Printer Friendly

Chemical and electrochemical characterization of [TiO.sub.2]/[Al.sub.2][O.sub.3] atomic layer depositions on AZ-31 magnesium alloy.

Abstract In this study, innovative [TiO.sub.2]/[Al.sub.2][O.sub.3] mono/ multilayers were applied by atomic layer depositions (ALD) on ASTM-AZ-31 magnesium/aluminum alloy to enhance its well-known scarce corrosion resistance. Four different configurations of ALD layers were tested: single [TiO.sub.2] layer, single [Al.sub.2][O.sub.3] layer, [Al.sub.2][O.sub.3]/ [TiO.sub.2] bilayer and [Al.sub.2][O.sub.3]/ [TiO.sub.2]/[Al.sub.2][O.sub.3]/ [TiO.sub.2] multilayer deposited using A1[[([CH.sup.3])].sup.3] (trimethylaluminum, TMA), and [TiCl.sup.4] and [H.sup.2]O precursors. All depositions were performed at 120[degrees]C to obtain an amorphous-like structure of both oxide layers. The four coatings were then investigated using different techniques, such as scanning electron microscope (SEM), stylus profilometer, glow discharge optical emission spectrometry (GDOES) and polarization curves in 0.05-M NaCl solution. The thickness of all the coatings was around 100 nm. The layers compositions were successfully investigated by the GDOES technique, although obtained data seem to be affected by substrate roughness and differences in sputtering rates between ceramic oxides and metallic magnesium alloy. Corrosion resistance showed to be strongly enhanced by the nanometric coatings, giving lower corrosion current densities in 0.05-M NaCl media with respect to the uncoated substrate (from [10.sup.-4] to [10.sup.-8] A/[cm.sup.2] for the single layers and from [10.sup.-4] to [10.sup.-8] A/[cm.sup.2] for the bi- and multilayers). All polarization curves on coated samples also showed a passive region, wider for the bi-layer (from -0.58 to -0.43 V with respect to Ag/AgCl) and multilayer (from -0.53 to -0.38 V with respect to Ag/AgCl) structures.

Keywords ALD, [TiO.sub.2], [Al.sub.2][O.sub.3], AZ-31, Magnesium, Corrosion protection, Polarization curves


Magnesium alloys have attracted great scientific and industrial interest for their low densities and potentially high strength/weight ratios and therefore provide promising alternatives to aluminum alloys for lightweight components in different applications, from automotive to aerospace and biomedical industries. The main limitations for wider industrial application of magnesium alloys are their low melting points (around 650[degrees]C) and their strong sensitivity to corrosion in different environments. Those that have been investigated include various concentrations of NaCl; (1) biological fluids; (2) aqueous sulfate solutions; (3) ethylene glycol based engine coolants; (4) urban, marine and even rural environments; (5) and even simple humid air. (6) In addition to general corrosion, magnesium alloys may undergo galvanic corrosion, (7) stress corrosion cracking and hydrogen embrittlement, (8) and pitting corrosion. (9) Depending on the specific working environment, a wide array of techniques can be used to protect magnesium alloys, most of which are under intensive study: anodic oxidations, (10) solgel depositions (11) (plasma enhanced), chemical vapor depositions, (12) electrodepo-sitions, (13) and conversion coatings. (14)

Atomic layer deposition (ALD) is a modern deposition technique that can be considered an extension of the atomic layer epitaxy (ALE) technique patented by Suntola in the late 1970s. (15) The ALD (as with the ALE) process involves a sequential set of self-limiting surface reactions. As evidenced for the first time in 1980 by Ahonen et al., (16) it is the self-limiting characteristic of each reaction step that differentiates ALE and ALD from other chemical vapor deposition technologies. In ALD, each deposition cycle is clearly divided into four steps: At first, a precursor is injected into the deposition chamber. The precursor is chosen so that it cannot react with itself, so a single monolayer is formed as a result of its reaction with the substrate. In the second step, the chamber is purged with nitrogen or argon gas to remove excess 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 lime. At the third step, the second precursor is injected into the chamber. In the case of metal oxide layers, this is an oxidant agent, usually simple [H.sub.2]O. The last step of the deposition cycle is a second purge to remove excess reactant with purging gas. Closed-loop repetitions of the four basic steps theoretically produce conformal deposits of any desired thickness, as atomic layer control of film growth can be obtained as fine as 0.1 A (10 pm) per monolayer by keeping the precursors separate throughout the whole coating process.

ALD was initially developed for thin him electroluminescent (TFEL) flat-panel displays, in which high-quality dielectric and luminescent films were required on large-area substrates. (17) Interest in ALD technology then rapidly grew, in particular since the electronics industry began to consider it for manufacturing of ultra-thin gate dielectrics for advanced metal-oxide semiconductor transistors in the late 1990s and early 2000s. (18) In 1999, Malero et al. (19) investigated for the first time the suitability of ALD for "making corrosion-protection coatings," expanding ALD applicability to a great number of industrial applications and creating even more interest in the technology.

Up to the present time, ALD technology has successfully been used to deposit several types of nanometric films, including in particular various metal oxides (e.g., [A1.sub.2][O.sub.3], CaO, CuO, [Er.sub.2][O.sub.3] [Ca.sub.2][O.sub.3], [HfO.sub.2], [La.sub.2][O.sub.3], MgO, [Nb.sub.2][O.sub.5], [Sc.sub.2][O.sub.3], [SiO.sub.2], [Ta.sub.2][O.sub.3], [TiO.sub.2], [Y.sub.2][O.sub.3], [Yb.sub.2][O.sub.3], ZnO and [ZrO.sub.2]), nitrides (e.g., TiN, TaN, AlN, GaN, WN and NbN), sulfides (e.g., SrS and ZnS), carbides (e.g., TaC and TiC), fluorides (e.g., [CaF.sub.2], [LaF.sub.3] and [MgF.sub.2]), pure metals (e.g., Ru, Ir and Pt), biomaterials (e.g., hydroxyapatite ([Ca.sub.10] [([PO.sub.4]).sub.6] [(OH).sub.2])) and even polymers (e.g., PMDA-DAH and PMDA-ODA). (20-22)

In this study, innovative nanometric mono/multilayer coatings of [A1.sub.2][O.sub.3] and [TiO.sub.2] are applied by low-temperature ALD (23) on rough magnesium alloy AZ-31 substrates to enhance corrosion resistance in low-concentration Nad aqueous solutions. A low deposition temperature (120[degrees]C) was selected to avoid any microstructural modification of the magnesium substrate. In fact, the stress-relieving temperature for hard-rolled AZ-31 alloy is 150[degrees]C. (24)

The magnesium alloy substrates were intentionally left rough by the grinding process, obtaining a final roughness similar to that of standard extruded and cold-rolled AZ-31, to investigate the properties of ALD depositions in conditions similar to industrial standards.

In the literature, ALD has already proved to be strongly effective for corrosion protection of different metallic alloys, (19) and in particular for microelectrome-chanical systems, (25) stainless steel, (26) sterling silver, (27), (28) CrN-coated stainless steel (29) and tool steel coated with physical vapor deposition TiAIN and TiCN, (30) usually with the deposited coating's thicknesses ranging from 4 to about 100 nm. A substantial reduction of corrosion currents and widening of passive region of the substrate materials was frequently observed. In the case of AISI M2 tool steel coated with TiAIN using PVD, (31) for example, the application of just 4 nm of ALD [A1.sub.2][O.sub.3] lowered the corrosion current density from [10.sup.-6] to [10.sup.-8] A/[cm.sup.2] and gave a 0.5 V large passive region. The sensibly higher corrosion protection over thickness ratio obtained with the investigated ALD layers with respect to conventional treatments and the great flexibility of the deposition process are clear indications that this technique is perfectly suitable for corrosion protection of industrial components.

Protective coatings for magnesium alloys with similar composition obtained by CVD have been presented in literature before. (32) In this patent, a method to deposit different protective films on magnesium alloys using CVD is described using temperatures ranging from 250 to 320[degrees]C and obtaining hard and adherent coatings with an overall coaling thickness around 1 [micro]m. All the deposits described in the patent required an adhesion promoter layer of metallic aluminum to efficiently protect the substrate. Smaller thickness (about one order of magnitude) and absence of adhesion promoter seem to be the two main advantages of ALD technology.


AZ-31 magnesium alloy sheets (Table 1) were obtained by standard industrial treatments, cold rolling and partial annealing. Square 4 x 4 cm samples were then mechanically dry cut with a low-speed handsaw and dry grinded to remove the natural dark gray opaque surface oxide layer, obtaining a coarse final roughness of about 1700 nm [R.sub.a]. The sample's surface was then decreased using ethanol in ultrasonic bath and dried in a dry heat sterilizer at a temperature of 50[degrees]C for 15 min. Square 1 x 1 cm samples were obtained with the same procedure and then polished using ethanol in substitution of water until a surface roughness of about 20 nm was obtained. Sample surfaces were then degreased using ethanol in ultrasonic bath and dried in heat sterilizer. These samples were only used for measuring chemical profiles by glow discharge optical emission spectrometry (GDOES) tests. Samples were then coated using one of the four different ALD strategies: Sample A was coated with about 100 nm of [TiO.sub.2] (1000 precursor cycles of [H.sub.2]O and [TiCl.sub.4] at a chamber temperature of 120[degrees]C), Sample B was coated with about 100 nm of [A1.sub.2][O.sub.3] (730 precursor cycles of [H.sub.2]O and TMA at a chamber temperature of 120[degrees]C), Sample C was coated with about 50 nm of [A1.sub.2][O.sub.3] and then 50 nm of [TiO.sub.2] (365 precursor cycles of H20 and TMA followed by 500 precursor cycles of [TiO.sub.2] and TMA at a chamber temperature of 120[degrees]C) and Sample D was coated with about 25 nm of [A1.sub.2][O.sub.3], 25 nm of [TiO.sub.2], again 25 nm of [A1.sub.2][O.sub.3] and again 25 nm of [TiO.sub.2] (153 cycles per layer for [H.sub.2]O, TMA precursors and 250 cycles per layer of [H.sub.2]O, [TiCl.sub.4] precursors at a chamber temperature of 120[degrees]C). The desired thickness was obtained in between 6 and 8 h of deposition, depending on the composition. For comparison, a bare AZ-31 sample was put under a vacuum in the deposition chamber for 6 h at the same temperature to obtain a similar heat treatment at 120[degrees]C, but without any ALD deposition. Microstructural investigations have been carried out by other research groups, (33), (34) proving that deposits obtained at about 120[degrees]C have an amorphous-like structure.
Table 1: Standard compositioin of AZ-31 magnesiun alloy

Element    Al       Zn       Mn     [Fe.sub.max]  [Si.sub.max]

   %     2.5-3.5  0.6-1.4  0.2-1.0          0.01          0.08

Element  [Cu.sub.max]  [Ni.sub.max]   Mg

   %             0.01         0.001  Bal.

For each deposition procedure, a special silicon wafer sample with a standard extremely low [R.sub.a] (less than 2 nm), partially covered with heat-resistant laboratory adhesive tape, was put into the deposition chamber with the magnesium alloy samples to perform profilometric and thickness analysis and obtain the average deposition rate per single cycle of deposition.

SEM (Zeiss Evo-40) images of samples before and after deposition were used to investigate the morphological aspect of the AZ-31 substrate. Secondary and backscattered electrons scanning were performed to determine a possible superficial presence of Mg-Al intermetallics, inclusions, dust or residual SiC sand particles from the grinding process. A stylus profilometer (Veeco Dektak) was used to obtain roughness values averaged on 2 x 2 mm square maps of the sample surface before and after deposition. Profilometric 200-[mu]m linear scans were also performed on the silicon reference samples after adhesive tape removal and ethanol ultrasonic bath cleaning (15') to measure the thickness of the different deposited layers on a completely flat and low-roughness surface. GDOES (Horiba Jobin-Yvon RF Profilometer) was used to evaluate the semi-quantitative composition and thickness of the ALD deposits. Ail GDOES graphs were obtained under the same working conditions, 25 W applied power and 650 Pa chamber pressure, after a flushing time of 60 s. Certified reference materials (CRM) and setting up samples (SUS) of strictly controlled composition were used to obtain a specific calibration of the GDOES method to perform analysis on the nanometric ALD layers, even if no specific samples are actually available on the market for amorphous ceramic materials.

Polarization curves were performed with an Autolab PGSTAT-20 potentioslat using a standard three electrode configuration: a pure Platinum wire (99.99% Pt) as a counter electrode and Ag/AgCl as reference electrode. The chosen medium was a 0.05-M NaCl solution in distilled water and the tests were performed at controlled room temperature (20[degrees]C). A scanning speed of 1 mV/s was used. Bare AZ-31 heat-treated samples with the same surface roughness were also tested in the same conditions and used as a reference. Impedance measurements were carried out at open circuit potential with AC voltage amplitude of 10 mV and frequency range from 100 kHz to 10 mHz (ten points for each frequency decade). All measurements were performed using an AUTOLAB PG-STAT 12 potentiostat equipped with a frequency response analyzer module.

Results and discussion

Preliminary post-deposition optical analysis showed that most of the applied ALD coatings seem to be perfectly conformal to the substrate. Only [TiO.sub.2]-coated samples (Samples A) showed different brownish stains on the coated surface, which were initially supposed to be deposition defects with successive uncoated substrate oxidation.

Low-roughness silicon reference samples partially coated with depositions type A and B were used to verify the deposition rate per cycle of the two different ALD ceramic materials [TiO.sub.2] and [A1.sub.2][O.sub.3], respectively. For each sample, the mean overall thickness of the deposited layer, obtained by the profilomctric linear scans, was divided by the number of deposition cycles to obtain the coating thickness deposited per single cycle. Results are shown in Table 2. The assumed deposition rates were obtained in previous works with the same ALD processes and were used to estimate the number of precursor cycles required to reach a total thickness of 100 nm.
Table 2: Deposition rates per cycle as obtained from profifometer
thickness measures on silicon comparison samples

Precursors               [TiCl.sub.4]-[H.sub.2]O  TMA-[H.sub.2]O

Deposition rate          ~ 1.04 A/cycle           ~1.47 A/cycle
Assumed deposition rate  ~1 A/cycle               ~1.5 A/cycle

SEM images obtained before and alter ALD deposition (Figs, la and lb, respectively) did not show any clear morphological differences between the two conditions. The oriented roughness caused by the cold rolling and the mechanical grinding is clearly visible and similar on both samples. No presence of dust or included particles was found by both secondary or backscattered electron analysis before and after ALD depositions. Up to the maximum resolution of the instrument, no macrometric or micrometric evidences of coating defects, partial deposilions or exposed substrate were visible on coated samples of SEM images, even for the apparently defective [TiO.sub.2] deposit. Only the uncoated sample shows some very small corrosion phenomena probably occurred at intermetallic regions because of the humidity of the air.


Stylus profilometer 2x2 mm square maps (Fig. 2) clearly showed the coarse surface of the samples, which had a roughness ([R.sub.a]) of about 1700 nm both before and after the ALD deposition. This is clearly not surprising, because the thickness of the coating is less than one tenth of the total roughness and also because this coating technology allows the deposition of almost perfectly conformal films. Surface SEM analysis even at really high magnifications showed no fractures, flaws or Hakes of nonadherent coating.


Stylus analysis on silicon reference samples (Fig. 3 for Sample A) showed that the thickness is close to the nominal value for all the coatings (Table 3). Because silicon wafers and magnesium alloys have completely different chemical compositions with a different reactivity with respect to the coating precursors, the measured thickness has to be considered only indicative but is particularly relevant to understanding the high reliability of the deposition process.
Table 3: Thickness evaluation of the ALD depositions

Method                       Thickness(nm)
                Sample A     Sample  Sample  Sample
                               B       C       D

Profilometer          104.1   108.6    98.9    97.3
on silicon

GDOES on low            110     150     108     105

The GDOES technique was then used to evaluate the thickness of magnesium alloy substrates with 1700 nm [R.sub.a] (Fig. 4). These measurements proved to be strongly influenced by both surface roughness and sputtering rate differences between oxide coatings and metallic substrates. A strong overlap of signals coming from the coating and the substrate made impossible any thickness evaluation or compositional analysis. In fact, using one of the possible conventions for measuring the coating thickness, for example, by using the intersection point between the signals coming from the substrate (Mg) and the sum of the signals coming from the coating (Ti, Al and O), the total layer thickness of a type D coating seems to be around 650 nm (point #3 in Fig. 4). This is a value six times higher than the expected one, as obtained on the reference silicon samples. Furthermore, the signals of titanium and aluminum seem to be "diffused" for more than 1500 nm. While titanium shows the expected two peaks of the four-layer configuration, aluminum shows just one large peak that goes deep into the substrate. Oxygen signal is strongly overestimated for the coating and even deep into the substrate. These roughness effects on GDOES accuracy were accurately investigated by Shimizu et al. (35) For the configurations considered in this study, the phenomenon is explained as follows: As the plasma sputtering takes away the ceramic coating, it occasionally reaches the "soft" magnesium alloy substrate in some preferential points before completely removing the ceramic coating. Because magnesium alloy is easily sputtered by the GDOES argon plasma, sputtering proceeds at increased speed in the bulk alloy, while slowly eroding the remaining part of the coating. This effect gives a great overestimation of the coating thickness and an underestimation of the coating elements' presence and percentages.



After this preliminary approach, it was possible to state that the GDOES technique can hardly be correctly used to analyze the very rough magnesium alloy substrates. Then, to prevent or at least strongly reduce that inconvenience, GDOES analyses (Figs. 5-8) were all performed on the magnesium alloy samples with the smoother surface finishing (about 20 nm [R.sub.a]) obtaining very interesting data. All presented GDOES data are an average between at least three different measurements.


Surface oxygen peaks are present in all GDOES graphs. This is caused by the combined effect of superficial roughness, absorbed humidity and hydrogen effect, as stated by Payling et al. (36)

GDOES profiles proved to have a good reliability in ALD coating thickness evaluation on polished samples, showing data very close to those obtained by profilometric techniques. The thickness was estimated to be in the range of 100-110 nm (Table 3) for all the samples apart from Sample B, differing from the nominal values obtained by profilometric scans because of the systematical error of the GDOES technique (up to 10%).

For sample A, the coating thickness was about 110 nm (point #1 on Fig. 5), even if coating signals can be found up to 170 nm (point #2 in Fig. 5). This is because of the fact that surface mean roughness (20 nm) is comparable with the overall coating thickness, generating a certain signal scatter.

For sample B, the same consideration can be easily made, as the coating thickness was about 150 nm (point #1 in Fig. 6) but coating signals can be found up to 200 nm (point #2 in Fig. 6).


The substrate roughness could have a greater effect when multilayer configurations are considered, because the single-layer thickness is even closer to the roughness value. For Sample C, the overall coating thickness is about 108 nm (point #2 in Fig. 7) and the coating signals are still present up to 170 nm (point #3 in Fig. 7). An interface between the two single layers of the coating can be observed at about 55 nm (point #1 in Fig. 7), but the signals coming from the two different layers have an overlapping region of more than 60 nm.


For Sample D, the overall coating thickness is about 105 nm (point #4 in Fig. 8) and the coating signals are present up to 160 nm (point #5 in Fig. 8). In this sample, the single-layer thickness is exactly in the range of the surface roughness and for this reason, layer interfaces are strongly overlapped. Interfaces can be observed at 30, 60 and 80 nm (points #1, #2 and #3 in Fig. 8).



Because no reference materials were present for amorphous ceramics, a specific GDOES calibration was required. The 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 semiquantitative and for this reason the composition of the coatings and in particular of [Al.sub.2][O.sub.3] layers is not stoichiometric.

Polarization curves performed on the different samples are shown in Fig. 9. Results showed that all ALD depositions clearly enhance the corrosion resistance of the AZ-31 magnesium alloy substrate, which usually is very sensitive to the chosen sodium chloride-containing media. The uncoated sample shows a corrosion potential ([E.sub.corr]) of about -1.48 V with respect to Ag/AgCl and a corrosion current density close to [10.sup.-4] A/[cm.sup.2] with no presence of passive region. All samples containing a superficial [TiO.sub.2] layer (Samples A, C and D) show similar [E.sub.corr] values, between -1.57 and -1.61 V with respect to Ag/AgCl, and more negative with respect to bare magnesium alloy, whereas the [Al.sub.2][O.sub.3]-coated Sample B shows an [pounds sterling]corr similar to the uncoated magnesium alloy. This is an index of the different electrochemical behaviors of the two ALD layers. [E.sub.corr] shifting in the presence of [TiO.sub.2] layers can be explained supposing that ALD [TiO.sub.2] layers have a semiconductor behavior. In this hypothesis, the anodic curve is related to corrosion reactions happening on the Mg substrate because of residual porosities, whereas the cathodic reactions are triggered at the [TiO.sub.2] surface resulting in an [E.sub.corr] shift. This phenomenon will be further investigated in a dedicated experimental setting in near future.

Sample A shows a good corrosion protection, with a corrosion current of about [10.sup.-6] A/[cm.sup.2] and a passive region from -1.61 to -1.45 V with respect to Ag/AgCl. Sample B shows a similar corrosion current density, but with almost no passive region (about 0.02 V) before pitting breakdown. Samples C and D show a rather similar behavior between each other, with a corrosion current density of about [10.sup.-8] A/[cm.sup.2] and a passive region, about from -1.59 to--1.44 V with respect to Ag/AgCl for Sample C and from -1.57 to -1.38 V with respect to Ag/AgCl for Sample D. Sample D also shows slightly lower current densities in the passive region.

Preliminary EIS tests were also performed on the coatings to obtain further information about the durability of the ALD-coated samples. The impedance modulus obtained at low frequencies (1 Hz) after 1 h of immersion in 0.05-M NaCl solution is shown in Table 4. The results confirm that both ALD oxides have a protective effect when applied on AZ-31 magnesium alloy, increasing the overall impedance modulus about one order of magnitude. As already observed on the polarization curves, ALD multilayers have an effect in further improving the corrosion resistance of the substrate, which shows a further improvement of the impedance modulus of more than two orders of magnitude, and even more in the case of the four-layers configuration (Sample D). Small diffused corrosion phenomena were only observed at the surface of Samples A and B after the EIS measurement, probably in correspondence of residual porosities. Taking into account that, at first approximation, low-frequency impedance modulus and corrosion current density are inversely proportional, a qualitative comparison between PC and EIS can be performed. For both single-layer ALD coatings, a difference in two orders of magnitude in corrosion current density corresponds to just one order of magnitude increase in impedance modulus, whereas for the two-layer configuration, a decrease of about four orders of magnitude in corrosion current density corresponds to a 2.5 orders of magnitude increase in impedance. The sample that showed the best electrochemical results is the four-layer configuration, with a four orders of magnitude decrease of corrosion current density corresponding to a three orders of magnitude increase in impedance modulus. An important feature of the impedance test is that it was carried out only after 1 h of immersion instead of just after immersion as in the case of the polarization curves. Therefore, it is possible to assume that during immersion, the coatings degradated, losing part of their protective ability. This may be caused by corrosion phenomena triggered at the residual porosities of the ALD coatings.
Table 4: Impedence modulus for all samples as obtained at1 Hz

                 Coating         Modulus ([mho])

Substrate  -                                 550

Sample A   [TiO.sub.2]                      5000

Sample B   [Al.sub.2][O.sub.3]              4000

Sample C   [Al.sub.2][O.sub.3]/            90000

Sample D   [Al.sub.2][O.sub.3]/           150000

The results of electrochemical tests performed in this study suggest that thin nanometric ALD coatings can protect AZ-31 magnesium alloys more efficiently than thicker micrometric CVD coatings of the same composition as the ones obtained by Dabosi et al. (32) The decrease in corrosion current density seems to be higher for ALD coatings, probably because of the lower density of defects in the thin amorphous ALD multilayers.

Every ALD deposition can be formed by hundreds of singular cycles that can be theoretically considered layers. As observed by Diaz et al., (33) ALD coatings improve corrosion resistance because they are formed by a high number of singular layers and thus can be considered multilayer structures. Furthermore, the better sealing effect shown by chemically alternated multilayer configurations may be explained by taking into account the different bounding reactions of TMA and [TiCl.sub.4] (37), (38) that may result in the nucleation of oxides at different locations of previous defects, improving the overall coverage.

Similar beneficial effects of multilayer configurations, by sealing defects or porosities of a layer with the subsequent one, were frequently reported in literature. An example is the work of Dobrzanski et al., where up to 150 PVD layers were overlapped to obtain a 6.5-[mu]m coating thickness. (39) Because the number of layers proved to be a critical parameter, new multilayer configurations will be tested in the future, as well as hybrid coatings obtained by different techniques.


Four different ALD coatings, both mono- and multilayer, were properly applied to magnesium AZ-31 alloy substrates.

A stylus profilometer map showed that the pristine coarse roughness (about 1700 nm [R.sub.a]) of the substrate is completely unaltered by ALD. The GDOES technique gave semi-quantitative information about thickness and composition. Polarization curves showed that ALD deposition can be successfully used to protect AZ-31 alloy against corrosion in aqueous solutions with low concentrations of NaCl. Multilayered structures proved to be more effective against corrosion, showing a lower corrosion current density and a wider passive region. These coatings can behave even better than thicker coatings obtained by PVD.

The results obtained in this study clearly show that ALD may be an interesting technology to be used for the corrosion protection of magnesium alloys, the deposition time being the main limitation: ALD deposition requires an intrinsically long lime to be performed. Nevertheless, following the Gray and Luan review of corrosion protection for magnesium alloys (2002), (40) experimental ALD coatings seem to have three advantages over the actual and consolidate commercial techniques investigated by these authors: (i) The reactive gases are not expensive, (ii) the deposited layers are nanometric and well controlled and (iii) the coating itself has a low residual porosity.


(1.) Pardo, A, Merino, MC, Coy, AE, Arrabal, R, Viejo, F, Matykina, E, "Effect of Mo and Mn Additions on the Corrosion Behaviour of AISI 304 and 316 Stainless Steels in [H.sub.2][SO.sub.4]." Corros. Sci., 50 823-834 (2008)

(2.) Alvarez-Lopez, M, Pereda, MD, del Valle, JA. Fernandez-Lorenzo. M, Garcia-Alonso, MC, Ruano. OA, Escudero, ML, "Corrosion Behaviour of AZ31 Magnesium Alloy with Different Grain Sizes in Simulated Biological Fluids." Acta Biomater.,6 1763-1771 (2010)

(3.) Yang, L-J, Wei, U-H, Hou, L-F, Zhang. D. "Corrosion Behaviour of Die-Cast AZ91D Magnesium Alloy in Aqueous Sulphate Solutions." Corros. ScL. 52 345-351 (2010)

(4.) Song, G, StJohn. D. "Corrosion Resistance of Aged Die Cast Magnesium Alloy AZ91B," Corms. Set., 461381-1399 (2004)

(5.) Jonsson, M, Persson, D, Leygraf, C, "Atmospheric Corrosion of Field-Exposed Magnesium Alloy AZ91D." Corros. Sci., 50 1406-1413 (2008)

(6.) Lindstreom, R, Johansson. LG, Thompson. GE, Skeldon. P, Svensson, JE, "Corrosion of Magnesium in Humid Air." Corms, 5d.. 46 1141-1158 (2004)

(7.) Song. G. Johannesson, B, Hapugoda, S, StJohn, D, "Galvanic Corrosion of Magnesium Alloy AZ91D in Contact with an Aluminium Alloy, Steel and Zinc." Corros. Sci. 46 955-977 (2004)

(8.) Songa, RG, Blawert, C, Dietzel, W, Atrens, A. "A Study on Stress Corrosion Cracking and Hydrogen Embrittlement of AZ31 Magnesium Alloy." Mater. Sci Eng., 399308-317 (2005)

(9.) Martin, HJ, Horstemeyer. MF, Wang. PT, "Comparison of Corrosion Pitting Under Immersion and Salt-Spray Environments on an As-Cast AE44 Magnesium Alloy." Corros. Sci, 52 3624-3638 (2010)

(10.) Wang, X, Zhu, L, Li, W, Liu, H, Li, % "Effects of Half-Wave and Full-Wave Power Source on the Anodic Oxidation Process on AZ91D Magnesium Alloy." Appi Surf. Sti, 255 5721-5728 (2009)

(11.) Galio, AF. Lamaka, SV, Zheludkevich, ML. Dick, LFP, Mullen 1L, Ferreira, MGS, "Inhibitor-Doped Sol-Gel Coatings for Corrosion Protection of Magnesium Alloy AZ31." Surf. Coat. Technol., 204 1479-1486 (2010)

(12.) Angelini, E, Grassini, S, Rosalbino, F, Fracassi, F, d'Agos-tino, R, "Electrochemical Impedance Spectroscopy Evaluation of the Corrosion Behaviour of Mg Alloy Coated with PECVD Organosilicon Thin Film." Prog. Org. Coat., 46 107-111 (2003)

(13.) Lei, T, Ouyang, C, Tang, W, Li, LF, Zhou. LS. "Preparation of MgO Coatings on Magnesium Alloys for Corrosion Protection." Surf. Coat. Technol., 204 3798-3803 (2010)

(14.) Ardelean, H, Frateur, I, Marcus, P, "Corrosion Protection of Magnesium Alloys by Cerium, Zirconium and Niobium-Based Conversion Coatings." Corros. ScL, 50 1907-1918 (2008)

(15.) Suntola, T, Antson, J, US Patent 4,058,430, 1977

(16.) Ahonen, M, Pessa, M, Suntola, T, "A Study of ZnTe Films Grown on Glass Substrates Using an Atomic Layer Evaporation Method." Thin Solid Films, 65 301-307 (1980)

(17.) Lappalainen. R, Jokinen, J, Li, W-M, Ritala. M, Leskeh, Ms Soininen, E, "Complementary Analysis of ALE-Grown SrS Based Thin Film Electroluminescent Structures with Ion Beam Methods." Nucl Instr. Meth. Phys. Res. B. 132685-696 (1997)

(18.) Lua, J, Aarik, J, Sundqvist, J, Kukli, K, Harsla, A. Carlsson. J-O, "Analytical TEM Characterization of the Interfacial Layer Between ALD [HfO.sub.2] Film and Silicon Substrate.".J Cryst. Growth, 273 510-514 (2005)

(19.) Matero, R, Ritala, M, Leskela, M, Salo, T, Aromaa, J, Forsen, O, "Atomic Layer Deposited Thin Films for Corrosion Protection." J. Phys. IV France, 9 493-499 (1999)

(20.) Bedair, SM, In: Bloor, D, Brook, RS, Flemings, MC, Mahajan, S (eds.) The Encyclopedia of Advanced Materials, Vol. 1, pp. 141-153. Pergamon, Oxford, 1994

(21.) Niisto, L, "Atomic Layer Epitaxy." Curr. Opin. Solid State Mater. ScL9 3 147-152 (1998)

(22.) Ritala, M, "Advanced ALE Processes of Amorphous and Polycrystalline Films." Appi Surf. Sci, 112 223-230 (1997)

(23.) Leskela, M, Ritala, M, "Atomic Layer Deposition (ALD): From Precursors to Thin Film Structures." Thin Solid Films, 409 138-146 (2002)

(24.) ASM Handbook, Vol. 4, Heat Treating. ASM International, pp. 2010-2030(1991)

(25.) Hoivik, ND, Elam, JW, Linderman, RJ, Bright, VM, George, SM, Lee, YC, "Atomic Layer Deposited Protective Coatings for Micro-Electromechanical Systems." Sens. Actuators A, 103 100-108 (2003)

(26.) Shan, CX, Hou, X, Choy, KL, "Corrosion Resistance of Ti02 Films Grown on Stainless Steel by Atomic Layer Deposition." Surf, Coat Technol., 202 2399-2402 (2008)

(27.) Guzman, L, Marin, E, Paussa, L, Isomaki, N, Fedrizzi, L, "Protection of Silver Surfaces Against Tarnishing by Means of Alumina/Titania-Nanolayers." Surf. Coat. Technol. (2011). doi:10.1016/j.surfcoat.2011.03.101

(28.) Marin, E, Lanzutti, A. Guzman, L, Fedrizzi, L, "Atomic Layer Depositions: State of the Art and Research/Industrial Perspectives." Corros. Rev. (2011). doi: 10.1515/COR-ROSRE.2011.0010.R1

(29.) Shan, CX, Hou, X, Choy, KL, Choquet, P, "Improvement in Corrosion Resistance of CrN Coated Stainless Steel by Conformal [TiO.sub.2] Deposition." Surf. Coat. Technol, 202 2147-2151 (2008)

(30.) Marin, E, Guzman, L, Lanzutti, A, Fedrizzi, L, Saikkonen, M, "Chemical and Electrochemical Characterization of Hybrid PVD/ALD Hard Coatings on Tool Steel." Electro-chem. Commun., 11 2060-2063 (2009)

(31.) Guzman, L, Marin, E, Lanzutti, A, Borin, A, Magnan, M Malinen, V, Fedrizzi, L, Proceedings of "E-MRS Spring Meeting," 2009

(32.) Dabosi, FJP, Morancho, R, Pounteau, D, US Patent 4980203, 1990

(33.) Diaz, B, Harkonen, E, Swiatowska, J, Maurice, V, Seyeux, A, Marcus, P, Ritala, M, "Low-Temperature Atomic Layer Deposition of [A1.sub.2][O.sub.3] Thin Coatings for Corrosion Protection of Steel: Surface and Electrochemical Analysis." Corros. ScL 53 2168-2175 (2011)

(34.) Aarik, J, Aidla, A, Mandar, H, Uustare, T, "Atomic Layer Deposition of Titanium Dioxide from TiCl4 and H20: Investigation of Growth Mechanism." Appl Surf. Sci., 172 148-158(2001)

(35.) Shirnizu, K, Brown, GM, Habazaki, H, Kobayashi, K, Skeldon, P, Thompson, GE, Wood, GC, "Influence of Surface Roughness on the Depth Resolution of GDOES Depth Profiling Analysis." Surf. Interf. Anal, 27153-156 (1999)

(36.) Payling, R, Jones, DG, Bengtson, A, Glow Discharge Optical Emission Spectrometry. John Wiley, Chichester, 1997

(37.) Granados-Aplizar, B, Muscat AJ, "Surface Reactions of [TiCl.sub.]4 and A1[([CH.sub.3]).sub.3] on GaAs(100) During the First Half-Cycle of Atomic Layer Deposition;' Surf. ScL, 605 1243-1248 (2011)

(38.) Ghosh, MK, Choi, CH, "Initial Adsorption Mechanisms of TiCl4 on OH/Si (100)-2 X L" Chem. Phys. Lett, 45769-73 (2008)

(39.) Dobrzanski, LA, Lukaszkowicz, K, Zarychta, A, Cunha, L, "Corrosion Resistance of Multilayer Coatings Deposited by PVD Techniques onto the Brass Substrate." J. Mater. Process. TechnoL, 164-165 816-821 (2005)

(40.) Gray, JE, Luan, B, "Protective Coatings on Magnesium and Its Alloys--A Critical Review." J. Alloys Compd, 336 88-213(2002)

E. Marin (*), A. Lanzutti L. Guzman, L. Fedrizzi

University of Udine, via Cotonificio 108, 33100 Udine, Italy


DOI 10.1007/sl1998-011-9372-8
COPYRIGHT 2012 American Coatings Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Marin, E.; Lanzutti, A.; Guzman, L.; Fedrizzi, L.
Publication:JCT Research
Article Type:Report
Date:May 1, 2012
Previous Article:Surfactant-induced growth of a calcium hydroxide coating at the concrete surface.
Next Article:Characterization and property of microarc oxidation coatings on open-cell aluminum foams.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters