Electrochimica Acta 112 (2013) 111–119
Contents lists available at ScienceDirect
Electrochimica
Acta
j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /e l e c t a c t
a
An investigation of ceramic coating growth mechanisms in plasma electrolytic oxidation (PEO)processing
R.O.Husin ∗,X.Nie,D.O.Northwood
Department of Mechanical,Automotive and Materials Engineering,University of Windsor,Windsor,ON,Canada N9B 3P4
a r t i c l e
i n f o
Article history:
Received 7May 2013
Received in revid form 20August 2013Accepted 22August 2013
益母草胶囊Available online 6 September 2013
Keywords:
Plasma electrolytic oxidation (PEO)AJ62Mg alloy野外灭绝
Coating growth mechanisms Oxygen diffusion
Optical emission spectroscopy
a b s t r a c t
To further our understanding of the plasma electrolytic oxidation (PEO)process,and to aid in the opti-
mization of the process,it is important to identify the mechanisms of coating formation.In the prent work,coatings up to 110m thick were produced on an AJ62Mg-alloy substrate using the PEO process.Optical emission spectroscopy (OES)was employed to follow the microdischarges and substrate and elec-trolyte elements prent in the plasma discharge during the coating growth,and to determine plasma electron temperatures.During PEO processing of magnesium,some of the metal cations are transferred outwards from the substrate and react with anions to form ceramic coatings.Also,due to the high elec-tric field in the discharge channels,oxygen anions transfer toward the magnesium substrate and react with Mg 2+cations to form a ceramic coating.In PEO process,the ceramic coating grows inwards to the alloy substrate and outwards to the coating surface simultaneously.The total coating thickness variation compared with the geometrical dimensions of the uncoated and coated samples were investigated.For the coating growth,there are three simultaneous process taking place,namely the electrochemical reactions,the plasma chemical reactions and thermal diffusion.Oxygen diffusion occurring during PEO processing is discusd in terms of coating growth mechanisms.
© 2013 Elvier Ltd. All rights rerved.
1.Introduction
PEO is a novel surface engineering technology,which is con-sidered to be one of the most cost-effective and environmentally friendly ways to improve the corrosion and wear resistance of magnesium and magnesium alloys [1–5].A driving force for the developments is the avoidance of expensive equipment required for competing vacuum-bad plasma technologies [5].Besides the process technology,the growth mechanisms of the PEO coating,and the coating’s microstructure and properties,are gaining more and more attention [6–9].The oxide layers obtained by PEO processing of magnesium alloys are typically thick,hard and well-adherent to the substrate.The coatings produced on magnesium alloys are compod of two sub-layers [10],a porous outer layer with poor mechanical characteristics (especially in terms of wear)[11]and an internal layer between the porous layer and the substrate,that has very good mechanical properties.This layer is usually described as the “compact layer”.
The formation mechanisms of the coating layer by PEO are complex due to the involvement of electro-,thermal-,and plasma-chemical reactions in the electrolyte [12,13].Over the last decade,significant progress has been made in both the understanding and
∗Corresponding author.Tel.:+15192528908.
E-mail address:husinr@uwindsor.ca (R.O.Husin).
utilization of the plasma in order to obtain coatings with desired properties.The effects of the process parameters on the characteris-tics of the PEO coatings on Mg and Mg-alloys have been investigated in a number of different studies.Magnesium alloy AJ62has recently been developed as a structural automotive power train material,namely,the transmission ca and the engine block.This alloy is characterized not only by a high strength-to-weight ratio synony-mous with magnesium alloys but also by relatively good properties at elevated temperatures [2].Some of the publications deal with the influence of the electrolyte,which is bad on an alkaline solution with additions of silicates,phosphates or sodium alumi-nate [14,15].Other studies have examined the effect of current mode and current density [16],the microdischarge phenomena [17],the chemical composition of the substrate materials [18],the microstructure and the properties of the PEO coatings [5],and many authors have concentrated on a study of the coating mechanisms [7,13].During plasma electrolytic oxidation,complex physical and chemical process occurs near the interface between electrolyte and the electrode.Due to the extreme nonlinearity of plasma discharge,monitoring the evolution of spectral signals and microstructure is helpful in understanding the mechanism underlying PEO process.Current density is one of the most important parameter during the PEO process [2].
The variation in geometrical dimensions during PEO processing is an important issue for understand
ing the coating growth mech-anisms.The main objective of this work is to develop a correlation
0013-4686/$–e front matter © 2013 Elvier Ltd. All rights rerved./10.1016/j.electacta.2013.08.137
112R.O.Husin et al./Electrochimica Acta112 (2013) 111–119 between the oxide ceramic layer formation process and the
growth rate during PEO coating.To do this,we examine the produc-
tion and diffusion of oxygen and its effect on the coating formation.
2.Experimental procedures
2.1.Materials and PEO processing method
A PEO coating system as described in Husin et al.[19]was
ud to produce the oxide coating on the samples.The PEO coat-
ing process was carried out using a bipolar current mode which
consists of two components,a positive component and a negative
component.A den,minimum defect coating can be achieved by
adjusting the positive to negative current ratio and their timing,
thereby eliminating or reducing the strongest plasma discharges.
Two different current generators were ud,which deliver cur-
rent to the substrate with an amplitude in the range of0–5A and
0–15A.The power supplies were specially designed to allow
independent control over the main pul parameters,such as pul
自我封闭
duration,amplitude and duty cycle,during both positive and neg-
ative biasing using a Spik2000A controller.During the coating
process,the positive I+and negative I−current density was main-
tained at0.07and0.06A/cm2respectively and the voltage was
incread gradually with time,as the coating thickness incread.
The duration of each pul(T+on and T−on,the period of positive and
negative pul respectively)and the resting gap(break)between
the positive and negative puls(T+
off and T−
off
,respectively)are
listed in Table1.The values of T+on,T−on,T+
off and T−
off
in Table1
correspond to a duration time of80%duty cycle.The charge ratio parameter,C R,which is the ratio of positive to the negative charge quantity[2],is given in Table1.Both pul duration and current density were chon bad on our previous work on AJ62Mg alloy [2],to get the most compact coatings.
AJ62magnesium alloy(MgA l6Mn0.3Sr2)disc coupons(˚25×5mm)were ud as the test samples in this study.They were prepared from the same cast ingot in order to minimize the differ-ences resulting from variations in composition and microstructure. The coupons were manually ground and polished on240,400,600, and1200grit silicon carbide(SiC)waterproof abrasive papers.The coating was obtained in an alkaline electrolyte containing sodium aluminate(10g/l Na2Al2O4)and potassium hydroxide(1g/l KOH). The temperature of the electrolyte was kept below25◦C by a water cooling system.The PEO processing parameters for the coated Mg samples are listed in Table1.
2.2.Optical emission spectroscopy
The main characterization of the micro-discharges was per-formed by means of optical emission spectroscopy(OES).Light emission of the discharges was collected using a spectrometer with four channel slots,each of which covers a certain wavelength region (effective range200nm to900nm).Th
e light emitted by the plasma was transmitted and focud to the entrance slit of the spectrom-eter.Since the discharges occur randomly,an integrated signal was ud which was collected from the total sample surface fac-ing thefiber optic.The emission intensity of the plasma species were monitored as a function of time using the OES system.For the spectroscopic analysis of the emitted radiation,the tables listed by Sansonetti and Martin[20]were ud to identify the atomic and ionic spectral lines.The relative intensities of spectral lines of the same atomic species can be ud to calculate the plasma elec-tron temperature[21].Four different spectral lines were recorded simultaneously,which eliminates discrepancies that may other-wi happen if the spectra are recorded at different times.The
A
n
o
d
i
c
a
n
d
c
a
t
h
o
d
i
c
v
o
l
t
a
g
e
(
V
)
Trea tment time (mi n)
Fig.1.Plots of anodic and cathodic atment time during the PEO pro-cess using bipolar current mode.
spectral lines(Table2)at285.2nm(Mg I),486.1nm(Hˇ),656.2nm (H˛),and777.2nm(O I)were recorded.
2.3.Coating characterization
Scanning electron microscopy(FEI Quanta200FEG with solid state backscattered detector operated at10kV)in the condary electron(SE)mode was ud to obrve both the coating surface morpholog
y and,through obrvation of sample cross ctions, coating thickness and integrity.The samples were cut to be about 4-mm-thick ctions normal to the tangential–radial surface and mounted with resin and polished to a mirrorfinish then sputtered with a goldfilm to make them conductive before SEM analysis. The coating thickness for different treatment times were deter-mined using a PosiTector6000coating thickness meter with N type probes:This instrument us the eddy current principle to measure the thickness of non-conductive coating(ceramic)on non-ferrous metal(magnesium).The dimensions of the magnesium alloy sam-ple before and after oxidation were measured using a Mitutoyo Absolute ID-S112spiral“absolute position”digital micrometer, from which the inner and outer thickness values at different treat-ment time were calculated.The meter was t to zero-position for the uncoated substrate then the total outer thickness measured after the coating,which reprents twice the thickness of the sam-ple dimension changes.
3.Results
3.1.Voltage behavior
Fig.1shows a typical output anodic(V A)and cathodic(−V C) voltage change during the120min of PEO treatment.By combin-ing the output voltage results with the OES emission intensities and plas
ma temperatures,Husin et al.[19]identified four discharge stages in the PEO process,namely:Stage I:In the early stage of the process which mainly involves the rapid electrochemical for-mation of an initial insulating oxidefilm,a sharp increa in the voltage was en.In this stage the breakdown voltage is not yet reached.Stage II:The rate of the voltage change decreas in this stage,which is characterized by numerous sparks moving rapidly over the whole sample surface area.This indicates a start of the breakdown of the oxide layer,an increa in temperature and, therefore,melting of the substrate metal.Stage III:In this stage the rate of voltage increa becomes slow;this stage is characterized by
R.O.Husin et al./Electrochimica Acta 112 (2013) 111–119
113
Table 1
PEO Process parameters for coating AJ62Mg alloy.
Sample
Current mode
Time (min)
I +(A)
I −(A)
T +on (s)
T −
on (s)
T +
off (s)
T −
off (s)
C R S1–S9
Bipolar
3–120
0.7
0.6
400
100
400
100
0.77
Table 2
Spectral lines ud in this experiment together with their wavelength ( ),transition,statistical weight of
the upper and lower state g k and g i (respectively),photon energy ( E )and the transition probabilities (A ki )[20].
Line
(nm)
Transition
g k
g i E (eV)
A ki ×108S −1
Mg I 285.23s3p 1P →3s 21S 31 4.33 5.0H ˇ486.14d 2D →2p 2P 42 2.550.172H ˛656.23d 2D →2p 2P 42 1.890.539O I
777.2
3p 5P →3s 5S
7
5
1.59
0.369
larger but slower moving discharges.As the oxide layer grows,its electrical resistance increas,therefore the nature of the plasma changes.Stage IV:In this stage the rate of voltage variation is even slower than that in stage III and concentrated discharges appear as relatively large and long lasting sparks.However,the occurrence of the strong discharges is less frequent than that in stage III due to the thicker coating causing more difficulty in the initiation of such discharges.For some cas,such strong discharges may cau irreversible damage to the coatings in stage IV.
3.2.Optical emission characterization
As the plasma coating process proceeds,the discharge appear-ance changes and the plasma emission intensities vary as shown in Fig.2.Fig.2shows the optical emission intensity profile of the Mg line (285.2nm)for a bipolar current mode for a total treatment time of 120min:This illustrates the ti
me evolution of a substrate element prent in the plasma discharge during the coating growth.As the PEO process proceeded,relatively strong variations in the micro-discharges were obrved,indicated by many parated spikes in Fig.2.The spikes correspond to the relatively strong discharges (B-type discharge)[2,19],which are initiated from the magnesium surface-coating interface.The spike height increas as the process proceeds,particularly after 40min processing time.Optical emission spectra of oxygen line at 777.2nm spectral line were ud to follow the time evolution and the behavior of the oxygen in the plasma discharge during the coating growth.The intensity spectrum of O I for three different treatment times 30,50and 120min,respectively are shown in Fig.3(a–c).For the 30min treatment time relatively low O I intensity signals are obrved as
20
40
60
哈勃定律
8010
120
1000
2000300040005000600070008000
900010000I n t e n s i t y (a .u )
Treatm ent ti me (min )
Mg I 2 85.2 n m
Fig.2.Typical time variation of the emission line intensity of Mg I during the PEO
process.
shown in Fig.3(a).The oxygen signal started to be noticeable for the period 40to 50min,e Fig.3(b).As the coating process pro-ceeds,the oxygen (O I)intensity signal generally incread with the treatment time (Fig.3(c));this is likely due to the incread amount of oxygen evolution as the process preceded,as has been previously obrved for aluminum by Snizhko et al.[6].By com-paring the optical emission intensity of oxygen Fig.3(c)with Fig.2for Mg,the evolution of oxygen is related to the intensity of the microdischarges,where after 40min processing time,both the O and Mg signals in
crea due to the existence of B-type discharges.However,the O signal densities are larger than that of Mg due to its relatively low excitation energy (1.59eV)compared to that of Mg (4.33eV)[20].
3.3.Plasma electron temperature
The intensity ratio of the recorded 656.2nm (H ˛)to 486.1nm
(H ˇ)spectra I H (3d 2D →2p 2P)/I H (4d 2D →2p 2P)(from the same ion-ization stage)were ud to determine plasma electron temperature (T e )which is prented in Fig.4.The averaged temperatures are also shown in Fig.4.The results are in good agreement with previous Te results for Mg AJ62alloy [2].
It is commonly accepted that the discharge in PEO occurs when the applied voltage reaches a certain critical value corresponding to the breakdown of the oxide layer (or at least of the barrier part of it)formed on the sample surface:This leads to the development of inten light emission generated at the numerous micro-discharge sites.Fig.4shows plasma temperature is initially around 5500K which corresponds to the early stage discharges where the density of the discharges is very high (accumulation of the individual dis-charge temperatures).Te then drops to about 4500K after about 10min and then incread to about 5000K at 25min after that drops to around 4500K at 40min.Then
the average electron tem-peratures curve started to gradually increa to reach up to 5000K.There were a high number of cloly-spaced temperature spikes some of them reaching up to 5800K.The spikes corresponded to relatively strong discharges initiated from the sample surface-coating interface,the so-called B-type discharges [22].
3.4.Microstructure of the coatings
Fig.5(a–f)prent SEM micrographs of the surface mor-phologies of coatings produced using different processing times.Projections and microporosity with different sizes and shapes were obrved on all the coated samples.Curly projections were found to be dominant on the surfaces of coatings.The surfaces of the coatings were dominated by a ‘pancake’shaped projections with open or aled microporosity in the center.Short processing 3min (Fig.5(a)),show the highest density of open (unaled)chan-nels at the centers of the ‘pancake’structures.The microporosity
114R.O.Husin et al./Electrochimica Acta 112 (2013) 111–119
10
20
30
40
1600
1700
1800
1900
2000
2100
I n t e n s i t y (a . u .)
Treatm ent time (min )
AJ62- OI 77 7.2 nm 30 mi
n
10
20
30色盲怎么治疗
40
50
60
1600
170
1800
190
2000
210
杨子莹I n t e n s i t y (a .u )
Treatment
time
(min)
AJ62- OI 77 7.2
nm 50 min
20
朗诵社团活动计划
40
60
8010
12
1500
兀立是什么意思2000
2500
3000
I n t e n s i t y (a . u .)
Treatme nt ti me (min)
AJ62- OI 777.2 nm
120 min
Fig.3.Typical time variation of the emission line intensity of O I during the PEO process at three different treatment times.
is considered to be “footprints”of the plasma discharge channels,
through which the Mg and Mg 2+from the substrate were likely ejected and reached the coating/electrolyte interface during the plasma-generated melting.The Mg and Mg 2+then reacted with O 2generated by electrolysis,and finally sintered and deposited on the coating surface,thus contributing to coating growth.
Fig.
6(a–f)prent SEM micrographs of cross-ctions of coated samples using different treatment times.The dashed-line on the micrograph indicates the approximate position of the original sur-face of the magnesium alloy specimens before PEO treatment.All coating-substrate interfaces had a wavy-jagged appearance,which may be the result of dissolution of the substrate in the early stages of processing [2]and/or the prence of inter-metallic phas at
T e (K )
Treatme nt time (mi n)
Fig.4.Plasma temperature as a function of treatment time (min)determined from
the intensity ratio of H ˛(656.3nm)/H ˇ(486.1nm).
the grain boundaries.The ␣-Mg grain boundaries in the AJ62alloy
are often decorated with the (Al,Mg)4Sr and Al 3Mg 13Sr inter-metallics [2].During the total PEO processing time,the coating is compod of two distinct layers,namely,an outer layer with a significant amount of connected porosity,cracks and other struc-tural defects and a more compact inner layer.The distributions of porosity and other defects were inhomogeneous in both layers but was more evident in the outer layer.However,the relative pro-portions of the two layers change with PEO processing time.It should also be emphasized again that the PEO processing param-eters were chon to produce a compact and adherent coating.During plasma discharges,process including melting,melt-flow and re-solidification continuously occur in the outer layer thus causing a fluctuating repetitive increa and decrea in surface temperature which leads to a porous structure.
The inner layer was den and adhered well to the substrate and exhibits excel-lent mechanical properties.A thin and very den coating/substrate interface layer is clearly shown for coatings with processing times longer than 15min.The oxide coating on the sample treated for 3min,Fig.6(a),was 5to 8m thick and was almost completely compod of a porous outer layer.This is consistent with the surface morphology,Fig.5(a),which showed a high density of open (un-aled)channels at the center of the ‘pancake’structures,which could extend to the coating substrate interface.As the processing time increas,the coating thickness increas.The thickness of the oxide layer was in the range of 70–90m and 100–115m for the samples treated for 90and 120min,respectively.
3.5.PEO coating thickness
The variation of the average total coating thickness,L T ,obtained by two different techniques,SEM cross ction and Eddy current,with PEO processing time is shown in Fig.7.There is good agree-ment between the two methods and that the total thickness of
R.O.Husin et al./Electrochimica Acta112 (2013) 111–119
115
Fig.5.SEM micrographs showing the surface morphology of oxide coatings on AJ62for different treat
ment times.
coatings increas linearly with processing time.The slope of the L T vs.time plot,which is determined by the positive current density[9],gives an average growth rate of0.9±0.1m/min:A coating with a thickness of110±5m is obtained after120min coating.The total thickness(L T)is compod of an outer L o and an inner L i layer,where L o is the coating thickness above the original surface of the sample before oxidation,and,L i is the coating thickness below the original surface toward the magnesium alloy substrate.The variations of L o and L i with PEO processing time are shown in Fig.8(a).In the initial stages up to about45min,the coating growth toward the coating surface,L o,is larger than that toward the magnesium alloy substrate,L i.After about50min,
the Fig.6.SEM micrographs of cross-ctions of coatings on AJ62for different treatment times.
