Materials Transactions, Vol. 44, No. 4 (23) pp. 539 to 545 Special Issue on Platform Science and Technology for Advanced Magnesium Alloys, II #23 The Japan Institute of Metals Microstructure and Voltage Current Characteristics of Anodic Films Formed on Magnesium in Electrolytes Containing Fluoride* 1 Sachiko Ono 1, Hideo Kijima* 2 and Noboru Masuko 2 1 Department of Applied Chemistry, Faculty of Engineering, Kogakuin University, Tokyo 163-8677, Japan 2 Department of Metallurgical Engineering, Faculty of Engineering, Chiba Institute of Technology, Narashino 275-8588, Japan Formation behavior of anodic oxide films on magnesium in various electrolytes including fluoride was investigated with attention to the effects of anodizing voltage, ph and aluminum content. In the range of formation voltage between 2 V and 1 V, porous film was formed in alkaline fluoride solution associated with high current density at around 5 V and at breakdown voltage. The critical voltage of breakdown to allow maximum current flow was approximately 6 V and relatively independent on substrate purity. Barrier type films or semi-barrier type films, which were composed of hydrated outer layer and inner layer, were formed at the other voltages. A peculiar phenomenon of high current density at around 5 V, which may be caused by trans-passive state, was not observed for anodizing in acidic fluoride solutions such as Dow17 and ammonium fluoride. In the case of AZ91D, the critical voltage increased to 7 V and peculiar phenomenon at 5 V was not observed, so that only barrier films were formed at less than the critical breakdown voltage. When AlO 2 ion was added in the electrolytes, the critical voltage remarkably increased and current density effectively decreased with increasing AlO 2 content. The passivation effect of aluminum addition in the electrolytes is more remarkable than the addition in magnesium substrates. The depth profiles of constituent elements showed that aluminum migrated into oxide film to reach near oxide/substrate interface. Atomic ratio of aluminum to magnesium increased with increasing voltage to attain.42 at 8 V and crystalline MgAl 2 O 4 and MgO were found in the film. (Received November 28, 22; Accepted January 2, 23) Keywords: magnesium, AZ alloy, anodic film, trans-passive state, structure analysis 1. Introduction Anodizing of magnesium is commercially used to provide a thick porous oxide layer and to improve corrosion resistance 1) similarly to that of aluminum. Recently, the practicalimportance of magnesium has remarkably increased, especially in the fields of automobile industry and mobile electric commodities such as personal computers and telephones with emphasizing a benefit of the recycling capability of magnesium itself. However, only limited information on the microstructure and fundamentalgrowth mechanism of surface oxide films on magnesium, for instance, naturaloxide films, 2 4) anodic films 5 11) and chemicalconversion coating films 12,13) is available in contrast to the numerous investigations on the oxide film formation on aluminum. Previously, the authors reported that the anodic films 7 11) as well as chemical conversion coating films 12,13) formed on pure magnesium and AZ die-casting alloys have cylindrical porous structure comparable to the Keller s model of anodic alumina. 7,14) However, the film growth mechanism by anodizing of magnesium in Dow17, which is a most commonly used commercial solution, was rather complicated in comparison with the well-known behavior of anodic film growth on aluminum. Namely, anodic film growth mainly proceeds by the formation of MgF 2 and magnesium oxyhydroxide at metal/film interface and the dissolution of the film at pore bases, followed by the crystal growth of MgF 2, NaMgF 3 and Cr 2 O 3. 7) The commercialelectrolyte such as Dow17 is an acidic solution and includes fluoride. According * 1 This Paper was Originally Published in Japanese in J. JILM 52 (22) 115 121. * 2 Undergraduate Student, Chiba Institute of Technology. Present address: CSK Chiba-system Co. Ltd., Chiba, Japan. to the Pourbaix diagram 15) however, oxide or hydroxide film is formed on magnesium only in an alkaline region. The anodic film growth in acidic fluoride solution must be brought about the characteristic affinity between fluorine and magnesium. Therefore, in the present study, the fundamental behavior of anodic film growth on magnesium and magnesium alloys in various types of electrolytes containing fluoride was investigated with focusing on the effects of formation voltage, substrate composition, ph and aluminum content of electrolytes. 2. Experimental The pure (99.95% and 99.6%) magnesium, AZ31B alloys and injection molded AZ91D alloys were used in this study. The former three types of substrates were prepared as rolled sheets. Chemicalcomposition of magnesium specimens is shown in Table 1. The magnesium sheets were degreased by dipping in an alkaline solution or subsequently etched in an acidic solution to obtain clear bright appearance of the surface. For the observation of substrate microstructure, specimens were etched by immersion in ethanol/nitric acid (24 : 1) mixture or 4% HF solution. Anodic films were formed at constant voltage for 1 min in an alkaline mixed solution of 3 moldm 3 KOH.6 moldm 3 KF.2 moldm 3 Na 3 PO 4 (ph13), an acidic.5 moldm 3 NH 4 FHF solution (ph2) at 298 K, and Dow17 solution (<ph1) containing 3 g of NH 4 HF 2, 1 g of Na 2 Cr 2 O 7 and 9 mlof H 3 PO 4 at 348 K. To verify the effect of aluminum in the electrolyte, KAlO 2 was added. Maximum current was limited at 1 Am 2 and anodizing was terminated when current density exceeded this value for 3 min. Anodic film structure was evaluated by using scanning electron microscopy (SEM, JEOL S-63), glow discharge optical emission spectroscopy (GD-OES, Jobin-Yvon JY5RF), electron
54 S. Ono, H. Kijima and N. Masuko Table 1 Chemical composition (mass%) of magnesium specimens. Element Mg Al Zn Mn Fe Cu Ni Si Ca Pb Sn Cr 99.95.1.1.1.3.2.1.1.3.3.1 99.6%Mg >99:6 :5> :5> >:1 >:5 >:2 >:2 >:5 >:2 AZ31B >95:7 2.83.83.48 >:5 >:2 >:2 >:5 >:2 AZ91D Bal9.6.17.4.15.1.5 (a) (b) (c) (d) Fig. 1 Microstructure of magnesium substrates after chemicaletching. (a), (b) 99.6%Mg, (c) AZ31B, (d) AZ91D. probe microanalysis (EPMA, JEOL JXA-88) and X-ray diffraction analysis (XRD, RIGAKU RINT-25). 3. Results and Discussion 3.1 Substrate microstructure Figure 1 shows microscopic structure of etched surfaces of four specimens. The size of crystalgrain was 3 5 mm for 99.95% Mg and AZ31B. While, the grain was 1 2 mm in sizes and showed dendrite-like structure for 99.6% Mg. For AZ91D, grains in the size of 5 1 mm and large grains indicating phase (Mg 17 Al 12 ) at the grain boundary were found. 3.2 Current-time curves Typicalcurrent-time curves at constant voltage anodizing of 99.95% Mg were shown in Fig. 2. Because of the low current density following the initial surge current observed at 1 8 5V 6 4 2V, 2V 5V 1V 2 1 2 3 4 5 6 Time t / s Fig. 2 Typicalcurrent-time curves of anodic film growth on 99.95% Mg at different voltage in KOH KF Na 3 PO 4 solution.
Microstructure and Voltage Current Characteristic of Anodic Films on Magnesium 541 2 V and also at the voltages ranged from 2 to 5 V, anodic films formed at these voltages appeared to be barrier type. Extremely high current density was observed at 5 V after 4 s rest period. Peculiar phenomenon of such high current density around 5 V followed by gas evolution was also reported in the anodizing in a sodium hydroxide solution. 16) The rest period before current increase observed at 5 V and 1 V suggests the formation of relatively protective barrier film. Sparking discharge accompanied by high current flow more than 1 Am 2 caused by electric breakdown was detected at the voltage higher than 6 V. Therefore, porous type films were formed when appreciable current flow occurred; namely at 5 V, 1 V and the voltage higher than 6 V. In the case of 99.6% Mg, I-t curves were similar to those of 99.95% Mg although the current density was slightly higher. The criticalvoltage of high current flow accompanied by sparking discharge was 6 V for 99.6% Mg similar to those for 99.95% Mg and AZ31B alloy. In the case of AZ91D alloy, current was depressed at all voltages, especially at 5 V. Therefore, no porous film was formed except at sparking voltages higher than 7 V. It suggests that aluminum contained in substrate effectively depressed current flow by the formation of aluminum oxide in the film to improve the surface passivity. 3.3 Surface appearance of the films Surface appearance of the films with the effects of anodizing voltage and substrate impurity was summarized in Table 2. Anodic films formed at 5 V and 1 V shows grayish appearance indicating relatively thick film formation accompanied by high current flow. The films formed at breakdown voltage are whitish gray suggesting thick porous film growth. The films formed at other voltages appeared to be uniform and whitish or transparent showing the barrier type film formation. It is noteworthy that the films formed on AZ91D were all uniform even when breakdown took place with sparking, suggesting the effect of aluminum on the uniformity of the film. 3.4 Current-voltage characteristics The relations between formation voltage and final current density after 1 min is shown in Figs. 3 and 4 with the effect of substrate composition. The shape of I-V relation is Table 2 Effects of anodizing voltage and substrate impurity on film appearance formed in KOH KF Na 3 PO 4 solution. Unif: Uniform, Hete: Heterogeneous, Lcal: Locally colored, Trns: Transparent, Whig: Whitish gray, Darg: Dark gray, Whit: White. Volts 99.6%Mg AZ31B AZ91D 2 V Unif-Trns Unif-Trns Unif-Trns Unif-Trns 5 V Hete-Whig Lcal-Darg Unif-Gray Unif-Whig 1 V Unif-Whig Unif-Whig Unif-Whig Unif-Whig 2 V Unif-Trns Unif-Trns Unif-Whig Unif-Trns 3 V Unif-Whig Unif-Whit Unif-Trns Unif-Whit 4 V Unif-Whig Unif-Whit Unif-Trns Unif-Whit 5 V Lcal-Whig Unif-Whit Unif-Trns Unif-Whit 6 V Lcal-Whit Lcal-Whig Lcal-Whig Unif-Whig 7 V Unif-Whig 12 1 8 6 4 2 99.6%Mg AZ31B 2 4 6 8 AZ91D Fig. 3 Voltage Current characteristics of anodizing in KOH KF Na 3 PO 4 solution with the effect of different Mg substrates. Anodizing was performed for 1 min at 298 K. 5 4 3 2 1 99.6%Mg AZ31B AZ91D 1 2 3 4 5 6 7 Fig. 4 Voltage Current characteristics at low current density region of Fig. 3. somewhat similar to the polarization curve observed at activated dissolution, passivation and trans-passive state for the ferrous group metals. However, the activated metal dissolution occurs at the potential slightly higher than redox potential, which is 2:36 V vs NHE for magnesium in acidic to neutral solution. Therefore, a peculiar phenomenon of high current density at around 5 V could be explained as transpassive state. The authors measured anodic polarization curves of various types of magnesium in sodium hydroxide solution. 16) The activated dissolution occurred at the potentialaround 1:4 V vs Ag/AgCl following a current retain by passivation. The peak current at 5 V for AZ91D significantly decreased. The criticalvoltage of high current flow over than 1 Am 2 accompanied by breakdown was relatively independent on substrate purity: namely, 6 V for 99.6% Mg, AZ31B and 99.95% Mg, and 7 V for AZ91D. This is different with the results obtained for NaOH solution. In the case of NaOH solution, the critical voltage accompanied by breakdown was strongly dependent on substrate purity: namely, 5 V for 99.6% Mg, 9 V for AZ31B, 1 V for AZ91D and 99.95% Mg. The current was generally depres-
542 S. Ono, H. Kijima and N. Masuko sed in fluoride solution than that associated with NaOH solution. These behaviors could be explained by the effect of fluorine to change the property of anodic film to form MgF 2 and AlF 3. Namely, fluoride formed in anodic oxide/hydroxide films predominantly control the film property and reduce the relative effect of substrate impurity. Figure 4 shows current voltage characteristics at low current density region of Fig. 3 and at lower than breakdown voltage. Minimum current density was observed at 4 V for every substrate. The current was higher in the following order: AZ31B > AZ91D > 99.6% Mg > 99.95% Mg. Remarkable depression of the peak current and increase in criticalvoltage for AZ91D is caused by the passivation ability of aluminum included in magnesium substrates. Aluminum appears to be enriched in anodic film and prevents high current flow due to substrate dissolution and oxygen gas evolution. Such passivation effect of aluminum seems to be insufficient in the case of AZ31B to allow high current flow. 3.5 Effect of aluminate ion addition in the electrolyte Subsequently, the effect of aluminum addition in the electrolyte was investigated. Figure 5 shows current voltage characteristics at anodizing of 99.95% Mg with the effect of AlO 2 concentration. The peak current at 5 V decreased with increasing aluminum content in the solution. The critical voltage of high current flow over than 1 Am 2 accompanied by breakdown was also increased with increasing AlO 2 concentration from 6 to 9 V. These results clearly indicate that AlO 2 ions incorporated into the films from the electrolyte to improve passivity to prevent high current flow caused by the trans-passive state as the similar manner to aluminum contained in the substrate. Thus, the passivation effect of aluminum addition in the fluoride electrolytes is more remarkable than the addition in magnesium substrates in contrast to that in the sodium hydroxide solution. 3.6 Effect of electrolyte ph Then, current voltage characteristic was measured in.5 moldm 3 NH 4 FHF acidic solution. As shown in Fig. 6, the peculiar phenomenon of high current density at around 5 V, which is caused by trans-passive state, was not observed 12 1 8 6 4 2 Without addition.2mol. dm 3 AlO 2.5mol. dm 3 AlO 2 1mol. dm 3 AlO 2 2 4 6 8 1 Fig. 5 Voltage Current characteristics of anodizing of 99.95% Mg in KOH KF Na 3 PO 4 KAlO 2 solution with the effect of aluminum addition. Anodizing was performed for 1 min at 298 K. 12 1 8 6 4 2 AZ31 2 4 6 8 1 Fig. 6 Voltage Current characteristics of anodizing of 99.95% Mg and AZ31B in.5 moldm 3 NH 4 FHF solution. Anodizing was performed for 1 min at 298 K. 2 Current density i / Am 12 1 8 6 4 2 2 4 6 8 1 Fig. 7 Voltage Current characteristics of anodizing of 99.95% Mg in Dow17 solution. Anodizing was performed for 1 min at 348 K. for anodizing in acidic fluoride solution, so that only barrier films were formed at the voltage less than 4 V. Although sparking was not observed in this electrolyte, the critical voltage of high current flow over than 1 Am 2 was 7 V. At 5 V and 6 V, relatively steady current transients with anodizing time was observed to imply porous film growth. In the case of AZ31, relatively low current flow indicating barrier film formation was observed at all voltage reached to 11 V. Figure 7 indicates current voltage characteristics measured in the Dow17 electrolyte. As it is clearly shown, no peculiar current peak at 5 V is observed. Electric breakdown accompanied by sparking appeared at the voltage higher than 5 V. However, the criticalvoltage of high current flow over than 1 Am 2 was 11 V. In the range from 6 to 1 V, thicker porous type films were obtained accompanying the current fluctuation caused by breakdown. Surface appearance of the film changed from yellow to dark green with the increase in film thickness according to the presence of Cr 3þ. 7) As shown in the above results, the peculiar phenomenon of high current density at around 5 V is not observed for anodizing in acidic fluoride solutions such as Dow17 and ammonium fluoride. The reason why trans-passive state accompanied by oxygen gas evolution is found only at
Microstructure and Voltage Current Characteristic of Anodic Films on Magnesium 543 (a) (b) (c) (d) (e) (f) Fig. 8 SEM images of the surfaces of (a) pre-treated, and anodized specimens of 99.95% Mg at (b) 2 V, (c) 5 V, (d) 1 V, (e) 4 V and (f) 6 V in KOH KF Na 3 PO 4 solution for 1 min at 298 K. alkaline solution is not clear, but it would be related to the change in nature of oxide film formed in alkaline and acid media. 3.7 SEM observation Figure 8 shows SEM images of the surfaces of (a) pretreated, and anodized specimens at (b) 2 V, (c) 5 V, (d) 1 V, (e) 4 V and (f) 6 V in KOH KF Na 3 PO 4 solution for 1 min at 298 K. The pre-treated specimen was covered with platelet like hydroxide described in the previous work 2) as shown in Fig. 8(a). The film formed at 2 V was a flat barrier type oxide having fine platelet of hydroxide at the surface, which appears to be rather similar to the pretreated surface. The same surface structure is observed on the film formed at 4 V as well as other barrier type films formed at the voltage range from 2 to 5 V. While, a relatively rough and porous surface is observed at the films formed at 5 V and 1 V in accordance with appreciable current flow. Holes in the diameter of 1 to 1 mm, presumably produced by local substrate dissolution were also shown. At the breakdown voltage, a lava like porous structure with dispersed pores in the size of approximately 1 mm, which is similar to those observed on aluminum and other valve metals, was detected. The surface structure of the films was not changed much when aluminate ions are added in the electrolyte, except at 1 V where current density was depressed (Fig. 9), so that flat and barrier type oxide was found. 3.8 Structure analysis by GD-OES, EPMA and XRD Depth profiles of constitute elements of the film formed on 99.95% Mg in alkaline fluoride solution was measured by GD-OED as shown in Fig. 1. It revealed that porous type films formed at 1 V were composed of hydrated outer layer, which was relatively resistant for argon sputtering, and inner layer. When aluminate ions were added in the electrolyte, film thickness decreased because of current depression. Aluminum incorporated and distributed in the whole thickness of the film. EPMA quantitative analysis of the film formed on 99.95% Mg in alkaline fluoride solution with 1 moldm 3 KAlO 2 at various voltages for 1 min is summarized in Table 3. The atomic ratio of Alto Mg for the film formed at 8 V is.42 (13/31) indicating notable high enrichment of Al compared to that of the film formed at 5 V. Mapping of elements obtained by EPMA indicated uniform distribution of Mg, O, F and Al at the surface. As indicated in Table 4, X-ray diffraction analysis of anodic film formed on 99.95% Mg in alkaline fluoride solution with 1 moldm 3 KAlO 2 at 8 V for 1 min clearly shows the presence of crystalline MgAl 2 O 4 and MgO in the film. It may be formed by heating effect of sparking caused by continuous electric breakdown. Thus, incorporated aluminum contributes to form spinel easily and to make the film property more protective than the magnesium oxy-hydroxide film itself. Driving force of such high enrichment of
544 S. Ono, H. Kijima and N. Masuko (a) (b) (c) (d) 2nm Fig. 9 SEM images of the surfaces of anodized specimens of 99.95% Mg at (a) 5 V, (b, c) 1 V and (d) 8 V in KOH KF Na 3 PO 4 KAlO 2 solution for 1 min at 298 K. Fig. 1 GD-OES depth profiles of constituent elements in anodic film formed on 99.95% Mg in (a) KOH KF Na 3 PO 4 and (b) KOH KF Na 3 PO 4 KAlO 2 solution at 1 V for 1 min at 298 K.
Microstructure and Voltage Current Characteristic of Anodic Films on Magnesium 545 Table 3 EPMA quantitative analysis of anodic film formed on 99.95% Mg in KOH KF Na 3 PO 4 solution including 1 moldm 3 KAlO 2 at various voltages for 1 min. Unit: at%. Element Mg Al O F P K Na 5 V 4. 1.79 55.4 2.37.23.9.15 1 V 82.2.74 15.8.66.4.11.49 8 V 31.1 12.9 53.6.96.94.37.16 Table 4 X-ray diffraction analysis of anodic film formed on 99.95% Mg in KOH KF Na 3 PO 4 KAlO 2 solution at 8 V for 1 min. d (nm) I=I Mg MgO MgAl 2 O 4.4766 44.466.2887 2.2858.264 1.265.2475 25.2452.2432.2437.2113 47.215.242 11.22.1921 8.19.1568 4.1555.1486 7.1473.148.1441 4.1428.1215 3.1215 aluminum in the film by migration from both substrate and electrolyte is assumed to be, in part, preferential spinel formation. 4. Conclusions Anodic film growth on 99.95% Mg, 99.6% Mg, AZ31B and AZ91D alloy sheets in various fluoride solutions was investigated with focusing on the effects of formation voltage, substrate composition, aluminate addition and ph of electrolytes, and following conclusions were obtained. (1) Porous film was formed in alkaline fluoride solution associated with high current density at around 5 V and at breakdown voltage. Except at those voltages, a barrier type film was formed in alkaline fluoride solution similar to that in sodium hydroxide solution. (2) The criticalvoltage of high current flow accompanied by breakdown was relatively independent on substrate purity, in discrepancy with that associated with sodium hydroxide solution. It may be due to the formation of MgF 2 /AlF 3 which predominantly control the film property to reduce the effect of substrate purity. The critical voltage remarkably increased with increasing AlO 2 content added in the electrolyte. (3) A peculiar phenomenon of high current density at around 5 V could be explained as trans-passive state. It was not observed for anodizing in acidic fluoride solutions such as Dow17 and ammonium fluoride. The peak current at 5 V depressed with increasing aluminate ion content added in the electrolyte. These phenomena are explicable by the effects of aluminum incorporation into the film to prevent dissolution and to promote passivation. The passivation effect of aluminum addition in the electrolytes is more remarkable than the addition in magnesium substrates. (4) The depth profiles of constituent elements showed that aluminum migrated into oxide film to reach near oxide/ substrate interface. Atomic ratio of aluminum to magnesium increased with increasing voltage to attain.42 at 8 V and crystalline MgAl 2 O 4 and MgO were found in the film. Acknowledgements Parts of this work were financially supported by grant-inaids from the Ministry of Education, Culture, Sports, Science and Technology, Japan and the Iketani Science and Technology Foundation. The authors are also grateful to Dr. K. Matsuzaka and Mr. R. Shimizu for their help. REFERENCES 1) E. F. Emely: Principles of Magnesium Technology, (Pergamon Press Ltd., New York, 1966) pp. 67 735. 2) J. H. Nordlien, S. Ono, N. Masuko and K. Nisancioglu: Corros. Sci. 39 (1997) 1397 1414. 3) J. H. Nordlien, K. Nisancioglu, S. Ono and N. Masuko: J. Electrochem. Soc. 143 (1996) 2564 2572. 4) K. Asami and S. Ono: J. Electrochem. Soc. 147 (2) 148 1413. 5) M. Takaya: J. Japan Inst. Light Metals 37 (1987) 581 586. 6) F. Sato, Y. Asakawa, T. Nakayama and H. Sato: J. Japan Inst. Light Metals 43 (1993) 65 71. 7) S. Ono, K. Asami, T. Osaka and N. Masuko: J. Electrochem. Soc. 143 (1996) L62 L63. 8) S. Ono, T. Osaka, K. Asami and N. Masuko: Corros. Rev. 16 (1998) 175 19. 9) O. Khaselev and J. Yahalom: J. Electrochem. Soc. 145 (1998) 19 193. 1) S. Ono and N. Masuko: Proc. of THERMEC 2 Las Vegas, Dec. 2:CDROM, Section A1, Vol. 117/3 Eds. T.Chandra et al., (Elsevier Science, UK, October 21). 11) S. Ono, M. Saito, M. Horiguchi, K. Terashima, K. Matsuzaka, A. Shida, T. Osaka and N. Masuko: J. Surf. Finish. Soc. Jpn. 47 (1996) 268 272. 12) S. Ono, K. Asami and N. Masuko: Mater. Trans. 42 (21) 1225 1231. 13) S. Ono, M. Saito, M. Horiguchi, K. Terashima, K. Matsuzaka, A. Shida, T. Osaka and N. Masuko: J. Surf. Finish. Soc. Jpn. 47 (1996) 263 267. 14) F. Keller, M. S. Hunter and D. L. Robinson: J. Electrochem. Soc. 1 (1953) 411 419. 15) M. Pourbaix: Atlas of Electrochemical Equilibria in Aqueous Solutions, (Pergamon Press, London, 1966). 16) S. Ono, N. Kijima and N. Masuko: J. Surf. Finish. Soc. Jpn. 53 (22) 46 412.