been studied [18]. In these methods adhesion of layers is calculated on the basis of a speed of the layer. The accuracy of the method depends mainly on the accuracy of measuring of velocity of the back sample’s surface. Due to a very short time of the process, from several to tens of nanoseconds, highly advanced measurement techniques are required [18,19]. A number of review articles on the LASAT measurement methods are published for specific applications and layer materials. Among others they refer to TiN layers [15,20], hydroxyapatite [22], thermal barrier coatings EB-PVD TBC with the use of shock wave propagation in two dimensions (LASAT 2D) [23] or carbon fibre reinforced composite CFRP [24]. Also the surface shapes of the substrate as well as configuration of samples are profoundly analysed [24,25]. Based on laser spallation technique new method of residual stress measurement was proposed by Ikeda at el. [26]. During delamination of thin films, a protuberance occurs in films that have compressive residual stresses [27,28]. In many thin films made by PVD and CVD methods compressive residual stress generate. The film and substrate are made of materials that has different thermal expansion coefficient thus during a cooling the film gets residual stress. To release this compression the film tends to fit off the substrate and fracture of interface can take place. Residual stress can be determined from a shape of protuberance without knowing elastic properties of film. Analytical solution for a pressurized membrane was applied taking into account the pressure and further neglecting residual stress release by the elastic deformation [26]. The advantage of this method compared to the X-ray diffraction is possibility of stress measuring of the film that has strong texture or poor crystallinity. In such case a conventional X-ray method cannot be used. The paper presents experimental results of residual stress measuring by a laser spallation technique. The technological thin films TiN and Ti, produced by PVD method on steel substrate, were studied. The residual stress value determined by the laser spallation technique were compared with stress value obtained by X-ray diffraction methods. The adhesion strength was also determined by LASAT method with use VISAR system. It was shown that shock wave induced by nanoseconds laser pulse can be suitable tool for determination of properties of technological thin films.
141
The PVD process consists of the following steps: vacuum generation, heating up to 450 K, ion etching, coating deposition and cooling. Nd:YAG Quantel YG 981E laser with a wavelength of 1.064 mm and pulse duration of 10 ns was applied for testing. The beam diameter was 2 mm. A diagram of the measurement system is shown in Fig. 1. The laser pulse (1) is directed through a glass (2) to the absorption layer (3) causing its evaporation and plasma generation. A pressure wave (4) is formed as a result of rapid expansion of a plasma plume and propagates into material (5). Graphite 5 mm thick was used as the absorption layer, while glass 1 mm thick was the inertial layer. At very short, nanosecond laser pulses, and a suitably selected type and thickness of the absorption layer, the thermal effects associated with the interaction between the beam and material are negligible [10,30]. This allows to examine such a case as pure mechanical interaction of the pressure wave with tested material. Different pulse energy levels and thicknesses of substrate were applied to obtain proper conditions for delamination of technological layers. The values of pulse energy were: 0.5, 0.7, 1.0 and 1.25 J while substrate thickness had three values: 1, 0.8 and 0.5 mm. Table 1 shows test parameters and presence of films delamination. In order to determine a strength of the interface a surface velocity was measured by a VISAR system. The studies were conducted at pulse energy 1.2 J for both films. Three thickness values of the steel substrate were applied. The parameters applied in the LASAT are denoted by the ‘‘⁄” mark in Table 1. The VISAR (Velocity Interferometer System for Any Reflector) determines the velocity of moving surface by measuring the Doppler shift of laser light reflected from the surface. It is sensitive to wavelength; therefore, it transforms changes in the wavelength to changes in the intensity of four output signals. Afterwards these intensities are converted by fast photodiodes to electrical signals recorded by an oscilloscope. Velocities can be determined in the range from m/s up to km/s and with sub-nanosecond time resolution with accuracy ±1%. The observed surface does not need to be mirror polished. Changes in its reflectivity or in the background light have no effect on derivation of velocity. The properties of the films were controlled before testing in order to confirm the producer declaration. The following properties were measured: roughness, thickness and hardness. A surface geometrical structure of the films was studied on scanning profilometer. The roughness parameters Ra, Rz were determined according to ISO 4287:1997. For selected samples the metallographic crosssections were made perpendicularly to the surface and thickness of the films was measured on a Scanning Electron Microscope SEM. A microhardness test of the films was carried out using the Vickers method at a load of 0.2 N (20 gf). The hardness values of
2. Experimental method The study of residual stress and adhesion of PVD thin films by laser spallation technique was carried out for typical commercial metals and thin films. As a substrate a stainless steel, EN X5CrNi18-10 1.4301 (304), was used. Two kinds of commercial thin films, TiN and Ti, deposited by PVD method in Surftech manufacture [29], were tested. Samples were prepared before the deposition process. Round samples of a diameter of 10 mm were cut by WEDM method out of sheets 0.5 mm thick. Surface of material was electrolytic polished and cleaned before the deposition. Multi-step degreasing was applied: first in tetra-chloroethyl activated by ultrasound, then in aqueous solution of a detergent with a rinse in water, afterwards cleaning in alcohol and acetone activated by ultrasound, finally in vapor of tetra-chloroethyl.
Fig. 1. Experimental scheme for testing of adhesion of thin films to the substrate: 1 – laser pulse; 2 – inert layer (glass); 3 – absorption layer (graphite); 4 – stress wave front; 5 – substrate; 6 – thin film; 7 – VISAR interferometer.
142
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Table 1 Delamination of tested films at different pulse energy and substrate thickness. ‘‘*” denotes - VISAR tests. Film
E [J]
0.5
0.7
1.0
Ti
0.5 0.7 1.0 1.2
No Yes Yes Yes*
No No Yes Yes*
No No No No*
1.0 1.2
Yes Yes*
No Yes*
No No*
TiN
Substrate thickness [mm]
The microhardness measurement were conducted on film surface, based on knowledge of films thickness and hardness of stainless steel substrate 220 HV. Hardness of the Ti film was estimated as 640 HV and 2500 HV for the TiN films. The measured values of hardness correspond to hardness given by the manufacturer [29] and confirm the proper realisation of the PVD process. 3.2. Residual stress measurement by LST
the films were determined based on knowledge of thickness of the films and hardness of the substrate according the method [31]. After delamination of the films by laser pulse the surface deformations were measured on a laser confocal microscope Keyence VK-X100. The diameter, d, and height, h, of the protuberance were determined based on a VK Analyzer program. The 3D views, maps and profiles of surface were generated. The height and diameter of protuberances were determined on profiles passing through the highest point of bulged area. To verify correctness of the stress measurement by the LST method the residual stress in the films were determined also by X-ray diffraction method (XRD). The studies were carried out on diffractometer Rigaku SmartLab 3 kW with use of the Cu tube and Ka radiations. The stress was calculated based on sin2w method. 3. Results and discussion 3.1. Properties of films Surface topography of the steel plate before deposition process was studied. The roughness of polished steel plate prior to the deposition was Ra = 0.018 mm, Rz = 0.096 mm. After the deposition roughness increased for both tested films. The parameters were Ra = 0.39 mm, Rz = 3.14 mm for Ti film, while for TiN: Ra = 0.26 mm, Rz = 2.91 mm The main reason of high roughness is a presence of particles, visible in Fig. 2a, created during the PVD process. Fig. 2a shows the 3D view after Ti film deposition, whereas Fig. 2b presents the representative surface profile. The thickness of the films was examined on the profilometer by step method and additionally using SEM on samples for which metallographic cross section were made. It was stated that thickness of the Ti film was 3.4 ± 0.6 mm whereas 1.8 ± 0.3 mm for the TiN film. Fig. 3a shows the cross section of stainless steel sample with the Ti film while Fig. 3b presents surface of the film. Due to relatively high roughness the film thickness is not uniform.
3.2.1. Ti film The surfaces of thin films processed by shock wave were examined on scanning profilometer. The presence of delamination was verified. Delamination of the films was observed for laser pulse energy 0.7 J and higher, and substrate thickness 0.8 mm and 0.5 mm (Table 2). For lower pulse energy and thicker substrate the shock wave was too weak to cause film delamination. Fig. 4 shows the 3D view of surface and profile of the Ti film after film debonding caused by the shock wave at energy level 0.7 J. Small protuberance is visible on the 3D view; its altitude is comparable with the height of surface roughness peak. The profile (Fig. 4b) shows that delamination of the film occurred and its maximum height is 1.4 mm. At higher pulse energy the protuberances were larger, the maximum height was 13.2 mm, diameter was 2.72 mm for laser pulse energy 1 J and substrate thickness 0.5 mm. Fig. 5 shows the map and the profile of surface after film debonding at the highest pulse energy 1.25 J. The process was repeated 3 times for each level of energy. The heights were in the range 1–13 mm, that is small when compared to diameter 2–3 mm of protuberance. For small film thickness and ratio of height to diameter of protuberance the value of residual stress can be calculated from a dimension according to relation [26]:
r¼
pr 2 ; 6th
ð1Þ
where p - atmospheric pressure, t - thickness of film; r, h – radius and height of protuberance, respectively. Residual stress can be released by elastic deformations of the protuberance. The value of released stress can be estimated assuming the plane state of stress from Eq. (2):
rd ¼
0
E l l ; 1m l
ð2Þ
where E – Young’s modulus m – Poisson’s ratio of the film and l diameter of protuberance, l0 - total length of protuberance. In case of Ti film the parameters are: E = 110 GPa, m ¼ 0:36 [32], thus a value of released stress rd = 4.9 MPa, while for TiN the film parameters are: E = 250 MPa, m ¼ 0:2, rd = 3.4 MPa. In both cases the values were small and they were neglected.
Length = 3.4 mm Pt = 4.26 µm Scale = 5 µm
µm 4 3.5 3 2.5 2 1.5 1 0.5 0 -0.5 0
a)
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
b)
Fig. 2. The 3D-view – a; and the roughness profile - b of the Ti film deposited by PVD method.
2.4
2.6
2.8
3
3.2
3.4 mm
143
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Fig. 3. Microstructure of the Ti film on 304 steel substrate - a (cross-section); b – surface of the Ti film. Confocal microscope.
compressive residual stress determined according to formula 1, for mean value, 3.4 mm, of film thickness was 1.01 ± 0.64 GPa.
Table 2 Laser pulse energy, height h and radius r of the delamination and determined residual stress for the Ti film calculated according to formula 1. E [J]
h [mm]
r [mm]
Residual stress [GPa]
0.7 0.7 0.7 1 1 1 1.2 1.2 1.2 1.25 1.25 1.25
0.0026 0.0032 0.0026 0.0026 0.0132 0.0127 0.0013 0.0037 0.0049 0.0062 0.0125 0.006
0.78 0.84 0.76 0.82 1.36 1.34 1.99 1.11 0.96 1.1 1.5 1.03
1.15 1.07 1.09 1.17 0.69 0.7 1.36 1.51 0.93 0.95 0.92 0.86
Mean value
3.2.2. TiN films The same study was conducted for TiN film. Delamination was observed at the laser pulse energy level 1.2 J. The size values of the protuberance for tested samples are: h = 0.0031 mm and r = 1.3 mm. Shapes of the bulges were different then compered to Ti films. The bulges were more flat. The mean value of compressive residual stress determined according to formula 1 was 6.3 ± 0.6 GPa (Table 3). Fig. 6 shows the map and photograph of a defect on film protuberance. Cracking was not observed, only a small collapse with irregular shape was detected. Localized instabilities were observed by Jin et al. [33] in systems consisting of a thin stiff film on an elastomeric substrate. The ridge formations were observed when bilayer system was compressed. In our case the film was stiff and had high compressive residual stress but instability was limited to a small area and was detected only in one case. No collapse was observed for plastic Ti films that have much lower value of compressive residual stress. Therefore, it suggests that instability can be connected with variations of the TiN film properties, adhesion or residual stress.
1.01
Size of the protuberances was different for the same energy level but calculated value of the residual stress was very close. Table 2 shows the dimensions of the protuberances and the calculated value of the residual stress for the Ti film. During the shape analysis of protuberance the problems occurred in exact determination of a radius due to absence of sharp boundary between delaminated and not delaminated parts of the film as well as relatively high surface roughness. This is probably the main reason of differences in the values of determined residual stress. Another reason is heterogeneity of the film thickness that varies from 2.8 to 4 lm. The thickness of the film affects residual stress thus the value can vary in different areas of the film. The mean value of
Alpha = 59°
3.3. Residual stress analysis by XRD method The recorded diffraction patterns for the Ti thin film is shown in Fig. 7. The film was polycrystalline exhibiting reflections related to hexagonal structure. The diffraction patterns indicated strong tex-
µm
Beta = 22°
8 7.5 7 6.5 6 5.5 5
8.06 µm
4.5 4 3.5 3 2.5 2
3.4 mm 3.94 mm
µm
Length = 4.24 mm Pt = 3.58 µm Scale = 5 µm
5 4 3 2
1.5 1
1
0.5 0
0 0
a)
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
3
b)
Fig. 4. The 3D view – a, and profile after debonding – b; the Ti film, pulse energy 0.7 J. Scanning profilometer.
3.2
3.4
3.6
3.8
4
4.2 mm
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
100 101
10000
Fig. 5. The map with a profile of surface after debonding of the Ti film at laser pulse energy 1.2 J, 10 ns. Confocal microscope.
E [J]
h [mm]
r [mm]
Residual stress [GPa]
1.2 1.2 1.2 1.25 1.25 1.25
0.0032 0.0033 0.0029 0.0062 0.0051 0.0015
1.21 1.25 1.38 1.85 1.84 1.15
4.9 8.6 6.4 5.0 6.2 6.7
60
80
100
300
212 120
140
2θ [deg] Fig. 7. The X-ray diffraction pattern of deposited Ti film and austenite crystalline phases. The star symbol marks the reflection from the austenite substrate, while hkl indexes mark the hexagonal Ti phase (red color indicates the reflection used to determine residual stress). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
100000
222
10000
*
*
**
* *
220
* 200
ture. A preferred orientation was a crystallographic plane 100 parallel to the surface. The residual stress was determined by sin2w method, the analysis was performed based on 300 reflection. For calculations the following data was assumed: Young’s modulus E = 110 GPa, Poisson’s ratio m = 0.36 [32]. The value of residual stress was r = 0.86 GPa, Dr ±0.07 GPa. Fig. 8 shows the recorded XRD patterns for TiN film on steel substrate. The TiN film was polycrystalline and the reflections indicate a cubic structure. A strong texture was stated. Almost all crystallites have preferential direction and they are orientated according to 111 crystallographic plane that is parallel to surface. Such anisotropy of microstructure of film limits the ability to perform measurements and calculations of stress because it gives large error, especially for the TiN layer. Nevertheless, an attempt was made and for calculating the following data: E = 250 GPa,
3 3 3/ 5 1 1
6.3
Intensity [cps]
Mean value
40
**
111
Table 3 Laser pulse energy, height h and radius r of the protuberance, and residual stress values determined according to formula 1.
20
211
* 1000
* 202
*
110 103 200 112 201
002
Intensity [cps]
* 102
144
1000
20
40
60
80
100
120
140
2θ [deg] Fig. 8. The X-ray diffraction pattern for deposited TiN film and austenite crystalline phases. The star symbol marks reflections from the austenite substrate, while the hkl indexes denote cubic TiN phase (red color indicates the reflection used to determine residual stress). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 6. The map with profile of the protuberance - a, photograph of defect - b, TiN film after debonding with laser pulse energy 1.25 J, 10 ns. Confocal microscope.
145
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
m = 0.20 were assumed. Value of the residual stress was calculated based on reflection 222. Compressive residual stress was estimated: r = 7.2 GPa, Dr ±1.7 GPa. Due to strong texture the value is approximate. Both applied methods, LSP and XRD, give similar values of residual stress. The value of stress for the Ti film obtained by LST, is several times smaller than for the TiN film. A difference in the stress value is similar to that obtained by the XRD method. 3.4. Interface strength by LASAT The adhesion of the films to the substrate was determined based on measurement of film velocity during delamination. Fig. 9a shows typical recorded signals from VISAR system made by shock waves. The signals were measured on back surface of steel samples, then the velocity profile were estimated based on them (Fig. 9b) [30]. The second peak of velocity corresponds to reflected pressure wave. The velocity of sound inside steel plate, estimated from experiment, is about 5.76 km/s, thus it is equal to the velocity of sound in steel AISI 304 with approximation of 3 %.
In Fig. 10 the signals from VISAR system for samples covered by TiN films are visible. The surface speed is lower and strong attenuation of wave pressure can be observed. This effect is caused by differences in acoustic impedance of substrate and films. Fig. 11a shows the signals from VISAR system for samples 0.7 mm thick covered by the Ti film for which delamination was observed, while the calculated velocity profile is shown in Fig. 11b. In Fig. 12 the data for non-delaminated film are presented. The shock wave that reaches the interface is divided into two waves: a reflected wave and a transmitted one. In case of weak shock waves (when shock stress is much lower than Young’s modulus of material) linear approximation may be used [34,35]. In our case shock wave pressure is in the order of 1 GPa and E = 110 GPa, E = 250 GPa, respectively for the Ti film and for the TiN film. Therefore, energy division between reflected and transmitted waves depends on a value of ratio of acoustic impedances of contacting materials. Factors of transformation are described by equations [34]:
P¼
2A ; 1þA
R¼
A1 ; Aþ1
A¼
q2 c 2 ; q1 c 1
Fig. 9. a - The VISAR interferometer signals recorded during deformation of steel plate 1 mm thick, diameter of 10 mm, b-velocity profile. Pulse energy pulse 1.2 J.
Fig. 10. The VISAR interferometer signals recorded during delamination of the TiN film, steel plate 0.5 mm thick, 10 mm of diameter. Pulse energy 1.2 J.
ð3Þ
146
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
Fig. 11. The VISAR interferometer signals recorder during delamination of the Ti film; steel plate 0.7 mm thick, 10 mm diameter. Pulse energy 1.2 J.
Fig. 12. The VISAR interferometer signals recorded during deformation of steel plate (1 mm thick, 10 mm diameter) with the Ti film. Pulse energy 1.2 J.
where A – ratio of the acoustic impedances; P – relative amplitude of the wave transmitted from medium 1 to medium 2; R – relative amplitude of the wave reflected from material boundaries. . – film and substrate density, c – sound speed. In case of the Ti film on steel substrate the ratio of acoustic impedances is 4.5 (for steel 47 (g/cm3km/s). Relative amplitude of the wave reflected from material boundaries is R = 0.43 while the relative amplitude of wave transmitted from steel to film is R = 0.57. For the TiN film we obtained A = 2.4 (TiN 5.22 g/cm3, c = 3.8 km/s) [11], R = 0.41 and P = 0.59. In both cases acoustic impedances of films were less than substrate thus less than 60% of shock wave energy was transmitted to the films. In Fig. 10 weaker reflected wave from surface of substrate is visible. The delamination of film takes place during impacts of the first peak of pressure. Sharp changes in velocity profile are not visible. Relatively smooth course of velocity profile indicates on viscous failure of interface. In a model case when incident and reflected waves meets at interface the strength of interface can be estimated by simple relation [14,16,36]:
1 2
r ¼ qcDu
ð4Þ
where q – film density, c – sound speed, Du – surface velocity changes of the films during delamination.
The Du values for the Ti film were: 25, 50, 40 and 50 m/s. For the TiN films only two proper signals from ViSAR system were registered and the average value of velocity changes during delamination was Du = 58 m/s. The spall strength of interface, estimated from relation 4 for the TiN film, was 550 MPa and for the Ti film was 450 ± 53 MPa. The higher adhesion of the TiN film compared to the Ti film agrees with level of pulse energy that caused delamination. The delamination was observed at 0.7 J for Ti film while for the TiN film it occurred at 1.2 J. 4. Conclusions The study shows that laser spallation technique can be successfully used for estimating the value of residual stress of technological thin films deposited by PVD method. Until now there are no suitable methods to measure residual stress in thin films with poor crystallization or strong texture. This method is simple and quick. It allows obtain indicative value of compressive residual stress even for films with high roughness. The value of residual stress (6.3 GPa) for TiN films measured by the proposed method is similar to that obtained for films of comparable thickness deposited by PVD method on steel substrate. In the work [37,38] residual stress measured by X-ray diffraction was 8 GPa while the value obtained by Carvalho et al. [39] was 5 GPa for 5 lm thick film. On hard substrate a lower film stress, 3.58 GPa, was stated [40]. For Ti film the value of residual stress obtained by the LST was 1.01 GPa, whereas by the XRD it was 0.86 GPa. The results are comparable to stress value got by Leoni et al. [38] by the XRD, 1 GPa for 0.2 lm thick film on the same
J. Radziejewska et al. / Optics and Laser Technology 104 (2018) 140–147
substrate. Therefore, the LST method gives reliable value of compressive residual stress in tested thin films. The accuracy of the LST depends on precision of measurement of size of the bulge. Several techniques can be adopted, for example scanning profilometry or confocal microscope. For technological thin films with strong adhesion to metal substrate the high level of laser pulse energy is required. The thickness of substrate is limited and should be matched to laser pulse energy. The study shows that in case of technological films with good adhesion to steel substrate the LASAT can be successfully applied. The interface strength for the Ti films was 450 MPa whereas for the TiN film it was 550 MPa that is similar to that found by Zhou et al. [17] (371 MPa) on steel substrate. For more precise determination of the interface strength the future study should include analysis of reflected waves from the surface of film. The presented experimental results show that shock wave induced by nanoseconds laser pulse can be suitable new tool for determination of properties of technological thin films. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.optlastec.2018.02. 014. References [1] J.Z. Malinowski, J.R. Klepaczko, A unified analytic and numerical approach to specimen behavior in the SHPB, Int. J. Mech. Sci. 28 (1986) 381–391. [2] W. Moc´ko, Z.L. Kowalewski, Application of FEM in assessments of phenomena associated with dynamic investigations on miniaturised DICT testing stand, Kovove Mater. 51 (2013) 71–82. [3] B.J. Koeppel, G. Subhash, Characteristic of residual plastic zone under static and dynamic Vickers indentations, Wear 224 (1999) 56–67. [4] W. Kohlhofer, R.K. Penny, Dynamic hardness testing of metals, Int. J. Press. Vessel. Pip. 61 (1995) 65–75. [5] G.A. Askaryon, E.M. Moroz, Pressure on evaporation of matter in a radiation beam, JETP Lett. 16 (1963) 1638–1644. [6] N.C. Anderholm, Laser generated stress waves, Appl. Phys. Lett 16 (1970) 113– 118. [7] D.W. Gregg, S.J. Thomas, Momentum transfer produced by focused laser giant pulses, J. Appl. Phys. 27 (1966) 2787–2789. [8] C.H. Skeen, C.M. York, Laser-Induced ‘‘blow-off” phenomenon, Appl. Phys. Lett. 12 (1968) 369–371. [9] L. Berthe, R. Fabbro, P. Peyre, L. Tollier, E. Bartnicki, Shock waves from a water confined laser-generated plasma, J. Appl. Phys. 82 (1997) 2826–2832. [10] J. Radziejewska, Application of a nanosecond laser pulse to evaluate dynamic hardness under ultra-high strain rate, Optics Laser Technol. 78 (2016) 125– 133. [11] J. Wang, R.L. Weaver, N.R. Sottos, A parametric study of laser induced thin film spallation, Exp. Mech. 42 (1) (2002) 74–83. [12] J.L. Vossen, Adhesion Measurement of Thin Films, Thick Films and Bulk Coatings, ASTM STP640, 1978, 122–133. [13] I.G. Epishin, V.V. Suslov, V.A. Yanushkevich, Determination of adhesion strength of film structures of components in electronic devices using laser shock waves, Fizika i Khimiya Obrabotki Materialov 22 (5) (1988) 80–84. [14] V. Gupta, A.S. Argon, D.M. Parks, J.A. Cornie, Measurement of interface strength by a laser spallation technique, J. Mech. Phys., Solids 40 (1992) 141–147. [15] C. Tang, J. Zhu, The measurement of interface strength of TiN coating/substrate by laser spallation, Int. J. Refract. Hard Metals 14 (1996) 203–206.
147
[16] A.G. Youtsos, M. Kiriakopoulos, T. Timke, Experimental and theoretical/ numerical investigations of thin films bonding strength, Theor. Appl. Fract. Mech. 31 (1999) 47–59. [17] M. Zhou, Y.K. Zhang, L. Cai, Adhesion measurement of thin films by a modified laser spallation technique: theoretical analysis and experimental investigation, Appl. Phys. A 74 (2002) 475–480. [18] J. Yuan, V. Gupta, Measurement of interface strength by the modified laser spallation technique. I. Experiment and simulation of the spallation process, J. Appl. Phys. 74 (1993) 2388–2396. [19] Berthe, R. Fabbro, P. Peyre, E. Bartnicki, Wavelength dependent of laser shockwave generation in the water-confinement regime, J. Appl. Phys. 85 (11) (1999) 7552–7555. [20] M.F. Othman, A.R. Bushroa, Wan Normimi, Roslini Abdullah, Evaluation techniques and improvements of adhesion strength for TiN coating in tool applications: a review, J. Adhesion Sci. Technol. 29 (7) (2015) 569–591. [21] D. Courapied, R. Kromer, L. Berthe, P. Peyre, S. Costil, J. Cormier, M. Boustie, X. Milhet, Laser adhesion test for thermal sprayed coatings on textured surface by laser, J. Laser Appl. 28 (2) (2016), https://doi.org/10.2351/1.4944451. [22] V. Guipont, M. Jeandin, S. Bansard, K.A. Khor, M. Nivard, L.t. Berthe, J.-P. CuqLelandais, M. Boustie, Bond strength determination of hydroxyapatite coatings on Ti-6Al-4V substrates using the Laser Shock Adhesion Test (LASAT), J. Biomed. Mater. Res. A 95A (4) (2010) 1096–1104. [23] G. Bégué, G. Fabre, V. Guipont, M. Jeandin, P. Bilhe, J.Y. Guédou, F. Lepoutre, Laser Shock Adhesion Test (LASAT) of EB-PVD TBCs: Towards an industrial application, Surf. Coat. Technol. 237 (2013) 305–312. [24] B. Ehrhart, R. Ecault, F. Touchard, M. Boustie, L. Berthe, et al., Development of a laser shock adhesion test for the assessment of weak adhesive bonded CFRP structures, Int. J. Adhesion Adhesives 52 (2014) 57–65. [25] M. Arrigoni, J.-P. Cuq-Lelandais, M. Boustie, E. Gay, L. Berthe, An industrial challenge based on the wave propagation: the shock adhesion test, in: Wave propagation, Academy Publish Ed., WY (USA), 2012. [26] R. Ikeda, T. Uchiyama, H. Cho, T. Ogawa, M. Takemoto, An advanced method for measuring the residual stress of deposited film utilizing laser spallation technique, Sci. Technol. Adv. Mater. 7 (2006) 90–96. [27] B. Audoly, Mode-dependent toughness and the delamination of compressed thin films, J. Mech. Phys. Solids 48 (2000) 2315–2332. [28] Yu. Hong-Hui, John W. Hutchinson, Influence of substrate compliance on buckling delamination of thin films, Int. J. Fracture 113 (2002) 39–55. [29] http://www.surftec.pl/firma/powlekanie-pvd/ [30] W. Mocko, J. Radziejewska, A. Sarzyn´ski, M. Strzelec, J. Marczak, A. Rycyk, Analysis of the plastic deformation of AISI 304 steel induced by nanosecond laser pulse, Optics Laser Technol. 90 (2017) 165–173. [31] S.R.J. Saunders, H.R. Vetters, Standardisation of test methods for the mechanical properties of thin coatings, Thin Solid Films 299 (1997) 82–87. [32] http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=mq304a. [33] L. Jin, A. Takei, J.W. Hutchinson, Mechanics of wrinkle/ridge transitions in thin film/substrate systems, J. Mech. Phys. Solids 81 (2015) 22–40. [34] H. Kolsky, Stress waves in solids, (tłum. ros.), U.Rokmcrbq, Bokys yagpa;eybq d ndeplsx nekax, Bplanek]cndo Byocnpayyoq Kbnepanyps, Mocrda, 1955. [35] T. de Rességuier, On the shock-based determination of the adhesive strength at a substrate-coating interface, J. Adhesion Sci. Technol. (2017), https://doi.org/ 10.1080/01694243.2017.1332502. [36] F. Bauer, PVDF shock compression sensors in shock wave physics, CP706, Shock Compression of Condensed Matter, in: M.D. Furnish, Y.M. Gupta, J.W. Forbes (Ed.), 2003, pp. 1121–1126. [37] L. Zhang, H. Yang, X. Pang, K. Gao, A.A. Volinsky, Microstructure, residual stress, and fracture of sputtered TiN films, Surf. Coat. Technol. 224 (2013) 120– 125. [38] M. Leoni, P. Scardi, S. Rossi, L. Fedrizzi, Y. Massiani, (Ti, Cr)N and Ti/TiN PVD coatings on 304 stainless steel substrates: Texture and residual stress, Thin Solid Films 345 (2) (1999) 263–269. [39] N.J.M. Carvalho, E. Zoestbergen, B.J. Kooi, J.Th.M. De Hosson, Stress analysis and microstructure of PVD monolayer TiN and multilayer TiNy(Ti, Al)N coatings, Thin Solid Films 429 (2003) 179–189. [40] P.C. Jindal, A.T. Santhanam, U. Schleinkofer, A.F. Shuster, Performance of PVD TiN TiCN, and TiAlN coated cemented carbide tools in turning, Int. J. Refractory Metals Hard Mater. 17 (1999) 163–170.