TiO2 Atomic Layer Deposition coatings for the corrosion protection of stainless steel

Thin Solid Films 522 (2012) 283–288

Contents lists available at SciVerse ScienceDirect

Thin Solid Films journal homepage: www.elsevier.com/locate/tsf

Multilayer Al2O3/TiO2 Atomic Layer Deposition coatings for the corrosion protection of stainless steel E. Marin a,⁎, L. Guzman a, A. Lanzutti a, W. Ensinger b, L. Fedrizzi a a b

University of Udine, 33100 Udine, Italy Darmstadt Technolnische Universität, 06151 Germany

a r t i c l e

i n f o

Article history: Received 10 November 2011 Received in revised form 9 August 2012 Accepted 9 August 2012 Available online 21 August 2012 Keywords: ALD TiO2 Al2O3 AISI 316 stainless steel Corrosion protection GDOES Polarization curves AFM

a b s t r a c t Atomic Layer Deposition (ALD) is used to deposit conformal nanometric layers onto different substrates. In this paper, characterization of different ALD layers has been carried out in order to evaluate the suitability of this deposition technolnique for the corrosion protection of stainless steel substrates. Al2O3, TiO2 and multilayer configurations, have been deposited on AISI 316 austenitic stainless steel and have then been investigated using atomic force microscopy (AFM), glow discharge optical emission spectrometry (GDOES), scanning electron microscopy (SEM), Vickers indentation and potentiodynamic polarizations (PP). AFM has been used to obtain a morphological characterization and to evaluate the thickness of the depositions. SEM has been used to investigate the presence of deposition defects. GDOES has been used to obtain a compositional profile. Vickers indentations were used in order to evaluate the resistance to delamination. PPs have been used in order to evaluate the corrosion protection. The results have showed that corrosion resistance can be effectively enhanced. Multilayer configuration proved to be more effective than single layers configurations. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Stainless steels are technolnological alloys actually employed in a number of applications thanks to their excellent combination of mechanical properties and corrosion resistance. As stainless steels are spontaneously protected by the formation of a surface nanometric chromium oxide layer [1], these alloys can be used for a wide range of applications even in mild aggressive environments, such as in hydroelectric energy production components [2], in marine environment [3] or in contact with body tissues [4]. Even so, stainless steels could show insufficient corrosion resistance in strongly aggressive media containing Cl− and S2− ions, usually in combination with high temperature or very high/low pH values [5–7]. For these reasons, in particular circumstances stainless steels may need a further improvement of corrosion protection. Conventional treatments, such as painting, are hardly applicable to stainless steel due to adhesion problems between paint and the metal substrate [8]. A great number of innovative treatments are nowadays under intensive study to improve stainless steel corrosion resistance, such as plasma detonation technolniques [9], arc-ion plating [10], sol–gel deposition [11,12], chemical conversion layers of cerium [13] chromium [14] or other elements, chemical vapor deposition

⁎ Corresponding author at: via Cotonificio 108, Udine 33100 (UD), Italy. Tel.: +39 0432 629974; fax: +39 0432 558802. E-mail address: [email protected] (E. Marin). 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tsf.2012.08.023

(CVD) [15], high-velocity oxy-fuel spray [16], plasma-nitriding [17], and lately Atomic Layer Deposition (ALD) [18–20]. The modern concept of ALD is an extension of the ALE (Atomic Layer Epitaxy), patented by Prof. Suntola [21]. Suntola's studies were mainly focused on switching effects in chalcogenide nanometric films for solid state electronic devices [22–24], and lately extended to a wide range of amorphous semi-conductive thin films [25]. ALD (as ALE) process involves a sequence of self-limiting surface reactions. As evidenced in 1980 by Ahonen et al. [26], the self-limiting characteristic of each reaction step differentiates ALE and ALD from other chemical vapor deposition technolnologies. In ALD each deposition cycle is clearly divided in four steps: in the first step a precursor is injected in the deposition chamber. The precursor is chosen so that its molecules will not react with each other at the deposition temperature. In ideal situations, a single monolayer is thus formed as a result of the reaction with the substrate. In the second step, the chamber is purged with nitrogen or argon gas in order to remove the excess of reactant and prevent “parasitic” CVD deposition on the substrate, which will eventually occur if two different precursors are present in the deposition chamber at the same time. In the third step, the second precursor is injected in the chamber. In the case of metal oxide layers, this is an oxidant agent, usually simple H2O. The last step of the deposition cycle is a second purge to remove the excess of reactant with purging gas. Closed-loop repetitions of the four basic steps theoretically allow obtaining perfectly conformal deposits of any desired thickness. By avoiding the contact between the precursors throughout the whole coating process, a film growth at

284

E. Marin et al. / Thin Solid Films 522 (2012) 283–288

atomic layer control, with a thickness control within ~10 pm, can be obtained. Interest in ALD has increased stepwise in the mid-1990's and 2000's, with the interest focused on silicon-based microelectronics [27]. Up to the present time, ALD processes have been used to deposit several types of nanometric films, including several chemical compounds (e.g. AsGa, CdSe,…), metal oxides (e.g. Al2O3, CaO, CuO, Er2O3, Ga2O3, HfO2, La2O3, MgO, Nb2O5, Sc2O3, SiO2, Ta2O5, TiO2, Y2O3, Yb2O3, ZnO, ZrO2), nitrides (e.g. TiN, TaN, AlN, GaN, WN, NbN), sulfides (e.g. SrS, ZnS), carbides (e.g. TaC, TiC), fluorides (e.g. CaF2, LaF3, MgF2), pure metals (e.g. Ru, Ir, Ta, Pt), biomaterials (e.g. hydroxyapatite (Ca10(PO4)6(OH)2)) and even polymers (e.g. Polyimides) [28,29]. Preliminary results on the corrosion protection on stainless steel by ALD TiO2 and Al2O3 layers were already obtained by Matero et al. (1999) [29], which supposed that the conformal ALD coatings could improve the corrosion resistance of different metal alloys. In 2007, Shan et al. [30] used TiO2 ALD layers to protect an undefined stainless steel, obtaining only a limited effect. In 2011, Marin et al. [31], Diaz et al. [17] and Potts et al. [32] clearly showed that the residual porosity of ALD layers decreases increasing the thickness of the layer thus improving the protection of the substrate. In most cases [31,17,32,33] the nanometric ALD layers clearly showed a corrosion protection similar, if not superior to conventional protective technolniques and thicker coatings, even if common industrial tests (salt spray) performed on Plasma Enhanced ALD by Potts et al. [32] clearly showed a time-limited corrosion protection. In this work, characterization of ALD coated AISI 316 L has been carried out, in order to determine the possible use of nanometric ceramic ALD coatings for the corrosion protection of stainless steel. 2. Experimental details Sheets of standard AISI 316 L stainless steel (chemical composition wt.%: C = 0.035; P b 0.04; S b 0.03; Mn = 2.0; Si = 0.75; Cr = 16.0 − 18.0; Ni = 10.0 − 15.0; Mo = 2.0 − 3.0) were obtained by cold rolling, pickling and skin-passing, obtaining a standard BS1449-2B finishing. Stainless steel sheets were then mechanically cut in 10 × 10 mm samples, grinded using SiC abrasive paper and polished with diamond suspension in order to obtain a surface roughness of about 20 nm Ra, which was considered to be suitable for testing, in particular in the case of Glow Discharge Optical Emission Spectrometry (GDOES) analyses. Sample surface was then cleaned using ethanol in ultrasonic bath for 10 min and dried in a dry heat sterilizer at a temperature of 50 °C for 15 min. Some samples were then partially masked with heat resistant laboratory tape to prevent ALD deposition on part of the substrate, thus creating a clear and sharp interface between coated and uncoated region after adhesive tape removal. The ALD coating was deposited using a TFS 500 reactor (Beneq Oy, Finland): Al2O3 layers were obtained using trimethylaluminium (Al(CH3)3) and H2O precursors and TiO2 layers were obtained using titanium tetrachloride (TiCl4) and H2O precursors. Both depositions were performed at a temperature of 120 °C. The low temperature processes were chosen in order to obtain an amorphous structure for both layers. The number of precursor cycles for each deposition was calculated using a growth rate per cycle (GPC) of ~ 0.1 nm/cycle for TiO2 and a GPC of ~ 0.15 nm/cycle for Al2O3. The samples were coated using different ALD configurations, with an overall coating thickness of about 300 nm, 100 nm, 30 nm and 10 nm (Table 1). For comparison, multilayer configurations were confronted with single layer Al2O3 deposits of the same thickness. 2.1. Morphology Morphological characterization was carried out using Veeco's Digital Instrument's Nanoscope IIIa atomic force microscope (AFM) in tapping

Table 1 AFM thickness results measured at the interface between coated and uncoated substrate and standard deviation between different AFM measurements. No. of layers

Composition

Theoretical total thickness

Measured total thickness

1 1 1 2

Al2O3 TiO2 Al2O3 Al2O3/TiO2 alternate Al2O3/TiO2 alternate Al2O3 Al2O3/TiO2 alternate Al2O3 Al2O3/TiO2 alternate

300 100 100 100

295 102 107 109

4 1 12 1 14

Standard deviation

nm nm nm nm

2% 2% 2% 1%

100

98 nm

2%

30 30

31 nm 34 nm

5% 6%

10 10

12 nm 13 nm

10% 12%

mode configuration, using a Bruker SCM-PIT tip (Antimony (n) doped Si, frequency: 60–100 kHz, elastic constant: 1–5 N/m, PtIr coated) and Carl Zeiss EVO-40 scanning electron microscope (SEM) with an operating voltage of 20 kV. Analyses were performed on the stainless steel substrate and on coated samples. In particular, SEM was used in order to scan the surface of the sample and investigate the presence of deposition defects and/or surface anomalies. AFM was mainly used to investigate the presence of surface morphological defects on the coating that were supposed to be hardly visible using SEM due to the coating transparency. AFM was also used at the interface regions between coated and uncoated substrate after adhesive tape removal in order to obtain information about the overall thickness of the deposits and confront it with the theoretical deposition rates values and the results obtained from GDOES analyses.

2.2. Composition In-depth compositional analyses were carried out using Horiba Yobin–Yvon's RF–GD Profiler GDOES. Due to the difficulties in calibration of GDOES for the analysis of titanium and aluminum oxides, only qualitative compositional analyses were performed, even if a keen calibration was performed in order to obtain reliable single layer thickness values. Just for the 10 nm thick coatings, the GDOES analyses were considered totally qualitative (both in thickness and composition) due to the fact that GDOES technolnique is strongly influenced by surface roughness, which was higher than the overall coating thickness. Another problem came from the fact that the multilayer 10 nm structure was composed by single layers of about 0.5 nm, which could not be correctly resolved by GDOES technolnique, which has a maximum theoretical resolution of about 2 nm.

2.3. Mechanical properties In order to evaluate the resistance to delamination of the different ALD configurations, Vickers indentations were applied to the samples under different load conditions (HV0.1–0.2–0.5–1.0–2.0) and the delaminated areas have been then measured using a specific image post-processing software (Wayne Rasband, ImageJ 1.44p). As all indentation hardness tests, Vickers indentation are dependant from the mechanical characteristics of the substrate and can only be used to compare different coatings applied on the same substrate and not results from different substrates. As adhesion is strictly connected to the surface roughness [33–35], Vickers adhesion tests will give reliable results only for substrates with similar surface finishing.

E. Marin et al. / Thin Solid Films 522 (2012) 283–288

285

2.4. Electrochemical properties Electrochemical characterization of the different samples was performed using Potendiodynamic Polarizations. An AUTOLAB PGSTAT20 potentiostat was used in a standard three electrodes configuration. The reference electrode was Ag/AgCl and the counter electrode was a 99.99% pure platinum wire. All measurements have been performed in a pH 6.5, 9 g/l solution of NaCl (physiological solution). All polarization curves were obtained using a scan speed of 0.2 mV/s after 10 min of immersion of the samples, in order to stabilize the Open Circuit Potential (OCP). The potential has been increased from −200 mV respect to the OCP to a measured current density of about 10−3 A/cm2. 3. Results and discussion 3.1. Morphology SEM resulted to be an inadequate technolnique for the analysis of ALD layers, since no morphological differences were found between images obtained on coated and uncoated regions and no morphological properties of the ALD layers could be correctly resolved using this technolnique. Information was only obtained at the interface regions between coated and uncoated substrates after adhesive tape removal. Fig. 1 shows the SEM image obtained at the interface region between stainless steel and a 100 nm TiO2 ALD coating. ALD seem to perfectly mirror the substrate roughness, even in the scratches caused by the surface polishing. Since the image features a sharp interface between coated and uncoated regions, it was supposed that these interface regions could be correctly resolved using AFM, thus allowing coating thickness evaluation. Fig. 2a shows the AFM map of the interface region between uncoated stainless steel and the 100 nm TiO2 layer. It could be observed that the surface of the TiO2 layer resulted to be completely covered by small blisters which were not resolved by SEM due to the transparency of the ALD layers to electron beams. Fig. 2b shows the interface region between stainless steel and the 100 nm Al2O3 layer. In the case of Al2O3, no blistering phenomenon was observed on the coated region, and for this reason it was considered less defective than TiO2. A certain number of blisters were also identified on the multilayer configurations, and it was considered to be caused by the presence of a defective TiO2 layer. Similar observations were obtained also on 300 nm, 30 nm and 10 nm thick samples. The overall coating thickness results obtained by AFM are presented in Table 1. In all cases, AFM thickness measurements resulted to be close to the theoretical thickness of the deposit.

Fig. 1. SEM image obtained at the interface region between stainless steel and a 100 nm TiO2 ALD in proximity of a cross section.

Fig. 2. (a) AFM map of the interface region between uncoated stainless steel and the 100 nm TiO2 layer (b) interface region between stainless steel and the 100 nm Al2O3 layer.

3.2. Composition GDOES thickness measurement accuracy was strongly influenced by the sharpness of the interface region between coating and substrate. Conventionally, in this work the coating thickness has been measured between the top surface and the intersection point between oxygen and iron signals. Since no reference materials were present for amorphous ceramics, a specific GDOES calibration was required. Sputtered crater's depth was measured using a stylus profilometer and the sputtering rate has been evaluated accordingly. Following this calibration GDOES results shall be considered semi-quantitative and for this reason the composition of the coatings and in particular of Al2O3 layers is not stoichiometric. Fig. 3 shows the results obtained by GDOES on three different ALD multilayer configurations, the 100 nm bilayer (Fig. 3a), the 100 nm multilayer (Fig. 3b) and the 30 nm multilayer (Fig. 3c). The thickness results for each layer of the different configurations as obtained by GDOES technolnique are summarized in Table 2. Fig. 3a shows the GDOES graph for the first 200 nm of the sample coated with about 50 nm of TiO2 and 50 nm of Al2O3. The thickness of the coating resulted to be around 98 nm, even if coating signals are present till 120 nm. An interface between the two different layers can be observed around 60 nm, even if the Al signal appears already at about 40 nm. Ti and Al signals overlap for over 80 nm, from about 40 nm to 120 nm. This is supposed to be an effect of the surface roughness of the sample, since the surface roughness value of the substrate has the same order of magnitude of the singular ALD layer. Fig. 3b shows the GDOES graph for the first 200 nm of the sample coated with four alternated layers of about 25 nm each. The thickness of the coating resulted to be around 100 nm, even if coating signals are present till 200 nm. The three layer interfaces can be found at about 30, 60 and 80 nm. In this sample, the effect of the surface roughness mainly influenced the two layers closer to the substrate, which show a not sharp interface and a wide distribution of the signal. This

286

E. Marin et al. / Thin Solid Films 522 (2012) 283–288

singular layers that compose the coating, which appears to be composed by a singular layer with a gradient composition, Ti-rich near the open surface and Al-rich near the coating/substrate interface. This is caused by the presence of two thicker layers (about 8 nm) composed by the two layers in these regions, respectively. Since the thickness results obtained by GDOES are close to the results obtained by AFM at the interface between coated and uncoated regions, GDOES can be considered a reliable technolnique when applied to measure the thickness of ALD nanometric coatings, even if it resulted to be able to give only qualitative results about the composition. As the substrate roughness resulted to be around 20 nm and the resolution of GDOES around 2 nm, analysis on the 10 nm coatings resulted to be only qualitative, giving an evident overestimation of the coating signals with a strong overlap between coating and substrate signals. 3.3. Mechanical properties

Fig. 3. Results obtained by GDOES on three different ALD multilayer configurations, the 100 nm bilayer (a), the 100 nm multilayer (b) and the 30 nm multilayer (c).

is easily explained as in this case the thickness of a single layer of the ALD configuration is almost equal to the surface roughness. Fig. 3c shows the GDOES graph for the first 80 nm of the sample coated with about 30 nm of multilayered Al2O3/TiO2. The thickness of the coating resulted to be close to 28 nm, even if signals from the coatings are still present at more than 40 nm, and the Al signal strong up to more than 30 nm. In this case the resolution of the GDOES instrument (about 2 nm) is not sufficient in order to discriminate the Table 2 GDOES thickness results measured at the interface between Fe and O signals. No. of layers

Composition

1st layer

2nd layer

3rd layer

4th layer

1 1 1 2

Al2O3 TiO2 Al2O3 Al2O3/TiO2 alternate Al2O3/TiO2 alternate Al2O3 Al2O3/TiO2 alternate Al2O3 Al2O3/TiO2 alternate

287 nm 98 nm 110 nm 55 nm

– – – 45 nm

– – – –

– – – –

3% 4% 5% 4%

30 nm

27 nm

25 nm

20 nm

4%

4 1 12 1 14

32 nm – – Overall thickness of 34 nm

Deviation

–

9% 11%

~14 nm – – – Overall thickness of ~16 nm

41% 55%

Vickers indentations performed on the coated sample surface were observed using SEM, as the Optical Microscope field depth resulted to be too small to sharply resolve all the delamination details on the deformed region in just one image. The dimension of the Vickers indentation is related to the applied load. Even if localized defects are hardly visible on the ALD coatings due to their transparency, larger areas with a complete coating removal resulted to be clearly visible on the SEM images. A clear example of how the delaminated areas have been evaluated is presented in Fig. 4a and b. Fig. 4a, in particular, shows a Vickers indentation as obtained using a 3 N load on the 100 nm TiO2 coated sample, while Fig. 4b shows the result of the image processing software, where the bright areas represent the delaminated region and the dark area represent the Vickers indentation area, which was not considered part of the delamination phenomena. The results obtained for the different samples are plotted in Fig. 5. All coatings showed a clear dependence on the applied load for the delaminated Area/Load ratio, which remains almost constant only at the highest indentation loads (9.8 N = HV1 and 19.6 N = HV2). For the 10 nm thickness coatings, the delaminated area could not be resolved using SEM due to the limited thickness. For the overall investigated load interval, the 300 nm monolayer coating showed the worst behavior, with delaminated areas more than three times higher than the other samples at the highest loads. This behavior is supposed to be caused by the relatively high thickness of the ALD deposition compared to the other samples. It appears clear that delamination is in fact strongly related to the overall coating thickness as all samples with 30 nm of ALD have shown better resistance to delamination that all the other coatings, while the 300 nm coating showed the worst resistance to delamination. Comparing single layer configurations, it seems that TiO2 layers have lower resistance to the delamination compared to Al2O3 layers, and this seems to affect also bi-layer configurations in which the coatings show an intermediate behavior. Comparing coatings with similar thickness, multilayer configurations usually show higher resistance to delamination compared to single layer configurations, with the only exception of the 100 nm Al2O3/TiO2 layer in which the relatively “thick” TiO2 layer probably played a role in reducing the overall coating resistance due the intrinsic defectiveness already observed on the SEM images. 3.4. Electrochemical properties Polarization curves for the different samples with a total coating thickness of about 100 nm are shown in Fig. 6a. Uncoated AISI 316 L clearly shows a passive behavior in the 9 g/l NaCl solution, with a corrosion current density between 10−6 and 10−7 A/cm2 and a passive region from 0.1 to 0.4 V with respect to Ag/AgCl. The OCP for uncoated AISI 316 L steel resulted to be about 0.1 V with respect to Ag/AgCl. All four 100 nm ALD configurations showed a shift of the OCP to about −0.3 V respect to Ag/AgCl. The TiO2 sample showed a reduction of the Corrosion

E. Marin et al. / Thin Solid Films 522 (2012) 283–288

287

Fig. 4. Vickers indentation on the 100 nm TiO2 coated sample (a) and image postprocessing results obtained on the same region (b).

Current Density to less than 10−7 A/cm2 and a passive region from −0.3 to 0.2 V respect to Ag/AgCl. The other three samples, Al2O3, bilayer and 4-layer configurations showed a stronger corrosion current density reduction, to less than 10−10 A/cm2 with wider passive regions and a shift of the breakdown potential quite over 0.6 V respect to Ag/AgCl. Polarization curves for the samples with a total coating thickness of 30 nm are shown in Fig. 6b. Both 30 nm configurations showed an improved corrosion resistance compared to uncoated AISI 316 L Stainless steel. In detail, both configurations showed a shift of the OCP to about − 0.25 with respect to Ag/AgCl. The sample coated with a single layer of Al2O3 showed a Corrosion Current Density decrease to about 10 −8 A/cm 2 and a passive region of about 0.65 V. The multilayer configuration showed a greater decrease of the Corrosion Current Density, to about 10 −9 A/cm 2 and a passive region of about 0.45 V. Polarization curves for the samples with a total coating thickness of 10 nm are shown in Fig. 5c. As already shown in 100 nm and 30 nm coated samples, both configurations showed an improved corrosion resistance compared to uncoated AISI 316 L stainless steel. Both configurations showed a shift of the OCP to about −0.05 respect to Ag/AgCl. The sample coated with a single layer of Al2O3 showed a Corrosion Current Density decrease to about 10 −8 A/cm2 and a passive region of more than 0.6 V. The multilayer configuration showed a greater decrease of the Corrosion Current Density, to about 10−9 A/cm 2 and a passive region of about 0.55 V. In Fig. 5d, a comparison between the corrosion resistance of the best configurations for each thickness value (300 nm, 100 nm, 30 nm and 10 nm) has been made. It appears clear that higher thickness multilayer

configurations give a higher corrosion protection of the substrate, with lower corrosion current densities and wider passive ranges. The efficacy of single layer Al2O3 ALD coatings for the corrosion protection of steel is confirmed by the experimental results of Dìaz et al. [18] and Potts et al. [32]. In particular, Dìaz et al. obtained similar current density reductions and increases of the breakdown potentials as a function of the coating thickness, clearly demonstrating that this effect is mainly caused by the reduction of residual porosity as the thickness is increased. Nevertheless, the results obtained in this paper clearly show that increasing the coating thickness is not the only efficient method to improve the corrosion protection, as multi-layer configurations proved to be more protective with respect to single layer's ones. The decrease of Ecorr observed on the coated samples with respect to the uncoated substrate may be related to the dissolution over time of the Al2O3 layers already observed in literature by Paussa et al. on silver substrates [36] and described in detail by Dìaz et al. on carbon steel substrates [37]. Following the results obtained in this work changing the coating thicknesses, optimized characteristics of both corrosion and delamination resistance cannot be achieved at the same time using just single layer configurations, as increasing the coating thickness will reduce its delamination resistance, while reducing the coating thickness will reduce its corrosion resistance. Multilayer configurations seem to overcome this limitation as increasing the number of layers clearly showed an improvement of both characteristics at the same time. In industrial practice, thinner ALD multilayers will probably be preferable in order to also obtain a sensible reduction of deposition time and thus bringing a cost reduction for the deposition process, also increasing the productivity.

4. Conclusions

Fig. 5. Vickers delamination test results, expressed as “delaminated areas per applied load” versus “applied load”.

Different mono/multilayer configurations of ALD coatings obtained using Al2O3 and TiO2 have been deposited on polished AISI 316 stainless steel substrates. No deposition discontinuities or cracks were observed by AFM and SEM analysis on the coated samples, but very small micro-blisters were detected on all samples containing TiO2 layers. AFM analysis at the coating/substrate interface showed that all depositions are close to their nominal thickness values and thus the final coating thickness can be precisely controlled by the number of deposition cycles. GDOES has proven to be a powerful tool for thickness evaluation of ALD layers. GDOES analysis proved to be strongly influenced by surface roughness and, for this reason, it was not possible to resolve correctly the 10 nm thick ALD configurations. GDOES also proved to be limited by the instrument resolution of 2 nm and GDOES technolnique, which

288

E. Marin et al. / Thin Solid Films 522 (2012) 283–288

Fig. 6. Polarization curves obtained in a 9 g/l solution of NaCl, for all different samples in comparison to uncoated AISI 316 stainless steel. (a) 100 nm thickness, (b) 30 nm thickness, (c) 10 nm thickness, (d) best configurations for each thickness value.

was not high enough to discriminate singular ALD layers in the 30 nm and 10 nm multilayer configuration. Adhesion tests performed using Vickers microhardness indentations proved to be a simple and reliable method for the evaluation of mechanical/adhesion properties of the ALD coatings. Adhesion proved to be directly dependant to the overall coating thickness. 10 nm coating showed almost no sign of delamination around the Vickers indentation, even under high loads (19.6 N), even if this may be caused by the fact that the coating could not be correctly resolved using SEM. 30 nm coatings showed little delamination, but only under relatively high loads. 100 nm coatings showed almost 5 times higher delamination compared to 30 nm coatings. Multilayer configurations proved to be more effective in the reduction of delamination, giving better results if compared to the relative monolayers. All different ALD coatings proved to be effective in the corrosion protection of the stainless steel substrate, reducing the corrosion current density and increasing the passive region. Corrosion resistance proved to be a function of the total coating thickness, but even so multilayer configuration proved to be effective in further reducing the corrosion current density if compared to monolayer of the same thickness. Combined characteristics of good corrosion and delamination resistance were obtained using thin multilayer strategies formed by a great number of alternate ALD layers, which are supposed to be more adequate for industrial applications thanks to the lower deposition costs resulting from the relatively short deposition time. References [1] M.G. Fontana, N.D. Greene, Corrosion Engineering, 3rd ed. McGraw Hill Higher Education, 1986. [2] K.M. Perkins, M.R. Bache, Int. J. Fatigue 22 (2005) 1499. [3] K. Asami, K. Hashimoto, Corros. Sci. 45 (10) (2003) 2263. [4] H. Amel-Farzad, M.T. Peivandi, S.M.R. Yusof-Sani, Eng. Fail. Anal. 14 (2007) 1205. [5] S.E. Ziemniak, M. Hanson, Corros. Sci. 44 (2002) 2209. [6] H. Zhang, Y.L. Zhao, Z.D. Jiang, Mater. Lett. 59 (2005) 3370. [7] D.H. Mesa, A. Toro, A. Sinatora, A.P. Tschiptschin, Wear 255 (2003) 139.

[8] J.D. Crossen, J.M. Sykes, G.A.D. Briggs, J.P. Lomas, in: Organic Coatings for Corrosion Control, 1998, p. 106. [9] P. Misaelides, A. Hatzidimitriou, F. Noli, A.D. Pogrebnjak, Y.N. Tyurin, S. Kosionidis, Surf. Coat. Technol. 180–181 (2004) 290. [10] C. Liu, G. Lin, D. Yang, M. Qi, Surf. Coat. Technol. 200 (2006) 4011. [11] V.H.V. Sarmento, M.G. Schiavetto, P. Hammer, A.V. Benedetti, C.S. Fugivara, P.H. Suegama, S.H. Pulcinelli, C.V. Santilli, Surf. Coat. Technol. 204 (2010) 2689. [12] R. Di Maggio, L. Fedrizzi, S. Rossi, P. Scardi, Thin Solid Films 286 (1996) 127. [13] C. Wang, F. Jiang, F. Wang, Corros. Sci. 46 (2004) 75. [14] C.R. Tomachuk, C.I. Elsner, A.R. Di Sarli, O.B. Ferraz, Mater. Chem. Phys. 119 (2010) 19. [15] D. Pech, P. Steyer, J.P. Millet, Corros. Sci. 50 (2008) 1492. [16] J. Kawakita, T. Fukushima, S. Kuroda, T. Kodama, Corros. Sci. 44 (2002) 2561. [17] C.X. Li, T. Bell, Corros. Sci. 48 (2006) 2036. [18] B. Díaz, J. Światowska, V. Maurice, A. Seyeux, B. Normand, E. Härkönen, M. Ritala, P. Marcus, Electrochem. Acta (2011), http://dx.doi.org/10.1016/j.electacta.2011.02.074. [19] E. Marin, A. Lanzutti, L. Guzman, L. Fedrizzi, J. Coat. Technol. Res. http://dx.doi.org/ 10.1007/s11998-011-9372-8. [20] C.X. Shan, X. Hou, K.-L. Choy, Surf. Coat. Technol. 202 (2008) 2399. [21] T. Suntola, J. Antson, U.S. Patent 4,058,430, 1977. [22] T. Suntola, Solid State Electron. 14 (1971) 933. [23] T. Stubb, T. Suntola, O.J.A. Tiainen, Solid State Electron. 15 (1972) 611. [24] T. Suntola, O.J.A. Tiainen, M. Valkiainen, Thin Solid Films (1972) 227. [25] T. Suntola, Thin Solid Films 34 (1976) 9. [26] M. Ahonen, M. Pessa, T. Suntola, Thin Solid Films 65 (1980) 301. [27] J. Lua, J. Aarik, J. Sundqvist, K. Kukli, A. Harsta, J.-O. Carlsson, J. Cryst. Growth 273 (2005) 510. [28] S.M. Bedair, in: D. Bloor, R.S. Brook, M.C. Flemings, S. Mahajan (Eds.), The Encyclopedia of Advanced Materials, vol. 1, Pergamon, Oxford, 1994, p. 141. [29] R. Matero, M. Ritala, M. Leskelä, T. Salo, J. Aromaa, O. Forsén, J. Phys. IV France 09 (1999) 493. [30] C.X. Shan, X. Hou, K.L. Choy, Surf. Coat. Technol. 202 (2008) 2147. [31] E. Marin, A. Lanzutti, L. Guzman, L. Fedrizzi, J. Coat. Technol. Res. (2011), http:// dx.doi.org/10.1007/s11998-011-9327-0. [32] S.E. Potts, L. Schmalz, M. Fenker, B. Dìaz, J. Światowska, V. Maurice, A. Seyeux, P. Marcus, G. Radnóczi, L. Tóth, M.C.M. van de Sanden, W.M.M. Kessels, J. Electrochem. Soc. 158 (2011) 132. [33] M. Ritala, Appl. Surf. Sci. 112 (1997) 223. [34] J. Takadoum, H.H. Bennani, Surf. Coat. Technol. 96 (1997) 272. [35] H.C. Barshilia, A. Ananth, J. Khan, G. Srinivas, Vacuum 86 (2012) 1165. [36] L. Paussa, L. Guzman, E. Marin, N. Isomaki, L. Fedrizzi, Surf. Coat. Technol. 206 (2011) 976. [37] B. Dìaz, E. Härkönen, V. Maurice, J. Swiatowska, A. Seyeux, M. Ritala, P. Marcus, Electrochem. Acta 56 (2011) 9609.