CORROSION RESISTANCE

Professor Dr. Loutfy H . Madkour

CORROSION RESISTANCE

Abstract

A Systematic Study and Characterization of Advanced Corrosion Resistance Materials and Their Applications for Plasma Etching Processes in Semiconductor Silicon Wafer Fabrication Hong Shih Etch Products Group, Lam Research Corporation, Fremont, California, USA

Figures (392)

Fig. 2. Aluminum metal film stack and common issues in etching processes [21, 22]. For etching process requirement, a metal etch film stack and common issues are shown in Fig. 2

Fig. 1. The relationship of chamber materials, etching processes, precision wet cleaning and etch byproducts in a plasma etching chamber [21, 22, 35, 36].

Fig. 3. Killer defects generated in aluminum metal etch processes.

In pattern A, chamber materials can be etched/sputtered by plasma uniformly. The etch rate can be very low or very high. The etch rate depends on the plasma chemistry, process recipe, and materials. For example, high purity Y2O3 has showed very high plasma resistance in both Metal and Silicon etch processes. A uniform corrosion/ erosion pattern is observed [21, 22, 25, 30]. For anodized aluminum, a very high corrosion/erosion rate is observed under BCl3-containing plasma during metal etch processes. In fact, an anodized aluminum film with a 75 um in thickness (hot deionized water sealed) can only hold up to 1,800 wafers in some etch process recipes in production. This became a severe problem on the lifetime of anodized aluminum in aluminum etch processes. For Silicon etch processes, the lifetime of anodized aluminum has no issue because there is no obvious attack of reactive gases to anodized aluminum in Silicon etch processes. The only concern is the formation of AIOF on anodized aluminum surface when SF¢ and NF; are used in the etching processes. The formed AIOF can either have chamber particle issue or cause etch process shift due to the surface impedance change on anodized aluminum surface. The wet cleaning to fully remove AIOF film on anodized aluminum surface is very critical to achieve a consistent and reliable etching performance on wafer fabrication. Fig. 5 shows an anodized aluminum metal etch chamber after 1,800 wafer fabrication in production. The anodized aluminum is fully removed under Cl2./BCl3 high density plasma [21, 22, 25, 30]. The major attack of anodized aluminum is due to the chemical reaction between BCl3 and AlsO3 under the high density plasma. The reaction rate of the attack to anodized aluminum highly depends on the gas concentration of BCl3 and the plasma density. Chamber erosion test indicates that Clz has little attack to anodized aluminum [21, 22, 25, 30]. The high density plasma reaction rate of BCl3; with anodized aluminum or high purity alumina at different flow is shown in Fig. 6. The high reaction rate occurs on chamber top window due to both high density plasma and gas flow. On the chamber wall, the reaction

Fig. 6. The maximum reaction rate of AlxO3 at different BCl3 gas flow under high density plasma [21, 22, 25, 30]. rate of BCl3 with Al:O3 is almost a liner relationship, but the reaction rate is much lower than that on the chamber top window. It also indicates that without BCl3 flow, the reaction rate of Clo plasma has almost no attack to anodized aluminum or to high purity alumina. In the plasma reaction rate study, the total flow is fixed as of 205 sccm. The Argon gas flow is fixed at 40 sccm. The test starts at 165 sccm Cl: flow and zero flow of BCls, then 155 sccm Clo flow and 10 sccm BCI; flow, until the final flow of Cl: is 85 sccm and BCl3 flow is 80 sccm. The test coupons are either on chamber top window or on the chamber wall. Nine different types of anodized aluminum and high purity alumina are used in the test [21, 22, 25, 30]. The reaction rate is in the unit of mils per RF hour.

Fig. 8. The ceramic surface of a used electrostatic chuck contains about 33 atom% fluorine on the surface film with a rainbow color. Fig. 7. A uniform AIOF film (rainbow color) covers the ceramic surface of a used electrostatic chuck after silicon etch processes.

Fig. 10 shows the test results of various materials obtained from worldwide suppliers. The letters of A, B, C, D et al represent the suppliers and their materials. Agreements were signed for not allowing to release the names of the worldwide suppliers and their materials. The plasma etching rate is in the unit of mils (1 mil = 25.4 1m). It is obvious that either YAG (solid solution of AlsO3 and Y2Os3) and solid Y2O3 can reduce the plasma etching rate at the order of 40-50 times in comparison with the previously used chamber materials such as high purity alumina. That is the reason why Y2O3 has been as one of the leading chamber materials in plasma etching tools in the past 10 years for the leading semiconductor etching equipment companies.

Fig. 12. Wet cleaning compatibility test of four configurations of boron carbide coated rings. The wet cleaning compatibility of four configurations is also tested by soaking the large size B,C coated rings in saturated AlCl; solution at pH=0 for 90 minutes, then put the rings in an environmental chamber to monitor the time when boron carbide coating starts to peel off. The test sequence is shown in Fig. 12 [25].

Fig. 11. After plasma etching for 200 RF hours, Boron carbide coated anodized aluminum sealed with HL126 sealant provides the best corrosion in all configuration. and fast in comparison with ASTM standard salt spray test method [45, 46]. The test results show that boron carbide coating on anodized aluminum and sealed with HL126 provide the best corrosion resistance among the four configurations as shown in Fig. 11 [25].

n order to qualify boron carbide coating as a new chamber material, many aspects have to e considered. One of the concerns is the impact to ICF (ion current flux). Three onfigurations are considered and compared in the etching chamber. The ICF of anodized luminum. chamber is used as the baseline. Boron carbide coatings on bare aluminum or on inodized aluminum are studied through ICF measurements. The results showed that the hree configurations have the compatible ICF. The results of ICF measurements are shown n Fig. 13 [21, 22, 25, 30].

There is no metal contamination introduced when a high purity boron carbide coating is introduced as the new chamber wall coating.

Al Etch Rate and Nonuniformity Fig. 16. Al etch rate and etch rate non-uniformity during a 3,000 wafer marathon.

Fig. 15. Gas-only and RF-on particles during a 3,000 wafer marathon. The up limit of allowance of defect and particles is defined as 50 adders/ per wafer.

Fig. 17. The boron carbide coated chamber shows an excellent aluminum etch performance on a feature size as of 0.15 um through a 2,000 wafer marathon.

Fig. 18. Particle adders as of 2 0.12 tm size in a 1,000 wafer marathon on a 300mm etch tool The boron carbide coating on anodized aluminum with HL126 sealant is used as the chamber coating to replace anodized aluminum.

Fig. 19. Production yield improvement at wafer fabrication when new chamber material, hardware and best-known method are implemented. Table 3. Gas-only and RF-on particles of old and new chamber configurations *: Unit in particle counts/ wafer and particle adders at 0.2 um or larger are recorded.

The new boron carbide coating has been introducing to worldwide wafer fabrication for over 10 years with over 1,000 chambers introduced to wafer fabrication in IC industry. The chamber lifetime has demonstrated to improve from the worse case as of 60 RF hours (1,800 wafers) under BCl3/Cl) etching plasma to over 4,000 RF hours or longer in semiconductor wafer fabrication in the world. It also demonstrates that the chamber materials play a critical role in semiconductor etching equipment, particularly, for the cost reduction. A short comparison of anodized aluminum and born carbide coating is highlighted in Table 4 [21, 22, 25, 30]. Anodized aluminum has been using as the major etching tools surface coatings since 1980. It still received a lot of applications in plasma etching tools because of its low cost, easy to manufacture, easy to make large or small sizes of the parts, wide applications, easy to refurbish, and achieving good quality control at different suppliers in the world. Therefore, the study of anodized aluminum has always been a major task for the major semiconductor etching tool manufacturers. For high purity Y2O3 thermal spray coating, it has been qualified and applied as one of the major chamber components in plasma etching tools in the past 10 years. It is still one of the major materials as coating or as a solid sintered material which is used in plasma etching tools. At Lam Research Corporation, great attentions have been paid in the improvements and the new development of anodized aluminum and Y2Os coatings.

Fig. 20. The typical interface model of anodized aluminum with a hot DI water seal. The principal crystallographic defects are (1) vacancies: Vo'and Vm* for MOx;2; (2) interstitials: O; and Mix*. In fact, oxide films can be described as exponentially-doped semiconductor junctions. The fundamentals and process optimization of anodized aluminum have been studied thoroughly by Brace, Thompson, Wood, Mansfeld, and recent years by Shih through the comprehensive studies of anodization of different aluminum alloys, different anodization processes, and different manufacturing processes [51 - 60]. The interface model of anodized aluminum with hot DIW seal has been described by Mansfeld, Kendig, Shih and others [61- 72] as shown in Fig.20.

The overall impedance and HCI bubble test results are shown in Table 5. EIS data of three test coupons after immersion in 365 days in 3.5wt% NaCl solution are analyzed using the software written by Shih and Mansfeld called “ANODAL” [63-69].

In Table 6, @ is called frequency dispersion which is related to surface inhomogeneties with different dimensions [83]. Chi-sq is the fitting error between the experimental data and fitted data at each frequency. The detailed calculation is shown as below [84] and is the sum of the fitting error at each frequency multiplying 100 and dividing the total data points.

A Systematic Study and Characterization of Advanced Corrosion Resistance Materials and Their Applications for Plasma Etching Processes in Semiconductor Silicon Wafer Fabrication Z

Fig. 24. Acidic corrosion resistance of two test coupons of anodized aluminum. On the left, anodized aluminum does not show any acidic corrosion in two hours and on the right, anodized aluminum shows severe acidic corrosion after only 10 minutes immersion in the acid. For acidic corrosion resistance of anodized aluminum, HCl bubble test is an easy and very effective method to obtain the corrosion resistance. From Fig. 24 below, one can see the good and poor anodized aluminum under the solution of 5wt% HCl solution (28, 31-33, 44, 78]. On the left of Fig.24, there is no any hydrogen bubble generated under the attack of a strong acid within 2 hours immersion. It indicates a high quality of anodized aluminum. On the right of Fig. 24, anodized aluminum generates a lot of hydrogen bubbles in 5wt% HCl solution only after 10 minutes immersion in the acid. It indicates a poor anodized aluminum.

Fig. 23. TEM pictures of a standard type III hard anodization (left) and a mixed acid anodization (right) [80]

The HCl bubble test can be processed at any position of etching chamber before and after etching process. In Fig. 25, one process chamber is studied on its corrosion resistance in 5.0wt% HCl solution [76]. This method has received a wide application for the corrosion resistance study of anodized aluminum.

Assuming that C., Cpo and Cp are capacitances which represent the capacitances of Y2O3 coating layer, porous layer of anodized aluminum and barrier layer of anodized aluminum, respectively. The interface parameters can be obtained and the coating quality can be monitored. The interface model can be described as the following equation [28, 78].

In order to study Y2O3 coating on anodized aluminum, the following electrochemical cell configuration is used to study the overall impedance and interface model of the coated samples or parts as shown in Fig. 27 [28, 78].

Table 7. Interface parameters of three spraycoated Y2O3 on anodized aluminum

Table 1. Chemical composition of the Zn and Al used to prepare the alloys.

Corrosion Resistance of Directionally Solidified Casting Zinc-Aluminum Matrix

Corrosion Resistance of Directionally Solidified Casting Zinc-Aluminum Matrix In all cases, the E / I response of the alloys shows the typical hysteresis indicates the phenomenon of pitting and found that the more susceptible are the composites than the alloys. The most susceptible are those containing neither SiC nor Al2O3 in the matrix; see Table 2 and Figure 5.

Corrosion Resistance of Directionally Solidified Casting Zinc-Aluminum Matrix The whole set of experimental impedance spectra can be discussed according to the following total transfer function.

High values of capacity confirm the formation of porous corrosion products, as can be seen in Figure 9. These high values of capacity may also be correlated with an increase in the area.

Fig. 9. Micrograph of ZA27 - 8%SiC sample after EIS test. The ZA27 alloy with different structures is less resistant to corrosion and its susceptibility to corrosion is dependent of the structure. The ZA27 - 15%CSi composite has the highest corrosion resistance. Also, discriminating by type of composite materials, the MMCs with SiC are more corrosion resistant than those MMCs prepared with alumina particles.

Table 3. Principal parameters obtained from the EIS analysis. 4. Conclusions

Finally, some typical HNS grades will be discussed with regards to their industrial application.

With ex : interaction coefficient, [%X]: Concentration of the elements X in % It is obvious as per table 1 that specific elements will increase the nitrogen solubility (e.g. manganese), while others will reduce the solubility (e.g. silicon). This has not only an impact on the nitrogen pick up at remelting but also on the precipitation of inter-metallic phases in the solid state.

Corrosion Resistance of High Nitrogen Steels

However - and this is a difference to the conventional nitrogen free alloy variations - one should consider that HNS-alloys have a specific precipitation behavior. This must be kept in mind so that potential difficulties at hot forming at heat treatment can be avoided. Additionally, any precipitation will affect the corrosion resistance so a good understanding of the alloy is mandatory to maintain the alloy characteristic.

The precipitation behavior of this carbide can only be prevented through a quick quench in the critical temperature range. Fig.8 shows the location of the precipitation with relation to the alloy composition. Fine carbides are beneficial with regards to the corrosion resistance as the local chrome depletion is less in comparison to coarser carbides. The Cr depletion can be balanced out through an extended homogenisation (i.e. holding time) within the precipitation area.

In the case of the austenitic steels it should be considered, that in the temperature range of approx. 500- 900 °C and in connection with the alloy composition a precipitation of nitrides (Type CroN) occurs. This nitrogen perlite identified microstructure raises significantly the susceptibility to cracking of the steel but can also support intergranular cracking. Depending on the alloy composition, the precipitation window for nitrogen perlite or other nitrides are adjusted to higher or lower temperatures. The figures 9 & 10 show exemplary an austenitic structure with beginning and advance nitrogen perlite precipitation. Clear to recognize at what speed that the precipitation occurs.

Fig. 10. Advanced precipitation of nitrogen perlit cold worked austenitic structure 1.3816. 800 °C/30 min.

Fig. 12. TTT-diagram for the beginning of Cr2N-precipitation at different nitrogen contents. Curve is based on X5CrMnN18-12. [Rashev et al., 2003] ‘ig. 11. Precipitation of CryN in 18Mn18Cr at various nitrogen levels. [Uggowitzer, 1991]

Fig. 13. Breakdown potential of HNS and commercial stainless steels in various electrolytes. [ETE-11] A synergism of nitrogen and molybdenum is suggested by many authors [Mudali & Rai, 2004], [Pickering, 1988], [Pedrazzoli & Speidel, 1991]. Molybdenum shifts the metal dissolution to higher potentials which will consequently lead to an increased enrichment of nitrogen at the metal/oxide-interface. In this case, nitrogen can lower the current density below the critical value for pitting corrosion [Mudali & Rai, 2004].

Fig. 14. Influence of nitrogen on Current-density depending of a given alloy. [Pedrazzoli & Speidel, 1991]

Fig. 16. Critical Pitting Temperature (CPT) as a function of PRE. [Pedrazzoli & Speidel, 1991]

Fig. 18. MARC equation to rank various alloying element in regards to the alloy pitting resistance. | Speidel & Theng-Cui, 2003]

weight percent alloying elements C, N Fig. 19. Influence of various alloying elements on the pitting potential. [Pedrazzoli & Speidel, 1991]

The NHy+ - formation will increase the pH-value which support the repassivation. [Mudali & Rai, 2004] reports that NHg* - ions have been confirmed by XPS within the passivation layer.

Fig. 22. Precipitation of M23C6 and y-phase of a given alloy (17Cr-13Ni-4,5Mo) depending on annealing times and temperatures. [Gillessen et al., 1991] IGC is mainly driven by the depletion of Cr at the grain boundaries and/or the precipitations of carbides, usually M23Cs. Therefore, Carbon is supposed to be the main driver for IGC but also grain size, cold working and heat treatment have a significant influence on IGC.

It is commonly known that high alloyed steels are generally sensitive to SCC. The role of nitrogen against SCC has been discussed over the years but it appears that the positive effect of nitrogen is to be seen in a more oblique way. As previously discussed, nitrogen delays the carbide precipitation and avoids a local Cr depletion. Additionally, the crack growth velocity does not only depend on the actual Cr content, but also on the C content. The crack growth is much higher with increased C levels. It has been reported that nitrogen doesn’t have any influence on crack growth velocity at C > 0,5 % . The impact of carbon is therefore higher compared to that of nitrogen. [Pedrazzoli & Speidel, 1991]

Fig. 24. High precision bearing of Cronidur 30 [ETE-11]. Since its development in the early 1990s, Cronidur 30 (X30CrMo 15-1 plus 0,4 % nitrogen) has been approved as high performace alloy for aerospace applications such as spindles, shafts and bearings. It is used as material for helicopter bearings, flap traps, fuel pumps etc. It excellent corrosion resistance combined with a high hardenability of 60 HRC makes it unique. The alloy has become an important material to industries as cutleries and knifes, general engineering, medicals (since it is free of nickel) and powder metallurgy.

Fig. 27. Retaining ring. X8CrMnN 18-18 [ETE-11]. Fig. 26. Generator shaft with retaining rings, X8CrMnN 18-18 [ETE-11].

Fig. 1. Example of a third body resulting from a tribological process during reciprocating sliding. Some abrasive grooves are also visible in the central area of the sliding track.

Fig. 2. Electrochemical noise at the open circuit potential Eoc of 316L stainless steel in artificial sea water for different tribological contact conditions. S.C.E.: saturated calomel electrode. Fn: normal force. V: sliding speed. Table 1. Variations of the open-circuit potential Eoc of 316L stainless steel in artificial se water as a function of varying tribological contact conditions. S.C.E.: saturated calomel electrode. Fy: normal force. V: sliding speed.

Fig. 5. Schematic evolution of the current transients obtained under reciprocating sliding. Methods based on analysis of potential or current transients (Ponthiaux et al., 1995) are particularly well suited to study reciprocating sliding tests (Mischler et al., 1997). These methods are used to study between successive contact events, the rebuild of damaged surface layers (oxide, passive film ...). Under imposed polarization e.g. in the passivation range, at each stop-start event, a transient variation of current is noticed with time (see Figure 5). The charge corresponding to this transition can be attributed to the re-growth of a uniform film in the damaged area.

mpedance measurements recorded on a non-rubbed and a rubbed Fe-31% nickel alloy are shown in Figure 6. The measurements were done in 0.5 M sulfuric acid at a potential of -675 nV / SSE located in the active region (Boutard et al., 1985). The impedance diagrams were ecorded in a limited range of measurement frequencies. In particular, under sliding, the mpedance was not measured at frequencies below 0.01 Hz, to limit the duration of the neasurements and to avoid in this way the influence of the long-term evolution of the electrochemical state of the surface due to wear - induced changes in the “rubbed 1rea/unrubbed area” ratio. By taking these precautions, we can consider that the condition of average electrochemical steady-state required for impedance measurements is fulfilled.

After selecting the set of appropriate test conditions, measurements are done to collect information on the electrochemical behaviour of a material fully covered by a passive film. This is done by electrochemical tests in absence of any sliding. After immersion in the electrolyte, the open circuit potential, Eo,, is measured versus a reference electrode. In general, a stable value of Exc is obtained after some time of immersion. From an electrochemical point of view, a stable E.: is obtained when the long-term fluctuations of Eoc are below 1 mV min during a minimum of 1 hour. The time necessary to reach such a stationary open circuit potential in the test electrolyte is an important characteristic of a passivating process, and is called in this protocol as the reaction time characteristic, treac. The evolution of Eo. from immersion time on provides useful information on the electrochemical reactivity of the tested material in the test electrolyte (see Figure 7).

Table 1. Data showing the enhancement of the mechanical properties of Pb-free solders[9, 10].

omposite solder by mechanically intermixing nano-Al2O3 particles into Sn0.7Cu Pb-f older, and this composite solder shows improved mechanical properties. The best ten: trength realized for the composite, which contains 1.5 wt.% alumina, far exceeds trength of the eutectic Sn-Pb solder. Many authors have studied the effect of adding sing valled carbon nanotubes [19] or multi-walled carbon nanotubes [20, 21] on the mechani roperties of nano-composite solders. The data on the enhancement of the mechani roperties of nano-composite solders collected from some of the literature are listed in Ta (13, 14, 16, 22, 23]. Here, it should be stressed that although the addition of nanopartic nto solder matrices can improve the creep behavior[24], the effects on the corros esistance and mechanical properties of the nano-composite solders cannot be ignored.

Fig. 3. Surface morphology changes of solder balls after the salt spray test for 96 hrs: (a) Sn Pb solder, and (b) SAC solder[39].

Unlike Sn-Pb joints, which have a dual phase structure and block the path of corrosion due to the existence of phase boundaries, the SAC305 joint is basically pure Sn with coarse islands of Ag3Sn and CugSns intermetallic precipitate (Fig. 5). A corrosion crack can propagate and lead to additional corrosion along the way, without interruption from the Sn phase structure. Although both materials show strong resistance to corrosion, the localized nature of the corroded area at critical locations causes significant degradation in Sn-Ag-Cu solder joints[40].

[41]. Thus, in the electronics industry, corrosion has become a significant factor in recent years because of the extremely complex systems that have been developed and the increasing demand on their reliability [42, 43]. For example, using Cu and Sn metals allows fine-pitch interconnections to be fabricated at relatively low cost. These features make Cu-Sn based SLID bonding very appealing for 3D stacked applications (Fig. 6) [44].

Table 3. Aemf values for metals commonly used in microelectronics[4].

corr. corrosion potential; Icorr.: corrosion current density; Op: breakdown potential; A® = Deorr. - Pp, Pp: passivation range of solder alloy; Ip: passivation current density at above 460 mVsce. Increasing the copper content, which reacts with Sn to form IMC, significantly improves the corrosion resistance of solders and increases the corrosion current density (Icorr), as shown in Fig. 8, 9 and Table 4. At above 460 mVsce, the passivation current densities of all specimens are around 101A/cm2, with the declining sequence of Sn37Pb = CueSns > CusSn > Cu.

The joining of materials with solders generally results in a multi-layer structure in which IMC are formed between substrate and solders. Such a structure in a flip chip package is a galvanic couple. The galvanic corrosion behavior of the solder bump structures have a great effect upon reliability[45]. For instance, the galvanic current densities of the Sn solder with respect to the IMC CucSns and CusSn, and base Cu have been investigated (Fig. 7). It appears that Sn solder has a greater galvanic current density and thus is very subject to corrosion, and it is especially so in coupling with the formation of Cu3Sn layers than with Cu¢Sns layers. The galvanic current densities of the Sn37Pb solders of CugSn, Cu, and Cu¢Sns are about 38, 16, and 5 (pA/cm2), respectively.

Fig. 8. The potentiodynamic polarization curves of Sn37Pb solder, CugSns IMC, CusSn IMC, and pure Cu samples in a 3.5 wt.% NaCl solution [45].

Ecorr - corrosion potential, Ip - passivation current density, Icorr - corrosion current density, Ep - passive potential.

thickness of the IMC layer was achieved with the addition of 1 wt% of nano-TiO, particles. The thickness of the CueSns IMC layer was reduced by 51%. The results indicate that the growth of the CugSns IMC layer at the solder/pad interfaces of Sn3.5Ag0.5Cu is depressed through the small addition of nano-TiO2 particles[58]. With the addition of 0.5- 1 wt% nano-TiO2 particles, fracture occurred in all of the solder joints as cracks propagated through the Sn3.5Ag0.5Cu composite solder balls, which ruptured mostly along the submicro Ag3Sn IMC and solder matrix, as shown in Fig. 13a, b. This phenomenon is similar to that occurring in Pb-free Sn0.7Cu composite solder BGA packages[59].

Fig. 11. SEM image of the (a) Sn3.5Ag0.5Cu solder and (b) Sn3.5Ag0.5Cu -1TiO, nano- composite solder[57].

Fig. 12. Morphology of intermetallic compounds formed at the interfaces of the as-reflowed solder joints: (a) Sn3.5Ag0.5Cu, (b) (a) magnifications ; (c) d Sn3.5Ag0.5Cu-0.75TiOz; (d) (c) magnifications [58].

To achieve high reliability, solder materials must have high resistance to corrosive conditions such as moisture, air pollutants from industry, and oceanic environments[54]. Although corrosion of solder alloys is not currently a major problem for electronic devices used in normal environments, it may be a problem when they are used in harsh environments, such as oceanic environments. However, there is a lack of information regarding the corrosion resistance of nano-composite solders in corrosive environments.

®corr : corrosion potential; Icorr: corrosion current density; Ob : breakdown potential; A® = Gcorr - Ob. resistibility or the stability of the passive film on the Sn3.5Ag0.5Cu composite alloy surface. The corrosion potential (@corr) of the Sn3.5Ag0.5Cu nano-composite solder is slightly more passive than that of the Sn3.5Ag0.5Cu solder. This implies that a finer grain size produces more grain boundaries, which act as corrosion barriers. On the other hand, the breakdown potential (®,) of the Sn3.5Ag0.5Cu nano-composite solders becomes much more passive with the addition of oxide nanoparticles. As Table 6 also indicates, the $n3.5Ag0.5Cu solders possess a higher pitting tendency (smaller A® value) than the Sn3.5Ag0.5Cu nano-composite solders. Rosalbino et al. reported that the pit formation at the surface of Sn-Ag-M alloys is due to the dissolution of the tin-rich phase [56]. In addition, the corrosion current densities were obtained by using the TAFEL extrapolation method. The corrosion current densities of the Sn3.5Ag0.5Cu solders and Sn3.5Ag0.5Cu nano-composite solders were very similar.

Surface element concentration (wt. %) Table 7. Surface element concentration of different solders after potentiodynamic polarization tests[60].

In addition, Li et al.[46] studied the corrosion properties of Sn-Ag, Sn—Ag-Cu, Sn-Cu, and SnPb solder in 3.5wt.% NaCl solution with different scanning rates, and their results showed that the corrosion product on the surface was tin oxide chloride hydroxide (Sn3O(OH)2Cl). In our case, the presence of such a surface layer, instead of a tin oxychloride layer, cannot be ruled out due to the detection limits of energy-dispersive spectroscopy. In order to understand the reaction during the corrosion products, XRD has been used to analyse the corrosion products on the surface after the polarization tests (Fig. 17). The results show that all the Sn3.5Ag0.5Cu and Sn3.5Ag0.5Cu solder materials have the same corrosion product, Sn30(OH)2Clz, which is a complex oxide chloride hydroxide of tin[67]. This further confirms that the corrosion product on the Sn3.5Ag0.5Cu composite solders is Sn30(OH)2Ch.

Fig. 17. XRD spectra of different solder materials after polarization tests[60].

Table 1. Composition (wt %) for alloys C-x and SS 304.

Fig. 1 shows the anodic dissolution behaviour of alloys in 0.5 M H2SQu, while Table 2 summarizes relevant data. This figure reveals a well-defined passive region of 0 Vsue to 1.2 Vsue in all curves. All curves, except for the one at x = 0.25 (C-0.25), show a secondary passive region at 0.15 Vsu. This passivation is attributed mainly to the further oxidation or hydroxidation of the passive oxide film, thus altering the valence of Cr [18,19]. Fig. 1 also indicates that the secondary passive regions of C-0.50 and C-1.00 are more prominent than those of C-0 and C-0.25. This observation is due to the selective dissolution in the duplex FCC-BCC structure for C-0.50 and in the BCC-ordered BCC structure for C-1.00, as compared with C-0 and C-0.25 which are single FCC phase. In the active-passive transition region, different compositions at different secondary passivation potentials reveal different dissolution rates owing to a selective dissolution. This observation resembles that observed in duplex phase stainless steel [20,21].

Fig. 1. Potentiodynamic polarization curve diagrams for alloys C-x and SS 304 at 25°C.

[he results of potentiodynamic polarization were compared via performing 15-day-dippins und weight loss experiments. In the 15-day-dipping and weight loss experiments, the orrosion rates for C-0.50 and C-1.00 were markedly higher than those of C-0 and C-0.28 Fig. 2). This observation differs substantially from the values of Ico obtained from oolarization experiment (Fig. 1), in which the two groups only differ slightly, despite the act that the trend is the same. A previous study found a similar deviation in corrosior current densities obtained from weight loss test and potentiodynamic polarization methoc D5). Fig. 1 shows potentiodynamic polarization diagrams for the Al,CoCrFeNi alloys and SS 304. The alloys have better overall general corrosion behaviour, with a larger Ecorr and smaller Icorty Teri, ANA Tpass than SS 304.

Table 4. The fits for A(x, T) and Ea(x) in Icorr(x, T) = A(x, T) exp(-Ea(x)/RT). Table 3. Fit data for Tafel slopes of alloys C-x in 20 °C - 65 °C.

@ Anodic Tafel slope Ba in mV/ decade, the measured Tafel regions are with 40~50 mV of overvoltage. ’ Cathodic Tafel slope B. in mV/ decade, the measured Tafel regions are with 150~170 mV of overvoltage. ae See ae Se ee GSE ee ee ee ORs FRR Sena Ree Fig. 3 shows polarization diagrams of Al,CoCrFeNi alloys at various temperatures. A rising temperature decreased the Tafel slopes of anode (Table 3), increased Icon, and increased Ecorr and E; (the transpassive potential) slightly. The corrosion rate is directly related to loom, according to Arrhenius equation, Ico = A exp(-E,/RT) [13,26], where the pre-exponential factor A is generally independent of temperature and is a constant of alloys, where R denotes the gas constant, T denotes temperature, and E, denotes activation energy for corrosion. In the case of small experimental temperature range, E, is assumed to be independent of T. Consequently, E, can be obtained from In(Icorr) vs. 1/T plot. Fig. 4 shows such plots for the alloys and SS 304, indicating that E, increases with x. This finding suggests that the corrosion rate is more sensitive to temperature for a larger Al content than for a smaller Al content. The In(Icorr) vs. 1/T curves intersect with each other in a range of 23°C - 27°C. Beyond this temperature range, [corr increases with x. The situation is reversed at temperatures lower than 23°C, which is inconsistent with a situation in which E,s for all alloys increase with x from 20°C to 65°C. Hence, Ez, i.e., an intrinsic property of metal, and A, i.e., a surface property of metal, are determinative factors of Ico. While E, depends only on x, A depends on both x and temperature (Table 4). Therefore, although E, increases with x, A also increases with x. Combining the effects of E;and A explains the different corrosion behaviours of the alloys with an increasing x at temperatures exceeding 27°C and lower than 23°C. Thus, the performance of passive films, when Al is added, at higher temperatures becomes inferior to that without addition of Al. In determining Icon, A is more important than E, at temperatures exceeding 27°C, while E, is more important than A at temperatures lower than 23°C.

Electrochemical Passive Properties of Al,CoCrFeNi (x = 0, 0.25, 0.50, 1.00) High-Entropy Alloys in Sulfuric Aci

Fig. 4. The Arrhenius plots for alloys C-x and SS 304 at 20°C - 65°C.

Fig. 6. EIS equivalent circuits for alloys C-0, C-0.25, C-0.50, and C-1.00.

Fig. 7. Impedance and relative thickness (1/Q,) of oxide layer vs. Al content x plots. Electrochemical Passive Properties of Al,CoCrFeNi (x = 0, 0.25, 0.50, 1.00) High-Entropy Alloys in Sulfuric Acids

Fig. 8. Impedance and Ipass of oxide layer vs. Al content x plots.

From an adsorption viewpoint, adsorption competition always prevails on the alloy surface between chloride ions and dissolved oxygen atoms. Notably, no oxide layer forms once chloride ions adsorb on the alloy surface, in which the metal ions readily dissolve. Therefore, the adsorption of chloride ions increases the reacting current density (as indicated by a comparison of Figs. 1 and 9), subsequently increasing the rate of metal dissolution.

Table 6. EDS analyses (at %) for alloys C-0, C-0.25, C-0.50, and C-1.00. Microstructures for not H25Q4-immersed alloys C-0, C-0.25, C-0.50, and C-1.00 are with single FCC, single FCC, duplex FCC-BCC, and BCC-ordered BCC phases, respectively [6]. Table 6 lists the EDS composition for each phase in different alloys.

Fig. 10. Histogram of Epi for alloys C-x and SS 304 in solution of different Cl- ion molarity (M). and C-1.00, decrease to 0.2-0.5 Vs at a chloride concentration of 0.25 M (Fig. 10). A higher Al concentration in the alloys implies a lower value of Epi. For C-0, deterioration of the passive layer is attributed to the evolution of oxygen. Meanwhile, for C-0.25, C-0.50, and C- 1.00, the deterioration of passive layer is attributed to the pitting process. An increasing chloride ion concentration causes the chloride ions to cluster at the defect sites of the passive layer and severely attack the passive layer. Consequently, Epi shifts to a more active region.

microstructure. The composition of BCC phase after alloy immersion is close to the overall alloy composition before immersion, indicating that the BCC phase is a corrosion-resistant phase. Moreover, the change in overall composition after immersion is attributed to the selective dissolution of Al and Ni in the ordered BCC phase of this alloy (Table 6). This selective corrosion in Al and Ni-rich phase in C-0.50 and C-1.00, which results in the corrosion attack on Al and Ni, is due to the large bonding in Al and Ni [37]. Alloys containing this bonding readily react with (OH)- and (SOu,)? to form Al and Ni complexes and dissolve in a sulfuric solution. Accordingly, after immersion, the remaining compound in the less corrosive-resistant Al and Ni-rich phase is an oxide, rich in Cr, in the residual passive film.

Fig. 13. Metallograph of alloy C-1.00. (a), before; (b) & (c), after immersion; and (d), EDS line-scan results of the location indicated in (c).

Fig. 14. (a) Iai, (b) Ipass, (c) Wioss, (d) Qt, (e) Re, and (i) Icorr values vs. Al content x plots. Ipass locate at the passive region of polarization curve. The same tendency for Ici, Ipass, ane Wiooss here indicates that the spontaneous passivation occurs for Al,CoCrFeNi, i.e., the oper circuit potential (OCP) is readily in the passive region of polarization curve. The above phenomenon can be attributed to the spontaneous passivation of pure Al in H2SOx [24]. EIS equivalent circuits reveal that the passive layers of AlxCoCrFeNi consist of an oxide layer and an adsorption layer mentioned in Section 3.3. Here, only parameters associated with the oxide layer, i.e., Q¢ and Rg, are discussed. The oxide layer thickness is evaluated by using the Helmholtz model mentioned above and denoted by d, as d = e&S/Q: where & denotes the permittivity of free space (8.85 x 10-74 F/cm), ¢ denotes the dielectric constant of the medium, and S denotes the surface area of the electrode. Assuming that ¢ and S for oxide layers of alloys are the same allows us to compare relative values of d for all samples by 1/Q;. Figs. 14(d)-(e), whose data were listed in Table 5, show the Q; and R; vs. x plot respectively. Both Qr and Ry decreases with x. This represents that d increases with A content x, and a thicker oxide layer implies a smaller value of impedance. Therefore, one car explain this phenomenon by considering both the thickness and the density of oxide layer Related study reported Al oxide easily forms a porous structure [25]. Hence, it is easily understood that in addition to causing a thicker oxide layer, Al promotes the dispersive anc porous oxide layer. In summary, Al has a negative effect to the passive parameters including Lori, Ipass, Wioss, and R¢, for AlxCoCrFeNi in H2SOu.

*General corrosion, **Selective dissolution in Al and Ni

Fig. 17. Icorr Values for alloys C-x (Al,) at various temperatures. As mentioned in Section 3.1, Icorr decreases with Al content at 25°C. However, this differs from temperatures to temperatures. Fig. 17 shows the [corr values of C-x at various temperatures. One can see that Icorr decreases with Al content x at low temperatures (< 27°C), and, conversely, at high temperatures (> 27°C). The EIS results (Section 3.1) indicate that more Al content x makes the oxide layers thicker and more dispersive. At low temperatures, the thicker oxide is the dominator for Icon; whereas, at high temperatures, the dispersive oxide dominates. Therefore, this special phenomenon occurs.

Fig. 1. Commercial spherical zinc dusts. D 50/50 average particle diameters: 4 jam (fine) and 8 pm (regular).

Table 2. Average values of the simultaneous statistical treatment of all variables. With the purpose of establishing the efficiency of each protective coating from an anticorrosive point of view, the average value of degree of rusting was calculated for areas with and without cutting. The results of Table 2 confirm the superior performance of water- based nano lithium silicate of 7.5/1.0 silica/alkali molar ratio in relation to solvent-based, partially hydrolyzed tetraethyl orthosilicate, the major efficiency of fine microzinc compared to regular one and the increasing efficiency of primers as the level of conducting reinforcement fibers increased (Ahmed et al., 2010). On this last variable of formulation, it is worth mentioning that primers with 2.0% conducting reinforcement fibers showed the best protective capacity, which would occur due to the improved electric contact between zinc particles and with metallic substrate. On the other hand, the quartz used as reinforcement fiber due to characteristic non-conductive showed a similar performance that the corresponding reference primer (without reinforcement fiber).

Reinforcement Fibers in Zinc-Rich Nano Lithiun Silicate Anticorrosive Coatings Table 3. Average values and standard deviation.

Fig. 2. Structure of 4,4’,4’,4’”’ (porphine-5,10,15,20-tetrayl)-tetrakis (benzeric sulfonic acid). 2. Experimental

Table 1. Results of cyclic voltammograms. Regarding the electrochemical studies of the corrosion resistance of the protective layers formed from the first set of porphyrin it has been demonstrated that the electrodes which have been treated with the system C (0.2 g of H2TPP porphyrin (5,10,15,20 tetrakis 4 phenyl- 21H,23H) disolved in 40 mL benzonitrile) gave the best results. The immersion time was 5 minutes.

Fig. 4. Tafel tests of coated electrode with system C.

We continued the studies, cyclic voltammetry and Tafel method for carbon steel electrodes treated in different ways.

Fig. 5. Cyclic voltammogram of uncoated electrodes.

Comparative Study of Porphyrin Systems Used as Corrosion Inhibitors

Fig. 8. Tafel tests of electrodes immersed for 5 minutes in system I

Fig. 9. Cyclic voltammogram of electrodes immersed for 5 minutes in system II

Fig. 10. Tafel tests of electrodes immersed for 5 minutes in system II

Fig. 11. Cyclic voltammogram of electrodes immersed for 60 minutes in system I. Comparative Study of Porphyrin Systems Used as Corrosion Inhibitors

Fig. 13. Cyclic voltammogram of electrodes immersed for 60 minutes in system II. Fig. 12. Tafel tests of electrodes immersed for 60 minutes in system I.

Fig. 14. Tafel tests of electrodes immersed for 60 minutes in system II.

Fig. 15. Cyclic voltammogram of electrodes coated with paint.

Fig. 16. Tafel tests of electrodes coated with paint

Table 3. Results obtained from cyclic voltammograms and Tafel tests. Cyclic voltammograms and Tafel tests for 5 minutes and 60 minutes immersion time are presented in Table 3. The notations from the table:ipeak - peak current density; €peak - peak potential; i cor corrosion

Properties of Graphite Sinters for Bipolar Plates in Fuel Cells Bipolar plates/interconnectors in fuel cells typically have channels on their surface to allow for the distribution of media to the electrodes [29]. The shape of channels and direction of flow of media might be different for the plate adjacent to anode compared to the plate near cathode. Media which flow in to both electrodes can be supplied by means of parallel channels, where media flow in one direction or channels where media are supplied to fuel cells with opposite directions. Another possible solution is that the media flow in with the direction transverse to the cell. The choice and optimization of the shape of the channels in bipolar plates affect the operation of the cell, particularly the degree of removal of products and distribution of gases to the surface of electrodes. The figure below presents bipolar plates with channels (Fig. 3). The essential effect on operation of the cell is from the depth of the channels, width of the channels, distance between spirals etc.

Bulk densities for the powders used in the study are contained in Table 2.

Table 1. Chemical composition of steel powder %. The material composites were obtained from commercial steel powder AISI 316LHD manufactured by Héganas (Belgium) sprayed with water and graphite powder Graphite FC (fiber carbon) manufactured by Schunk Kohlenstofftechnik GmbH (Germany). Chemical composition of the powder 316LHD is presented in Table 1.

Fig. 3. Bipolar plate in fuel cell with channels which supply media.

Fig. 4. Powders morphology, magnification x500.

Fig. 4 presents the morphology of the powders used for preparation of graphite-steel composites. The values of statistical parameters of the particles of steel and graphite powders are presented in Fig. 5 and Fig. 6 in the form of histograms. Table 3 contains statistical parameters of stereological values of the used powders.

Table 3. Statistical parameters of stereological values for the powders used in the study.

Fig. 6. Histograms of: a) particle surface; b) particle perimeter; c) mean particle diameter; d) roundness of the particle in graphite powder.

Fig. 8. Diffractograms of graphite-steel composites.

Fig. 9 presents microstructures in graphite-steel composites.

Table 4. Density and hardness of graphite-steel composites. Table 4 contains the values of density for 316L steel sinter and graphite-steel composites. Addition of graphite to the material reduces density from the level of 7.16 g cm® for steel to the level of 2.35 g cm? for the sinter modified with 80% of graphite. Modification of steel sinter 316L with graphite allows for obtaining materials with reduced density. With respect to future applications of these materials for bipolar plates in fuel cells, the use of light materials will allow for achievement of one of the most essential goals of hydrogen technologies, i.e. reduction in generator weight. Addition of graphite to steel sinter impacts also on material hardness. Change in sinter hardness with concentration of graphite in the composite is presented in Table 4.

20% 316L + 80% graphite Fig. 9. Microstructures of graphite-steel composites, magnitude 100 x.

Table 5. Mean values of grain parameters in graphite-steel composites. Table 5 contains mean cross-sectional surface area, mean number of grains per mm/ of the surface and mean number of grains per mm$ of graphite-steel composites. Based on the data contained in the table, one should note that no effect of chemical composition of the sinter on mean grain size is observed. Mean grain diameter varies from 48 to 68 mm, whereas the greatest grain diameters are observed for the sinter with 50% proportion of graphite. The data are also presented in Fig. 10.

Fig. 11. Hysteresis of mercury intrusion/extrusion in graphite sinters and distribution of pores depending on the proportion of stainless steel in the sinter.

Table 6. Comparison of porosity in graphite-steel composites. Properties of Graphite Sinters for Bipolar Plates in Fuel Cells

Fig. 12. Diagram of wetting hydrophobic and hydrophilic materials.

Table 7. Values of contact angles for graphite-steel composites.

Fig. 13. Contact angle evaluated for graphite-steel composites.

Fig. 14. Effect of porosity on contact angle in graphite-steel composites. Properties of Graphite Sinters for Bipolar Plates in Fuel Cells

Table 8 contains the values of height and longitudinal parameters for 316L sintered steel and graphite-steel composites. The substantial impact on surface geometry in sinters is from the presence of graphite. In the case of rough surface, an insignificant contribution of contact surface is observed. It is essential for fuel cells that the surface of the material for these parts is smooth, which is obtained through polishing or covering the surface with a coating [59]. Graphite, as a material with high porosity is subjected to polishing in order to obtain the smooth surface.

Table 8. Parameters of surface geometry in graphite-steel composites. As results from the data contained in Table 8, addition of 20% of graphite to the composite considerably increases surface roughness. Further addition of graphite to the steel insignificantly reduces roughness, but it is still higher than roughness in 100% steel sinter 316L. This fact should be closely associated with the porosity revealed for individual composites.

In order to determine the effect of composition of a composite on contact resistance of interfacial contact, the measurements of contact resistance between graphite-steel composites and the diffusion layer were carried out (Fig. 7). An increase in stress value causes the decrease in contact resistance, whereas at high values of pressure force, contact resistance does not change. The values of contact resistance are the lowest for the system of 316L steel sinter - diffusion layer (Fig. 16). Addition of graphite to steel sinter elevates contact resistance by nearly 40 mQ cm? in the case of a composite 80% 316L + 20% graphite.

Table 9. Corrosion parameters of graphite - stainless steel composites. Values of corrosion parameters estimated based on potentiokinetic curves are contained in Tab. 9.

Fig. 17. Potentiokinetic curves recorded for graphite-steel composites. With regard to application of the used materials, one should determine corrosion resistance in the sinters. Fig. 17 presents the patterns of potentiokinetic curves recorded under conditions of work of fuel cell. The sinter 100% 316L is subjected to passivation both under conditions of cathode operation and under conditions of anode operation. Corrosion potential of the sinter 100% 316L in the analyzed environment does not change whether the solution was saturated with oxygen or hydrogen and amounts to -0.30 V vs. NEK. In the case of composite graphite-steel materials, addition of graphite caused an increase in corrosion resistance of the sinter. As results from the profile of the potentiokinetic curves, value of current density in the anode range is decreased even by two orders of magnitude. The value of corrosion potential in 316L+graphite sinters is insignificantly changed or remains at the same level compared to Exo, for the sinter of 100% 316L. It should be noted that the value of corrosion potential for the sinter of 100% graphite is shifted towards positive values and amounts to ca. 0.09 V vs. NEK in the solution saturated with Oz, and ca. -0.02 V vs. NEK in the solution saturated with Ho.

Fig. 2. Wagner’s model of the transition from the internal to external oxidation of alloy A-B under the condition where only B can oxidize: a) B is less than the critical amount of B required for the transition and b) B is higher than the critical content of B required for the transition from the internal to external oxidation [44,45, 110-112].

For Fe-Cr alloy, X’ is calculated based on available data listed in the Table 1 (extrapolated from the literature [114,115] with do assumed to be 5 micron, and plotted it as a function of grain size in Fig 4. Figure 4 clearly shows that Cr required for such transition decreases substantially when grain size is below 100 nm.

Table 2. Parabolic oxidation rate constants (k) values in g’cm“s*[12]

Fig. 5a. Oxidation kinetics of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys at 300°C as represented by weight-gain vs time plots for 3120 minutes [12,37,39]. Inset shows the zoom of the region representing initial periods of oxidation (up to 240 minutes of oxidation).

Oxidation Resistance of Nanocrystalline Alloys

Fig. 5c. Oxidation kinetics of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys oxidised at 300°C: weight-gain2 with time, suggesting parabolic kinetics for mc alloy but departure from parabolic kinetics for nc alloy [12,37,39].

Fig. 6a. Oxidation kinetics (weight-gain vs time plot) of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys, during oxidation at 350°C for 3120 min in air [12,39]. Inset shows the zoom of the region showing initial periods of oxidation (up to 240 minutes of oxidation).

Fig. 6b. Oxidation kinetics of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys oxidised at 350°C: weight-gain2 with time (up to 240 minutes) suggesting parabolic kinetic for both mc and nc alloys [12,39] Oxidation Resistance of Nanocrystalline Alloys

Fig. 7. Oxidation kinetics (weight-gain vs time plot) of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys, during oxidation at 400°C for 3120 min in air [12,39]. Inset shows a zoom of the region of initial periods of oxidation (up to 240 minutes of oxidation).

Fig. 8. SIMS depth profiles for Cr the in the oxide scale developed during oxidation of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys at 300°C for in air for 3120 minutes, using a Cameca ims (5f) dynamic SIMS instrument. SIMS parameters were: Cs* io primary beam (10 nA), depth profiling of craters of 250 um x 250 um area [12,39]. Oxidation Resistance of Nanocrystalline Alloys

Fig. 9. SIMS depth profiles for Cr the in the oxide scale developed during oxidation of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys at 350°C for in air for 3120 min, using a Cameca ims (5f) dynamic SIMS instrument. SIMS parameters were: Cs* ion primary beam (10 nA), depth profiling of craters of 250 ym x 250 um area [12,39].

Fig. 10. SIMS depth profiles for Cr the in the oxide scale developed during oxidation of nanocrystalline (nc) and microcrystalline (mc) Fe-10Cr alloys at 400°C for in air for 3120 min, using a Cameca ims (5f) dynamic SIMS instrument. SIMS parameters were: Cs* ion primary beam (10 nA), depth profiling of craters of 250 um x 250 um area [12,39]. SIMS analysis as carried out in our previous work [12,39] provides a qualitative analysis of Cr enrichment of the surface. Based on such qualitative analysis of Cr content, a Cr2O03 oxide layer was proposed to develop in nanocrystalline alloy, whereas, it was proposed that a mixed Fe-Cr oxide layer forms in case of microcrystalline alloy. A Future study quantifying the Cr, Fe and O contents of oxide layer and their oxidation states using techniques such as X-ray photoelectron spectroscopy (XPS) must provide a better understanding of the effect of nanocrystalline structure on the chemical composition of oxide laver.

region of potentials somewhere between the passivation potential (Ep) and the potential c transpassivity (E;) (or the potential of pitting formation (Epit)), in which the increase of th effectiveness of the cathodic process is leading not to enhancement but rather to abatemen of the corrosion rate (for example in the course of the transition from one cathodic proces Kk, to another one Ks) one can observe a system more stable to corrosion (easily passivatin: system). Obviously in this case of minimal corrosion currents there will appear a cross-poin between the anodic and the cathodic curves of the corrosion diagram within the zone o stable passive state. Under these conditions it is quite possible that a smaller corrosiot current is corresponding to a more efficient cathodic process in comparison to the syster displaying a lower cathodic efficiency (if we exclude the conditions where the potential c the system is reaching the value of the potential of transpassivation and the potential o pitting formation - K4). Therefore one can conclude that during the occurring of an efficien cathodic process the system will pass over spontaneously to a stable passive state and it wil be corroding at a much lower rate, corresponding to the current of complete passivatior The stationary corrosion potential of such a system will be more positive than the potentie of complete passivation (E-p) and at the same time more negative than the potential of brea through the passive film and the potential of transpassivation. In this way the rate c corrosion can be reduced to a considerable extent by the correct use of the phenomeno: “passivation”. “passivation” . So, in order to create a system stable to corrosion and to decrease the rate of corrosion, it is necessary to find a way to promote the cathodic effectiveness (for example as it is in this specific case of investigations, carried out by us, to modify the steel surface with CeO2-Ce20:

Figure 2 represents an example of a corrosion diagram, illustrating the changes in tk behavior of the corrosion system upon increasing the surface concentration of the effectiv cathodic coating (for example in the case of modifying the steel surface with cerium oxides It follows from the diagram that the shift in the corrosion potential of the system - associated with the increase in the concentration of the cathodic depolarizer (the ceriu1 oxides), which will facilitate the transition from active state into passive state of the systen under the conditions of disturbed passivity. At concentration of the cerium oxide corresponding to the cathodic curve C; (Fig. 2), the rate of dissolution of the steel wi

Figure 3 shows a typical experimentally obtained corrosion diagram E-lg i, illustrating the kinetics of the cathodic and anodic processes on the studied steel in the absence of electrochemically deposited cerium oxides film (the curves 2) and after the deposition of thin oxide films with different surface concentrations of Ce (curves 3-5).

Corrosion Behavior of Stainless Steels Modified by Cerium Oxides Layers

Fig. 6. SEM images of stainless steel before (a) and after thermally treatment (b).

Figure 5 illustrates the analogous E-t dependences at open circuit, obtained upon immersion of SSx in 0.1 N H2SOx, solution, to which various concentrations of Ce‘* ions have been added. It was important to find out what is the influence of cerium ions on the corrosion behavior of the samples of thermally treated steel, when the character of the corrosion process is changed as a consequence of the thermal treatment of the steel. It should be reminded at this point that the stationary corrosion potentials (Est) of SS and SS,.,in 0.1N H2SOx solution have values for the non-thermally treated steel Es= ~ -0.300 V, while for the thermally treated steel this value is E.= ~ -0.980 V (Fig. 4)). The registered negative values of E, in the absence of cerium ions for the SS,;, samples, in our opinion, are connected with the strong cracking of the natural passive film on the surface of SS (Fig. 6). The most probable reason for this loss of the “stainless character” of the steel surface are the revealed sections, determining a several times higher concentration of iron-containing agglomerates in the surface layer (Table 2).

Fig. 7. Potentiodynamic E-lgi curves of SS; at different concentrations of Ce** in 0.1 N H2SO, solution. of a reduction reaction Ce** <> Ce** in the redox system Ce**/ Ce>*. Therefore the Ce** ion: as component of the corrosion medium, are acting as inhibitor, exerting an oxidative effec It follows from (Fig. 7), that the increase in the concentration of the inhibitor in the corrosio environment leads to a substantial decrease in the corrosion current - from 1.10% (é inhibitor concentration of 0.1 ppm) - to 1. 108 A.cm2 (at inhibitor concentration ~0.9 ppm What is making impression is the fact that the further increase in the concentration of th Ce4+ ions in the corrosion medium from ~0.9 ppm (which could be accepted as “critical”) u to 1000 ppm influences to a smaller extent the rate of corrosion. It is necessary also to poit out that with the increase in the concentration of cerium ions the corrosion potential shifted in positive direction, whereupon its values remain more positive than the potenti of complete passivation and more negative than the potential of depassivation of the ste: within the entire interval of studied concentrations - an effect analogous to the one alread established for the systems Ce2O3-CeO2/SS,,in 0.1 N H2SO, solution.

Table 3. Electrochemical parameters characterizing corrosion behaviour of SSi,.. Corrosion Resistance It is seen from the table that upon increasing the concentration of Ce ions in the corrosion medium the degree of protection reaches values up to 99% for the samples of thermally treated steel. The obtained data about the promotion in the efficiency of the inhibiting action with the increase in the concentration of Ce*+ in the corrosion medium for the thermally treated steel (Table 3) supposes an interconnection between the inhibitor concentration and the degree of surface coverage, Q following the equation (8):

and in the absence of inhibitor and judging from the measurements of the stationary corrosion potential of the steel, depending on the concentration of the inhibitor at open circuit we could suppose that under the conditions of internal anodic polarization in the presence of inhibitor the nature of the passive layers remains the same as in the case of external anodic polarization. As far as we can judge the specific action of the inhibitor is manifested in the formation of an adsorption layer, which is transformed into bulk phase, on the active anodic sections of the surface of the metal. In view of the XPS analyses ( Guergova, 2011) after 500-hour interval of staying of the thermally treated steel in the aggressive medium in the presence of Ce** ions (25 ppm), on the surface of the studied film in the region of the Ce3d XPS band one can observe the appearance of a certain amount of cerium (1.5 at. %) in the form of Ce2O3. The cerium is most probably incorporated into the surface film as a result of the stay of the steel sample in the inhibited corrosion medium, which leads to its modifying as a consequence of the formation of mixed oxides of the type of cerium aluminates and chromates (Burroughs et al., 1976; Hoang et al., 1993). In support of such hypotheses comes the absence of visible corrosion damages on the surface of SSi4. exposed for 500 h in 0.1N H2SOs, solution in the presence of Ce** ions (Fig. 9). Of course, from purely electrochemical point of view, the ability of the inhibitor to define strongly positive oxidation-reduction potential of the steel is connected with the proceeding of reduction of Ce** into Ce+. In order to investigate the kinetics of reduction of the Ce** ions to Ce5+, in the region of potentials, characteristic of the passive state of the steel under consideration, we plotted the anodic and the cathodic potentiodynamic curves, characterizing the behavior of the oxidative- reductive couple Ce#*/Ce3+ at various concentrations of Cet in 0.1 N H2SOx, solution, on an inert support of platinum (Fig. 10). Such an approach (Tomashov & Chernova, 1965), to our mind, enables the complete elucidation of the mechanism of inhibitory action of the cerium ions. It allows direct juxtaposition of the changes in the values of the corrosion potentials (respectively the corrosion currents) of the steel in their presence with the values of the reversible redox potentials (respectively the exchange currents) of the couple Ce**/ Ce%* at comparable concentration levels.

Fig. 10. Potentiodynamic E-lgi curves of Pt at different concentrations of Ce** in 0.1 N H2SO, solution.

Table 4. Reversible redox potentials E,, and equilibrium currents io, of the system Ce** /Ce** on Pt at different concentrations of the Ce*+ in the corrosion medium. corrosion process (i.e. oxidative depolarization), respectively in the establishment of oxidative-reductive potential of the medium more positive than the potential of complete passivation of the steel. This effect, in its turn, defines the value of the stationary corrosion potential of the steel to be more positive than the potential of complete passivation.

In order to prove the integral nature of cerium oxides films as efficient cathodic coating, involved directly in the corrosion process and the role of cerium ions as inhibitors possessing oxidative effect, participating also directly in the corrosion process and leading to the formation of phase layer of cerium oxides on the active cathodic sections of the steel surface, we compared the dependences E-lgi for the systems Ce2O3-CeO2/SS;;. as well as for the system Ce**/Ce3+/Pt. - Fig. 11. It is seen in Fig. 11 that the corrosion currents of the systems Ce203-CeO2/SS;;.are close in value to the exchange current of the oxidation- reduction current of the couple Ce4*/Ce3+ on Pt. The differences between the corrosion potential of the systems Ce203-CeO2/SS;; and of the equilibrium oxidation-reduction potential of the couple Ce*+/Ce3+ can be explained by the discrepancies in the surface and bulk phase concentrations of the components.

Table 5. Concentration of the elements (in at. %) on the surface layers of the system CeO2- Ce203/SS after thermal treatment and after corrosion test in 0.1 N H2SO.. In confirmation of these results and the conclusions come also the data of the XPS analyses of the samples, having electrochemically deposited cerium oxide films, characterizing the changes in the chemical state and in the composition of the surface film, depending on the time interval of the immersion stay of the samples in 0.1NH2SOx, solution (Stoyanova et al., 2010). Table 5 represents the results for the sample with surface concentration of electrochemically deposited cerium oxide layer 45.1 at.% .It is seen that after 1000 hours of exposure to the corrosion medium the surface concentration of cerium is decreased from 45.lar.% down to 0.2 at.%. This result is convincing evidence for the occurring of depolarizing reaction involving the participation of the rich in CeO? sections of the surface, acting as effective cathodes, in accordance with the equations (1-4). It becomes evident that the presence of cerium oxide film determines the establishment of more positive stationary corrosion potential of the system, due to the proceeding of the reactions 1- 6, the surface passive film will become modified, whereupon its composition, respectively the ratio Cr/Fe, will become different.

Fig. 12. Ce3d and Ols XPS spectra of CeO2-Ce203/SS sample after 250 h exposition to 0.1 N H2SOx, Panel A and B correspond to high BE and high BE portion of Ce3d and O1s spectral regions taken at different step of corrosion test. (1) as-deposited; (2) after 18 h ; (3) after 200 h; (4) after 250 h; (5) after 400 h; (6) after 1000 h exposition to 0.1 N H2SO, solution. obvious that the chemical state of Ce could be evaluated based on the percentage of the area of the u’”’ peak, located at 916.8 eV with respect to the total Ce3d area. So if the percentage of the u’’”’ peak related to the total Ce3d area varies from 0 to 14%, then the Ce** percentage related to the total amount of Ce varies from 0 to 100%. In our case the u’”’% amounts change as a result of dipping the as-deposited sample into 0.1 N H2SQ, solution, so we observed also change in the percentage of Ce** and Ce%* on the surface (Table 6). The obtained Ols X-ray photoelectron spectra, recorded after different time intervals of exposure, are quite complicated. These spectra had to undergo de-convolution procedure to analyze the contribution of the separate components in them.

the respectively detected high-energy peak in the spectrum of oxygen for these samples is growing up initially, while afterwards it decreases its intensity. This effect, in our opinion, is owing to consecutive enrichment and then impoverishment of the surface layer of the film in OH: groups i.e. adsorbed water molecules (Fig.12, Table 6). The obtained results gives us the reason to draw the conclusion that the surface passive film under these conditions at this stage consists of CeO, Ce2xO3, Ce(OH)4 as well as CeO(OH)2 and Ce(OH)3, whose existence has been ascertained also by some other authors (Huang et al., 2008).

propagation and spreading further to give total corrosion (Fig.13 a,b). For the sake of comparison Fig. 13 shows the same surface in the presence of cerium oxide coating after 50 hours of exposure to the corrosion medium. In this case no corrosion damages can be observed on the surface not only after the 50 hour, but even after 1000 hours of exposure (Fig.13) to the corrosion medium, as a result of its modifying, already discussed above.

Fig. 1. Schematic for corrosion of a material system with the components involved in the corrosion process. There are different types of corrosion that can take place on a material system and they could be uniform type or localized type. In uniform corrosion, as the name suggests, corrosion takes place all over the surface. On the other hand, localized corrosion can be several kinds, such as galvanic corrosion, pitting corrosion, selective attack, stray current corrosion, microbial corrosion, intergranular corrosion, crevice corrosion, thermo galvanic corrosion, corrosion due to fatigue, fretting corrosion, stress corrosion, hydrogen damage etc. (Jones, 1992). For more details on each process, the reader is suggested to refer to any review articles or books on corrosion science.

As listed above there are several approaches available for processing metal - oxide systems and their corrosion properties are going to be dependent on the processing techniqu: employed too. For example, the processing defects like porosity, improper bonding betweet the matrix and the oxide dispersoids, and their interfacial properties can influence the corrosion behavior quite extensively. Wetting of the oxide particles becomes a critical facto: in some of the processing approaches to deal with the particle - matrix bonding. Fig. + shows a schematic for interfacial bonding of the second phase particles with matrix alons the grain boundaries and triple junctions. In addition, high temperatures in some of the processing techniques may cause an interfacial reaction between the metal matrix and the dispersed second phase particles, thereby the interfacial stability and its properties play a1 increasingly important role in the corrosion. It is also possible that the interfaces coulc become prone to corrosion attack by providing preferential sites. In spray depositior approach splat boundaries, porosity, and distribution of the oxide particles may play a1 important role in deciding the corrosion properties. Added to that, the microstructures o the composites could also vary from process to process. The effect of some of thes parameters on the corrosion of different metal - oxide systems is discussed in brief in the following sections.

Following the above mentioned approach of inflight synthesis, different oxide ceramic particles were introduced into the NiCr (55/45 wt.%) alloy coating. The following coatings were deposited onto mild steel coupons for characterization: (a) NiCr only, (b) NiCr + Cr2O3, (c) NiCr + AlxO3, and (d) NiCr + SiO2. Along with these coatings, NiCr coatings using a twin wire arc spray process (TAFA 3830, Praxair Surface Technologies, Indianapolis, IN) were also deposited for comparison purposes. The arc current and voltage for both the processes were kept at 100 amps and 36 volts, respectively. The HVOF gas pressures were

Fig. 5. (a) Schematic of hybrid gun and (b) hybrid gun in operation. The hybrid process offers all benefits of wire stock and productivity of electric arc spraying combined with noticeably improved coating density of HVOF. In addition to introducing material through arcing mechanism, if desired, powder/liquid/gas precursors can also be fed through the HVOF coaxial feed line (Fig. 5a). This enables us to tailor the composition inflight by introducing particles into the HVOF jet, to cater to specific property requirements of a composite coating. This unique capability completely eliminates the necessity of processing and handling of the ultrafine particulates prior to feeding them into the hybrid gun. Synthesizing and introducing ultrafine and nano dispersoids inflight in a functional manner to produce composite coatings by the hybrid technique is quite unique in terms of simplicity compared to any other processes. A comparative picture of the steps involved in processing of particulate reinforced composites by conventional routes versus our approach is presented in Table 1.

Fig. 7. (a) SEM picture of NiCr coating with dispersed Al,O3 particulates and (b) higher magnification SEM picture of NiCr coating with dispersed Al,O3 particulates. Fig. 7a, presents the general cross section microstructure of a NiCr coatings with embedded alumina particles produced by the hybrid spray process. The coating is very dense and exhibits the characteristics of an HVOF coating rather than of an arc sprayed coating. The hybrid spray process is unique in the sense that while it yields comparable density to that of the HVOF process, the deposition rate is closer to that of an arc spray process. The observed density is advantageous for high temperature corrosion and erosion performance of the coatings. Details on the corrosion and erosion performance of the coatings are discussed in the forthcoming sections. The dispersion of the alumina particles (dark phase) in the NiCr matrix is shown in Fig. 7b.

Fig. 8. TEM picture of NiCr coating with dispersed SiO? particulates. The atomization of the liquid precursor (for oxide particles) prior to the injection into the combustion jet plays an important role on the size as well as on the distribution of the particles in the final coating. The requirements for the atomization system include: controlled and uniform flow, ability to operate against a back pressure of 30 psi pressure that exists in HVOF flame at the point of injection and the ability to generate mono- dispersed micron sized droplets. Details on the atomization and optimization of parameters could be found elsewhere (Mohanty et al., 2010). From Fig. 7(b), it is apparent that the distribution of the particles was uniform across the cross section. Similar observations were made in the case of NiCr + Cr2O3 and NiCr + SiOz systems also. It is to be noted that composites made from premixed powders commonly exhibit large clusters of nanoparticles. Fig. 8 presents the TEM picture of a NiCr coating with embedded silica particles. Many fine particles are observed in the matrix, as well as along the grain boundaries. For enhanced creep resistance resulting from grain boundary pinning, the particles must be small and coherent with the matrix. Especially alloys with very high chrome content can substantially benefit from such ultrafine particle embedment as observed in Fig. 8.

Fig. 9. Weight gain measured using TGA for oxidation studies. undergo changes upon reheating. If CrO2, which has higher oxygen content compared to Cr,O3, forms during the spray process; it can undergo stoichiometric changes to a stable oxide (Cr,O3) upon reheating and this could lead to an initial weight loss in the coatings. However, part of the initial weight loss could also be attributed to the evaporation of moisture absorbed by porosity in the coatings. According to Lars Mikkelsen (2003), the specimens may also lose weight due to vaporization of chromium containing species from the chromia scale. Whereas the oxidation of pure Cr to Cr,O3 and also the transformation of CrO and Cr3O, to Cr2O3 will lead to weight gain because of increasing oxygen content in the coatings.

Fig. 11. Wet corrosion of thermally sprayed NiCr coatings in NaCl solution at room temperature. The corrosion currents measured from the electrochemical tests are shown in Fig. 11. At zero hours, although the hybrid coating showed less current compared to the arc sprayed coating;

0 gm of grit being used. Samples tested included arc sprayed coatings, plain hy atings, and hybrid coatings with alumina, chromia and silica, respectively. The result e tests, shown in Fig. 10, indicate that the hybrid coatings are up to 30% more resistar osion than the arc sprayed coatings at 750°C and this is thought to be due to the his nsity of the hybrid coatings. However, the weight loss was slightly higher in the cas ide particulate embedded coatings. This is contrary to the observation of Jiang Xu e u et al., 2008), who have reported improved erosion resistance with the additio1 moparticles in Ni based alloys. Especially, in the case of chromia embedment, fference was evident. This may be linked to the large bubble shaped features with inte ids that were observed in chromia particles (which are not shown here).

Fig. 12. Weight loss measured in hot corrosion test. The hot corrosion test results are shown in Fig. 12. This chart compares the weight loss da obtained on weld overlay coating (with and without salt), 304 stainless steel (304 SS) samt and the coatings - NiCr by arc spray, NiCr and NiCr + SiOz by hybrid gun . The chrom stabilized alumina embedded coatings were not included in the test due to their unfavorak oxidation results presented in Fig. 9. NiCr + SiO2 coatings showed the lowest weight lc compared to all the other samples. Plain NiCr coating by hybrid spray also exhibited less weight loss compared to the arc spray coating and this could be attributed to the improv density of the hybrid spray coatings. The superior corrosion resistance of the NiCr + Sil coating is possibly due to the enhanced stability of the chromia scale and the improv oxidation resistance caused by SiOz. Weld overlay coating showed least weight loss in t absence of the salt; however, when salt was present, it showed poor corrosion resistan compared to the hybrid spray coatings.

The case structure of nitrided steel depends on its type, concentration of alloying elements and particular conditions of nitriding treatment. The diffusion zone is the original core microstructure with the addition of nitride precipitates and nitrogen solid solution. The surface compound zone is the region where y' (FeaN) and ¢ (Fez3N) intermetallics are formed. The corrosion resistance of steel varies with nitrided layer structure. The surface “white layer” can contain ¢ nitride, y' nitride or a two phase mixture e+y', below that is the diffusion zone. In acid solutions the iron nitrides corrode more slowly than iron and when the “white zone” is formed on the steel surface the improvement of the corrosion resistance is a fact. In Fig. 1 is shown the typical structure of nitrided plain carbon steel (Minkevich, 1965).

Table 1. Gas nitriding process conditions for improvement of the corrosion resistance for plain carbon steels and free-cutting steels. [he gas nitriding for improvement of corrosion resistance of plain carbon steels and low- alloy steels can be carried out for shorter times at elevated temperatures (Minkevich, 1965). The purpose is to obtain on the steel surface non-etched nitrided layer without pores and thickness about 0.015 - 0.030 mm. In Table 1 are given the conditions of this process for some plain carbon steels and free-cutting steels.

Table 2. Comparison of the white layer structures of gas and ion nitrided GOST 10 steel and reduced distortion of the workpiece. The white layer on the surface of the ion nitrided medium-carbon steel contains mainly from y' iron nitride with a small amount ¢ iron nitride. With the control of the process can be obtained layer from single y' nitride phase. In Table 2 is made a comparison of the white layer structures of gas and ion nitrided GOST 10 steel (Buchkov & Toshkov, 1990).

Fig. 2. X-ray diffraction patterns of ion nitrided surface layer on steel 31CrMoV9. Our investigations on the microstructure of ion nitrided EN 31CrMoV9 steel show that on the surface is formed nitrided white layer consists from e-nitride and y'-nitride (Krastev et al., 2010) - Fig. 2. The thickness of the nitride white layer is 10 - 12 zm and in depth follows 150 - 200 pm diffusion zone with nitride precipitates - fig. 3.

Improvement of Corrosion Resistance of Steels by Surface Modification By contrast to plain carbon steels, the corrosion resistance of stainless steels can be reduced by nitriding, due to breakdown of the surface chrome oxide barrier to enable nitrogen diffusion into the steel. The stainless steels exhibit generally poor tribological properties, because of that treatments such as nitriding can enhance the surface hardness and improve the wear resistance. Plasma nitriding can be carried out for this purpose at temperatures from 350 to 500 °C (Castaletti et al., 2008). While giving significant improvement in wear resistance, the higher treatment temperatures tend to adversely affect on the corrosion performance of the stainless steels in result of formation of CrN. Improved corrosion resistance of plasma-nitrided layers on stainless steels are observed when the nitriding process is carried out at a lower temperature (400 °C) with presence in the layer of “S - phase”, which is supersaturated with nitrogen austenite. The expanded austenite layer in nitrided YB 1Cr18Ni9Ti steel has a good pitting corrosion resistance in 1 % NaCl solution and an equivalent homogeneous corrosion resistance in 1 N H2SO, solution, compared with the original stainless steel (Lei & Zhang, 1997).

Dn. the surface of the boronized steels generally a boron compounds layer is formed. It can oe a single-phase or double-phase layer of borides with definite composition. The single ohase boride layer consists of Fe2B, while the double-phase layer consists of an outer phase of FeB and an inner phase of Fe2B. The FeB phase is brittle and harder, forms a surface under igh tensile stress and has a higher coefficient of thermal expansion. The Fe2B phase is oreferred because it is less brittle and forms a surface with a high compressive stress, the oreferred stress state for a high-hardness, low-ductility case. Although small amounts of FeB are present in most boride layers, they are not detrimental if they are not continuous. Continuous layers of FeB can be minimized by diffusion annealing after boride formation. in Fig. 4 is shown the typical microstructure of borided layer on the surface of plain carbon steel (Schatt, 1998).

Paste boriding is an attractive technique for producing of boride diffusion coatings on steels surface because of lower cost and less difficulty in comparison with pack boriding. It is carried out usually in a paste from BC as a boriding agent, Na3AIF4 as an activator, fluxes, and binding agent for the paste formation. The temperature of the process is from 800 to 1000 °C and heating is mostly inductively or resistively. A layer in excess of 50 jam thickness may be obtained after inductively or resistively heating to 1000 °C for 20 min. The relationship between the boride layer thickness and time for iron and steel boronized with ByC-Na2ByO7-Na3AlF¢ based paste at 1000 °C is shown in Fig. 5 (ASTM Handbook, 1991).

The borided steels have a higher corrosion resistance together with the high hardness and wear resistance. In Fig. 7 is given a comparison in the corrosion resistance of 0.45 % C plain carbon steel before and after boronizing (ASTM Handbook, 1991).

Fig. 6. Microstructure of boride layers on C60 steel obtained for 2 (a) and 6 (b) hours pack boriding at 950 °C in powder mixture containing 3 % NasAIFe (x150). EN C60 steel at 950 °C and time from two to six hours. The results show that it is possible to exchange the traditional for the process NaBF, with the inexpensive Na3AIF¢ as an activator, and amount of 3 % is enough to provide boride layer with the required thickness, structure and hardness. In Fig. 6 are given the structures of borided surface of steel C60 obtained for 2 and 6 h in powder mixture containing 3 % Na3AIF». The XRD analysis shows that the boride layers consist mainly from Fe2B with a small amount of FeB. The microhardness of the boride diffusion coatings is 1600 - 1800 HV.

After boronizing, the corrosion rate of boronized low carbon steel AISI 1018 is about 100 times lower than the corrosion rate of unboronized one based on the electrochemical measurement (Suwattananont, 2005). The boronized high strength alloy steel AISI 4340 and austenitic stainless steel AISI 304 have corrosion rate about several times lower than the corrosion rate of unboronized steels. The comparison of the tafel plots between boronized and unboronized AISI 1018 steel is shown in Fig. 8. After boronizing the steel becomes nobler and has lower corrosion rate than the unboronized one which proves that the corrosion resistance of the boronized steels is improved.

Fig. 9. The microstructure of nitrocarbonized EN C15 steel at 570 °C for 2 hours. Ferritic nitrocarburizing is this thermochemical treatment which is realized at temperatures completely within the ferrite phase field. The primary object of such treatments is usually to improve the anti-scuffing characteristic of ferrous engineering components by producing a compound layer in the surface which has good tribological characteristics. In addition, the fatigue characteristics can be considerably improved, particularly when nitrogen is retained in solid solution in the diffusion zone beneath the compound layer. This is normally achieved by quenching into oil or water from the treatment temperature, usually 570 °C. The obtained at these temperatures compound white layer provided the enhancing of the corrosion resistance of the nitrocarburized surface. The compound layer produced by ferritic nitrocarburizing consists mainly from ¢ carbonitride because of low carbon solubility in y nitride. In Fig. 9 is shown the typical microstructure of nitrocarburized surface of low carbon steel (Chatterjee-Fischer, 1986).

Fig. 10. Diffusion coatings on C45 steel produced by chromizing (a) x150, and aluminizing (b) x250. Our research group has been carried out investigations on high-temperature corrosion and abrasive resistance of chromium and aluminium diffusion coatings on EN C45 steel as parts of sintering machine in an agglomeration process of iron ores. The specimens are produced by pack-cementation process at the optimal characteristics for the both heat treatment techniques. The structure of the obtained diffusion coatings is given in Fig. 10. The comparison shows that better behaviour in these work conditions have the chromium diffusion coatings despite of their smaller thickness.

Fig. 11. Microstructure of EDM modified steel surface. The recast layer is also named white layer and it crystallizes from the liquid metal cooled at high rate in the dielectric fluid. The depth of this top melted zone depends on the pulse energy and duration. Below the top white layer is the heat affected zone with changes in the average chemical composition and possible phase changes. In Fig. 11 is shown the typical microstructure of EDM modified steel surface.

The workpieces are made from high speed steel HS 6-5-2 with structure after the typical heat treatment for tools of this steel and hardness about of 950 HV. The choice of high-alloy steel is founded on the opportunity for higher effectiveness of treatment on structure and properties of modified surfaces after the nonequilibrium phase transformations from liquid state.

The high rate of the recasting process gives opportunities for formation of metastable phase: and considerable decreasing of grain size. The electrolyte type is of great importance for the chemical composition, microstructure and properties of the modified layer. By thes« experiments the electrolyte is on water basis and contains boron or silicon compounds. A short times of treatment it is not observed diffusion of elements from the electrolyte in the modified surface, but it is available diffusion process inside the workpiece between the white layer and the matrix - Table 3. The strong carbide-formed elements such as Mo, W. and V diffuse from the white layer to the matrix and Cr, Co in the opposite side.

Fig. 13. Microstructure of recast layer obtained by electrical discharge treatment in electrolyte of HS 6-5-2 steel, x800. The electric discharges generate an enormous amount of heat, causing local melting on the workpiece surface and thereupon it is rapidly quenched from the liquid state by the electrolyte. This recast area has a specific structure which is composed of several microscopic metallurgical layers, depending of machining conditions. In Fig. 13 is shown an optical micrograph of the modified surface of high speed steel obtained by electrical discharge treatment in electrolyte.

The efforts have been made in the past decades to modify the surfaces of these materials to improve their surface hardness, wear resistance as well as corrosion resistance which is shown in Fig. 1.

2.1 Thermochemical and diffusion surface engineering treatments

Fig. 3. Concentrations and concentration gradients of nitrogen and carbon. (Christiansen & Somers, 2005) Once in the metal, the transport of the absorbed substance takes place by diffusion, and follows Fick’s laws:

Table 1. Possible configurations of (a) nitriding, (b) carburizing. To reduce further the potential of distortion and to avoid structural modifications of the substrate, and without repeating the quench and tempering treatments, these carburizing and nitriding treatments have evolved, in the past few years, towards lower temperature processes (350-450°C for austenitic stainless steels). This reduction in the treatment temperatures had to include specific treatments for removing oxide layers, which act as a barrier to the diffusion of nitrogen and carbon.

The use of fluidized bed furnace in heat treating operation has been introduced by Reynoldson which offers several advantages, including faster treatment time, precise control of treatment parameters, despite its economic benefits of low investment and operationa! cost (Reynoldson, 1995; Haruman & Sun, 2005). The schematic picture of fluidized bed furnace is shown on Fig. 4. Recent work has shown that low temperature nitriding of austenitic stainless steel is possible in a fluidized bed furnace (Haruman & Sun, 2005).

Table 2. Treatment conditions and their corresponding layer thicknesses. Nitriding, carburizing, and hybrid treatments were performed at 450°C in a fluidized bed furnace having particulate alumina as fluidized particles which flow inside the chamber due to the flow of nitriding or carburizing gases.

Fig. 5. Schematic diagram of the flat cell used for polarization corrosion test (Li & Bell, 2004).

Fig. 6. Thickness of treated layers measured from micrographs.

he fluidized bed process, where the disruption of the native oxide film to cause the iitriding reactions, has to be by thermal dissociation. The higher the temperature, the faster s the dissociation of the oxide film and thus the shorter the incubation time. According to his hypothesis it is understood why a very thin discontinuous layer was formed after a 2h reatment time ; the layer was about 8 um thick after 8 h nitriding, which gave a bright \ppearance and was resistant to the etchant used to reveal the microstructure of the substrate. The treated layers for 4C-4N consist actually of two separate zones with a omewhat diffused interface which was clearly observed under microscope, but not evealed in the micrographs. The outer zone is yn and the inner zone is yc. Conversely, the iitrocarburised sample shows a distinct separation of the yn and yc layers, with the yc layer losest to the austenite substrate (as indicated in the micrograph in Fig. 7b). The X-ray liffractograms of carburised and nitrided AISI 316L show that two different types of xpanded austenite are present (Fig. 6).

Fig. 8. Depth profiles of Microhardness. 4C-4N specimen.

The wear properties of the low temperature surface treatment specimens as weight loss under dry sliding friction are presented in Fig. 11 along with an untreated specimen for comparison purpose. The results suggest that the fluidized bed thermochemical-treated specimens have excellent wear resistance. The 8N specimen has the highest wear resistance compared to the values of 4C-4N, 8(C+N), and 8C specimens.

caused primarily by the dissolution of nitrogen which causes a dilation of the fcc lattice (hence the name expanded austenite), although residual stress and stacking faults also play a role in this respect (Somers, 2005). The X-ray diffraction pattern of carburised AISI 316L is shown in Fig. 10. yc is identified as the only phase present in the surface adjacent region, i.e. within the information depth for the probing X-ray beam. A marked difference is observed as compared to nitrided AISI 316L; a smaller shift of the austenite peaks to lower 20, which indicates a substantially lower content of the interstitially dissolved atoms, provided that nitrogen and carbon induce a similar distortion in the fcc lattice. The asymmetrical (200) austenite peak in Fig. 10 indicates a depth-gradient of the carbon content in the near surface zone. The distinct peaks for the carburised sample indicate a smooth concentration gradient and lower defect density in yc layers as compared to yn layers.

Although, the corrosion behaviour of low temperature nitrided, carburized and hybrid stainless steel thermochemical treatment have been investigated by several investigators Zhang et al, 1985; Rie et al, 1995; Sun et al, 1999), the reason for improvement of corrosior ‘esistance due to nitrogen and carbon supersaturation in austenite has not been fully understood. A possible mechanism is that supersaturation of nitrogen and carbon promotes he improvement of the passivation ability of austenite, and this effect seems to give oeneficial with increasing degree of supersaturation (Munther et al, 2004). Thus, the mprovement of corrosion resistance for hybrid-NCT treated material may be attributed tc he extremely large supersaturation of the upper part of nitrogen-enriched layer with both uitrogen and carbon. This would be contribute to the observed higher hardness and better orrosion resistance as compared to those achieved by individual nitriding-NT and -arburizing-CT.

Fig. 1. Temperature variation as a function of the gas composition. 8.4x10? mbar. The sample was heated mainly by the rf field. The sample temperature was easured during the rf plasma process by a Chromel-Alumel thermocouple, attached to the mple holder. As shown in Fig. 1, the substrate temperature was influenced by the effect of is compositions. It was found that the temperature gradually increases from 475 °C for ire nitriding up to 550 °C for carbonitriding (50% C:H2, 50% Nz) and raises up to 600 °C for ire carburizing. Grazing incidence X-ray diffraction (GIXRD) with Cu Ka radiation was ed to determine the phases, present in the treated layers. For the chosen incidence angle of the (1/e)-penetration depth of the X-rays was approximately 700 nm. The recorded ffraction pattern shows therefore mainly the structure of the phases formed in this ar-surface region. In this paper we concentrate on the study of corrosion resistance, surface orphology before and after corrosion, and tribological properties of the treated samples. 1e surface roughness was measured by use of the rough machine (Dektak 8000, Veeco strument GmbH). Wear and friction measurement were performed at room temperature in boratory air with low humidity of 16 to 24 % using an oscillating ball-on-disk type bometer wear tester without lubrication. The 3-mm ball of cobalt tungsten carbide was oved at mean sliding speed of 15 mm/sec with different normal loads of 3, 5 and 8 N. The rrosion properties were evaluated using the electrochemical testing technique. The corrosion sts were performed in a 1 wt. % NaCl solution by application of the potentiodynamic larization method. A three-electrode electrochemical cell has been used, counter id reference electrode were related to Pt and saturated calomel electrode, respectively.

Fig. 2. X-ray diffraction pattern obtained at 2° grazing incidence from 304 stainless steel of untreated samples and samples treated at different CoH2/N2 gas pressure ratios. The sliding wear behaviour of the untreated and treated samples was assessed using oscillating ball-on-disk type tribometer. The depth of the wear tracks of examined samples as a function of wear path at a load of 3 N is shown in Fig. 5. Generally, the wear resistance of the untreated samples in comparison to the treated is extremely poor. For all treated samples, examined up to 320 m wear path, maximum one micrometer wear depth has been observed and the wear depth slowly increases with increasing wear path. Otherwise the wear rates have been accounted as total volume loss in mm3 divided by the total sliding distance in meters. The wear rates for the untreated material were accounted in order to know the improvement in the wear rates for treated one. The wear rate for the untreated 304 austenitic stainless steel was 2.4 x 10-¢ mm3/m at 20000 numbers of tracks (80 m wear path).

Fig. 3. Relative surface roughness for 304 ASS samples treated at different CoH2/N2 gas ressure ratios.

Corrosion Performance and Tribological Properties of Carbonitrided 304 Stainless Steel

Fig. 6 presents the variation in the wear rate at different number of tracks as a function of gas composition ratios for a load of 3 N. The wear rate after treatment was reduced by more than two orders in comparison with the untreated material. At 20000 wear tracks, a decrease of the wear rate was observed as much as the CoH2 gas ratio increases. However, this improvement was continued up to 75 % CsH2 and 50 % C2H2 at 40000 and 80000 numbers of tracks, respectively. The decrease in the wear rate by increasing the C2H2 content is related to some improvement in the friction coefficient of treated layers due to fine carbon precipitations which work as solid lubricant on the first few hundred nanometers. However, for samples treated at high carbon content, a small increase in the wear rate can be seen with increasing the sliding distance.

Fig. 6. Wear rate of treated samples as a function of gas composition for a load of 3 N.

Fig. 8. Anodic polarization curves for untreated and treated samples at different CoH2/No gas pressure ratios obtained in 1 wt. % NaCl solution.

and/or strain-induced martensitic transformation can occur. The group of high-manganese steels includes alloys with 15-30% Mn content. Two mean chemical composition strategies had been worked out so far. The first includes alloys with different Mn concentration and 0.5 to 0.8% C (Ghayad et al., 2006; Jimenez & Frommeyer, 2010). The function of carbon is stabilization of yphase and hardening of solid solution. In the second group, the concentration of carbon is decreased below 0.1%, whereby there is an addition up to 4% Al and/or 4% Si (Frommeyer et al., 2003; Graessel et al., 2000). The solid solution strengthening caused by Al and Si compensates smaller C content. Sometimes, the steels contain chromium (Hamada, 2007; Mujica Roncery et al., 2010) or microadditions of Nb, Ti and B (Bleck & Phiu-on, 2005; Grajcar et al., 2009; Huang et al., 2006).

Fig. 2. Schematic drawing of the effects of temperature and chemical composition on the stacking fault energy (SFE) of austenite and the correlation of SFE with a main deformation mechanism in high-Mn alloys. Plasticity) effect is considered as the major deformation mechanism (Frommeyer & Bruex, 2006). High impact on the dominating deformation mechanism have also the temperature, strain rate and grain size (Dini et al., 2010; Frommeyer et al., 2003; Graessel et al., 2000). The key to obtain the mechanical properties regime in Fig. 1 is the high work hardening rate characterizing the plastic deformation of high-Mn alloys. The high level of ductility is a result of delaying necking during straining. In case of the local presence of necking, strain- induced martensitic transformation occurs in such places (in TRIP steels) or deformation twins are preferably generated in locally deformed areas (in TWIP steels). It leads to intensive local strain hardening of the steel and further plastic strain proceeds in less strain- hardened adjacent zones. The situation is repeated in many regions of the sample what finally leads to delaying necking in a macro scale and high uniform and total elongation. The shear band formation accompanied by dislocation glide occurs in deformed areas of TRIPLEX steels and the SIP effect is sustained by the uniform arrangement of nano size k- carbides coherent to the austenitic matrix (Frommeyer & Bruex, 2006).

Corrosion Resistance of High-Mn Austenitic Steels for the Automotive Industry 3.2 Immersion tests The immersion tests were used to assess the corrosion resistance of the steels at the initial state (after the thermo-mechanical rolling) and after cold deformation. The corrosion resistance was investigated in two solutions: IN H2SO, and 3.5wt% NaCl. Prior to the corrosion tests, three samples of the each steel with the size 3.2x10x15 mm were ground to the 1000-grit finish and then they were washed in distilled water, ultrasonically cleaned in acetone and finally rinsed with ethanol and dried. The specimens were weighed with the accuracy of 0.001g and put into the solution for 100 hours at the temperature of 23+1°C. After the test the specimens were weighed and analysed using optical microscopy and SEM. Corrosion loss was calculated in a simple way as the difference between final and initial mass of the samples. Percentage mass decrement was also calculated. Cold deformation was applied by bending at room temperature. Samples with a size of 10x15 mm and a thickness of 3.2 mm were bent to an angle of 90°, with a bending radius of 3 mm. The immersion tests were used to assess the corrosion resistance of the steels at the initial

The results of the immersion tests in two media are given in Table 2. After 100 hours immersion in 1N H2SO, both steels showed a significant percentage mass decrement, among 38 and 41%. Mass loss of samples dipped in 3.5wt% NaCl is about 100 times lower. The difference is due to different corrosion mechanisms. When the solution is acidic, the corrosion process is running according to hydrogen depolarization, whereas in chloride media the specimens are corroding with oxygen depolarization.

Fig. 6. Corrosion pits in 25Mn-3Si-1.5Al-Nb-Ti steel after the immersion test in 3.5wt% NaCl.

Fig. 7. Austenitic matrix containing plates of « martensite and microhardness test results of 25Mn-3Si-1.5Al-Nb-Ti steel.

Corrosion Resistance of High-Mn Austenitic Steels for the Automotive Industry

Fig. 9. Craters created as a result of corrosion pitting and bursted corrosion products layer in 25Mn-3Si-1.5Al-Nb-Ti steel after the immersion test in 1N H2SO,.

Fig. 16. Scaled surface and corrosion pits of the specimen after corrosion tests in the 0.5N H2SOa.. Fig. 15. Scaled and partially cracked surface of the specimen after electrochemical tests in the 0.5N H)SOu.

Besides non-metallic inclusions, especially privileged to hydrogen accumulation are lamellar areas of ¢ martensite. Absorbed atomic hydrogen penetrating the steel, accumulates in places with non-metallic inclusions, lamellar precipitations of the second phase, microcracks and other structural defects, where convenient conditions for recombining of atomic hydrogen to molecular H2 exist. The recombination of atomic hydrogen to molecular state is a very exothermic reaction, which provides a pressure increase in formed H2 bubbles as well as nucleation and growth of microcracks in a surface region of the sample (Fig. 20).

Table 3. Mean percentage mass loss of cold-deformed samples after the immersion tests, % After bending and 100 hours immersion in IN HSO, both steels show a meaningful percentage mass decrement, among 47 and 49% and 2 orders of magnitude lower in chloride solution (Table 3). Cold deformation rises slightly the mass decrement in acidic medium, in comparison with the specimens investigated in undeformed state (Table 2). The opposite is true for chloride solution. However, the differences are not significant. Comparable mass loss of two steels both in non-deformed and plastically deformed states indicates that the initial structure does not have meaningful impact on the corrosion progress.

Fig. 20. Regions of hydrogen failures and hydrogen microcracks in 25Mn-3Si-1.5AI-Nb-Ti steel, plastically deformed and immersed in IN H2SOg (transverse section).

Fig. 21. Deep corrosion decrements and banding-like arrangement of corrosion products in 26Mn-3Si-3AI-Nb steel, plastically deformed and immersed in 1N H2SO,. Corrosion Resistance of High-Mn Austenitic Steels for the Automotive Industry

Fig. 25. Deep corrosion cracks in 25Mn-3Si-1.5Al-Nb-Ti steel after bending and tl immersion in 3.5wt% NaCl. Fig. 24. Scaled and cracked layer of corrosion products in 25Mn-3Si-1.5Al-Nb-Ti steel, plastically deformed and immersed in 3.5wt% NaCl.

Table 1. The chemical-composition of Al-Mg-Si alloy 3.2 Test solution

Seawater predominantly consists of about 3.5% of sodium chloride (NaCl) and many other ions. Chloride ions are very strong and could easily penetrate the passive film. Thus, dissolution of the aluminium substrate occurs and results in corrosion. The adsorption of the corrosion inhibitor competes with anions such as chloride. By assuming that the corrosion inhibitor molecules preferentially react with Al3+ to form a precipitate of salt or complex on the surface of the aluminum substrate, the anodic and cathodic processes subsequently suppressed by inhibitor molecules. Thus, this result suggests that the protective film that was formed comprise aluminium hydroxide, oxide and salts or complexes of the corrosion inhibitor anions.

The values of Re for the alloy in inhibited solution with NH were enhanced up to 10 times higher as compared to that of the value of Ra in uninhibited solution. Meanwhile, VL and TS shown better performance in improving the value of Ra for Al-Mg-Si alloy in studied aggressive solution, where they were increased the values up to 13 and 14 times higher, respectively. It should be noted that while Ro values increase with the addition of corrosion inhibitor, the capacitance, Cg values decrease indicating the formation of a surface film. Thus, effective corrosion resistance is associated with high Ra and low Ca values (Yagan et al., 2006). Increase in Rj values and decrease in Ca values by NH, VL and TS indicated that the studied inhibitors inhibit the corrosion of Al-Mg-Si alloy in seawater by adsorption mechanism (Noor, 2009) and the thickness of the adsorbed layer increases with the increase of inhibitor concentration.

The equivalent circuit fitting for these experimental data is a Randles circuit. The Randles equivalent circuit is one of the simplest and most common circuit models of electrochemical impedance. It includes a solution resistance, R, in series to a parallel combination of resistor, Re, representing the charge transfer (corrosion) resistance and a double layer capacitor, Cy, representing the electrode capacitance (Badawy et al., 1999). In this case, the value of R, can be neglected because the value is too small as compared to that of the value of Ry. The equivalent circuit for the Randles cell is shown in Figure 2.

Table 4. Rot and Ca of Al-Mg-Si alloy in seawater obtained using impedance method

The values of the result are presented in Table 5. All these parameters showed a similar trend. In all cases, increasing the inhibitors concentration is accompanied by an increase in the IE (%) and maximum for 1000 ppm. The JE (%) for all the measurements obtained from three different methods; PP, LPR and EIS are in good agreement. The inhibitive properties of the studied natural products can be given by the following order: NH < VL < TS.

Table 5. Values of IE (%) for Al-Mg-Si alloy at various concentrations of NH, VL and TS

4.3.2 Inhibition mechanism of VL Improvement of Corrosion Resistance of Aluminium Alloy by Natural Products

Fig. 4. FTIR spectrum of VL. The inhibition process of vanillin could be explained as follows: FTIR spectrum illustrate that vanillin is an aromatic aldehyde containing carbonyl, methoxy, and hydroxyl groups arranged around the aromatic ring (Figure 4). The bands at about 1153.7 to 1199.9 cm= and 2362.3 cm=! in the spectrum are assigned to carbonyl group, meanwhile the bands located between 1429.8 to 1664.9 cm~ are refers to hydroxyl group and aromatic compound (benzene ring).

The inhibition process of tapioca starch can be explained as follows: FTIR spectrum shows that TS is composed of mixture of two molecular entities (polysaccharides), a linear fraction, amylase and highly branched fraction, amylopectin (Figure 5). Both of them are polymers of glucose. Amylose is constituted by glucose monomer units joined to one another head to tail forming alpha-1, 4 linkages; these are linked with the ring oxygen atoms all on the same side. Amylopectin differs from amylase in that branching occurs, with an alpha-1, 6 linkages every 24-30 glucose monomer units. Amylopectin has phosphate groups attached to some hydroxyl group (Wu et al., 2009).

Fig. 6. SEM analysis of the unexposed Al-Mg-Si alloy surface. The SEM micrograph of the unexposed Al-Mg-Si alloy is shown in Figure 6. It shows that the surface of the metal is absolutely free from any pits and cracks. Polishing scratches are also visible.

Figure 7 corresponds to the SEM of the specimen surface after 60 days of immersion in seawater. Flakes showing corrosion products like metal hydroxides and its oxides can be observed, however no pits or cracks were noticed (Gao et al., 2008). The figures also show the presence of micro organisms (plankton) on the surface of the specimen which contributes to the corrosion process. The corrosion process in deep seawaters occurs under very specific conditions and is characterized mainly by high chloride contents, the presence of CO» and H2S and micro organisms (Yagan eft al., 2006).

Figure 9 depicts the SEM of the specimen surface after 60 days of immersion in seawater with the addition of 1000 ppm of NH. It can be seen that the flakes in the surface of the specimens decreased as compared to that of the micrograph in Figure 8. The specimen is covered with the inhibitor molecules giving a protection against corrosion, where a thin layer developed on the specimen surfaces.

The EDS spectrum in Figure 8 presents the elements exist in Al-Mg-Si alloy after immersion test. It shows the presence of sodium (Na) and chlorine (Cl) in the specimen surface. Seawater predominantly consists about 3.5% of sodium chloride (NaCl) and many other ions. Chloride ions are very strong, and could easily penetrate the passive film, and dissolution of the aluminium substrate occurs and results in corrosion.

Fig. 10. EDS analysis of Al-Mg-Si alloy immersed in seawater containing 1000 ppm of NH. The EDS spectrum in Figure 10 shows the existence of carbon (C), oxygen (O) and sulphur (S); due to the carbon, oxygen and sulphur atoms of the NH. These data show that carbonaceous material containing O and S atoms has covered the specimen surface. This layer is absolutely due to the inhibitor, because the carbon signal and the high contribution of the oxygen signal are absent on the specimen surface exposed to uninhibited seawater. It can be seen obviously from the spectra, the Al peaks are dramatically suppressed relative to the samples in uninhibited seawater (Figure 8). This is due to the overlying inhibitor film on the specimen surface area (Amin et al., 2009).

Fig. 11. SEM of specimen surface after immersion in seawater with presence of 1000 ppm VL Figure 11 portrays the SEM micrograph of Al-Mg-Si alloy immersed in seawater with the presence of 1000 ppm of VL. It was observed that compounds of VL were precipitated on the alloy surface.

Figure 13 depicts the SEM of specimen surface after 60 days of immersion in seawater with the addition of 600 ppm of tapioca starch. TS had shown similar characterization as NH and VL. From the micrograph, it can be seen that the flakes in the surface of the specimen are lessen when compared with that of the micrograph of uninhibited specimen.

Fig. 12. EDS analysis of Al-Mg-Si alloy immersed in seawater containing 1000 ppm of VL. The observations mentioned above were confirmed by EDS analysis (Figure 12). The EDS spectra showed an additional line characteristic for the existence of C (due to the carbon atoms of the VL). These data show that carbonaceous material has covered the specimen surface. This layer is absolutely due to the VL compound, because the carbon signal is absent on the specimen surface exposed to uninhibited seawater.

Fig. 14. EDS analysis of Al-Mg-Si alloy immersed in seawater containing 1000 ppm of TS

Fig. 1. Three-electrode system for corrosion studies: a) measuring system, b) circuit diagram. steady state corrosion by applying the sinusoidal alternating potential signal of small amplitude, but in a wide range of frequencies and the automatic recording of current intensity responses of the system. Investigations of corrosion by those methods, based on a change in the relationship between potential and current were implemented in the three- electrode system (Fig. 1).

Fig. 2. Block diagram of the electrochemical corrosion tests. In the three-electrode system, used in the current study, the examined metal takes the role of the active electrode. A calomel electrode, Hg/Hg2Clo/KCl characterized by the potential +244 mV, was used as the reference electrode. Auxiliary electrode was made of platinum (Pt). The tests were carried out in the corrosive environment of 0.5M NaCl solution at pH = 7 and a temperature of 293K (Trzaska &Trzaska, 2010).

Fig. 3. The set of appliances for the electrochemical corrosion testing by EIS technique.

Potentiodynamic polarization distortion of the steady state technique at the interface of both Al as well as AlsO3 with 0.5M NaCl solution, was applied with the change in the potential ranging from -780mV to -450mV. The rate of the potential change during the test was 0.2mV/s. Current characteristics j=f(E) of test materials in the form of potentiodynamic polarization curves are shown in Fig. 5.

Corrosion test has been applied to technical aluminum (99.9%) and to aluminum with AlO3 surface layer produced by hard anodic oxidation and sealed in boiling-hot deionized water. Images of morphology and topography of the surface layer of aluminum and Al,Os3 before corrosion tests using scanning electron microscope (SEM) are shown in Fig. 4.

For further characterization of electrochemical corrosion processes at the interface between the environment of 0.5M NaCl solution and Al and AlO3, the electrochemical impedance spectroscopy method was applied. As stated above this method allows considering the corrosion process as a combination of equivalent electric circuits. In the case of Al the study was catried out with the amplitude of the forcing sinusoidal signal of 10mV. However, in the case of AloOs layer the amplitude of perturbing signal in the corrosion balance was fixed at 20mV. The study was conducted in the frequency range 23kHz + 16mHz. Measured impedance spectra of Al and Al,O3 layer in the corrosive environment of 0.5M NaCl solution are presented in the form of Nyquist and Bode diagrams. Equivalent electrical circuits obtained by minimizing the mean square error were further used for the analysis of experimentally identified frequency characteristics and a description of corrosive processes in the systems under investigation. A simple electric circuit consisting of three elements of type R and C with a single time constant was adopted as system model for Al. Figs. 6 and 7 show the circuits modeling respectively the

Fig. 7. Layout of the corrosive system for aluminum with a protective layer of AloO3 in 0.5M NaCl environment and its equivalent circuit.

The frequency characteristics of corrosion systems of bulk Al and AlOs layer in 0.5M NaCl solution in the form of Nyquist and Bode plots obtained by the measurements and calculations based on adopted equivalent electrical circuits are shown in Figs. 8 and 9, respectively.

Table 2. Parameters of equivalent electrical circuit for corrosion processes of bulk Al and AlOs layer in 0.5M NaCl solution.

Electrochemical corrosion characteristics of iron were determined by potentiodynamic and impedance spectroscopy techniques. Tests were applied to chemically pure iron Fe made by electrocrystalization method and to carbon steel S235JR with the chemical composition shown in Table 3.

The corrosion test parameters of iron and S235JR steel in 0.5M NaCl solution are summarized in Table 4.

Fig. 11. Topography and surface morphology of Fe and 5235JR steel prior corrosion tests. The characteristics j = f(E) for iron Fe and S235JR steel in 0.5M NaCl solution obtained in the above potential ranges are shown in Fig. 12.

Nyquist diagrams (Fig. 14) in the shape of the characteristic semi-circles indicate the activation process control during corrosive material tests. Much larger diameter of the semi- circle in the case of electrochemically generated iron shows high electrical resistance at the interface metal-solution, which is the result of oxidation of iron and Fe*? ions passing into the solution. This indicates a greater corrosion resistance of iron compared to steel S235JR in

Fig. 16. Images of the surfaces of Fe and S235JR steel after corrosion tests. Images of destruction of corrosion test samples of Fe and S235JR steel after corrosion tests are shown in Fig. 16.

Fig. 15. Bode amplitude and phase spectra of circuit impedance and corrosion systems of Fe and $235]R steel in 0.5M NaCl solution determined experimentally (point line) and as a result of calculations (solid lines).

inhibits the processes of corrosion of nickel in a certain range of the potential. However, differences in the corrosion resistance of pure nickel and its alloy result from both the additive contained in the material alloy, as well as different material structures.

Table 6. Corrosion parameters of nickel and alloy NiP in 0.5M NaCl solution.

Fig. 22. Nyquist diagrams of impedance spectra of investigated corrosion systems of nickel and its alloy in the environment of 0.5M NaCl solution determined experimentally (point line) and as a result of calculations (solid lines).

In the case of the microcrystalline structure nickel and NiP alloy in corrosive environment of 0.5M NaCl solution, a pickling of their internal structures occurred over the entire surface exposed and even its internal structures was revealed. On the other hand, corrosion of the nanocrystalline nickel in this environment takes the form of uneven local corrosion.

plastic, or epoxy coating), and positive or negative experiences with these materials could have influenced their responses. The questions included two stimuli, which are compared simultaneously. Each respondent rated each of the two alternatives on a scale from 1 to 9. The scale value of 1 indicates the plumbing material is not preferred, while 9 indicates an extremely preferred plumbing system. The material attributes are listed in Table 1. @Names of the plumbing materials were not revealed to the study participants

Respondents with pinhole leaks had somewhat older homes compared to respondents without leaks (Table 2). Fifty-three percent of respondents without leaks lived in homes built since 2000 compared to 4% of respondents with leaks. Five percent of respondents without leaks lived in homes built before 1990 compared to 23% of respondents with leaks.

lable 3. Pinhole leaks occurring in cold or hot water pipes.

aMultiple choices per respondent were accepted. Percent = number reported divided by the total number of respondents with leaks (212).

¢Numbers do not sum to 100 due to rounding. Table 4. Method of leak repair.

Sixty-seven percent of respondents use some type of water treatment for purposes other than pinhole leak prevention (Table 7). The most common treatment was a refrigerator filter, used by 63%. Thirty-two percent of respondents reported that they purchase drinking water. The most common reasons given for using water treatment devices are to improve taste or smell of drinking water (mentioned by 45% of respondents), and to improve safety of drinking water (mentioned by 33% of respondents).

Households’ Preferences for Plumbing Materials

Totals do not sum to 100 due to rounding. Table 7. Use of water treatment for purposes other than corrosion prevention. Fig. 2. Satisfaction with home drinking water quality.

Table 8. Preference valuation of plumbing materials.

Table 9. Plumbing material chosen as most preferred. Households’ Preferences for Plumbing Materials 6.2 Empirical analysis results

@The number of observations included in the model is 1086. Independent variables take form of dummy variables with value of one when the characteristic was present in the plumbing material profile and zero otherwise. To avoid a dummy variable trap, one of the attribute levels was excluded from the analysis. The omitted characteristics represent Material C (copper) descriptions. Coefficient estimates show how much increase in the likelihood of being in a higher category results from a one unit increase in the independent variable. ¢Standard error represents the variation of the estimate. d Wald statistics and p-value represent the significance level. © Model Statistics: Log-likelihood value is 182.641 with chi-square of 114.136 and p-value of 0.000; Nagelkerke’s R-square is 0.084. Besides evaluating the directional impact of the independent variables on the preference level of the households, the impact of the statistically significant independent variables on the preference category is evaluated for all three plumbing materials. As the attribute levels describing each of the three hypothetical materials are known, the regression results can be organized by plumbing materials. For example, Material A is described by attribute level called ‘corrosion proof’ as well as ‘installation takes around 4 days’. The coefficient estimates for the statistically significant attribute levels are employed to compute preference valuation categories for each material type. In case of the Material A (epoxy coating) computation of the preference valuation category called ‘Moderately Preferred’, the following represents the estimate computation: 1.315 - 0.654 - 0.559= 0.102, where 1.315 is the moderately preferred coefficient, 0.654 is ‘the corrosion proof’ coefficient, and 0.559 is ‘the convenience of installation’ coefficient; and the odds ratio computation: exp(0.102) = 1.107.

has a larger probability estimate for ‘Strongly Preferred’ to ‘Extremely Preferred’ category preference. On the other hand, Material C has a higher probability estimates for categories ‘Not Preferred’ and ‘Moderately Preferred’. All three materials have the highest frequency of estimates falling into “Not Preferred’ and ‘Moderately Preferred’ categories. Based on the overall results, Material A (epoxy coating) is the most preferred material followed by Material B (plastic). Material C (copper) is the least preferred plumbing material.

2 The number of observations included in the model is 1072. Independent variables take form of dumm variables with value of one when the characteristic was present in the plumbing material profile and zero otherwise. To avoid a dummy variable trap, one of the attribute levels was excluded from the analysis. The omitted characteristics represent Material C (copper) descriptions. > Coefficient estimates show how much increase in the likelihood of being in a higher category results from a one unit increase in the independent variable. ¢ Standard error represents the variation of the estimate. 4 Wald statistics and p-value represent the significance level. © Model Statistics: Log-likelihood value is 1565.522 with chi-square of 119.384 and p-value of 0.000; Nagelkerke’s R-square is 0.101. level called ‘corrosion proof’ and ‘installation takes about 4 days’. The coefficient estimates for the statistically significant attribute levels are employed in the material based preference category computation. In case of Material A (epoxy coating) the computation for preference valuation category of ‘Moderately Preferred’, the following represents the estimate computation: 1.300- 1.145- 0.575 -0.0001*$522 = -0.472; and the odds ratio computation: exp(- 0.472) = 0.624 (Table 15).

Log In

CORROSION RESISTANCE

Sign up for access to the world's latest research

Abstract

Related papers

Related topics