Influence of in-situ and postannealing technique on tribological performance of NiTi SMA thin films

Influence of in-situ and postannealing technique on tribological performance of NiTi SMA thin films

Surface & Coatings Technology 276 (2015) 286–295 Contents lists available at ScienceDirect Surface & Coatings Technology journal homepage: www.elsev...

3MB Sizes 0 Downloads 18 Views

Surface & Coatings Technology 276 (2015) 286–295

Contents lists available at ScienceDirect

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

Influence of in-situ and postannealing technique on tribological performance of NiTi SMA thin films Wolfgang Tillmann ⁎, Soroush Momeni Institute of Materials Engineering, Technische Universität Dortmund, Leonhard-Euler-Str 2, 44227 Dortmund, Germany

a r t i c l e

i n f o

Article history: Received 15 May 2015 Revised 6 July 2015 Accepted in revised form 8 July 2015 Available online 14 July 2015 Keywords: Thin films NiTi Sputtering Annealing Tool steel Wear

a b s t r a c t Magnetron sputtered NiTi thin films were crystallized through two convenient techniques: (i) postannealing and (ii) in-situ annealing during the deposition. The annealing parameters (temperature and time) were kept constant by employing each technique. The microstructure, morphology, phase transformation behavior, mechanical and tribological properties of these thin films were investigated using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), 4-point probe resistivity measurement, nanoindentation test, pin-ondisc, scratch test and three dimensional (3D) optical microscopy. The results show how postannealing and insitu annealing techniques can differently affect properties of NiTi thin films in spite of employing similar annealing temperature and time. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Due to their unique properties such as shape memory effect (SME) and superelasticity (SE), NiTi shape memory alloy (SMA) thin films have been widely used in micro-to-millimeter scale microelectromechanical systems (MEMS) and miniature robotics. These films have the additional potential for batch fabrication. They are suitable candidates for microfabrication and integration in micro-miniature systems composed of actuators, sensors and mechanical elements deposited on one chip. NiTi SMA thin films can be heated resistively and they possess high power density and large actuation force. They are attractive materials to be used in microactuators, such as micropumps, microgrippers, microvalves, and micropositioners [1]. Previous studies shows capability of these thin films in thermomechanical data storage technology [2]. The load bearing capacity and superelasticity of NiTi SMA thin films make them ideal materials for depositing excellent tribological thin films [3] and cavitation erosion resistant coatings [4]. A standard technique for fabrication of NiTi thin films is sputtering deposition onto an unheated substrate. The as-deposited films are amorphous and SME and SE do not occur. An appropriate thermal annealing process is required in order to crystallize the films and obtain desired microstructure, shape memory effect and superelasticity. ⁎ Corresponding author. E-mail addresses: [email protected] (W. Tillmann), [email protected] (S. Momeni).

http://dx.doi.org/10.1016/j.surfcoat.2015.07.012 0257-8972/© 2015 Elsevier B.V. All rights reserved.

During this postannealing process, the film is heated above its crystallization temperature allowing the amorphous film to crystallize [5]. On the other hand, a sputtering onto a heated substrate leads to formation of crystallized NiTi SMA coatings without a need for a postannealing process [6]. This in-situ annealing process could be beneficial in terms of conserving thermal processing budgets because lower temperatures (above 350 °C) are required to obtain crystallized NiTi thin films with good shape memory effects [7]. It reduces the possibility of film-to-substrate reactions and NiTi oxidation. By employing in-situ annealing technique, it is possible to deposit a protective layer on NiTi SMA thin films in only one-shot coating process [8]. In recent years, many high quality research works performed separately on post and in-situ annealed NiTi thin films which provide valuable information to show a link between annealing parameters and film microstructure, morphology, phase transformation and mechanical properties [9–14]. For example, the authors of this paper have recently reported the microstructure development of NiTi thin films during insitu annealing at various temperatures [15]. In spite of these well performed research works, a systematic comparison of postannealed and in-situ annealed NiTi films has not been investigated up to now. During postannealing process, the crystallization develops from the amorphous NiTi film while during in-situ annealing the crystallization development occurs simultaneously with NiTi film deposition. Different modes of crystallization results in different microstructure which can drastically affect phase transformation behavior, mechanical, physical and tribological properties of NiTi SMA thin films. As a result, it is worthwhile to investigate the influence of the post and in-situ annealing techniques

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

on the properties of NiTi SMA thin films. Such an investigation could be even more interesting when examines the effect of annealing techniques on tribological performance of NiTi thin films. The link between the sort of annealing technique and tribological characteristics of NiTi thin films has not been investigated up to now. Tribo-mechanical behavior of these thin films is an important issue of concern. It directly affects the service life and functionality of NiTi coating materials. The aim of the present paper is to analyze mechanical and tribological behavior of NiTi SMA thin films deposited under same processing parameters and subjected to same annealing parameters (duration and temperature) but different annealing types: postannealing and in-situ annealing. Within the current study, the term of in-situ annealing refers to annealing during the sputtering which results in formation of crystallized coatings at the end of the deposition process (growth of films at raised temperature onto heated substrates). The postannealing or ex-situ annealing term implies postdeposition heat treatment following film growth at room temperature. The obtained results present an appropriate annealing mode for fulfilling the requirements put forth by the development of durable devices based on NiTi SMA thin films. 2. Experimental The deposition of NiTi thin films was carried out on plasma nitrided hot work tool steel (X38CrMoV51) substrates by means of an industrial magnetron sputtering device (Cemecon MLsinox800, Germany). The plasma nitriding process of the substrates is explained in these references [16,17]. The size of the sputtering targets was 200 mm × 88 mm. NiTi alloy targets (Ni-51.82 at.%Ti) with high purity (99.99%) were employed at sputtering power of 1400 W for deposition of near equiatomic NiTi thin films. During in-situ annealing treatment, the NiTi thin films were sputtered for 3600 s on the heated substrate at the temperatures of 425 °C and 525 °C. The post annealing treatment was performed on as-deposited thin films after breaking the vacuum inside of the coating chamber. In this case, the deposition at the heating power of 0 W was performed without an intentional heating of the substrates. However, using the thermocouple, some substrate heating was detected around 80 °C during the deposition process. By post annealing treatment, NiTi thin films were annealed for 3600 s at the temperatures of 425 °C and 525 °C inside of the coating chamber and under the same vacuum conditions with in-situ annealing treatments. All substrates were ultrasonically cleaned in an ethanol bath for 15 min prior to deposition. The sample holder was rotated on a horizontal table during the sputtering process in order to achieve a uniform film composition. The target-to-substrate distance was fixed at approximately 9.5 mm. All coating parameters were maintained constant during the deposition of the NiTi thin films. The parameters as well as the thickness of the coatings are listed in Table 1. The phase structure and crystallographic planes of the thin films were analyzed by employing X-ray diffraction, using a Cu Kα radiation and a 9° incident angle (D8 Advance, BRUKER AXS, and Germany). The morphology and thickness of the thin film were analyzed by means of a field emission scanning electron microscope (Joel JXA840, JSM 35, Japan) on a fracture cross-section of the coated samples. As it Table 1 Processing parameters and thickness of the deposited NiTi thin films. Substrate bias voltage Gas Argon flow rate Sputtering chamber pressure Heating power Substrate pre etching time Substrate rotation speed Thickness of postannealed films at 425 °C Thickness of in-situ annealed films at 425 °C Thickness of postannealed films at 525 °C Thickness of in-situ annealed films at 525 °C

−75 V Argon 320 ml/min 350 mPa 1000 W 3600 S 5 rpm 3 μm 3.3 μm 3 μm 3.1 μm

287

was mentioned in Table 1, the thickness of the deposited thin films was about 3 μm. The composition of the coatings was determined using an energy-dispersive X-ray spectrometry (EDX) with an electron acceleration voltage of 20 kV, and a beam current of 15 nA. All samplings were conducted by analyzing an area and not a point composition. The phase transformation behavior was characterized by four-point probe electrical measurement during thermal cycling at a heating/cooling rate of 5 K-min− 1. The nanoindentation tests were performed by a depth-controlled nanoindenter XP (MTS Nano instrument, USA) with a penetration depth of 10% of the coating thickness in order to minimize the effect of the substrate hardness on the measurement. During nanoindentation measurements, a Berkovic tip, made of a conductive diamond, was employed. The method used in this research work is called “G-series CSM Hardness, Modulus for Thin Films.” The indenter penetrates the surface at a rate determined by the strain rate target of (0.05 1/S). When the surface penetration reaches the depth limit (300 nm), the load on the indenter is held constant for 100 s. The indenter is then partly withdrawn from the sample at a rate equal to the maximum loading rate. When the load on the sample reaches 10% of the maximum load on the sample, the load on the sample is held constant for 100 s. The indenter is then completely withdrawn from the sample which is moved into position for the next indentation. For each sample, 49 indentations (7 × 7 square arrays) were performed. The ball-on-disc tests were performed to examine the wear resistance of the coatings. During this test, a ball with a diameter of 6 mm, made of Al2O3, was employed as a counterpart for the coated rotating discs. The track radius and normal force were adjusted to 5 mm and 1 N for 7500 rotations. The cross-section areas of the wear tracks were measured at four different points by means of an Alicona 3D-analyzer (infinite focus) to calculate the average volume loss of the coating for each sample and the corresponding standard deviation. The specific wear rate defined as the volume loss per sliding distance per applied normal load. To evaluate the adhesion of the coatings, a micro-scratch test was performed using the single pass scratch mode with a diamond stylus topped as a conical. During the test, the force was increased linearly up to 100 N. Scratch tests were performed four times on each sample to calculate the arithmetic median and standard deviation. 3. Results 3.1. Composition and microstructure The compositions of the in-situ and postannealed NiTi thin films are mentioned in Table 2. The atomic ratios of Ni to Ti in the samples are all close to 1, indicating their ability to present shape memory effect and superelasticity around room temperature. The as-deposited and insitu annealed thin films were not oxidized. However, breaking the initial vacuum and performing postannealing treatment inside of the coating chamber resulted in minor oxidation of the deposited thin films. It occurred in spite of the fact that the same vacuum conditions were employed inside of the coating chamber during post and in-situ annealing processes. This, in fact, shows the difficulty in preventing the oxidation phenomenon by employing post deposition annealing treatments. The XRD pattern of the in-situ annealed thin films at 425 °C is shown in Fig. 1. The sharp peak with high intensity at 2θ of 42.3° corresponds with (110) diffraction pattern for B2 austenite phase of NiTi thin films. The other major diffraction peaks of this phase, (200) and (211), were detected at 2θ of 62° and 78.2°, respectively. The XRD pattern of NiTi thin films deposited at 525 °C is shown in Fig. 2. Although this XRD pattern presents the existence of dominantly B2 austenite phase, the intensity of (110) diffraction pattern is up to 30% lower than that observed in the XRD spectra of deposited films at 425 °C. Generally speaking, the intensity of XRD pecks can be lowered because of smaller degree of crystallization (less amounts of well crystallized materials) and/or overlapping of several diffraction patterns. Increasing the temperature accelerates the degree of crystallization [14] and therefore, the

288

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

Table 2 The compositions of the deposited NiTi films recorded by EDX. Element

Composition of in-situ annealed coatings at 425 °C (at.%)

Composition of in-situ annealed coatings at 525 °C (at.%)

Composition of postannealed coatings at 425 °C (at.%)

Composition of postannealed coatings at 525 °C (at.%)

Ni Ti O

50.65 49.35 –

50.61 49.39 –

42.23 40.85 16.92

42.92 41.11 15.97

reduction in intensity of (110) is a consequence of an overlapping with diffraction patterns of Ni4Ti3 precipitations. These precipitations were enlarged and further crystallized by increasing in-situ annealing temperature. Inset of Fig. 1 shows the magnified XRD patterns of NiTi thin films deposited at 425 °C. As it can be seen, there is a weak and broad peak at 2θ = 37.8°, corresponding with (131) diffraction pattern of Ni4Ti3 precipitations. The major diffraction peak of Ni4Ti3 precipitations was indexed as (122) at 2θ = 43.1° which is quite close to (110) diffraction pattern of NiTi B2 austenite phase. The magnified XRD spectrum of NiTi thin films deposited at 525 °C is shown inset of Fig. 2. It is obvious within this spectra that the peak intensity of (131) diffraction pattern for Ni4Ti3 precipitations was increased and also new diffraction patterns of these precipitations were appeared at 2θ of 49.4° indexed as (312) and 55° indexed as (232). The XRD diffractogram reveals acceleration of non-equilibrium precipitation reaction of Ni4Ti3 phases by increasing the in-situ annealing temperature. This finding during in-situ annealing of NiTi thin films corresponds with the research work of Zhang et al. on postannealing of NiTi thin films at various temperatures [9]. They reported that increasing the postannealing temperature rises the crystallization degree of NiTi thin films and cause enlargement of the content of precipitation phases. D profiles of the NiTi thin films postannealed and in-situ annealed at 425 °C are shown concurrently in Fig. 3. It can be seen that the postannealed NiTi thin films at 425 °C were not properly crystallized and have mostly amorphous microstructure. On the other hand, insitu annealing of NiTi thin films at 425° resulted in full crystallization of these thin films. This discrepancy clearly evidences the advantage of in-situ annealing process in faster formation of well crystallized coating materials. Fig. 4 simultaneously shows the XRD patterns of NiTi thin films post and in-situ annealed at 525 °C. The XRD patterns of these thin films presented sharp peaks, indicating their well crystallized microstructure during the postannealing and in-situ annealing treatments at 525 °C However, the peak intensity of (110) diffraction pattern for postannealed thin films at 525 °C is smaller than that observed in the XRD pattern of in-situ annealed thin films at this temperature.

According to the XRD spectra of post and in-situ annealed thin films at 425°, the crystallization process happens slower within postdeposition thermal annealing treatment compared with in-situ annealing. Thus, it is deducible to assume that lower intensity of (110) peaks in the XRD spectrum of postannealed coatings at 525° is a consequence of lower degree of crystallization and/or less amounts of well crystallized coating materials within these thin films. In terms of the qualitative analysis of crystallization reaction kinetics, the reaction rate is related to the activation energy (Q) and the temperature (T) in the Arrhenius exponential relationship, i.e., proportional to exp (− Q/KT), where K is the Boltzmann constant [18]. Annealing during sputtering (in-situ annealing) will speed up the inter-diffusion of the sputtered atoms at specific temperature to overcome the activation energy of thin film crystallization that is lower than in the conventional postannealing treatments. In other words, the activation energy of the reaction will be reduced by employing in-situ annealing treatment. This reduced activation energy results in a faster crystallization and also faster precipitation formation within the coating materials. By employing post-annealing technique, more thermal energy or higher atomic mobility at a higher temperature is required to initiate the crystallization (nucleation and growth) process. FESEM images of the deposited thin films in Fig. 5 clearly show differences on surface morphology of the deposited thin films in spite of the employing similar annealing conditions. The postannealed thin films at 425 °C present agglomerated hemispherical surface grains that were uniformly distributed across the surface while the in-situ annealed thin films at 425 °C shows well crystallized polyhedral surface grains. The postannealing of thin films at 525 °C leads to expanding and uniform distribution of small surface grains as a consequence of facing to a proper crystallization temperature. The in-situ annealed thin films at this temperature present again polyhedral surface grains that were agglomerated in some areas.

Fig. 1. X-ray diffraction patterns of the in-situ annealed NiTi thin films at 425 °C.

Fig. 2. X-ray diffraction patterns of the in-situ annealed NiTi thin films at 525 °C.

3.2. Phase transformation It is known that the total resistivity of crystalline metallic materials is the sum of the resistivity due to the electron scattering by phonons and

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

289

was reported by Kumar et al. [19] for RT plot of NiTi thin films sputtered at a substrate temperature of 350 °C. They assumed that such phenomenon could be attributed to the following reasons: (a) presence of intrinsic defects; and (b) Austenite ↔ R-phase transition. A large hysteresis can be clearly seen upon heating and cooling in the RT plot of postannealed coatings at 525 °C. This represents low quality of the phase transformation behavior perhaps duo to the minor oxidation of these coatings which causes a reduction in volume of the material available for phase transformation. On the other hand, the in-situ annealed coatings present a narrow hysteresis upon heating and cooling. The existence of such a narrow hysteresis interestingly corresponds with differential scanning calorimetry (DSC) measurements of these coatings which were previously reported by the authors of this paper [15]. More information about phase transformation behavior of these in-situ annealed coating is presented in Ref. [8] and [15]. 3.3. Nanoindentation Fig. 3. X-ray diffraction patterns of the in-situ and postannealed NiTi thin films at 425 °C.

the resistivity due to the electron scattering in lattice imperfections and impurities [19]. Since each crystallographic structure has its own electrical properties due to its own lattice imperfection, measuring electrical resistance upon variation in temperature could be an effective method for detecting phase transformation temperatures in shape memory alloys. The postannealed NiTi thin films at 525 °C as well as the in-situ annealed NiTi thin films at 425 °C and 525 °C were well crystallized during the thermal annealing process and are therefore, capable of presenting phase transformation behavior. It was concluded by analyzing the resistances versus temperature (RT) plots of in-situ and postannealed NiTi films at 425 °C and 525 °C (Fig. 6). The RT plot of the post annealed thin films at 425 °C doesn't show any sign of phase transformation behavior within the coating materials. It demonstrates higher resistivity than the other coating due to its dominantly amorphous microstructure which was earlier found in the XRD analysis (Fig. 3). The atomic scale disorder present in an amorphous structure causes its electrical conductivity to be lower than the conductivity of the corresponding crystalline metal. This is due to the susceptibility of the structural disorder to impede the motion of the mobile electrons. Electrical resistance of metallic materials shows a linear behavior during temperature changes (it increases upon heating and decreases upon cooling). As it can be seen, the RT plots of the postannealed thin films at 525 °C as well as in-situ annealed thin films at 425 °C and 525 °C show a nonmetallic behavior. However, the observed trend in resistivity changes upon changing temperature is a clear evidence for the presence of phase transformation behavior within these coatings. A similar result

An interesting contribution to the investigation of thermal annealing treatment is to investigate the effect of the sort of annealing technique on mechanical properties of thin films. The obtained numerical values for hardness and Young's modulus of the deposited films are mentioned inset of Fig. 7. Fig. 7a and b shows representative load–displacement curves for nanoindentation into NiTi thin films post and in-situ annealed at 425 °C, respectively. Due to the higher hardness of the postannealed thin films, larger force is required to reach the maximum indentation depth. This corresponds with the obtained numerical values for hardness and Young's modulus of the deposited films. It was shown in the XRD analysis that the in-situ annealed films were well crystallized and possess dominantly austenite phase while the postannealed thin films at 425° are mainly amorphous. The crystalline NiTi thin films are softer than the amorphous films. Limei [4], X. Cao et al. [20], and the authors of this study [21] have already reported the capability of B2 austenite NiTi thin films in lowering the hardness values within the nanoindentation tests. This is a consequence of their inherent superelastic effect, originated from isothermal phase transformation from austenite to martensite (upon loading) and vice versa (upon unloading). The load displacement curves of the post and in-situ annealed thin films at 525 °C were shown in Fig. 7c and d. The calculated numerical values for hardness of these thin films are somehow close to each other while the Young's modulus of in-situ annealed thin films is considerably higher than that for post annealed thin films. This could be explained on the basis of the formation of Ni4Ti3 precipitations during in-situ annealing of the coatings at 525 °C. According to the XRD analysis, different crystallization evolution of post and in-situ annealed thin films at 525 °C resulted in a faster Ni4Ti3 precipitation formation within the in-situ annealed coating materials. Since Ni4Ti3 precipitates possess a higher elastic stiffness than the surrounding parent B2 matrix [22], it can effectively increase the Young's modulus values in nanoindentation response of in-situ annealed NiTi thin films at 525 °C. 3.4. Adhesion

Fig. 4. X-ray diffraction patterns of the in-situ and postannealed NiTi thin films at 525 °C.

The sort of annealing technique, either post or in-situ annealing, can drastically affect the adhesion of NiTi thin films to their substrates. The minimum load at which an adhesive failure happens is called critical load (LC) and is representative of the coating to substrate adhesion. The higher critical load represents the larger adhesion strength between the coating and substrates. A comprehensive explanation for calculating the critical load of the NiTi coatings was reported earlier in Ref. [23]. Fig. 8 shows the calculated numerical values for critical load of the deposited thin films. This result obviously illustrates the poorer adhesion of the post annealed NiTi thin films compared to the in-situ annealed ones. Interfacial bonding between the coating and substrate is known as one of the key parameters affecting the coating adhesion [24]. By employing in-situ annealing technique, the high temperature of the

290

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

Fig. 5. FESEM images from the surface of the NiTi thin films (a) postannealed at 425 °C (b) in-situ annealed at 425 °C (c) postannealed at 525 °C (d) in-situ annealed at 525 °C.

substrate enhances the inter-diffusion between the sputtered atoms and the substrates. This, in turn, could strength the interfacial bonding between the coating and substrate which results in improvement of the coating to substrate adhesion. Furthermore, the thermal annealing treatment can adversely affect the residual stress of the coatings [25]

and subsequently their adhesion strength to the substrates [26]. During in-situ annealing, NiTi atoms arrive on a heated substrate and have more thermal energy to rearrange themselves and relieve the stress [27]. It is perceptible to assume that the lower residual stress of the in-situ annealed thin films is another contributing factor to their higher

Fig. 6. Resistivity plots of the NiTi thin films (a) postannealed at 425 °C (b) in-situ annealed at 425 °C (c) postannealed at 525 °C (d) in-situ annealed at 525 °C.

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

291

Fig. 7. Load–displacement plots of the NiTi thin films (a) postannealed at 425 °C (b) in-situ annealed at 425 °C (c) postannealed at 525 °C (d) in-situ annealed at 525 °C.

adhesion strength when compared to the post annealed thin films. However, the residual stress of the coatings is a complex phenomenon and is not the scope of this research work. In order to have a more precise analysis, the scratches morphologies were analyzed by means of optical and electron microscopes. Generally speaking, various failure modes are observed during the scratch test such as, coating detachment, through thickness cracking and plastic deformation or cracking in the coating or substrate. What happens usually is simultaneous occurrence of several different failure modes which makes the result of the test difficult to interpret. The coating–substrate

Fig. 8. Critical load of the in-situ and postannealed NiTi thin films at 425 °C and 525 °C.

duo behaves as a single mechanical system that faces to deformations created by external loads under service conditions [28], in scratch test, the action of diamond stylus. By concerning this point, mechanical failure of the duo can be originated from one or both of these two factors: (i) Cohesive failure (failures in the coatings or in the substrate) which occurs by tensile stress behind the stylus. (ii) Adhesive failure (failures at the interface of the coating–substrate duo) as a consequence of compressive stress. It results in separation of the coating from the substrate either by cracking and lifting (buckling) or by full separation (spallation; chipping). The scratch test is most effective if the substrate does not plastically deform to any great extent. In such cases, the coating is effectively detached and the uncovering of substrate itself can be used as a guide to adhesion [29,30]. The most reproducible scratch morphology results could be obtained from the end of the scratch lines because the scratch faces to the highest progressive load values of the stylus in this area. Fig. 9 displays the optical micrographs from the end of the scratch lines of the post and in-situ annealed NiTi thin films. The postannealed thin films at 425 °C (Fig. 9a) and 525 °C (Fig. 9c) show combination of both of cohesive and adhesive failures. Within the scratch grooves, they are straight micro-cracks that are open to the direction of the scratch and formed behind the stylus. Such brittle tensile cracking is known as chevron cracks. These figures obviously illustrate the adhesive failure of the postannealed thin films due to the large detached regions of the coatings around the scratch lines. Such a damage feature is known as gross spallation and is common in coatings with low adhesion strength. The higher adhesion strength of the in-situ annealed thin films could be also concluded from their optical micrographs in Fig. 9b and d. In these figures, there is no coating spallation around the end of scratch lines. Thus, the in-situ annealed thin films didn't face to any adhesive failure. Within the end of scratch line of the in-situ annealed NiTi coatings at 425 °C, the observed feature damage of the cohesive failure corresponds to chevron cracks as a

292

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

Fig. 9. Optical micrographs at the end of the scratch lines of the NiTi thin films (a) postannealed at 425 °C (b) in-situ annealed at 425 °C (c) postannealed at 525 °C (d) in-situ annealed at 525 °C.

consequence of brittle tensile cracking across the substrate. The end of the scratch line of the in-situ annealed thin films at 525 °C doesn't show any specific damage feature or cracks. Fig. 10a and c shows low magnification FESEM images of the scratch lines on surface of the postannealed NiTi thin films. The poor adhesion strength of these thin films is quite apparent whereas the coating delamination occurred very fast after initiation of the scratch test. Fig. 10b and d illustrates FESEM images from the failure points of the

in-situ annealed thin films at 425 °C and 525 °C, respectively. As it can be seen, the scratch line after the failure points of the in-situ annealed thin films is even and smooth without any sign of large damages or cracks within the exposed substrate area. The same morphology was also found at the end of the scratch line of in-situ annealed thin films at 525 °C which could be attributed to higher adhesion strength of the in-situ annealed thin films at this temperature. The failure point of the in-situ annealed thin films at 525 °C was approximately at the middle

Fig. 10. Low magnified FESEM from the failure points of NiTi thin films (a) postannealed at 425 °C (b) in-situ annealed at 425 °C (c) postannealed at 525 °C (d) in-situ annealed at 525 °C.

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

of the scratch line. Since the coating materials could sustain properly during the scratch test, the progressive loading of the stylus couldn't adversely damage the substrate materials. 3.5. Wear The 3D optical micrographs from the worn surfaces of the postannealed thin films are shown in Fig. 11. As it can be seen, these films present an abrasive wear behavior since the wear track is smooth, the coating materials were evenly removed by the wear ball and the substrate were exposed. This is due to the high brittleness and weak coating-to-substrate adhesion of these thin films. The in-situ annealed NiTi thin films present a dominant wear mechanism of adhesive wear probably due to their well crystallized microstructure and high adhesion strength. Plastic deformation of the NiTi thin films can be clearly observed within the 3D surface topography of the worn surfaces for these thin films in Fig. 12. It is quite obvious in the worn surfaces that the wear topography is not so homogeneous and the coating materials were not removed evenly by the wear ball. In some areas the coating materials were even not detached away and they are nearly intact. The specific wear rates of the deposited thin films were shown in Fig. 13. From tribological point of view, this result clearly shows the advantage of in-situ annealing over post-annealing technique. In order to have a precise interpretation for the wear performance, the result of the pin-on-disc test were integrated by the mechanical properties of thin films obtained by nanoindentation test. Generally speaking, there is a close relationship between mechanical properties of the coatings and their wear resistance, particularly; the ratio of hardness to Young's modulus (H/E). This ratio is known as one of the key parameters controlling wear. According to the previous studies, the coatings with larger H/E have better wear resistance [3]. However, a comparison of the wear behavior between the post and in-situ annealed films at the same temperature shows an incapability of this assumption for analyzing wear performance of NiTi thin films. Irrespective of smaller H/E values of the in-situ annealed thin films at 425 °C and 525 °C, they have significantly a better wear resistance. The specific wear rates of the in-situ annealed thin films are much smaller than the postannealed thin films. The poor wear resistance of the postannealed films at 425° is due to their amorphous microstructure which is unable to present

293

superelasticity as a consequence of load-induced martensitic phase transformation. The in-situ annealed thin films at 425 °C were sufficiently crystallized. The superelasticity of these films can enhance their load bearing capacity and therefore, their wear resistance [31]. In contrast to the thermal annealing treatment of NiTi thin films at 425 °C, the thermal annealing treatments at 525 °C could properly crystallize the post and in-situ annealed thin films and resulted in formation of B2 austenite phase. Although these thin films are capable of presenting superelasticity and consequently a good wear resistance, their wear performances are significantly different. The specific wear rate of the postannealed thin films at 525 °C is considerably larger than the wear rate of the in-situ annealed thin films at the same temperature. There are two assumptions for excellent wear resistance of in-situ annealed thin films at 525 °C compared to the postannealed ones. The first assumption is about the existence of Ni4Ti3 precipitations. These precipitations are known as an effective factor for strengthening the matrix phase and thus improving the resistance of parent phase (B2 austenite phase) to deformation [20]. The second assumption is based on the close relation between the coating-to-substrate adhesions and wear resistance of the coatings. Increasing the coating-to-substrate adhesion is known as an important factor for improving wear performance of the coatings [32]. The excellent wear resistance of in-situ annealed thin films at 525 °C could be attributed to their higher adhesion strength to the substrates, as it was mentioned in Section 4. An interesting point here is that the specific wear rate of the postannealed films at 525 °C with well crystallized microstructure is even larger than the postannealed thin films at 425° with dominantly amorphous microstructure. This poor wear resistance could be ascribed to the higher residual stress as well as internal imperfections which were induced by increasing postannealing temperature from 425 °C to 525 °C. 4. Conclusion The mechanical and tribological behavior of NiTi thin films were investigated after post and in-situ annealing under the same annealing conditions (duration and time). It was found that the in-situ annealed NiTi SMA thin films have much better quality in terms of tribomechanical performance and shape memory behavior. The highlighted results are listed as follow.

Fig. 11. 3D optical micrograph of the worn tracks on the surface of the postannealed NiTi thin films (a) postannealed at 425 °C (b) postannealed at 525 °C.

294

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295

Fig. 12. 3D optical micrograph of the worn tracks on the surface of the postannealed NiTi thin films (a) In-situ annealed at 425 °C (b) In-situ annealed at 525 °C.

1. Crystallization reaction rate is much higher during in-situ annealing treatment which leads to faster formation of well crystallized NiTi SMA thin films. 2. Employing in-situ annealing technique could effectively prevent oxidation of NiTi coatings. Lack of oxidation and large degree of crystallization enables in-situ annealed coatings to present better quality of the phase transformation behavior (narrow hysteresis in the transformation temperatures). 3. Due to their well crystallized microstructure, hardness and Young's modulus of the in-situ annealed thin films at 425 °C is much smaller than the postannealed films at this temperature. The hardness of post and in-situ annealed thin films at 525 °C is quite close to each other. The Young's modulus of the in-situ annealed thin films is higher due to the formation of Ni4Ti3 precipitations. 4. The adhesion of postannealed thin films to the substrates is really poor due to their higher residual stress. The in-situ annealed thin

Fig. 13. Correlation between mechanical properties and tribological performance of the post and in-situ annealed NiTi thin films.

films have a remarkable coating-to-substrate adhesion due to the inter-diffusion of the sputtered atoms and the substrate as well as the lower residual stress. 5. The wear resistance of the in-situ annealed thin films is much better than the postannealed films due to their better adhesion to the substrates and their crystalline microstructure. References [1] Y. Fu, H. Du, W. Husang, S. Zhang, M. Hu, TiNi-based thin films in MEMS applications: a review, Sensors Actuators A Phys. 112 (2004) 395–408. [2] Wendy C. Crone, Gordon A. Shaw, Applying NiTi shape-memory thin films to thermomechanical data storage technology, MRS Proc. 855 (2004) W1.7, http:// dx.doi.org/10.1557/PROC-855-W1.7. [3] W. Ni, Y.T. Cheng, M. Luktisch, A.M. Weitner, L.C. Lev, D.S. Grummon, Novel layered tribological coatings using a superelastic NiTi interlayer, Wear 259 (2005) 842–848. [4] L. Yang, Cavitation Erosion Resistance of NiTi Thin Films Produced by Filtered Arc DepositionDoctor of Philosophy Thesis School of Mechanical, Materials and Mechatronic, University of Wollongong, 2010. (http://ro.uow.edu.au/theses/3220). [5] J.J. Kim, P. Moine, D.A. Stevenson, Crystallization behavior of amorphous NiTi alloys prepared by sputter deposition, Scr. Metall. 20 (1986) 243–248. [6] J.Z. Shi, C.Z. Chen, X. Dang, Magnetron sputtering applied in shape memory alloys preparation, Appl. Mech. Mater. 66–68 (2011) 882–887. [7] Y.Q. Fu, H.J. Du, Effects of film composition and annealing on residual stress evolution for shape memory TiNi film, Mater. Sci. Eng. A 342 (2003) 236. [8] W. Tillmann, S. Momeni, Deposition of superelastic composite NiTi based films, Vacuum 104 (2014) 41–46. [9] L. Zhang, C. Xie, J. Wu, Effect of annealing temperature on surface morphology and mechanical properties of sputter-deposited Ti–Ni films, Alloys Compd. 427 (2007) 238–243. [10] A. Kumar, S.K. Sharma, S. Bysakh, S.V. Kamat, Effect of substrate and annealing temperatures on mechanical properties of Ti-rich NiTi films, J. Mater. Sci. Technol. 26 (11) (2010) 961–966. [11] G. Satoh, A. Birnbaum, Y.L. Yao, Effect of annealing parameters on the shape memory properties of NiTi thin films, in: ICALEO 2008 Congress Proceedings, Poster presentation gallery, 2008. 100–167. [12] J. Birnbaum, U.-J. Chung, X. Huang, J.S. Im, A.G. Ramirez, Y.L. Yao, Substrate temperature effect on laser crystallized NiTi thin films, J. Appl. Phys. 105 (2009) 073502. [13] S.K. Sharma, H.S. Vijaya, S. Mohan, Influence of substrate temperature and deposition rate on structural and mechanical properties of shape memory NiTi films, Phys. Procedia 10 (2010) 44–51. [14] Xu Huang, J. San Juan, A.G. Ramirez, Evolution of phase transformation behavior and mechanical properties with crystallization in NiTi thin films, Scr. Mater. 63 (1) (Jul. 2010) 16–19. [15] W. Tillmann, S. Momeni, In-situ annealing of NiTi thin films at different temperatures, Sensors Actuators A Phys. 221 (1) (January 2015) 9–14.

W. Tillmann, S. Momeni / Surface & Coatings Technology 276 (2015) 286–295 [16] W. Tillmann, E. Vogil, S. Momeni, Improvemt of press dies used for the production of diamond composites by means of DOUPLEX-PVD coatings, Surf. Coat. Technol. 205 (5) (2010) 1571–1577. [17] W. Tillmann, S. Momeni, F. Hoffmann, A study of mechanical and tribological properties of self-lubricating TiAlVN coatings at elevated temperatures, Tribol. Int. 66 (2013) 324–329. [18] M. Ståhl, B.L. Åslund, A.C. Rasmuson, Reaction crystallization kinetics of benzoic acid, AICHE J. 47 (2001) 1544–1560. [19] A. Kumar, D. Singh, D. Kaur, Grain size effect on structural, electrical and mechanical properties of NiTi thin films deposited by magnetron co-sputtering, 203 (12) (Mar. 15 2009) 1596–1603, http://dx.doi.org/10.1016/j.surfcoat.2008.12.005. [20] X.H. Cao, X.L. Cao, Q. Zhang, Nanoscale indentation behavior of pseudo-elastic Ti–Ni thin films, J. Alloys Compd. 465 (2008) 491–496. [21] W. Tillmann, S. Momeni, Comparison of NiTi thin films sputtered from separate elemental targets and Ti-rich alloy targets, J. Mater. Process. Technol. 220 (2015) 184–190. [22] M.F.-X. Wagner, W. Windl, Elastic anisotropy of Ni4Ti3 from first principle, Scriptamaterialia 60 (2009) 207–210. [23] W. Tillmann, S. Momeni, Tribological performance of near equiatomic and Ti-rich NiTi shape memory alloy thin films, Acta Mater. 92 (15) (June 2015) 189–196. [24] D.M. Mattox, Surface effects on the growth, adhesion and properties of reactively deposited hard coatings, Surf. Coat. 81 (1996) 8–16, http://dx.doi.org/10.1016/ 0257-8972(95)02652-5.

295

[25] M.S. Ahmed, X. Zhao, Z.F. Zhou, P. Munroe, N. Chen-Tan, L.K.Y. Li, Z. Xie, Effect of thermal annealing upon residual stress and mechanical properties of nanostructured TiSiN coatings on steel substrates, J. Am. Ceram. Soc. 94 (2011) 1546–1551, http://dx.doi.org/10.1111/j.1551-2916.2010.04280.x. [26] Y. Fu, H. Du, C.Q. Sun, Interfacial structure, residual stress and adhesion of diamond coatings deposited on titanium, Thin Solid Films 424 (2003) 107–114. [27] J.J. Gill, D.T. Chang, L.A. Momada, G.P. Garman, Manufacturing issues of thin film NiTi microwrapper, Sensors Actuators 63 (2001) 148–156. [28] R. Viana, A.R. Machado, Influence of adhesion between coating and substrate on the performance of coated HSS twist drills, J. Braz. Soc. Mech. Sci. Eng. Rio de Janeiro 31 (4) (2009) 327–332. [29] S.J. Bull, E.G. Berasetegui, An overview of the potential of quantitative coating adhesion measured by scratch testing, Tribol. Int. 39 (2006) 99–114. [30] As Tm C1624-05, Standard test method for adhesion strength and mechanical failure modes of ceramic coatings by quantitative single point scratch testing, ASTM international, ASTM international, West Conshohocken, PA, 2010.http://dx.doi.org/10. 1520/C1624-05R10. [31] Feng, X.Q. Feng, L. Qian, W. Yan, Q. Sun, et al., Wearless scratch on NiTi shape memory alloy due to phase transformational shakedown, Appl. Phys. Lett. 92 (2008) 121909, http://dx.doi.org/10.1063/12903106. [32] L. Rapoport, A. Moshkovich, V. Perfilyev, A. Gedanken, Yu. Koltypin, E. Sominski, G. Halperin, I. Etsion, Wear life and adhesion of solid lubricant films on lasertextured steel surfaces, Wear 267 (2009) 1203–1207.