Al 2 O 3 and Pt Atomic Layer Deposition for Surface Modiﬁcation of NiTi Shape Memory Films

: Pt coatings on NiTi ﬁlm micro-actuators and / or sensors can add some useful properties, e.g., they may improve the NiTi anticorrosion and thermomechanical characteristics or activate surface properties beneﬁcial for a speciﬁc application (e.g., functionalized surfaces for biomedical applications). Pt coatings prepared via atomic layer deposition (ALD) may help reduce cost due to the nanometric thickness. However, no authors have reported preparation of Pt ALD coatings on NiTi ﬁlms, perhaps due to the challenge of the concurrent NiTi ﬁlm oxidation during the Pt ALD process. In the present study, Al 2 O 3 and Pt ALD coatings were applied to NiTi thin ﬁlms. The ALD coating properties were studied using electron and atomic force microscopies and X-ray photoelectron spectroscopy (XPS). Potential structural changes of NiTi due to the ALD process were evaluated using electron microscopy and X-ray di ﬀ raction. The presented ALD process resulted in well-controllable preparation of Pt nanoparticles on ultrathin Al 2 O 3 seed layer and a change of the transformation temperatures of the NiTi ﬁlms.


Introduction
Atomic layer deposition (ALD) is a kind of chemical vapor deposition (CVD) technique that typically uses two precursor gas reactants sent into the reaction zone in an alternating sequence of precisely controlled pulses. Any excessive amount of unreacted precursors in the deposition chamber is purged with an inert gas. This purging results in the self-limiting character of the surface chemical reactions on the substrate and in this way not more than one precursor monolayer is left on the surface after purging. In ideal cases, the ALD may provide researchers with an excellent thickness control at the Angstrom level, and it forms conformal coating on complex 3D shapes [1][2][3][4]. It is worth mentioning that an additional advantage of the ALD is also compositional control demonstrated, e.g., in references [5][6][7][8][9]. Due to low deposition rates (100-300 nm/h [1,10] depending on the size of the deposition chamber and the aspect ratio of the substrate), the ALD is suitable mainly for nanometric coatings. However, successful attempts have been made to increase the deposition rate significantly [11]. When the ALD is applied to coat NiTi alloys, then the deposition temperature (substrate temperature, or specifically, temperature at the NiTi surface) is a concern. High deposition temperature may result in unwanted NiTi grain growth and/or NiTi surface oxidation. Although some ALD processes can take place at low The ALD process for Pt on NiTi is not straightforward. The usually used precursors are (trimethyl)-methylcyclopentadienyl-platinum (IV) (MeCpPtMe 3 ) [43][44][45][46] or Pt(acac) 2 (acac = acetylacetonate) [47,48]. Precursor dimethyl(N,N-dimethyl-3-butene-1-amine-N)platinum (DDAP, C 8 H 19 NPt) has also been previously explored in some detail [49].
The co-reacting precursor is usually oxygen or dry air. The temperature window in the case of the MeCpPtMe 3 precursor is 200-300 • C [47]. The Pt ALD deposition rate is low on TiO 2 [50] and TiO 2 is regularly present on the NiTi alloy surface because of NiTi oxidation, even at room temperature [51]. In addition, the Pt ALD process (about 100 ALD cycles) on TiO 2 results in a low spreading density of Pt NPs [50]. The deposition rate of Pt on Al 2 O 3 is superior to the rate of Pt on TiO 2 . Therefore, it is suggested that an Al 2 O 3 underlayer may facilitate initial ALD growth better than the original TiO 2 on NiTi. It is worth mentioning that the authors of reference [39] (also reference [52]) achieved a thin ALD Pt compact layer by depositing an additional W ALD layer on Al 2 O 3 as an adhesion layer.
In the present study, the growth of Pt coating on an Al 2 O 3 underlayer is evaluated using various techniques such as transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS) and atomic force microscope (AFM). The thermal influence of the ALD processes on the grain size and transformation temperatures of NiTi films is examined as well.

Samples
Transforming NiTi films with a thickness of about 1 µm deposited on Si substrates (5 × 10 mm 2 ) were obtained from Acquandas GmbH. In reference [53], experimental work showing determination of the transformation temperatures of the as-purchased NiTi films was described. Before the ALD, the as-purchased NiTi samples were cleaned in ethanol and deionized water and finally dried using N 2 gas.

ALD
The deposition was performed in a home-built ALD reactor assembled (including all accessory parts) in Taiwan Instrument Research Institute (TIRI). The precursors used for the Al 2 O 3 (or Pt) ALD process were trimethyl-aluminum (TMA) and water vapor (or (methylcyclopentadienyl)trimethylplatinum (MeCpPtMe 3 ), a product of Strem Chemicals and dry air as a source of O 2 ). First, Al 2 O 3 /NiTi/Si samples with 10 Al 2 O 3 ALD cycles and subsequently Pt/Al 2 O 3 /NiTi/Si samples with 100 and 200 Pt ALD cycles were prepared. The substrate temperature during the Al 2 O 3 (or Pt) ALD process was 100 • C (or 300 • C). N 2 gas was exclusively used for purging. The base pressure was about 3.33 Pa. The typical cycle times for Al 2 O 3 (or Pt) ALD growth experiment were as follows: 0.2 s TMA pulse, 10 s N 2 purge, 0.2 s water vapor pulse, and 10 s N 2 purge (or 0.5 s MeCpPtMe 3 pulse, 10 s N 2 purge, 0.5 s air pulse, and 10 s N 2 purge). Hence, the Pt ALD process took about 35 min (Pt 100 ALD cycles) and 70 min (Pt 200 ALD cycles). In the initial Pt ALD cycle, hydrogen plasma was used (power output 300 W, deposition temperature 200 • C, H 2 pulse duration 5 s) instead of dry air [39]. Finally, Table 1 lists all the used film samples with their names.

Microscopic Observations
Surface observations of samples Pt100 and Pt200 were made using a scanning electron microscope (SEM) TESCAN, FERA3 GM (Brno, Czech Republic). The NiTi grain size distributions in samples NiTi and Pt100 were obtained based on electron backscattering diffraction (EBSD) method using the EDAX system (EDAX, Mahwah, NJ, USA) with DigiView IV camera. It was possible to get the EBSD signal in NiTi and Pt100 samples, whereas no signal was obtained in sample Pt200, because the layer covering the NiTi layer was too thick.
A thin cross-sectional lamella was cut out from sample Pt100 (using a Ga focused ion beam in an FEI Quanta 3D Dual-Beam SEM, FEI, Hillsboro, OR, USA) and observed with a Fei Tecnai F20 field emission gun transmission electron microscope (TEM, FEI, Hillsboro, OR, USA) operated at 200 kV. Selected area electron diffraction and energy dispersive X-ray spectroscopy (EDS) were performed to identify phases and chemical elements, respectively.
Surface morphology and local conductivity of samples NiTi, Alu10 and Pt100 were investigated using Dimension Icon AFM (Bruker, Billerica, MA, USA) with an SCM-PTSI platinum silicide coated probe with a tip radius of ca. 15 nm. All the measurements were performed at room temperature under ambient conditions. The topography images were measured in PeakForce mode whereas the maps of local current were measured either in contact mode or PeakForce mode (PeakForce mode was used in the case of the simultaneous measurement of topography and local conductivity). The velocity of scanning varied from 250 to 1000 nm/s. The cantilever deflection was detected by a red laser diode (685 nm).

XPS and XRD (X-ray Diffraction) Measurements
The XPS spectra were obtained on a Kratos Axis Supra spectrometer (Kratos Analytical, Manchester, UK) with monochromatic Al Kα (1486.6 eV) X-ray radiation. The pressure in the analysis chamber was maintained near 10 −9 Pa. For surface cleaning and depth profiling, a multi-mode Ar gas cluster ion source (GCIS) was used with an acceleration voltage of 5 keV and the beam current about 5.8 nA. The scanned area, dwell time in one step and number of steps were 2 × 2 mm 2 , 10 min and 6 steps, respectively. The mean cluster size was around 2000 (or 1000) argon atoms/cluster in the first four steps (or in the last two steps). It was estimated that 10 min of Ar ion bombardment corresponds to the depth ranging from few to several nm in the ALD layers. The signal was taken from a circular area with a diameter of 100 µm on sample Pt200. The XPS input data were compiled and analyzed using Gaussian-Lorentzian functions and Shirley background implemented in CasaXPS software (version 2.3) [54].
In the present study, the M s transformation temperatures of samples Pt100 and Pt200 were determined using XRD in order to find out any effect of the Pt and Al 2 O 3 ALD processes applied on NiTi/Si samples. The XRD signal was registered from the film surface using an X-ray diffractometer (X'Pert PRO from PANalytical, Malvern, UK) with a thermal chamber. Before any XRD measurement, either the NiTi film was first heated above the A f (austenite finish) temperature and then cooled down to a test temperature. The XRD patterns were registered at various test temperatures, which were kept constant throughout each XRD measurement. The X-ray source was Co Kα radiation (λ Kα1 = 0.178901 nm, λ Kα2 = 0.17929 nm). The M s temperature was determined based on the change of the full width at half maximum of (110) B2 peak, using the fact that during the martensitic transformation, peak (110) B2 splits into two, or more neighboring peaks. Determination of the M s temperature (of the as-purchased NiTi film) was shown in reference [53].

Microscopic Observations
The microscopy results are expected to reveal (i) whether the Al 2 O 3 ALD layer in samples Alu10, Pt100 and Pt200 is continuous and (ii) whether Pt in samples Pt100 and Pt200 forms nanoparticles (NPs) Coatings 2020, 10, 746 5 of 14 or a continuous layer and, finally, (iii) whether the grain size distribution in the NiTi layer changes after the Al 2 O 3 and Pt ALD processes. Figure 1a,b shows tiny Pt NPs, at the limit of SEM resolution, on the surfaces of samples Pt100 (a) and Pt200 (b). The areal density and mean size of the Pt NPs in Pt200 (Figure 1b) were higher than those in Pt100 (Figure 1a). The sizes of the Pt NPs in Figure 1a,b were 9-17 and 12-20 nm, respectively.
Coatings 2020, 10, x FOR PEER REVIEW 5 of 14 (NPs) or a continuous layer and, finally, (iii) whether the grain size distribution in the NiTi layer changes after the Al2O3 and Pt ALD processes. Figure 1a,b shows tiny Pt NPs, at the limit of SEM resolution, on the surfaces of samples Pt100 (a) and Pt200 (b). The areal density and mean size of the Pt NPs in Pt200 (Figure 1b) were higher than those in Pt100 (Figure 1a). The sizes of the Pt NPs in Figure 1a,b were 9-17 and 12-20 nm, respectively. After the Al2O3 and Pt ALD processes, the grain size and grain distribution in the NiTi layer changed (due to the Pt deposition temperature up to 300 °C) as shown in Figure 2. Figure 2 compares grain distributions in the NiTi layers of samples NiTi (before the ALD depositions) and Pt100 (after the ALD depositions).  After the Al 2 O 3 and Pt ALD processes, the grain size and grain distribution in the NiTi layer changed (due to the Pt deposition temperature up to 300 • C) as shown in Figure 2. Figure 2 compares grain distributions in the NiTi layers of samples NiTi (before the ALD depositions) and Pt100 (after the ALD depositions).
Coatings 2020, 10, x FOR PEER REVIEW 5 of 14 (NPs) or a continuous layer and, finally, (iii) whether the grain size distribution in the NiTi layer changes after the Al2O3 and Pt ALD processes. Figure 1a,b shows tiny Pt NPs, at the limit of SEM resolution, on the surfaces of samples Pt100 (a) and Pt200 (b). The areal density and mean size of the Pt NPs in Pt200 (Figure 1b) were higher than those in Pt100 (Figure 1a). The sizes of the Pt NPs in Figure 1a,b were 9-17 and 12-20 nm, respectively. After the Al2O3 and Pt ALD processes, the grain size and grain distribution in the NiTi layer changed (due to the Pt deposition temperature up to 300 °C) as shown in Figure 2. Figure 2 compares grain distributions in the NiTi layers of samples NiTi (before the ALD depositions) and Pt100 (after the ALD depositions).    Figure 3a,b shows the TEM cross-sectional micrographs of sample Pt100, with low (a) and high magnification (b). The Si substrate in the low-magnification image is seen in a light color whereas the W layer (deposited via gas injection system (GIS) to manipulate Pt100 lamella) is seen in a dark color. The high-angle annular dark-field scanning transmission electron microscopy (HAADF STEM) image shows the upper layers in the cross-section (Figure 3c). The path of the line scan with a fixed point denoted as point 1 are shown in the image whereas the results of the EDS analysis along the line are shown in Figure 3d. The presence of Ga, Cu and W elements originates from Ga focused ion beam used to cut out the examined lamella, a Cu ring used as a support for the lamella and GIS material, respectively. The electron diffraction pattern taken from the SAED (selected area electron diffraction) inside of a NiTi grain is shown in Figure 3e. The pattern corresponds to the R-phase (the rhombohedral structure); the zone axis was [−121].
Coatings 2020, 10, x FOR PEER REVIEW 6 of 14 Figure 3a,b shows the TEM cross-sectional micrographs of sample Pt100, with low (a) and high magnification (b). The Si substrate in the low-magnification image is seen in a light color whereas the W layer (deposited via gas injection system (GIS) to manipulate Pt100 lamella) is seen in a dark color. The high-angle annular dark-field scanning transmission electron microscopy (HAADF STEM) image shows the upper layers in the cross-section (Figure 3c). The path of the line scan with a fixed point denoted as point 1 are shown in the image whereas the results of the EDS analysis along the line are shown in Figure 3d. The presence of Ga, Cu and W elements originates from Ga focused ion beam used to cut out the examined lamella, a Cu ring used as a support for the lamella and GIS material, respectively. The electron diffraction pattern taken from the SAED (selected area electron diffraction) inside of a NiTi grain is shown in Figure 3e. The pattern corresponds to the R-phase (the rhombohedral structure); the zone axis was [-121].   The surface patterns in Figure 4a,b (belonging to samples NiTi and Alu10) are regular. The round objects in Figure 4a,b are 10-30 nm in size. Figure 4c shows surface of sample Pt100 containing two kinds of round objects-the small object (with high occurrence) and the large objects (with low occurrence) with diameters 10-17 and 30-50 nm, respectively. The cross section analysis in Figure 4d corresponds to the white line in Figure 4a. The maps of local current obtained by conductive AFM (C-AFM) for samples NiTi (a), Alu10 (b) and Pt100 (c) are shown in Figure 5a-c. The scan direction is from the top downward. Figure 5a shows high currents about −40 pA at the top and currents of few pA at the bottom. A similar distribution of local currents is shown in Figure 5b but with much lower values of local currents. Figure 5c shows some conductive round objects with diameters ranging 10-17 nm. The cross section analysis in Figure 5d corresponds to the white line in Figure 5a.  Figure 4a,b (belonging to samples NiTi and Alu10) are regular. The round objects in Figure 4a,b are 10-30 nm in size. Figure 4c shows surface of sample Pt100 containing two kinds of round objects-the small object (with high occurrence) and the large objects (with low occurrence) with diameters 10-17 and 30-50 nm, respectively. The cross section analysis in Figure 4d corresponds to the white line in Figure 4a. The maps of local current obtained by conductive AFM (C-AFM) for samples NiTi (a), Alu10 (b) and Pt100 (c) are shown in Figure 5a-c. The scan direction is from the top downward. Figure 5a shows high currents about −40 pA at the top and currents of few pA at the bottom. A similar distribution of local currents is shown in Figure 5b but with much lower values of local currents. Figure 5c shows some conductive round objects with diameters ranging 10-17 nm. The cross section analysis in Figure 5d corresponds to the white line in Figure 5a.     Figure 4a,b (belonging to samples NiTi and Alu10) are regular. The round objects in Figure 4a,b are 10-30 nm in size. Figure 4c shows surface of sample Pt100 containing two kinds of round objects-the small object (with high occurrence) and the large objects (with low occurrence) with diameters 10-17 and 30-50 nm, respectively. The cross section analysis in Figure 4d corresponds to the white line in Figure 4a. The maps of local current obtained by conductive AFM (C-AFM) for samples NiTi (a), Alu10 (b) and Pt100 (c) are shown in Figure 5a-c. The scan direction is from the top downward. Figure 5a shows high currents about −40 pA at the top and currents of few pA at the bottom. A similar distribution of local currents is shown in Figure 5b but with much lower values of local currents. Figure 5c shows some conductive round objects with diameters ranging 10-17 nm. The cross section analysis in Figure 5d corresponds to the white line in Figure 5a.

XPS and XRD Measurements
The diagrams of measured and fitted XPS spectra expressed as counts per second (CPS) versus binding energy (BE) for several individual elements after various Ar n+ etching times are shown in Figures 6-8. Specifically, Figure 6a,b shows the diagrams of CPS versus BE attributed to C before Arn + etching (0 min) and after 30 min of Arn + etching, respectively. The diagrams of CPS versus BE, attributed to C, after etching times 10, 20, 40, 50, and 60 min (not shown here) are similar to the diagram in Figure  6b. Figure 6b does not contain any fitted XPS spectra because the measured XPS data are too noisy due to a relatively small amount of C, in various forms. Figure 7a,b shows the diagrams of CPS (relative offset) versus BE attributed to Al and Pt after various Arn + etching times, respectively. Figure  8a,b shows the diagrams of CPS versus BE attributed to O before Arn + etching (0 min) and after 30 min of Arn + etching, respectively. Again, the diagrams of CPS versus BE, attributed to O, after the other etching times (not shown here) are similar to the diagram in Figure 8b. Atomic percentages of the elements at the surface of sample Pt200 determined from the XPS after various times of Arn + etch are summarized in Table 2.     Figure 6a,b shows the diagrams of CPS versus BE attributed to C before Arn + etching (0 min) and after 30 min of Arn + etching, respectively. The diagrams of CPS versus BE, attributed to C, after etching times 10, 20, 40, 50, and 60 min (not shown here) are similar to the diagram in Figure  6b. Figure 6b does not contain any fitted XPS spectra because the measured XPS data are too noisy due to a relatively small amount of C, in various forms. Figure 7a,b shows the diagrams of CPS (relative offset) versus BE attributed to Al and Pt after various Arn + etching times, respectively. Figure  8a,b shows the diagrams of CPS versus BE attributed to O before Arn + etching (0 min) and after 30 min of Arn + etching, respectively. Again, the diagrams of CPS versus BE, attributed to O, after the other etching times (not shown here) are similar to the diagram in Figure 8b. Atomic percentages of the elements at the surface of sample Pt200 determined from the XPS after various times of Arn + etch are summarized in Table 2.     Specifically, Figure 6a,b shows the diagrams of CPS versus BE attributed to C before Ar n + etching (0 min) and after 30 min of Ar n + etching, respectively. The diagrams of CPS versus BE, attributed to C, after etching times 10, 20, 40, 50, and 60 min (not shown here) are similar to the diagram in Figure 6b. Figure 6b does not contain any fitted XPS spectra because the measured XPS data are too noisy due to a relatively small amount of C, in various forms. Figure 7a,b shows the diagrams of CPS (relative offset) versus BE attributed to Al and Pt after various Ar n + etching times, respectively. Figure 8a, Figure 8b. Atomic percentages of the elements at the surface of sample Pt200 determined from the XPS after various times of Ar n + etch are summarized in Table 2. The transformation temperatures of sample NiTi and the M s temperatures of samples Pt100 and Pt200 are summarized in Table 3. At room temperature each sample may contain both phases, austenite and martensite with crystal structures B19 and B2, respectively. The diffractograms for the as-purchased NiTi film measured at various temperatures are shown in reference [53]. The diffractograms for samples Pt100 and Pt200 measured in the (110) B2 neighborhood at various temperatures between 25 and 90 • C are shown in Figure 9. Diffraction peaks (110) B2 are not shown for all the test temperatures but only for few selected temperatures in order to avoid accumulating too much data. The M s temperatures for samples Pt100 and Pt200 are determined from the diagrams of FWHM (full width at half maximum) versus temperature (see the insets of Figure 9)  The transformation temperatures of sample NiTi and the Ms temperatures of samples Pt100 and Pt200 are summarized in Table 3. At room temperature each sample may contain both phases, austenite and martensite with crystal structures B19′ and B2, respectively. The diffractograms for the as-purchased NiTi film measured at various temperatures are shown in reference [53]. The diffractograms for samples Pt100 and Pt200 measured in the (110)B2 neighborhood at various temperatures between 25 and 90 °C are shown in Figure 9. Diffraction peaks (110)B2 are not shown for all the test temperatures but only for few selected temperatures in order to avoid accumulating too much data. The Ms temperatures for samples Pt100 and Pt200 are determined from the diagrams of FWHM (full width at half maximum) versus temperature (see the insets of Figure 9)    [53]) and the M s temperatures attributed to samples Pt100 and Pt200 and determined from the XRD patterns registered at various temperatures.

Sample
Transformation Temperature AFM XRD

Discussion
The discussion is separated into three parts related to (i) properties of Pt coating on samples Pt100 and Pt200, (ii) properties of Al 2 O 3 coating on samples Alu10, Pt100, Pt200 and (iii) the effect of the Pt and Al 2 O 3 ALD processes on the properties of the NiTi layer.

Pt Phase
Pt in samples Pt100 and Pt200 did not form a continuous layer but NPs. The sizes of the Pt NPs on the surfaces of samples Pt100 and Pt200 were 9-17 nm (determined from Figures Figures 1a and 4c) and 12-21 nm (determined from Figure 1b), respectively. The areal densities of Pt NPs in samples Pt100 and Pt200 were about 800-1000 NPs per 1 µm 2 (determined from Figures Figures 1a and 4c) and about 1700 NPs per 1 µm 2 (determined from Figure 1b), respectively. As for the Pt NP size range and the Pt NP areal density on sample Pt100, there is a fair agreement between the results obtained from the SEM (Figure 1) and AFM images (Figure 5c). Due to the relatively low density of Pt NPs on sample Pt100, the EDS elemental analysis (in the chamber of TEM) along a line across the top surface did not show any presence of Pt (Figure 3d). There is no clear interface between the individual deposited layers in Figure 3c. Therefore, the locations of the individual layers were hinted at in Figure 3c. The Pt NPs are conductive ( Figure 5c) and do not contain any significant amount of PtO or PtO 2 phases as implied from the values of XPS binding energies corresponding to the peaks in Figure 7b. Specifically, according to reference [55], binding energies corresponding to Pt 4f 7/2 for Pt, PtO and PtO 2 are 71.1 ± 0.3, 72.2 ± 0.3, and 74.2 ± 0.3 eV, respectively. The XPS binding energy registered from sample Pt200 and corresponding to Pt 4f 7/2 was 71.4 eV (Figure 7b), a value close to 71.1 eV (a BE value characteristic for Pt) measured in [55]. Some organic residua may remain on the surfaces of samples Pt100 and Pt200 as seen from Figure 6a. Specifically, Figure 6a shows carbon peaks with BE corresponding to C-C, C-OH and C=O bounds (the bounds characteristic for organic material). In addition, Table 2 shows that the non-etched surface of Pt200 contains about 38 at.% of carbon. After the Arn+ etch, the content of carbon significantly dropped to values around 10 at.%. The origin of the organic residua is likely in the Pt organic precursor. However, the distinct advantage of the Pt ALD process is the controllability of Pt NP size and the areal density of Pt NPs by controlling the number of Pt ALD cycles (as shown also in reference [56]).

Al 2 O 3 Phase
In general, Al 2 O 3 ALD processes and properties of Al 2 O 3 ALD layers are discussed in references [57][58][59]. Thin Al 2 O 3 ALD layers grown below 600 • C are amorphous regardless of the type of the used substrate [58]. Hence, the Al 2 O 3 coating in the present study was amorphous. In our previous study [29], an almost identical Al 2 O 3 ALD process applied to NiTi plate samples resulted in the deposition rate of the 0.1 nm per ALD cycle. Therefore, it is assumed that if the Al 2 O 3 coating in samples Alu10, Pt100 and Pt200 is continuous then its thickness is about 1 nm.
After comparing Figure 5a,b one can see that the surface of sample NiTi is much more conductive than the surface of sample Alu10. Since the map of the local current in Figure 5b is quite uniform (no islands are obvious), it can be deduced that the Al 2 O 3 coating (10 ALD cycles) in sample Alu10 forms an insulating and continuous layer. It is assumed that the Al 2 O 3 layers in samples Pt100 and Pt200 are also insulating and continuous. Some other properties of the Al 2 O 3 layer can be found from the XPS measurement (Figure 7a). Figure 7a shows that the individual Al 2s peaks shift with the changing Ar n + etch time. This effect may be an artifact from the Ar-etch.
Here, the Al 2 O 3 layer has a double function: (i) Al 2 O 3 as a barrier of further oxidation of NiTi at higher temperatures (the Pt ALD process requires high temperatures on the samples' surfaces, about 300 • C) and (ii) Al 2 O 3 as a layer allowing the Pt growth rate to be higher than that on TiO 2 .
Apart from the properties of the Pt and Al 2 O 3 ALD coatings, it is worth mentioning the absence of Ni (in any form) on the top surface of sample Pt200 (Table 2). Generally, Ni from a NiTi alloy may be released in a corrosive environment in the form of Ni ions, which is of concern when the NiTi alloy is used as an implant in the human body. Furthermore, as the content of TiO 2 is concerned, Table 2 and Figure 8a,b show the difference between the surface of sample Pt200 before and after the Ar n + etch.
There is less TiO 2 before the Ar n + etch than after the etch. The presence of TiO 2 is due to the oxidation of NiTi layer at room temperature and during the Pt and Al 2 O 3 ALD processes.

The Effect of the Pt and Al 2 O 3 ALD Processes on Properties of the NiTi Layer
The deposition temperatures of the Pt and Al 2 O 3 ALD processes were 300 and 100 • C, respectively. The Al 2 O 3 ALD process was relatively short (only 10 ALD cycles) with the relatively low deposition temperature. Therefore, the Pt ALD process predominantly influenced the properties of the NiTi layer in samples Pt100 and Pt200. Properties of NiTi alloys and other SMAs depend to a large extent on the following factors: (i) chemical composition, (ii) grain size, (iii) texture, (iv) internal stresses, etc. Each factor (of factors (i)-(iv)) may change due to annealing of an SMA in question. The effects of factors (i)-(iv) may be inter-related and/or unevenly distributed in an SMA sample and it may be hard to discriminate contributions of the individual factors as, e.g., in the case of Fe-30 at.% Pd SMA melt-spun ribbons [60]. Since the NiTi layers in samples Pt100 and Pt200 were firmly attached to the substrate it was not easy to study mechanical properties of the NiTi layers but it was convenient to determine the M s transformation temperatures (an important parameter of SMAs). Table 3 indicates an increase of the M s temperature of the NiTi layers in samples Pt100 and Pt200 by 14 • C, due mainly to the Pt ALD process. Both the Pt (100 ALD cycles) and Pt (200 ALD cycles) processes resulted in the identical shift of the M s temperature implying that in samples Pt100 and Pt200, the exposure time (about 35 and 70 min) did not matter as much as the exposure temperature (300 • C). In our previous work, a similar effect on the transformation temperatures was observed when exposing a thin NiTi plate to a polymerization temperature of 200 • C for 2 h during manufacturing the NiTi-PI composite. The PI polymerization process in the NiTi-PI composite resulted in the drop of the M s temperature by about 3 • C [61]. In both cases (NiTi-PI and Pt100/200 samples) it is hard to find the exact cause of the M s shift. In the case of Pt100/200 samples, oxidation of NiTi, precipitation (Ni 4 Ti 3 ), stress due to the difference in the coefficients of thermal expansion of the substrate and the NiTi alloy might play a certain role in the increase of M s .
As for the grain size distribution change due to the Pt and Al 2 O 3 ALD processes, there was a minor increase of NiTi grains in sample Pt100 (Figure 2a-d). However, the examined areas in samples NiTi and Pt100 were too small to be statistically significant. The main purpose of the EBSD measurements was to make sure that the exposure of the NiTi film to temperatures as high as 300 • C did not result in a remarkable change of the grain size distribution.

Conclusions
The Al 2 O 3 ALD process consisting of ten ALD cycles and applied to NiTi films results in a continuous and insulating layer. The following Pt ALD coating is not continuous but forms NPs with sizes 9-17 nm (Pt100) and 12-21 nm (Pt200). The areal densities of Pt NPs are about 800-1000 NPs per 1 µm 2 (Pt100) and about 1700 NPs per 1 µm 2 (Pt200). The Pt NPs are conductive and do not contain any significant amount of PtO or PtO 2 phases. The distinct advantage of the Pt ALD is the controllability of Pt NP size and the areal density of Pt NPs by controlling the number of Pt ALD cycles.
The Al 2 O 3 and Pt ALD processes cause non-significant grain size growth and an increase of the M s transformation temperature by 14 • C.