Characterization of soft magnetic spinel ferrite coating prepared by plasma spray

Surface & Coatings Technology xxx (2014) xxx–xxx

SCT-19749; No of Pages 11

Contents lists available at ScienceDirect

Surface & Coatings Technology

j ourna l homepage: www.e lsev ie r .com/ locate /sur fcoat

Characterization of soft magnetic spinel ferrite coating prepared byplasma spray

Yi Liu, Shicheng Wei ⁎, Haoliang Tian, Hui Tong, Binshi XuNational Key Laboratory for Remanufacturing, Academy of Armored Forces Engineering, Beijing 100072, China

⁎ Corresponding author. Tel.: +86 10 66718541; fax: +E-mail address: [email protected] (S. Wei).

http://dx.doi.org/10.1016/j.surfcoat.2014.09.0290257-8972/© 2014 Elsevier B.V. All rights reserved.

Please cite this article as: Y. Liu, et al., Surf. C

a b s t r a c t
a r t i c l e i n f o
Article history:Received 9 April 2014Accepted in revised form 12 September 2014Available online xxxx

Keywords:Plasma spraySpinel ferriteTribology

The primary objective of this work is to investigate themicrostructure and tribological performance of magneticnickel–zinc–magnesium ferrite coating prepared throughplasma spray. The phase composition,morphology andvalence state of the ferrite coating are studied by X-ray diffraction (XRD), scanning electron microscopy (SEM)and X-ray photoelectron spectroscopy (XPS). It is found that the ferrite coating shows a spinel structure. Theas-sprayed coating exhibits nanocrystalline with the grain size of 100–200 nm. According to the XPS analysis,the ferrite coating contains both Fe2+ and Fe3+, a fewNi ions are reduced tometallic Ni. Tribological performanceof the ferrite coating is tested in dry friction and 3.5 wt.% NaCl solution using a ball-on-disk test. The wear loss ofthe ferrite coating is lower than the substrate in both dry friction andNaCl solution under the samewear test con-ditions. The wearmechanism of the ferrite coating is dominated by delamination and brittle fracture. Microwaveabsorbing property of the ferrite shows a high value of magnetic loss factor in the frequency range of 1–12 GHz.The results suggest that plasma spray is a promising technology for the production of magnetic ferrite coatings.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The majority applications of thermal spray are in the field of protec-tive coatings and themain functions of the overlay coating are to protectthe underlying substrate fromheat, wear or corrosion [1]. Newopportu-nities are now emerging in advanced functional materials such as mag-netic materials, catalytic materials and bioactive materials. Thermalspray provides advantages in the efficiency for the preparation of suchmaterials. In recent years, thermal spray deposition of microwave ab-sorbing materials has attracted attention because of the relative higheradhesion strength and thinner thickness of the coating compared withthe traditional polymer composite method. Among the few examples,Lisjak et al. and Begard et al. focused on the atmosphere plasma sprayto prepare barium hexagonal ferrites as electromagnetic absorbers [2,3]. Other researchers also made correlation studies in this field. Bobzinet al. prepared SrFe12O19 coatings through atmospheric plasma sprayusing two kinds of feedstock powders (type A and type B). The feed-stock of type A consisted of spray-dried spherical agglomerates ofmicrometric SrFe12O19 particles while the feedstock of type B was syn-thesized by reactively sintered agglomerates of SrCO3 and Fe2O3. Theyfound that type B agglomerates exhibited magnetic properties, whichwere promising for electromagnetic wave absorption applications [4].Furthermore, Zhao et al. investigated themicrowave absorption proper-ties and the complex permittivity of Fe/FeAl2O4 coatings deposited byreactive plasma spray using Al/Fe2O3 powders. They found that the

86 10 66717144.

oat. Technol. (2014), http://d

real part and the imaginary part of the complex permittivity increasedwith Al concentration in the frequency range of 8.2–12.4 GHz [5].Most of these papers focused on the preparation procedure and givetrials for thermal spray in the application of microwave absorptiontechnology.

Nickel−zinc−magnesium (NZM) spinel ferrites are soft magneticmaterials and can be used for radar absorbing materials (RAMs) whichare mostly applied in the frequency range of 1–18 GHz [6–8]. TheNZM ferrites have various forms in products, such as sheet, paints,films and powders. Apart from these products, the ferrite can beblended with organic binders to produce microwave absorbing coat-ings. Recently much effort has been made to improve the performanceof RAMs and simplify the preparation procedure [9]. Plasma spray is apromising technique tomeet this need. If NZM ferrites can be fabricatedby plasma spray with proper magnetic properties, the preparation effi-ciency can be improved. However, the deposition of NZM ferrite coatingthrough plasma spray has not been reported so far.

RAMs usually serve in various environments. The RAMs particularlycoated on the ships or tanks is often partly scratched by the obstacle inthe environment. It can be attributed to the mechanical damage. In ad-dition, the RAMs also can be scratched by the flying glass when a bombwent off [10]. Surface damage generated by abrasive contact will limitthe life of RAMs and reduce their reliability. Therefore, tribology perfor-mance is also necessary for the RAMs. Unfortunately, the studies on thetribological behavior of RAMs are scarce up to now. The aim of thisworkis to prepare NZM ferrite coating through plasma spray. Small amountsof lanthanumare added to theNZM ferrite in order to improve themag-netic property and refine the crystal size [11–13]. Tribological tests are

x.doi.org/10.1016/j.surfcoat.2014.09.029

http://dx.doi.org/10.1016/j.surfcoat.2014.09.029

mailto:[email protected]


http://www.sciencedirect.com/science/journal/02578972

www.elsevier.com/locate/surfcoat


Table 1Plasma spray parameters of the coating.

Voltage(V)

Current(A)

Ar(m3/h)

H2

(m3/h)N2

(m3/h)Spray distance(mm)

Powder feeding rate(g/min)

120 380 3.2 0.4 0.6 100 30

2 Y. Liu et al. / Surface & Coatings Technology xxx (2014) xxx–xxx

conducted in dry friction and 3.5 wt.% NaCl solution considering thatequipments coated with RAMs sometimes serve in a marine environ-ment. The effect of environmental conditions on the tribological behav-ior is discussed.

2. Material and methods

2.1. Powder production

The feedstock powder consisted of spray-dried spherical agglomer-ates of micrometric spinel ferrite particles with a composition ofNi0.5Zn0.4Mg0.1La0.05Fe1.95O4. Stoichiometric amounts of MgO, ZnO,NiO, La2O3 and Fe2O3 as starting materials were mixed and homoge-nized in a ball mill. The particles were synthesized by a solid state reac-tion. These startingmaterials were sintered at 1300 °C for 2 h to achievereactive sintering, so as tomake the original reagents convert to a spinelphase. During this process, the size of reagents got bigger and

0 50 100 150 200 250 300 350 400 450 500 550-0.00050.00000.00050.00100.00150.00200.00250.00300.00350.00400.00450.00500.00550.00600.00650.00700.00750.00800.00850.0090

Substrate

0.0

2.0x10-6

4.0x10-6

6.0x10-6

8.0x10-6

1.0x10-5

1.2x10-5

1.4x10-5

1.6x10-5

1.8x10-5

(a)

(c)

Fig. 1. SEM image of the ferrite coating: (a) Cross-section morphology, (b) im

Please cite this article as: Y. Liu, et al., Surf. Coat. Technol. (2014), http://d

distributed inhomogeneously. In order to achieve fine products, thesintering products were broken down to powders again. The powderswere heat-treated at 900 °C for 1 h after the shattering process, whichis used to reduce the internal stress and defects. At last, the powderswere cooled down to room temperature naturally. However, thesepowders were not suitable for plasma spray due to their poorflowability. In order to prepare spherical agglomerates and improvetheir flowability, spray-drying process was selected to produce plasmaspray materials. The preparation procedure was as follows: the ferritepowders were used as rawmaterials and polyvinyl alcohol as the bind-er. The particles were agglomerated into spherical granules, using aspray-drier attached to a cyclone; the entry and exit temperatureswere 240 °C and 140 °C, respectively. The spray dried powders weresieved between 200 mesh and 400 mesh to separate the bigger andfiner ones, so that the size of agglomerates fell within the scope of plas-ma spray. The size distribution of the particles finally used in plasmaspray was in the range of 200 mesh to 400 mesh.

0 50 100 150 200 250 300 350 400 450 500 550-0.00050.00000.00050.00100.00150.00200.00250.00300.00350.00400.00450.00500.00550.00600.00650.00700.00750.0080

0.0

2.0x10-6

4.0x10-6

6.0x10-6

8.0x10-6

1.0x10-5

1.2x10-5

1.4x10-5

Coating

(d)

(b)

age of the interface, (c) CTE of the substrate and (d) CTE of the coating.



Fig. 2. TEM image of the ferrite coating.

20 30 40 50 60 70 80

2 Theta (degree)

∗

coating

(444)

(440)

(511)

(422)(400)

(220)

(311)

∗

∗

∗ LaFeO3

Intensity(a.u.)

powder

Fig. 3. XRD patterns of the ferrite samples, one phase is the spinel phase, and the secondphase is the LaFeO3.

3Y. Liu et al. / Surface & Coatings Technology xxx (2014) xxx–xxx

2.2. Coating deposition

Mild steel substrate was used in this study. The substrate wasground with abrasive papers up to 600 grit, cleaned with acetone, andthen sandblasted using brown alumina grit prior to plasma spray. Theferrite coatings were prepared through the high efficiency supersonicplasma spraying equipment developed by National Key Laboratory forRemanufacturing. The equipment consumed lower energy and gasflow rate [14]. The torch used a single anode Laval nozzle. The anode–nozzle internal diameter was 4.5 mm. Mechanical compression wasthe primary mechanism of the torch. It achieved a high reliability andheavy energy density [15]. The powders were injected through radialposition and the internal diameter of the injector was 3 mm. Detailedoperating parameters were listed in Table 1. The nitrogen was used asa powder carrier gas. The carrier gas flow rate was 0.6 m3/h. It was de-termined by the powder feeding rate and plasma jet power.

2.3. Characterization

The phase structures of the ferrite powder and coating were studiedby D8 advance X-ray diffraction (XRD) with λ= 0.154 nm Cu-Kα radi-ation. The cross sectional samples were prepared by metallographiccutting, hot-mounting in resin, grinding with abrasive papers (up to2000 mesh) and polishingwith a diamond slurry. The prepared samplewas etched by the Nital. The morphology patterns of powder and coat-ing samples were observed using a field emission scanning electronmi-croscopy (FEI, Navo NanoSEM 450) coupled with an energy dispersivespectra (EDS, OXFORD Feature Max). The microstructure of the ferritenanocrystalline was studied by a transmission electron microscope(TEM, JEOL JEM 2000FX). The valence states of elements in ferritecoating were analyzed by X-ray photoelectron spectroscopy (XPS, VGScientific ESCALAB 220i-XL, USA). The thermal expansion coefficientsof both substrate and ferrite coating were measured from 30 °C to500 °C using a NETSCH Dilatometer Model DIL 402C. The complexpermeability and complex permittivity were measured in the rangeof 1–18 GHz by an HP8722ES network analyzer. For this purpose,the ferrite powders were homogeneously dispersed into wax matrixand compacted into rings for the permeability and permittivity mea-surements. The size of the ring was 7 mm in outer diameter, 3 mm ininner diameter and 2 mm in thickness. The ferrite–wax compositescontained 80 wt.% of ferrite.

2.4. Tribological test procedure

Tribological tests were carried out in a reciprocating mode using aball-on-disk sliding wear testing apparatus (CETR UMT-3). The counterpartwas a Si3N4 ball. The diameter of the counterpartwas 4mmand theoverall test duration timewas 30min. The friction testswere carried outat room temperature with a frequency of 10 Hz, a normal load of 10 Nand a wear track length of 4 mm. The wet friction test was performedin 3.5 wt.% NaCl solution. The wear volume was determined using a3D non-contact surface mapping profiler (LEXT OLS4000, Olympus). Itshould be noted that the surface state was one of the key factors to beconsidered. The as-sprayed coatings exhibited a significantly differentsurface roughness. In order to decrease the friction coefficient, the sur-face roughness should be as low as possible [16]. Therefore, before tri-bological tests, the samples were ground using SiC papers (up to1000 mesh) and then polished with diamond slurries until the averagesurface roughness in the same order of magnitude.

3. Results and discussion

3.1. Morphological study

Fig. 1a shows the typical cross-section image of the ferrite coating. Itcan be seen that smooth and dense coatings are fabricated by plasma


spray. The coating is constituted of well flattened splats and partialmol-ten particles. The partial molten particles aremarked by the arrow. Dur-ing the process of plasma spray, a large amount of microparticles aremelted in the high-temperature flames (more than 104 K). These mol-ten droplets strike the substrate surface and form amicron sized lamel-lar structure under the action of high-speed flame. There is an obviousinterface between the coating and the substrate. The magnification




image of the interface is showed in Fig. 1b. The interface shows a looseconnection and is accompanied by someunmelted particles (Fig. 1b, cir-cle) and cracks (Fig. 1b, arrow). It is attributed to excessive shrinkagerates of the large number of molten particles deposited on the coldmetal substrate at a rapid cooling rate. As is known to all, a lamellarstructure is built up after molten particles impact, flatten, adhere, andsolidify on the target substrate [17]. The CTE of the substrate and the fer-rite coating was shown in Fig. 1c and d, respectively. It can be seen thatthe CTE of ferrite coating is much smaller than the substrate. Thus it re-sults in themismatch of thermal stress. Because of the high cooling rateand mismatch of thermal stress between the deposit coating and thesteel substrate, defects such as pores and microcracks easily developat the interface of the two different materials, which may degrade themechanical property [18]. Moreover the wetting property maybeinfluence by the poor connection between the coating and the substrateand inappropriate contact angles of the molten particles lead tothe porosity between the coating and substrate. Therefore, suitablewettability–temperature relation between the molten particles andthe cold substrate is required to produce good quality coatings [19]. Inthis paper, the bond layer is not used for the connection of the metalsubstrate and ceramic coating, whichmay also result in a loose connec-tion between the two heterogeneous material interfaces. Fig. 2 showsthe TEM image of the as-sprayed ferrite coating. A number of ferritegrain boundaries can be obviously seen. The crystals exhibit an equiaxialrather than a columnar shape because the sample is cut from the axialposition of coating surface. The observed grains of the ferrite coatingsare nanocrystalline (as marked by an arrow). The grain size is in the

740 730 720 710 70010000

15000

20000

25000

30000

35000

40000

Intensity(cps)

Binding Energy (eV)

1050 1045 1040 1035 1030 1025 1020 1015

48000

50000

52000

54000

56000

58000

60000

62000

64000

Intensity(cps)

Binding Energy (eV)

Zn

(a)

(c)

Fe 2p1/2

FeFe 2p3/2

Fig. 4.XPS spectra of the ferrite coating: (a) Fe, (b) Ni, (c) Zn and (d)Mg. (For interpretation of t


range of 100–200 nm. The white parts in the TEM picture are also theferrite grains. They show different contrasts with the black grains. Theformation of the nanocrystal can be considered as the result of the fastsolidification process. Plasma spray is a rapid process; the residencetime of the ferrite powders inside the plasma jet is less than 10−3 s.Therefore, there is little time for themolten particles to grow, which re-sults in the formation of nano-crystals [20]. In addition, micro-struc-tures may also be present in the coating.

3.2. XRD analysis

Fig. 3 shows the XRD patterns of sintered ferrite powder and the fer-rite coating. The phase analysis shows that the sintered powder samplesare well crystalline. The sharp and strong diffraction peak around (311)plane provides a clear evidence for the formation of the cubic spinelstructure of the ferrite as reported everywhere [21]. The several intensepeaks correspond to planes of (220), (311), (400), (422), (511) and(440). Furthermore, the diffraction peaks of the ferrite coating becomebroad as compared with the powder. There are two mechanisms forthe XRD peak broadening. Firstly, the background noise and broadnessof the peaks are characteristic of particles with nanometer dimensions.It happens in the nano-sized particles because there are insufficient dif-fraction centers that cause the line broadening [22]. This is confirmed bythe TEM image. Secondly, residual stress and crystal defects in thermalspray coatings also have an effect on XRD peak broadening and signal/background ratio, particularly when only a proportion of particles aremelted [23,24]. In addition, the XRD peaks of the ferrite coating are

875 870 865 860 855 850 845

40000

42000

44000

46000

48000

50000

Intensity(cps)

Binding Energy (eV)

1308 1306 1304 1302 1300 1298 129650000

52000

54000

56000

58000

60000

62000

64000

66000

68000

Intensity(cps)

Binding Energy (eV)

Mg

(b)

(d)

NiNi 2p3/2

Ni 2p1/2

he references to color in this figure, the reader is referred to theweb version of this article.)



0

2

4

6

8

10

12

14

16

Volumeloss(10-2 mm3 )

Test sample

substrate in dry frictionsubstrate in solutioncoating in dry frictioncoating in solution

Fig. 6. Variations in the volume loss of the samples.


clearly shifted to lower angles compared with the ferrite powder. It isdue to zinc loss during the plasma spray process. Yan et al. conductedextensive fundamental studies to improve the understanding of thezinc loss on different sizes of splats. They determined XRD spectrafrom different sizes of splats and also found Bragg peaks exhibit a shiftto lower 2θwith increased size of splats [25]. Moreover, a little amountof the secondary phase (LaFeO3) is observed in both the sintered ferriteand the coating. The secondary phase indicates that lanthanumdoes notform a solid solutionwith spinel ferrite or it has a limited solid solubilityin the lattice. The ionic radii of metal ions depend on their coordinationnumber. The ionic radii of the studied ions such as Ni2+ (0.69 Å), Zn2+

(0.74Å), Fe2+ (0.74Å) and Fe3+ (0.64 Å) are similar. However, the ionicradius of La3+ (1.06 Å) is much larger than the other ions. Lanthanumhas a very low solubility in spinel lattice due to its bigger ionic radii.The amount of La3+ ions is limited when they are used to substitutefor Fe3+ ions in spinel lattice, as the substitution of La exceeds thesolid solubility, the redundant La3+ ions aggregate on the grain bound-aries and produce the secondary phase. During structure refinement, itis found that the occupancy of the octahedral site tends to increase forLa-doped compositions, which allows few of La3+ ions to enter intothe lattice [26]. Lanthanum is a non-magnetic element; the substitutionof La can modify the magnetic properties of the ferrite. On the otherhand, the redundant La forms a secondary phase that segregates ingrain boundaries, which prevents the grain growth and refines the crys-tal size. In many cases on thermal sprayed ferrites, the breaking of theferrite structure is reported, which usually happens in the hexagonalferrite materials [27]. The required time for the crystallization processof hexagonal ferrite which involves various peritectic reactions is toolong compared to the very short solidification process. A few hexaferritephases can be retained in the coatings [4]. In this research, spinel ferriteinstead of hexagonal ferrite is selected. The spinel phase is a quick for-mation process and it is the first phase developed from the liquid [28].Therefore the spinel phase is retained and no structure breaking of thecoating is observed.

3.3. XPS analysis

Fig. 4 represents XPS core level spectra of Fe 2p, Mg 1s, Ni 2p and Zn2p of the ferrite coating sample. The Fe 2p binding energy is shown inFig. 4a. It is reported that Fe 2p photoelectron peaks of oxidized ironare associated with satellite peaks. The analysis of Fe 2p3/2 photoelec-tron peak is complicated due to a high temperature spray process andion distribution in the ferrite system. Therefore, the satellite peaks ofFe3+ and Fe2+ are important to distinguish their chemical states. It is lo-calized on 6 eV and 8 eV for FeO and Fe2O3 above their basic

0 300 600 900 1200 1500 18000.3

0.4

0.5

0.6

0.7

0.8

0.9

substratedry3.5% NaCl

Time/s

Frictioncoefficient

(a) (b

Fig. 5. Friction coefficients of the samp


photoelectron peak [29]. According to the previous reports [30], the Fe2p3/2 binding energy of FeO and Fe2O3 is in the range of 709.3–709.7 eV and 710.4–711.2 eV, respectively. As shown in the figure, theblack line is the measured spectrum. The as-synthesized ferrite coatingsample shows a main peak at 710.1 eV. A satellite peak is observed at716.2 eV, which is about 6 eV above the main peak. Therefore, it canbe concluded that Fe2+ ions exist in the ferrite coating. However, thebinding energy of the main peak locates between FeO and Fe2O3. Thus,the ferrite coating contains both its trivalent state and divalent state.The deconvoluted Fe 2p peak is shown in the color solid lines. The redline stands for the totalfitting chemical states of iron. The photoelectronpeak at 709.3 eV in the pink line stands for Fe2+ in FeO while the pho-toelectron peak at 711.2 eV in the green line stands for Fe3+ in Fe2O3.The blue line stands for the Fe2+ satellite spectrum. Besides, a peak lo-cates in the range of 720–725 eV, which corresponds to Fe 2p1/2 photo-electron peak.

The XPS core level spectrum of Ni is shown in Fig. 4b. The black solidline is themeasured curve. The red line is the fitting results of all the dif-ferent chemical states of Ni. The binding energy of Ni 2p3/2 is associatedwith three peaks, which correspond to the Ni3+ in Ni2O3, Ni2+ in NiOand metallic Ni, respectively. The peak value at 852.4 eV in the greenline corresponds to the metallic Ni while the peak value at 854.7 eV inthe pink solid line corresponds to the Ni2+ in NiO [31]. In addition, thepeak value at 855.9 eV in the blue line corresponds to the Ni3+ inNi2O3. Another peak is observed at the binding energy of 872.3 eV,

0 300 600 900 1200 1500 1800

0.5

0.6

0.7

0.8

0.9

1.0

coatingdry3.5% NaCl

Frictioncoefficient

Time/s

)

les: (a) substrate and (b) coating.




which is typical of Ni 2p1/2 in NiO. According to the handbook of XPS[32], the peak of Ni2+ in NiO is observed at 853.8 eV. However, thesame peak appears at 854.7 eV in the ferrite coating. The binding energyof photoelectron peak not only depends on the chemical state, but alsorelies on its molecular environment. The increase in the binding energyof Ni2+ may be due to the cation distribution in the spinel structure.Nickel prefers the octahedral sites and it occupies the octahedral siteduring the ferrite growth. Moreover, there is an interesting phenome-non in the XPS spectrum. The metallic Ni is discovered, which indicatesthat as the ferrite powder is sprayed on the substrate during plasmaspray, some Ni ions are reduced to metallic Ni. It may be related to thereductive atmosphere of the Ar/H2 plasma and the carrier gas(nitrogen). This phenomenon is also observed in other reports. Tao pre-paredNiFe2O4 throughmolten salt synthesis in KCl flux using Fe2O3 andNiO. The powders containing 1.2 wt.% polyvinyl alcohols were sinteredin air and nitrogen atmospheres. Metallic nickel was detected in thesamples that sintered in nitrogen atmosphere [33]. Furthermore, Yanet al. also discovered a reductive process for metal ions. The oxidationstate of Fe changed from Fe3+ to Fe2+ during the spray process [25].

Fig. 4c shows the XPS core level spectra of Zn 2p. Zn2+ ions preferthe tetrahedral site in the spinel structure. According to literature [34],there is some difference for Zn2+ in the nanocrystalline materials. IfZn2+ ions occupy octahedral sites, the binding energy of Zn 2p peakwill be slightly higher. The binding energy of Zn 2p3/2 in the tetrahedralsite is seen at 1021.4 eV while the binding energy of Zn 2p3/2 in the oc-tahedral site is seen at 1023.2 eV. In this work, Zn 2p3/2 peak is observedat 1021.5 eV, which indicates that Zn2+ ions occupy the tetrahedral sitein the coating. Another peak at 1044.7 eV stands for Zn 2p1/2.

Mg2+ ions usually prefer the octahedral sites in oxygen lattice. Thus,in ferrites, when cations distribute in two different sites, two XPS peaksor a peakbroadening of the photoelectron peak is expected.However, inthis research, there is only one peak for Mg 1s. The binding energy is1303.9 eV (Fig. 4d), which corresponds to the Mg2+. The result showsthat Mg2+ ions distribute in the octahedral site.

3.4. Wear behavior of the samples

The friction coefficient is an important parameter to evaluate thetribology property [35]. Fig. 5 shows the variations of friction coeffi-cient of both the substrate and the ferrite coating with a frequency of10 Hz and normal load of 10 N. Initially, the friction coefficient of the

Fig. 7. Worn surface of the substrate: (a


substrate (Fig. 5a) increases rapidly until it reaches a peak value, andthen a gradual decrease is followed to a steady state value. Thesteady friction coefficient is about 0.82 in dry friction. However, thefriction coefficient of the substrate was lower in 3.5 wt.% NaCl solu-tion than that in dry friction and the steady value is 0.75.With regardto the ferrite coating (Fig. 5b), friction coefficient suddenly increasesfrom 0.5 to the peak value in a very short time and then shows agradual decrease to the end of the test. The friction coefficient ofthe ferrite coating finally exhibits a steady value of 0.88 in dry condi-tion and 0.76 in solution, respectively. The decrease of friction coef-ficient is due to the lubrication action of the solution. The evolutionof friction coefficient exhibits two stages in the tribology test. Thefirst stage is the running-in period. There is a rise in the friction coef-ficient plot, which represents the formation of wear debris. The nextstage is the steady wear state. A steady and lower friction coefficientis measured. The lower friction coefficient can be attributed to thatdebris is either trapped in the wear track and roll under the indentor pushed out of the ball path [36]. The volume wear loss of the sub-strate and the coating in dry friction and NaCl solution is showed inFig. 6. The volume loss of the substrate is larger than the ferrite coat-ing in both dry friction and the solution under the same condition. Asthe substrate sliding in NaCl solution, the volume wear loss is largerthan in dry friction. The steel substrate exhibits poor corrosion resis-tance in NaCl solution. Accordingly, the corrosive wear aggravatesthe volume wear loss of the substrate. However, the result of the fer-rite coating is contrary to the substrate. The volume loss of the coat-ing in solution is lower than in dry friction. Both the ferrite coatingand the counterpart belong to ceramic materials, which have bettercorrosion resistance in NaCl solution. The addition of solution playsa role in a lubricant, which reduces the volume loss of the coating.

Fig. 7a shows the worn surface of the substrate in dry friction. Hugematerial accumulation and apparent plow scratches can be foundaround thewear track. It is accepted that during theunlubricated slidingwear process, oxidation reaction occurs first in the real contact area dueto the synthetic effect of contact stress and frictional heat. Fine debrisgenerates and detaches from the oxidation regions under repeated cy-clic stress [37]. It is seen that the worn surface of the substrate showssigns of abrasive wear and adhesion wear when sliding against a Si3N4

ball. However, the worn surface of the substrate in 3.5 wt.% NaCl solu-tion exhibits a severe damage. A large number of apparent abrasivegrooves and pits (generated by the delamination of the materials) are

) dry friction and (b) wet friction.




identified on the worn surface (Fig. 7b). In summary, the wear scar ofthe substrate immersing in 3.5 wt.% NaCl solution is associated withabrasion, generation of cracks, material and debris removal. The mor-phology analysis of wear scar is in accordance with the fluctuations offriction coefficient. Furthermore, the transparent solution subsequentlychanges into a red turbid one, which suggests that corrosion plays animportant role in the wear test. The wear mechanisms are abrasivewear and corrosive wear. The corrosion reactions of the substrate in3.5 wt.% NaCl solution include the following steps [38]:

0:5O2 þH2Oþ 2e→2OH− ð1Þ

Feþ OH− þ Cl−→FeClOH− þ e ð2Þ

FeClOH−→FeOHþ þ Cl− þ e ð3Þ

FeOHþ þ Hþ→Fe2þ þ H2O ð4Þ

Fig. 8.Worn surface of the ferrite coating: (a) polished coating surface before the tribology test, (and (d) EDS analysis of the selected area.


Fe2þ þ 2OH−→Fe OHð Þ2 ð5Þ

Fe OHð Þ2 þ 0:5O2 þH2Oþ e→Fe OHð Þ3 þ OH−: ð6Þ

As the steel dissolves, the corrosion products play the role of abra-sive particles. These particles are taken away from the worn surface bysolution. The fresh substrate is continually exposed to the corrosivemedia, which speeds up the corrosion process. Compared with dryfriction, the substrate is severely damaged and shows large wearscratches. It is notable that the lower friction coefficient is not equal tolower wear volume loss in the wet tribology test. The number of corro-sion pits and cracks continually develops due to corrosive wear. Theworn surface of the sample gets rough under corrosive attack, which re-duces the energy of abrasive wear and accelerates the wear damage.

In order to understand better the coating damage induced by the tri-bology test, a polished surface of the plasma spray coating before the tri-bology test is shown in Fig. 8a. It can be seen that the coating surface is

b)worn surface of the coating in dry friction, (c)worn surface of the coating inwet friction




porous. The worn surface of the ferrite coating under a normal loadof 10 N in dry friction is shown in Fig. 8b. Extensive brittle fractureoccurs, leading to splat detachment on the worn surface in dry fric-tion. The coating material is pressed and distorted under the normalload. This process increases the coating stress, decreases the adhe-sive strength, aggravates the coating deformation and produces alarge number of debris. The worn surface of the ferrite coating inwet friction is shown in Fig. 8c. SEM micrograph exhibits a largeamount of pits. Small chip can be found on the surface. The wearmechanism is brittle spalling. According to the previous research[39], friction can induce the initiation of cracks between two splatsof the same lamella. The cracks propagate through the splat bound-ary under cyclic load, the solution quickly fills in the cracks, whichaccelerates the crack growth rate and results in the final exfoliationon the worn surface. During the friction test in NaCl solution, hydro-dynamic effects such as enhanced cooling of the surface and flushingof the debris also play a role of reducing the friction coefficient andvolume wear loss. The thicker corrosion product and the bigger pits

Fig. 9. SEM images of wear debris: (a) substrate in dry friction, (b) s


filled with liquid, the better lubrication and less contact area can beprovided [40]. EDS of the scar (Fig. 8d) shows that the worn surfaceconsists of Ni, Zn, Mg, La, Fe, O, Si, Na and Cl.

The wear debris is accumulated after the tribology test. Fig. 9ashows the SEM image of substrate debris in the dry friction test.The debris is primarily in the form of flakes, which suggests it is con-tinually flattened and deformed by the counterpart. In order to re-veal the distinct variations of chemical composition, the weardebris is investigated by EDS. The debris is composed of O, Si andFe. The mole fraction of oxygen is 57.4 at.%, which indicates thatthe debris is heavily oxidized. However, the morphology of the sub-strate debris in NaCl solution is quite different from that in dry fric-tion (see Fig. 9b). The debris is surrounded by the NaCl particles.The composition is Fe, Cl, Si, Na and O. The silicon comes from thecounterpart (Si3N4 ball). Fig. 9(c) shows thewear debris of the ferritecoating in dry friction. The debris also exhibits a flaky structure.Composition analysis shows that it is composed of Ni, Si, O and Fe.The Fe, O and Si are the major composition. Among these elements,

ubstrate in NaCl solution and (c) ferrite coating in dry friction.




Fe and O come from the ferrite coating while Si comes from thecounterpart. It has to be noted that the oxygen can also be a resultfrom the oxidation during the sliding wear. Unfortunately, the debrisof the ferrite coating in NaCl solution is too few that it is hard to col-lect accurately.

The worn surface of the counterpart (Si3N4 ball) is also investi-gated. Fig. 10a shows the counter surface of the substrate in dryfriction. A large number of apparent parallel plows exhibit in thepicture. According to EDS analysis, the worn surface is composedof Si and Fe. The presence of Fe suggests the adhesion of the sub-strate. The SEM image of the counterpart against the substrate inNaCl solution is shown in Fig. 10b. It can be seen that the surfacesof a Si3N4 ball show obvious grooves and adhesion slice. EDS anal-ysis reveals that the worn surface is composed of Si, N, O and Fe.The presence of Fe and O also confirms the adhesion regime onthe worn surface. Interestingly, the worn surface of the counterparthas a higher content of Fe when it slides against the substrate inNaCl solution. This result corresponds to severe adhesion of thesubstrate in wet friction. The detection of Fe on the counterpart

Fig. 10. Worn surface of the counterpart ball sliding against: (a) substrate in dry friction, (b) su


suggests a continuous tribo-film layer on the silicon nitride ball.The worn surface of the counterpart ball against ferrite coating indry friction and solution is shown in Fig. 10c and d. Apart fromsome shallow plows on the surface of the counterpart ball in dryfriction, the wear track of the counterpart shows a smooth surface.Little debris is present on the worn surface. The debris most likelycomes from the ferrite coating.

3.5. Microwave absorbing property

Fig. 11(a) and (b) displays a frequency dependence of the complexpermeability and complex permittivity of the ferrite, respectively. Itcan be seen the real part of the permeability (μ′) decreases rapidlywith increasing frequency until it reach the minimum value at the fre-quency of 7.8 GHz. Then it increases with increasing frequency. Theimaginary part of permeability (μ″) decreases with the increasing fre-quency below 13 GHz and then it is almost constant. As shown inFig. 11(b), both the real part (ε′) and the imaginary part (ε″) of permit-tivity are found almost constant within the entire frequency range. The

bstrate in NaCl solution, (c) ferrite coating in dry friction and (d) ferrite in NaCl solution.



2 4 6 8 10 12 14 16 18

0.0

0.3

0.6

0.9

1.2

1.5

1.8

2.1

2.4

μ'

μ''

μ'μ''

μ

Frequency/GHz2 4 6 8 10 12 14 16 18

-2

-1

0

1

2

3

4

5

6

7ε'

ε''

Frequency/GHz

ε

ε'ε''

2 4 6 8 10 12 14 16 18

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

tanμ

tanε

Electromagneticlossfactor

Frequency/GHz

(a) (b)

(c)

Fig. 11. Frequency dependence of the ferrite powders: (a) complex permeability, (b) complex permittivity and (c) electromagnetic loss factors.


value of ε′ is in the range of 5.5–5.9 and the value of ε″ is about 0.05–0.25. It is well known that microwave absorbing property can be de-scribed by electromagnetic loss factor ( tan δ). It is calculated fromthe formulas [41]:

tan δμ ¼ μ ″μ 0 ð7Þ

tan δε ¼ε″ε0

: ð8Þ

The electromagnetic loss factor is illustrated in Fig. 11c. As can beseen, the magnetic loss factor shows a very high value in the frequencyrange of 1–12 GHz and the peak value is about 1.21 at the frequency of5.25 GHz.Moreover themagnetic loss factor ismuch higher than the di-electric loss factor when the frequency is below 13 GHz; however theyare almost same in the frequency of 13–18GHz. The dielectric loss factoris very low in the investigated frequency. The peak value of tan δε is just0.04. The result indicates thatmagnetic loss plays a dominant role in theabsorbing mechanism.

4. Conclusions

NZM spinel ferrite coating is successfully synthesized by plasmaspray using the spray dried ferrite powders. The coating exhibits acubic spinel structure with a secondary phase of LaFeO3. The diffractionpeaks of the ferrite coating become broader as compared with the pow-ders. Cross section image shows an obvious interface between the


coating and the substrate. TEM morphology analysis indicates that theferrite coating is composed of nanoparticles. The average size of thesenanoparticles is in the range of 100–200 nm. During plasma spray, a re-ductive process is discovered for Ni ions. Some Ni ions are reduced tometallic Ni due to the reductive atmosphere. Tribology behavior of thesubstrate and the ferrite coating sliding against Si3N4 counterpart ballis investigated in dry friction and NaCl solution. The friction coefficientof both the substrate and coating samples is lower in solution. The sub-strate suffers adhesion wear and abrasive wear. The ferrite coatingshows a better wear resistance than the substrate. The wear process ofthe coating is dominated by delamination and brittle fracture. Micro-wave absorbing property of the ferrite shows a high value of magneticloss factor in the frequency range of 1–12 GHz and the peak value isabout 1.21 at the frequency of 5.25 GHz.

Acknowledgments

This paper is financially supported by the National Natural ScienceFoundation of China (No. 51222510) and 973 Project (2011CB013403).

References

[1] Sanjay Sampath, J. Therm. Spray Technol. 19 (2010) 921–949.[2] D. Lisjak, K. Bobzin, K. Richardt, Marion Bégard, Giovanni Bolelli, Luca Lusvarghi,

et al., J. Eur. Ceram. Soc. 29 (2009) 2333–2341.[3] M. Begard, K. Bobzin, G. Bolelli, A. Hujanen, P. Lintunen, D. Lisjak, et al., Surf. Coat.

Technol. 203 (2009) 3312–3319.[4] K. Bobzin, G. Bolelli, M. Bruehl, A. Hujanen, P. Lintunen, D. Lisjak, et al., J. Eur. Ceram.

Soc. 31 (2011) 1439–1449.


http://refhub.elsevier.com/S0257-8972(14)00838-X/rf0005









[5] D. Zhao, F. Luo, W.C. Zhou, D.M. Zhu, Surf. Coat. Technol. 205 (2011) 4254–4259.[6] C.H. Peng, C.C. Hwang, J. Wan, J.S. Tsai, S.Y. Chen, Mater. Sci. Eng. B 117 (2005)

27–36.[7] A.R. Bueno, M.L. Gregori, M.C.S. Nóbrega, J. Magn. Magn. Mater. 320 (2008)

864–870.[8] N. Singh, A. Agarwal, S. Sanghi, Paramjeet Singh, Physica B 406 (2011) 687–692.[9] H.L. Fan, W. Yang, Z.M. Chao, Compos. Sci. Technol. 67 (2007) 3472–3479.

[10] L. Yang, Y.F. Wang, Y.S. Zhang, Y.Z. Lv, New Technol. New Process 1 (2009) 97–100(in Chinese).

[11] Ankush Thakur, R.R. Singh, P.B. Barman, J. Magn. Magn. Mater. 326 (2013) 35–40.[12] N. Chen, K. Yang, M.Y. Gu, J. Alloys Compd. 490 (2010) 609–612.[13] L.W. Deng, L. Ding, K.S. Zhou, S.X. Huang, Z.W. Hu, B.C. Yang, J. Magn. Magn. Mater.

323 (2011) 1895–1898.[14] B.S. Xu, H.J. Wang, S. Zhu, X.B. Liang, Manuf. Technol. Mach. Tool 2 (2003) 30–33 (in

Chinese).[15] H.J. Wang, B.S. Xu, P. Zhang, Z.P. Wang, C.S. Zhai, The spray gun of ultrasonic plasma

spray, Patent, 01101077.0 (China).[16] G. Darut, H. Ageorges, A. Denoirjean, P. Fauchais, Surf. Coat. Technol. 217 (2013)

172–180.[17] S.Y. Tao, Z.J. Yin, X.M. Zhou, C.X. Ding, Tribol. Int. 43 (2010) 69–75.[18] H.Y. Zhu, Y.R. Niu, C.C. Lin, L.P. Huang, H. Ji, X.B. Zheng, Ceram. Int. 39 (2013)

101–110.[19] R. Comesaña, F. Quintero, F. Lusquiños, M.J. Pascual, M. Boutinguiza, A. Durán, et al.,

Acta Biomater. 6 (2010) 953–961.[20] Y. Zhu, M. Huang, J. Huang, et al., J. Therm. Spray Technol. 8 (1999) 219–222.[21] T.H. Ting, R.P. Yu, Y.N. Jau, Mater. Chem. Phys. 126 (2011) 364–368.[22] M.A. Gabal, Abdullah M. Asiri, Y.M. AlAngari, Ceram. Int. 37 (2011) 2625–2630.


[23] T.C. Totemeler, R.N. Wright, W.D. Swank, Metall. Mater. Trans. A 34 (2003)2223–2231.

[24] X.B. Liu, B. Zhang, J. Mater. Sci. 37 (2002) 3229–3239.[25] Q. Yan, R.J. Gambino, S. Sampath, L.H. Lewis, L. Li, E. Baumberger, et al., Acta Mater.

52 (2004) 3347–3353.[26] P.K. Roy, J. Bera, Mater. Res. Bull. 42 (2007) 77–83.[27] K. Bobzin, T. Schlaefer, M. Begard, et al., Surf. Coat. Technol. 205 (2010) 1015–1020.[28] N. Langhof, D. Seifert, M. Göbbels, J. Töpfer, J. Solid State Chem. 182 (2009)

2409–2416.[29] V.K. Mittal, P. Chandramohan, Santanu Bera, M.P. Srinivasan, S. Velmurugan, S.V.

Narasimhan, Solid State Commun. 137 (2006) 6–10.[30] Marina V. Bukhtiyarova, Aleksandra S. Lvanova, Elena M. Slavinskaya, Lyudmila M.

Plyasova, Vladimir A. Rogov, Vasily V. Kaichev, et al., Fuel 90 (2011) 1245–1256.[31] V.V. Kaichev, A.Yu. Gladky, I.P. Prosvirin, A.A. Saraev, M. Hävecker, A. Knop-Gericke,

et al., Surf. Sci. 609 (2013) 113–118.[32] John F. Moulder, F. StickleWilliam, Peter E. Sobol, D. Bomben Kenneth, Handbook of

X-ray Photoelectron Spectroscopy, Perkin-Elmer cooperation, Eden Prairie, 1992.[33] Y.Q. Tao, Z.Y. Li, K.C. Zhou, Ceram. Int. 39 (2013) 865–869.[34] L. Zhang, Y. Wang, Q.Q. Ni, Mater. Chem. Phys. 124 (2010) 1029–1033.[35] J.B. Cheng, X.B. Liang, Z.H. Wang, B.S. Xu, Tribol. Int. 60 (2013) 140–146.[36] P. Henty, M.-J. Pac, C. Rousselot, M.-H. Tuilier, Surf. Coat. Technol. 223 (2013) 79–86.[37] C. Guo, J.S. Zhou, J.M. Chen, J.R. Zhao, Y.J. Yu, H.D. Zhou, Wear 270 (2011) 492–498.[38] Q. Bao, D. Zhang, D.D. Lv, P. Wang, Corros. Sci. 65 (2012) 405–413.[39] P.P. Psyllaki, M. Jeandin, D.I. Pantelis, Mater. Lett. 47 (2001) 77–82.[40] H.B. Lee, D.S. Wuu, C.Y. Lee, C.S. Lin, Tribol. Int. 44 (2011) 1603–1609.[41] X.G. Huang, J. Zhang, L.X. Wang, et al., J. Alloys Compd. 540 (2012) 137–140.






















































Documents

Characterization of soft magnetic spinel ferrite coating prepared by plasma spray