Corrosion-erosion processes of the AA 5083 (Al-Mg) alloy in seawater

Procesos de corrosión-erosión de la aleación AA 5083 (Al-Mg) en agua de mar

A. Aballe¹, M. Bethencourt^1*, F.J. Botana¹, M. Marcos² and J.M. Sánchez-Amaya¹

¹ Departamento de Ciencia de los Materiales e Ingeniería Metalúrgica y Química Inorgánica, Facultad de Ciencias del Mar, Universidad de Cádiz, Ave. República Saharaui s/n, Puerto Real, 11510-Cádiz, Spain. * E-mail: manuel.bethencourt@uca.es

Recibido en octubre de 2001;
aceptado en septiembre de 2002.

Abstract

Fourier Transforms are applied to analyse surface roughness profiles recorded on samples coming from corrosion-erosion assays. The information retrieved using this method clearly complements that revealed by the more classical roughness amplitude parameters. The analysis procedure here proposed can be applied not only to characterise the surface of corroded samples but, in general, to evaluate the quality of any surface after application of finishing treatments.

Key words: corrosion-erosion, roughness, Fourier Transform, surface quality.

Resumen

Palabras clave: corrosión-erosión, rugosidad, Transformadas de Fourier, calidad superficial.

Introduction

When a material is exposed to an erosion-corrosion process it suffers transformations which involve changes in its superficial texture. These transformations are related to the mechanism through which this process has been verified (Stack et al., 1996).

The study of the mechanism of erosion-corrosion processes is limited, due to the lack of instrumental techniques which allow to make a quantitative analysis of these textural changes.

The use of the superficial roughness technique is proposed in the present work to characterise the topography of samples proceeding from erosion-corrosion assays at a quantitative level. Fourier Transform has focused special attention on the analysis of roughness profiles.

Experimental

Roughness measurements included in this work were made using a M4Pi/CNOMO Perthometer from Perthen. A profile resolution of 12 µm and a scanned distance range from 250 µm to 16 mm are attainable with this instrument. A PC running software was developed in our lab for the control of this device.

Theoretic considerations about roughness measurements

Surface roughness measurement is considered to be a basic technique to evaluate the quality of any surface after application of a finishing treatment (García de la Chica, 1993). In the common use of this technique, after recording the surface roughness profile, it is possible to calculate normalised parameters, which provide quantitative information on the analysed surface. In the literature, two families of parameters have been classically described, one of them related with the amplitude and the other with the shape of the profiles.

No doubt, amplitude parameters (R_a, R_z...) are used more frequently, while shape parameters are yet seldom at use. In fact, until very recently these parameters have been employed qualitatively and only since 1990 quantitative methods have been proposed for their evaluation (DIN, 1990).

One of the methods recommended in DIN 4776 to evaluate the shape of the profiles is based on the bearing ratio curves or Abbott-Firestone curve. As an example, in figure 1 these curves have been represented for three surfaces with extreme texture characteristics. The first one (fig. 1a), corresponds to a surface with homogeneous undulations. The second one (fig. 1b) corresponds to a surface with exceptional valleys of great depth with respect to the mean surface and, the third (fig. 1c) corresponds to a regular surface with some peaks of great altitude. The characterisation of actual surfaces can be obtained by means of a qualitative comparison of Abbott curves with one of the models represented in figure 1.

A second method to evaluate the shape of the profiles is based on the procurement of the Amplitude Density Function in the roughness profile (AENOR, 1986), equivalent to the Probability Density Function used in statistics. From these curves it is possible to calculate the skewness parameter of the distribution S_k (AENOR, 1986). In figure 1 we have indicated the value of the skew corresponding to each model.

Results and discussion

In this section we presennt the results obtained from the application of the surface roughness technique to the characterisation of samples coming from erosion-corrosion tests. At this point, it is important to stand out that this technique has been scarcely used in the bibliography to characterize corroded samples and, if used, it was limited to the calculation of some amplitude parameters (Tomlinson and Matthews, 1994; Rao and Buckley, 1985).

In figure 2, the roughness profiles recorded on tested samples at different experimental conditions have been represented. In short, the S-1 sample has been subjected to the action of a NaCl dissolution moving at a velocity of 20 m/s, while the S-2 sample has been tested at 36 m/s. In this figure clear differences that exist between both profiles can be observed, as far as the amplitude and the shape are concerned. The differences in amplitude that exist between both profiles have been revealed through the corresponding values of R_a, which were 5.4 and 7.0 respectively.

With the aim of evaluating the differences that exist in the shape of the profiles, in figure 3 the Abbott-Firestone curves have been plotted and the skew values obtained from the roughness profiles included in figure 2. The obtained results show no significant differences neither regarding the Abbott curves nor in the calculated values of the skew. It means that, if we use the parameters recently proposed in the literature to evaluate the shape of the roughness profiles, it is not possible to detect differences between the two studied samples. Therefore, these results suggest that no modifications in the texture of eroded samples were observed by increasing the fluid velocity in the assays.

The results disagree with the results obtained through studying the eroded samples by means of Scanning Electronic Microscopy (SEM). In figure 4, the SEM images acquired in the eroded zones of the samples S-1 and S-2 have been presented. In this figure we can observe the existence of large differences between both samples. Therefore, these results show that conventional shape parameters are not sensible enough to detect the changes produced in the erosion-corrosion tests.

In the present work, in order to improve the characterisation of the eroded samples, the application of a Fourier Transform treatment to the roughness profiles has been proposed. So, it is possible to obtain complementary quantitative information to the conventional shape and amplitude parameters provided.

Roughness profiles, as those represented in figure 2, can be mathematically treated as if they were the result of adding sinusoidal modulations with varying frequencies, phases and amplitudes. The Fast Fourier Transform operation provides values of frequencies, phases and amplitudes of each component present in the profile analysed (Champeney, 1973). These data are usually plotted as power spectra (PSD), which represents the amplitude of each component present in the profile versus the spatial frequency. In figure 5, the power spectrum obtained applying the Fast Fourier Transform (FFT) to the experimental profiles in figure 2, is presented.

In the power spectra corresponding to the sample S-1 (fig. 5b), we can observe that the larger intensity components appear in 0.01, 0.015 and 0.02 µm^-1. This indicates that the profile corresponding to sample S-1 can be considered to be formed by the addition of three sinusoidal modulations whose amplitudes and spatial frequency are shown in the power spectra.

The main difference between the PSD of the samples S-1 and S-2 is that in the sample assayed at high velocity (fig. 5b), the modulation of low frequency, centered at 0.005 µm^-1, increases considerable its amplitude.

As indicated above, it is possible to calculate amplitude and phase for the wave centered in 0.005 µm^-1 from the PSD in figure 5b. This wave developed when the fluid velocity increased in the erosion-corrosion tests and is represented in figure 6 together with the roughness profiles for samples S-1 and S-2. In this figure it can be verified that the roughness profile for the sample assayed at high velocity is dominated by the 0.005 µm^-1 wave, while such wave is absent in the profile of the sample S-1.

These results indicate a clear relation between the increasing velocity of the corrosive medium producing textural changes on the surface of the samples. This changes create modulations of lower frequency than those developed in the test at low velocity. The wavelength of such undulations is similar to the biggest diameter pits observed in the SEM images in figure 4b.

Conclusions

This study showed that superficial roughness measurements can evaluate textural modifications in eroded samples due to varying velocities of the fluid. Nevertheless, if conventional parameters are used, only changes related to the amplitude of the profiles can be detected.

By using FFT to the experimental signals, we have been able to identify the frequency, amplitude and phase of each of the sine modulations generated in the surfaces as a result of the corrosive medium action. The results obtained indicate that the increase in the fluid velocity creates low frequency undulations in the surface. This kind of information is of interest to delve into the knowledge of corrosion-erosion process. Nevertheless, if correlations are to be established among the kind of modulation, the surrounding conditions of the tests and the mechanism of the process, it is necessary to apply this method to a higher number of tested samples in different experimental conditions.

Finally, is ought to that the Fourier analysis on roughness signals can be used to evaluate the quality of any surface after the application of finishing treatments. The results obtained in this work show that this kind of analysis provides complementary information to that given by conventional shape parameters.

Acknowledgements

This work has been financed by the Comisión Interministerial de Ciencia y Tecnología (CICYT), Project MAT2001-3477.

References

Asociación Española de Normalización y Certificación (AENOR). (1986). Rugosidad superficial. Parámetros, sus valores y las reglas generales para la determinación de las especificaciones. Norma UNE 82-301-86. pp. 2-3.         [ Links ]

Bethencourt, M., Botana, F.J., Calvino, J.J., Cauqui, M.A., Marcos, M. and Rodríguez-Chacón, M.A. (1996). Quantitative evaluation of erosion-corrosion effects in a jet submerged device. Proceedings of the 13^th International Corrosion Congress, Vol. II, paper 198. Australasian Corrosion Association. pp. 1-8.         [ Links ]

Champeney, D.C. (1973) Fourier Transforms and Their Physical Applications. Academic Press, London. pp. 59-61.         [ Links ]

Deutsches Institute für Normung (DIN). (1990). Parameters Rk, Rpk, Rvk, Mrl, Mr2 for the description of the material portion in the roughness profile; measuring conditions and evaluation procedures. Norma DIN 4776. pp. 4-5.         [ Links ]

García de la Chica, A. (1993). Metrología del Acabado Superficial: Rugosidad, Colección Consejos para la Práctica Metrológica. Asociación Española para la Calidad, Madrid, pp. 16-17.         [ Links ]

Rao, B.C.S. and Buckley, D.H. (1985). Erosion of aluminium 601-T6 under cavitation attack in mineral oil and water. Wear, 105: 171-182.         [ Links ]

Stack, M.M., Zhou S. and Newman, R.C. (1996). An approach to defining erosion-corrosion regimes for erosion corrosion in aqueous slurries. Proceedings of 13^th International Corrosion Congress, Vol. II, paper 191. Australasian Corrosion Association. pp. 1-4.         [ Links ]

Tomlinson, W.J. and Matthews S. J. (1994). Cavitation erosion of structural ceramics. J. Mat. Sci., 29: 1101-1108.         [ Links ]