JOURNAL OF RARE EARTHS, Vol. 27, No. 6, Dec. 2009, p. 1003

Foundation item: Project supported by the Government of Pakistan-IICS

Corresponding author: A. Ali (E-mail: amjadali130@gmail.com; Tel.: +92-300-5106983)

DOI: 10.1016/S1002-0721(08)60357-9

Multilayer ceramic coating for impeding corrosion of sintered NdFeB magnets A. Ali1, A. Ahmad1, K. M. Deen2

(1. Department of Metallurgical & Materials Engineering University of Engineering & Technology, Lahore, 54890 Pakistan; 2. Department of Metallurgy and Materials Engineering CEET, University of the Punjab, Lahore, 54590 Pakistan)

Received 23 April 2009; revised 3 August 2009

Abstract: Sintered NdFeB magnets have complex microstructure that makes them susceptible to corrosion in active environments. The cur-rent paper evaluated the anticorrosion characteristics of multilayer titanium nitride ceramic coating applied through cathodic arc physical va-pour deposition (CAPVD) for protection of sintered NdFeB permanent magnets. The performance of ceramic coating was compared to the electrodeposited nickel coating having a copper interlayer. Electrochemical impedance spectroscopy (EIS) and cyclic polarization in simulated marine environment were employed to determine the rates of coatings degradation and passivation behaviour respectively. The coating mor-phologies and surface chemistry were studied with scanning electron microscope (SEM). X-ray diffraction (XRD) was used for identification of component phases in the coatings and the substrate. The results showed that the polarization resistance of ceramic coating increased with the exposure time. The rate of degradation of Rp for the ceramic coating had an extraordinary negative slope followed by a stable duration, before declining towards the coating failure. In comparison the nickel coating with copper interlayer degraded sharply. The vapour deposited ceramic coating was found to have permeable defects that tended to “re-passivate” during exposure providing prolonged corrosion protection to the NdFeB substrate. The magnetic properties were unaffected and remained at par with the nickel coating having copper interlayer.

Keywords: sintered NdFeB magnets; multilayer ceramic coating; permeable defects; rare earths

The NdFeB permanent magnets have become important technological materials due to the exceptionally advanta-geous magnetic properties since their discovery in 1983[1–3]. At ambient temperature, neo magnets have highest energy product and thus known for the efficient utilization of elec-trical energy. Over the years, neo magnets have occupied a leading position among the strong permanent magnetic ma-terials for variety of engineering applications, namely com-puter peripherals, automation, automobile, aerodynamic, magnetic resonance, biomedical, acoustics and consumer electronics[4,5]. However, the presence of electrochemically active phases in the microstructure of the sintered NdFeB magnets[6] deteriorate their efficiency as they corrode during exposure to various environments[7,8].

Efforts have been made to control the corrosion of sin-tered neo magnets either by alloying additions to alter the electrochemical potential of active microstructural phas-es[9–12] or by tailoring the surface to incorporate mechanical barrier coatings like epoxy, metallic or alloy coatings[13–15]. Improving corrosion resistance by alloying additions did not serve the purpose as only a few alloying elements margin-ally improved the corrosion resistance at a substantial com-promise of magnetic properties[16–18]. The surface treatments

such as electrodeposited nickel or zinc, electroless nickel, electroless nickel-phosphide or nickel- cobalt-phosphide, hot dip zinc, aluminizing, electrophoresis, chromate passivated aluminium coating are known corrosion-proofs for neo magnets but each has its own limitations. So, new develop-ments are always on cards. Currently, the cathodic arc physical vapour deposited transition metal nitride coating has been reported to cater the corrosion protection of sin-tered neo magnets[19]. In CAPVD process the plasma as-sisted high energy cathodic arc ejects the metal vapours from the solid metal at ambient temperature in an evacuated chamber. The applied potential gradient accelerates the me-tallic ions in the vapour flux towards the substrate with ki-netic energies in the range of 10–100 eV. The metallic ions encounter the methane or nitrogen gas introduced in the chamber, before their impact to the substrate surface. The vapour deposited transition metal nitride coatings normally have microscopic defects such as pores, pin holes and/or voids[20,21] that provide path for the environmental species to access the substrate surface. Such permeable defects can not be completely eliminated but their size and density can be reduced by depositing thicker coatings[22,23], by incorporat-ing noble and dense interlayer[24], by interrupting columnar

1004 JOURNAL OF RARE EARTHS, Vol. 27, No. 6, Dec. 2009

growth with multilayered coatings[25] or by post deposition sealing of defects with polymeric deposits[26].

The present work aims at depositing the multilayered tita-nium nitride coating by CAPVD process to impede the cor-rosion of sintered neo magnets. It also aims at studying the degradation mechanism of the coating//substrate system during exposure to simulated marine environment by elec-trochemical impedance spectroscopy (EIS). The properties of the titanium-nitride//neo-magnet system were compared to Ni-Cu-Ni//neo-magnet system.

1 Experimental

1.1 Materials

Sintered NdFeB magnets, commercially produced through powder metallurgical route were used as substrate material in this study. The neo magnets had the hydrostatic density of 7.58 g/cm3, while their theoretical density is 7.60 g/cm3. The chemical composition of the magnets (Table 1), was determined through wet analysis since low atomic number elements like boron having low electronic transition energies can not be detected by commercially available EDX analyser.

Table 1 Chemical composition of sintered NdFeB permanent magnets

Elements B Nd Fe

w/wt.% 1.26 ±0.15 32.90 ±0.33 Balance

x/at.% 8.0 15.0 77.0

1.2 Specimen preparation

Disc shaped neo magnets with 12 mm diameter and 2.5 mm thickness were polished metallographically to 1.0 m alumina suspension and ultrasonically cleaned in ultrasonic soap solution at ambient temperature. To reveal the micro-structure the specimens were etched in 2% Nital.

1.3 Coating set up for CAPVD

The prepared substrates were mounted on a rotary sample holder at an angle of 30° with the normal to the cathode. The coating set up had a disc shaped cathode (titanium target) mounted on water cooled copper stage. The anode was also a disc, fixed at a perpendicular distance of 300 mm on top of the cathode. A double walled stainless steel jacket was used to enclose the whole assembly. The hollow cathode argon plasma discharge was used to clean the chamber after evacuation. Ground copper wire was used to trigger the arc. Automatic microprocessor controlled feeding system was used to introduce nitrogen into the chamber with controlled

partial pressure. Keeping other parameters constant as re-ported earlier[19], three coating cycles of 40 min each were run to deposit the multilayer ceramic coating in an attempt to deposit a thick and dense coating with low density of permeable defects.

1.4 Electrodeposition of nickel with copper interlayer

Electrodeposition of nickel and copper was carried out with the Watt’s solution and copper sulphate bath respec-tively. The bath composition and coating parameters have already been reported[19]. A nickel strike layer was deposited followed by the copper interlayer and again nickel layer above all. The coated surfaces were buffed before interme-diate shifting between the baths and after final layer of nickel.

1.5 Characterization

After gently cleaning with muslin cloth the coated speci-mens were fixed in the specimen holder for exposure to aer-ated aqueous solution containing 3.5% NaCl. The neo sub-strate was made working electrode where as graphite rod was used as counter electrode. The Ag/AgCl electrode was used as a reference. The EIS measurements were carried out at ambient temperature from high frequency to low fre-quency i.e. from 105 to 10–2 Hz. The amplitude of ac poten-tial was 10 mV. The EG&G Potentiostat 273A was em-ployed for DC cyclic polarization measurements. The range of potential scan was –250 to +250 mV with respect to open circuit potential (OCP). The potential scan rate was 5 mV/s. Gamry Potentiostat and Echem Analyst software was used to measure and plot the AC impedance of the coat-ing//substrate systems. The qualitative phase analyses were carried out with Siemens D-500 X-ray diffractometer using Fe filtered Co K radiations and Origin-5 graphic software. Jeol scanning electron microscope (SEM) equipped with EDS analyzer was used to study the back scattered electron images. The magnetic properties were measured with the help of Riken Denshi B-H curve tracer. The experimental measurements were repeated three times to verify the re-producibility and consistency of results.

2 Results and discussion

The chemical composition of the NdFeB magnet is given in Table 1. The comparison of phase components and sec-tioned views of coating//substrate systems is shown in Fig. 1.

The diffraction pattern of the substrate had the signatures of Nd2Fe14B tetragonal phase while the CAPVD coating had a single reflection at 42.34º Bragg’s angle corresponding to Ti2N tetragonal phase. The titanium nitride ceramic coating

A. Ali et al., Multilayer ceramic coating for impeding corrosion of sintered NdFeB magnets 1005

Fig. 1 Comparison of diffraction patterns and SEM back scattered electron images of sectioned specimens of sintered NdFeB substrate (a), multilayer Ti2N ceramic coated neo magnet (b) and NiCuNi coated neo magnet (c)

had a mean thickness of 1.6 m which was not sufficient to completely block the reflections from the substrate. On the other hand, the mean thickness of electrodeposited nickel- copper-nickel coating was 17 m and the diffraction pattern showed two phases with no reflection from the substrate. Obviously the major phase was nickel and the Cu3.8Ni was expected at the interface between the nickel and copper lay-ers. The multilayered structure of ceramic coating is visible in Fig. 2 which also shows some defects in the coating structure namely pin holes, craters, single layered or re-moved top layer regions. The pin holes or voids are perme-able defects in the multilayered structure that straight away provide environmental access to the substrate surface. The craters are the region that have very thin layer of ceramic

coating and during exposure to the environment these de-fects degrade much earlier than the multilayered regions. Similarly the regions that are not multilayered will also de-grade at a faster rate than the multilayered regions during exposure to the active environment.

The coatings degradation was measured with EIS at sys-tematic exposure intervals in simulated marine environment, that is, aerated 3.5% NaCl solution at ambient temperature. The neo substrate was made working electrode whereas graphite rod was used as counter electrode. The results of the coating’s impedances are shown in Fig. 3 to 5 as Nyquist plots. The impedance of the multilayered ceramic coating increased with time till 48 h exposure, see Fig. 3.

This is an extraordinary result, because ordinarily the coat-ing either maintains its resistance for some time and then start degrading or it starts degrading straight away as exposed to the active environment. It happened with the nickel-cop-per-nickel coating that started degrading straight away and collapsed to failure within 24 h of exposure, see Fig. 4.

The multilayered structure however, started degrading somewhere between 48 to 72 h of exposure and collapsed to complete failure after 84 h exposure, see Fig. 5.

Fig. 6 shows the graph of Rp or the polarization resistance as a function of exposure time for the two coatings. The Rp of NiCuNi coating at 10 MHz was 4.28 kohm·cm2 just after 1 h of exposure that decreased sharply to 1.61 kohm·cm2

during 24 h of exposure with complete coating failure, while the Rp of ceramic coating increased from 1.68 kohm·cm2 to 3.18 kohm·cm2 for the same duration of exposure. The Rp of ceramic coating almost remained constant till 48 h and after

Fig. 2 SEM back scattered electron images showing sectioned views for multilayer Ti2N ceramic coating on sintered neo magnet (a), re-

moved top layer (b), crater with thin layer of ceramic coating (c) and pin hole in the ceramic coating (d)

1006 JOURNAL OF RARE EARTHS, Vol. 27, No. 6, Dec. 2009

that started decreasing till complete failure between 84 to 90 h of exposure. It means that the rate of degradation of ceramic coating is at least three times less than that of nickel-copper- nickel coating. In other words, the corrosion resistance of the CAPVD multilayered ceramic coating for sintered NdFeB magnets is three times better than the electrodepos-ited nickel-copper-nickel coating. Importantly, the magnetic properties of sintered NdFeB magnets remained unaffected by the CAPVD ceramic coating see Table 2.

Fig. 3 Nyquist plot showing ac impedance arcs at systematic expo-

sure intervals for the multilayer ceramic coating

Fig. 4 Nyquist plot showing ac impedance curves at systematic ex-

posure intervals for the nickel-copper-nickel coating

Fig. 5 Nyquist plot showing ac impedance arcs for post 48 h expo-

sure of the multilayer ceramic coating

This exceptional behaviour could be related to the forma-tion and growth of a passive film or re-passivation of pin holes/voids in the vapour deposited coating. To support the evidence and understand the mechanism involved, the ce-ramic coating was subjected to cyclic polarization test. The CP results are shown in Fig. 7, the anode appeared to pas-sivate initially during forward potential scan, but pitting started around –738 mV (EPit) and the current density in-creased to 70 mA/cm2 with a slight increase in the potential.

The reversed potential scan showed re-passivation of pits or pin holes around –778 mV. The narrow hysteresis loop for the forward and reversed potential scans and EProt above the ECorr indicates that the vapour deposited coating has the tendency to “re-passivate”. So this re-passivation of coating defects appears to be the reason behind the exceptional negative rate of degradation of the vapour deposited ceramic coating. Corrosion initiates through the pin holes and craters that tend to re-passivate during exposure and thus improve the coating impedance with exposure time. It might also be

Fig. 6 Comparison for the rate of change of polarization resistances

of two types of coatings on sintered neo magnets

Table 2 Magnetic properties of coated sintered NdFeB perma-nent magnets

Coating BBr/T iHC/T BHmax/(kJ/m3)

Ti2N 1.39 1.52 380

Ni-Cu-Ni 1.3 1.53 337

Fig. 7 Cyclic polarization of sintered NdFeB magnet with Ti2N ce-

ramic coating in aerated 3.5% NaCl aqueous solution

A. Ali et al., Multilayer ceramic coating for impeding corrosion of sintered NdFeB magnets 1007

due to the clogging of voids/pin holes by the corrosion products. Based on these results, it is assumed that a post deposition sealing treatment or a metallic strike layer is likely to improve the corrosion resistance of CAPVD ce-ramic coated sintered NdFeB magnets. The results of post deposition treatment or a metallic strike layer shall be pub-lished separately.

3 Conclusions

The CAPVD multilayered ceramic coating for sintered NdFeB magnets significantly enhanced the corrosion resis-tance without affecting the magnetic properties. The reduced rate of degradation of polarization resistance compared to nickel-copper-nickel ensures longer and better exploitation of the magnetic energy.

A post deposition sealing treatment or a metallic strike layer or an undercoat is being proposed for improving the corrosion characteristics of the multilayered ce-ramic-coating//neo-substrate system.

Acknowledgements: The authors would like to thank Mr. Azmat Hussain, Mr. Irfan Ahmad and Mr. Khawar Shoaib for their assis-tance during coating and characterization of magnets.

References:

[1] Hadjipanayis G C, Hazelton R C, Lawless K R. Cobalt-free permanent magnet materials based on iron-rare-earth alloys (invited). J. Appl. Phys., 1984, 55(6): 2073.

[2] Croat J J, Herbst J F, Lee R W, Pinkerton F E. Pr-Fe and Nd-Fe-based materials: A new class of high performance permanent magnets (invited). J. Appl. Phys., 1984, 55(6): 2079.

[3] Sagawa M, Fujimura S, Togawa N, Yamamoro H, Matsura Y. New materials for permanent magnets based on Nd and Fe (Invited). J. Appl. Phys., 1984, 55(6): 2083.

[4] Mitchell P. Corrosion Protection of NdFeB magnets. IEEE Trans. Magn., 1990, 26: 1933.

[5] Szymura S, Bala H, Rabinovich Yu M, Sergeev V V, Paw-lowska G. Structure, magnetic properties and corrosion be-haviour of sintered Nd16Fe76-xCrxB8 magnets. J. Magn. Magn. Mater., 1991, 94: 113.

[6] Williams C J, Narasimhan K S V L. Corrosion characteristics of RE-Fe-B permanent magnets. J. Appl. Phys., 1987, 61: 3766.

[7] Mummert K, Aziz A M El, Barkleit G, Rodawald W, Shultz L. Corrosion behaviour of NdFeB permanent magnets. Mater. Corr., 2000, 51: 13.

[8] Gurappa I. Corrosion characteristics of permanent magnets in acidic environments. J. Alloys Compd., 2003, 360: 236.

[9] Sagawa M, Tenaud P, Vial F, Hiraga K. High coercivity Nd-Fe-B sintered magnet containing vanadium with new mi-

crostructure. IEEE Trans. Magn., 1990, 26: 1957. [10] Rozendall E. Coercivity dependency of sintered NdFeB mag-

nets on Al content and the alloying route. IEEE Trans. Magn., 1990, 26: 2631.

[11] Tokuhara K, Hirosawa S. Corrosion resistance of NdFeB sin-tered magnets. J. Appl. Phys., 1991, 69: 5521.

[12] Hirosawa S, Mino S, Tomizawa H. Improved corrosion resis-tance and magnetic properties of Nd-Fe-B type sintered mag-nets with Mo and Co. J. Appl. Phys., 1991, 70: 5844.

[13] Yu Shengxue, Ling C. Preparation technology and perform-ances of Zn-Cr coating on sintered NdFeB permanent magnets.J. Rare Earths, 2006, 24: 223.

[14] Tamborim S M, Azamboja D S, A M Sliba Silva, Costa I. Corrosion protection of NdFeB magnets by phosphating with tungsten incorporation. J. Surf. Coat. Technol., 2006, 200: 6826.

[15] Zhang H, Song Y W, Song Z L. Electrodeposited nickel/alu-mina composite coating on NdFeB permanent magnets. J. Mater. Corr., 2008, 59: 324.

[16] Sagawa M, Hirosawa S, Tokunaga T, Yamamoto H, Fujimura S, Tsubukawa Y, Shimizu R. Dependence of coercivity on the anisotropy field in the Nd2Fe14B type sintered magnets. J. Appl. Phys., 1987, 61: 3559.

[17] Szymura S, Bala H, Pawlowska G, Rabinovich Yu M, Sergeev V V, Pokrovski D V J. Modification of the magnetic proper-ties and corrosion resistance of NdFeB permanent magnets with addition of cobalt. J. Less-Common Met., 1991, 175: 185.

[18] Femengel W, Rodewald W, Blank R, Schrey P, Katter M, Wall B. The influence of Co on the corrosion resistance of sin-tered Nd-Fe-B magnets. J. Magn. Magn. Mater., 1999, 196: 288.

[19] A Ali, A Ahmad. Corrosion protection of sintered NdFeB magnets by CAPVD Ti2N coating. J. Mater. Corr., 2009, 60(5): 372.

[20] Morita R, Azuma K, Inoue S, Miyano R, Takikawa H, Koba-yashi A, Fujiwara E, Uchida H, Yatsuzuka M. Corrosion re-sistance of TiN coatings produced by various dry processes.Surf. Coat. Technol., 2001, 136: 207.

[21] Ahn S H, Lee J H, Kim J G, Han J G. Localized corrosion mechanisms of the multilayered coatings related to coating defects. Surf. Coat. Techol., 2004, 638: 177.

[22] Uchida H, Inoue S, Koterazawa K. Electrochemical evaluation of pin hole defects in TiN films prepared by r.f. reactive sput-tering. Mater. Sci. Eng. A, 1997, 649: 234.

[23] Okamoto T, Fushima M, Takizawa K. Corr. Eng., 1996, 45: 425.

[24] Creus J, Idrissi H, Mazille H, Sanchett F, Jacquot P. Im-provement of the corrosion resistance of CrN coated steel by an interlayer. Surf. Coat. Technol., 1998, 107: 183.

[25] Liu C, Bi Q, Ziegele H, Leyland A, Mattews A. Structure and corrosion properties of PVD Cr-N coatings. J. Vac. Sci. Tecnol. A, 2002, 20: 772.

[26] Yang Q, Cai F, Zhao L R, Huang X. Improving corrosion re-sistance of CrTiAlN coating by post deposition treatments.Surf. Coat. Technol., 2008, 203: 606.