The effect of silica NPs incorporation on protective properties of oxide layers formed by PEO on Mg₉₇Y₂Zn₁ alloy with LPSO-phase

Abstract

The inherited chemical inhomogeneity in oxide layers obtained by plasma electrolytic oxidation (PEO) on the magnesium alloy Mg₉₇Y₂Zn₁ is associated to long-period stacking-ordered (LPSO) phase present in the treated alloy. This heterogeneity results in decrease of corrosion resistance and adhesion strength. The problem was solved by adding silica nanoparticles (NPs) into the electrolyte under PEO. According to the model developed, NPs which are harder than the oxide layer and being electrically charged, can be accelerated by an electric field and penetrate deep into the layer. The near-surface incorporation of NPs results in branching of the local breakdowns of vapor-gas bubbles that leads to an increase of the volumes of oxide layer and improvement of its properties.

Highlights

• •PEO coating was formed on Mg-LPSO alloy under the adding SiO₂ NPs into electrolyte.
• •SiO₂ NPs demonstrated inert mechanism of incorporation into the coating during PEO.
• •Incorporation of SiO₂ NPs increased adhesion strength and corrosion resistance.

1. Introduction

Outstanding progress was achieved recently in the engineering of magnesium alloys with a long-period stacking-ordered phase (LPSO-phase): the ultimate tensile strength and ductility of Mg-LPSO alloys were increased to the values ∼600–800 MPa [1,2] and ∼15–20 % and higher [2,3], respectively. However, the LPSO-phase led to the electric potential differences of the order of 60–250 mV between the α-Mg and the LPSO-phase [4,5]. This provokes the occurrence of microgalvanic corrosion in Mg-LPSO alloys and necessitates their surface protection, which can be solved by deposition of different coatings [6].

Plasma electrolytic oxidation (PEO) makes it possible to produce the wear- and corrosion-resistant oxide layers on the surface, coatings with high photocatalytic activity and polymer-ceramic coatings, and solves a wide range of tasks to ensure the required surface properties [[7], [8], [9], [10], [11], [12], [13], [14], [15], [16]]. However, PEO of Mg-LPSO alloys has been used very rarely because of poor results under conventional regimes. Increase of PEO frequency up to >10 kHz reduced the porosity and corrosion current density about one order of magnitude or more compared to PEO at frequency 500 Hz or less on Mg-Gd-Y-Zr alloy [17]. The Ce-sealing after the PEO (sealing by CeO₂ release from the chemical solution) of Mg–Y–Zn alloy with LPSO-phase gave good corrosion protection; the lowest porosity (<2 %) and the highest anticorrosive properties of PEO layer were achieved [5].

It is well known that the use of various substances in the form of particles as a dispersion phase of the electrolyte under PEO is one of the most effective ways to improve the protective and functional properties of oxide coatings, as well as to increase the productivity and efficiency of the PEO technology [[18], [19], [20], [21], [22], [23], [24]]. The using of SiO₂, ZrO₂, TiN, Si₃N₄, TiO₂, TiC, MoS₂ particles, carbon nanotubes or halloysite nanotubes as additions to the electrolyte during PEO has led to significant improvement of wear- and corrosion resistance of formed oxide layers [21,[25], [26], [27], [28], [29], [30]]. Silicon dioxide in a nanoscale form is one of the most proven substances as an additive to the electrolyte in the PEO of aluminum, magnesium and titanium alloys. It provides a significant increase in the wear resistance and corrosion resistance of oxide layers [21,25,[31], [32], [33], [34], [35]], is biocompatible [[36], [37], [38]] and, at the same time, relatively inexpensive. In recent years modifying PEO by adding various ceramic particles to the electrolyte has been widely studied for Mg-alloys such as Mg–Al–Zn, Mg–Mn–Ce and others [18,21,24,39,40]. However, as far as we know, the effects of using micro- or nanoparticles as dispersed phase of the electrolyte in PEO of Mg-LPSO alloys have not yet been investigated.

The behavior of nanoparticles (NPs) captured by vapor-gas bubbles (VGBs) prior the microarc discharge and their interaction with the surface of the oxide layer depend on various factors [28,41]. Most ceramic NPs are capable of adsorbing negative ions and electrons on their surface [[42], [43], [44]]. Getting into VGB the negatively charged nanoparticles accelerate by the applied electric field to very high velocities depending on their size: the smaller the particle the higher its final velocity [41]. This, as well as the ratio of the hardness of NPs and the oxide layer, leads to different scenarios of their interaction [28,41].

This work is aimed at revealing effects of adding silica NPs in the electrolyte on the PEO process of Mg₉₇Y₂Zn₁ alloy with LPSO-phase and elucidation of specific interaction mechanisms of silicon dioxide NPs and an oxide layer obtained during PEO.

2. Experimental procedures

Oxide layers were formed by PEO on flat specimens with sizes 20 × 60 × 6 mm³ of Mg₉₇Y₂Zn₁ alloy with Mg₁₂YZn LPSO-phase in alkaline-phosphate-fluoride-aluminate electrolyte at pulse frequency 1 kHz and duty cycle 50 % under constant current density 8 A/dm² within 10 min. PEO was provided without (specimens marked PEO (1 kHz)) and with addition of amorphous silica NPs (SiO₂ NPs, n-SiO₂) produced by Plasmotherm LLC (Moscow, Russia) with an average size of $100nm$ (size range – $56÷188nm$) at a concentration 3 g/L (specimens PEO(1 kHz) + n-SiO₂). The concentration of NPs addition in the range of 2–3 g/L and specifically 3 g/L [39,41,[45], [46], [47], [48]] has proven to be one of the most effective for achieving the best functional characteristics of oxide layers. The used silica NPs were characterized in detail in our previous work [41].

The aqueous solution of potassium hydroxide (KOH, 1 g/L), sodium phosphate disubstituted twelvehydrate (Na₂HPO₄ × 12H₂O, 10 g/L), sodium phosphate (NaF, 10 g/L) and sodium aluminate (NaAlO₂, (10 g/L) was used as the base electrolyte. All reagents were of high chemical purity. The chemical composition of the alloy and preparatory procedures were described in Ref. [49].

The microstructure, chemical (elemental) and phase composition, thickness (T, μm), surface porosity (P_S, %) and effective porosity (P, %), hardness (HV_0.01, MPa), adhesion strength (specific critical load F_c, N/μm) and anticorrosive properties (corrosion potential E_corr, current density i_corr, polarization resistance R_p, impedance modulus |Z|_f = 0.01Hz) were investigated by scanning electron microscopy (SEM), confocal laser scanning microscopy (CLSM), energy dispersive X-ray microanalysis (EDX), X-ray diffraction analysis (XRD), potentiodynamic polarization (PDP), electrochemical impedance spectroscopy (EIS) and instrumental methods using techniques described in Refs. [28,41,49].

3. Results and discussion

3.1. The evolution of forming voltages during PEO

The time evolution of anodic and cathodic forming voltages during the PEO process is shown on Fig. 1 for both electolytes used in the experiment. The electrolytes are marked as PEO and PEO+n-SiO₂ for electrolyte without and with silica NPs addition, respectively. In the first 100 s of PEO treatment, a rapid jump in the anodic voltage to 270 … 300 V was observed for both electrolytes. Then a slight increase in the anodic voltage was observed (up to values of 390 … 400 V at the end of the process), while the cathodic voltages were significantly smaller than the anodic ones. In the first ∼60 s, the cathode amplitude voltage was only ∼1–2 V, then increased to 50–60 V, and then gradually rose to 80–90 V at the end of the PEO treatment. It can be assumed that the formation and compaction of the barrier layer at the magnesium alloy/electrolyte interface is the reason for the observed evolution of the anodic and cathodic voltages at the initial stage of the PEO process (Fig. 1). After ∼100–120 s, the barrier layer probably acquires sufficient dielectric strength, which leads to a significant increase in the breakdown voltage, and then this characteristic changes slightly during the PEO process, and the corresponding increase in the forming voltages is relatively small.

Fig. 1.

Time dependencies of anodic (U_a) and cathodic (U_c) forming voltages under PEO treatment of Mg₉₇Y₂Zn₁ alloy in the case of base electrolyte (without n-SiO₂) and with addition of SiO₂ NPs.

It is important to note while the voltages behavior for both electrolytes during most of the PEO process is very similar, the forming voltages in the electrolyte with addition of silica NPs were clearly smaller than those in the base case (without addition of silica NPs to the electrolyte). The observed differences in the anodic voltages were about ∼15–20 V after ∼120 s and to ∼5–10 V after ∼450 s of the PEO treatment, while the difference in the values of cathodic voltages was ∼5–10 V during most of the time processing. After an initial time of ∼70–80 s of the PEO at voltages of about 240–250 V, visual sparking begins on the specimens in the bath, and then the brightness of microarc discharges and their number increase. Furthermore, the microarcs on the specimen in the electrolyte with the addition of SiO₂ NPs seemed to be brighter and sharper despite some turbidity of the electrolyte due to a suspension of silica NPs, which indicates their greater power compared to the sparks during PEO in the base electrolyte (without NPs). Finally the $n‐Si{O}_2$ led to a decrease in the breakdown voltage during PEO in the anode and cathode half-periods at the end of the PEO process from 398 ± 2 V and 93 ± 1 V to 387 ± 2 V and 81 ± 1 V, respectively.

3.2. Structure, surface morphology and chemical composition of the oxide layers

It was found by phase contrast in SEM BSE mode and EDX, that the LPSO-phase is poorly oxidized that results in the hereditary chemical inhomogeneity (bright areas on the cross-section images – Fig. 2) of the formed layer, as well as in increase of porosity and discontinuities at the Mg-alloy/oxide layer interface (Fig. 2a and b). This phenomenon is confirmed by EDX-analysis showing that the bright areas are enriched by yttrium and zinc (Fig. 3a and b) transferred to the oxide layer from the LPSO-phase of the substrate. Furthermore, the presence of yttrium and zinc was detected on the surfaces of both specimens (Fig. 3c and d). A similar effect is known for PEO of Al–Si alloys as the inheritance of chemical inhomogeneity in the oxide layer from Si-particles of silumin [50].

Fig. 2.

Cross-section structure (a, b) and morphology of the surface (SEM, BSE) (c, d) of the layers obtained by PEO without (a, c) and with (b, d) n-SiO₂ addition.

Fig. 3.

The distribution maps of the main elements over the cross-sections (a, b) and on the surfaces (c, d) of the specimens obtained by PEO without (a, с) and with (b, d) n-SiO₂ addition.

The surface morphology of the obtained oxide layers has a strong unevenness, and contains cellular type zones, with numerous nanosized and microsized pan-cake elements, traces of hot gases, solidified bubbles, cracks and craters (Fig. 2c and d). Different micron and submicron objects resulting from micro breakdowns and subsequent thermal relaxation processes are seen in both cases.

Study of the oxide layer structure by SEM BSE and EDX methods revealed a nonuniform distribution of elements such as Zn and Y in the oxide layer cross-section which indicates LPSO-phase Mg₁₂YZn (Fig. 3a and b), as well as fluorine F, which is detected predominantly in the inner zones of the obtained specimens (Fig. 3a and b). Other elements (Mg, Al, O and P, as well as Si in the case of addition of n-SiO₂ NPs in the electrolyte) are distributed quite uniformly over the cross-section structures (Fig. 3a and b) as well as on the surfaces of the specimens, without pronounced localizations near pores or on numerous plateaus of oxide layers (Fig. 3c and d).

3.3. Coatings phase composition

The quantitative EDX analysis showed the presence of Si in the oxide layer (Fig. 2, Fig. 3) in the amount of 2,18 ± 0,27 wt%, but no peaks of crystalline SiO₂ or products of chemical reactions of NPs with the oxide layer were detected by XRD (Fig. 4a). The both oxide layers consist of a mixture of magnesium oxide MgO (Hexagonal, P6₃/mmc) and magnesium aluminate MgAl₂O₄ (spinel, Cubic, Fd-3m) (Fig. 4a). It should be noted, that the broad peak (amorphous halo) in the range (25–40)° of the 2θ values (Fig. 4b) was observed which indicates the presence of a significant fraction of the amorphous phase being a part of the oxide layer.

Fig. 4.

XRD patterns of the Mg₉₇Y₂Zn₁ magnesium alloy samples with the oxide layers obtained by PEO in the base electrolyte (without silica NPs) and with n-SiO₂ additions to the electrolyte (a) and corresponding pattern's zones with broad peak (amorphous halos marked by grey) (b).

A joint analysis of EDX and XRD results suggests that the silica NPs are incorporated in the coating in the initial chemical and structural state, without chemical reactions and/or recrystallization. Indeed, the incorporation of a silicon-containing substance into the oxide layer was established by EDX with a silicon content of ∼2.2 wt % (which is equivalent to 4.7 wt% SiO₂). At the same time, XRD analysis did not show the presence of any reflections of silicon-containing crystalline phases, while showing an increase in the background signal (amorphous halo) and an increase in the fraction of the X-ray amorphous phase by ∼4.4 wt% (Fig. 4, Table 1) that is near to addition of 4.7 % n-SiO₂. That is, the increase in the proportion of the amorphous phase practically coincides with the amount of SiO₂ (based on the conversion of Si detected by EDX to SiO₂). At the same time, the amount of magnesium oxide in the layer was at the same level of ∼58.5 wt% for both cases (Table 1), which indicates the absence of reactions of magnesium oxide with SiO₂ NPs and is also confirmed by the absence of peaks of any of the magnesium silicate phases. Thus, according to detected XRD patterns and EDX data and calculated proportions of substances and phase compositions in the samples, a conclusion was made about the inert incorporation of SiO₂ nanoparticles into the layer. One can assume that some silica particles entered into reactions with the substances of the oxide layer and formed additional amorphous compounds in the composition of the amorphous phase. However, the proportion of such compounds is probably insignificant. In addition, it can be concluded, that phosphorus P and fluorine F are part of the X-ray amorphous phase, since no peaks of P-containing or F-containing phases were detected by XRD, but the presence of phosphorus and fluorine was detected by the EDX.

Table 1.

Results of quantitative XRD analysis^a of the obtained coatings.

Oxide Layer	Phase, wt%
	MgO	MgAl2O4	Amorphous
PEO(1 kHz)	58.21	31.35	10.44
PEO(1 kHz) + n-SiO2	58.66	26.54	14.8

^aThe estimated error is less than 5 % (by factor wRp in the Jana 2006 software).

3.4. Protective properties of oxide layers

The addition of $n‐Si{O}_2$ to the electrolyte partially reduced the effect of hereditary chemical inhomogeneity (Fig. 2b), reduced the porosity P measured by PDP from ∼0.19 % to ∼0.07 %, surface porosity P_S measured by CLSM from ∼1.5 % to ∼1.2 % and roughness measured by CLSM from ∼2.75 μm to ∼2.60 μm, as well as provided an increase in the average coating thickness from ∼22 μm to ∼28 μm (Fig. 2, Table 2, Table 3).

Table 2.

Measured characteristics of the formed oxide layers.

Oxide Layer	Thickness, T, μm	Roughness Sa, μm	Surface porosity, PS, %	HV0.01, MPa	Fc, N/μm
Bare alloy	–	–	–	67 ± 0,2 HB	–
PEO	22 ± 2.6	2.74 ± 0.26	1.48 ± 0.09	610 ± 120	1.1 ± 0.2
PEO + n-SiO2	28 ± 3.5	2.58 ± 0.21	1.21 ± 0.12	600 ± 130	1.5 ± 0.2

Table 3.

Corrosion properties of the bare Mg₉₇Y₂Zn₁ alloy and the alloy specimens with PEO layers obtained without and with silica NPs addition.

Sample	OCP, V	Ecorr, V	icorr, nA⋅cm−2	βc, V/dec	βa, V/dec	Rp, kΩ⋅cm2	P, %
Bare alloy	−1.58 ± 0.03	−1.47 ± 0.03	4600 ± 3641	0.10 ± 0.04	0.17 ± 0.09	7.6 ± 4.2	–
PEO	−1.53 ± 0.02	−1.54 ± 0.02	25.7 ± 11.5	0.18 ± 0.02	0.15 ± 0.06	(1.6 ± 0.8)a103	0.19 ± 0.08
PEO + n-SiO2	−1.50 ± 0.06	−1.48 ± 0.01	4.7 ± 0.3	0.13 ± 0.04	0.24 ± 0.08	(9.8 ± 1.6)a103	0.07 ± 0.04

^a±2σ is provided as an error.

The mechanical and corrosion tests showed that incorporation of n-SiO₂ in the oxide layer provided the increase in the adhesion strength and anticorrosion properties – the specific critical load F_c, polarization resistance R_p and impedance modulus |Z|_f = 0.01Hz increased by ∼1.4, 6.3 and 2.5 times (Table 2, Table 3, Table 4, Fig. 5), respectively, while corrosion current i_corr decreased by ∼5.5 times (to ∼4.7 nA/cm²). The i_corr = 4,7 ± 0,3 nA/cm² (present research) is comparable to one of the best results achieved by PEO of the Mg-LPSO alloy followed by Ce-sealing post-treatment (2 nA/cm² [5]) and ∼7 times less than the corrosion current obtained under ultra-high frequency PEO without NPs (∼34 nA/cm² [17]).

Table 4.

Calculated parameters of the equivalent electrical circuits for the Mg₉₇Y₂Zn₁ samples with various types of oxide layers.

Sample	Rout, kΩ⋅cm2	CPEout		Rin, MΩ⋅cm2	CPEin		|Z|f→0.01Hz, MΩ·cm2
		Qout, Ω−1·cm−2·sn	nout		Qin, Ω−1·cm−2·sn	nin
Bare alloy	–	–	–	(1 ± 0.3)·10−3	(1.6 ± 0.1)·10−5	0.94 ± 0.01	(1.8 ± 0.2)·10−3
PEO	10.5 ± 0.6	(3.0 ± 0.3)·10−7	0.79 ± 0.01	1.27 ± 0.30	(1.5 ± 0.2)·10−7	0.91 ± 0.03	1.34 ± 0.30
PEO + n-SiO2	17.5 ± 2.3	(2.2 ± 0.1)·10−7	0.80 ± 0.01	3.19 ± 1.10	(0.8 ± 0.1)·10−7	0.95 ± 0.01	3.33 ± 0.13

*±2σ is provided as an error.

Fig. 5.

Tafel plots (a), Nyquist plots (b), Bode plots (c) and phase angle plots (d) of the obtained specimens.

The obtained Tafel plots (Fig. 5a), Nyquist plots (Fig. 5b), Bode plots (Fig. 5c) and phase angle plots (Fig. 5d) characterize the anticorrosive properties of the oxide layers. The results of processing the obtained dependencies are given in Table 3 (PDP) and Table 4 (EIS). It was found, that the presence of silica NPs in the electrolyte led to an increase of anticorrosive properties. It resulted in a decrease in the corrosion current i_corr by ∼5.5 times (Table 3) and an increase in the anodic slope of the Tafel curve by ∼ 1.6 times (Fig. 5a, Table 3), which together led to a high value of R_p according to Stern-Geary equation. The measured by EIS charge transfer resistance R_in through the barrier layer increased by ∼2.5 times (Table 4). In addition, a decrease in the rate of anodic dissolution of magnesium and an increase in the passivity area by ∼50–75 mV (some semblance of initial self-passivation) (while maintaining the value of catodic Tafel slope, β_c) can be observed (Fig. 5a, Table 3). This anodic passivation is probably responsible for the observed difference (more than 2.5 times) between R_p and |Z|_f = 0.01Hz (Table 3, Table 4). However, an additional investigation of this effect is needed and it will be a subject of our further research. In addition, the value of i_corr under maximum polarization voltage (+1.0 V vs OCP) in the case of PEO + n-SiO₂ specimen is about two order less against the base PEO specimen (Fig. 5a), which indicates its higher protective properties even in the case of high anodic polarization voltages. It can be assumed that the observed improvement of anticorrosive properties of coating obtained with addition of n-SiO₂ to the electrolyte can be partly due to incorporation of SiO₂ NPs into the barrier layer which leads to thickening and compaction of the barrier layer of the coating.

3.5. The mechanism of the NP-oxide layer interaction during PEO

The following colliding interaction mechanism is suggested in order to explain an inert incorporation of NPs into the oxide layer during PEO. The negatively charged in the electrolyte NPs are captured by vapor-gas bubbles (VGBs), accelerated by an applied electric field to very high velocities and they hit forcibly the surface of the oxide layer prior the microarc discharge. The final velocity V achieved by the NP depends on the potential difference U in the way that the particle passes before colliding with the surface. For a spherical particle of mass $m$, diameter D and a charge q, the final velocity is the following:

$$V=\sqrt{\frac{2Uq}{m}}=\sqrt{\frac{12U{\rho }_q}{D·\rho }}$$ (1)

where ${\rho }_q$ is the surface charge density, ρ is the NP material density. If the particle is harder than the oxide coating, as in the case of silica NPs and magnesium oxide layer, it can penetrate deep into the layer almost without deformation. This leads to creation of a new nanochannel with a new surface in the layer. Kinetic energy of the particle is spent on creating this surface and heating the NP and the nanochannel walls which is associated with the friction of the particle against the new surface. In the case of a spherical particle the new surface will be approximately equal to the surface area of cylinder with a particle diameter $D$. The work to create this surface is

$$A_f=\gamma \pi DL$$ (2)

where γ is the effective surface energy taking into account a plastic deformation and heating of the layer material, L is the depth of the particle penetration into the oxide layer. Thus, the energy balance equation can be written as following:

$$\frac{m{V}^2}{2}=mC\Delta T+\gamma \pi DL$$ (3)

where C is the heat capacity of the NP substance, $\Delta T$ is the heat-up value of the particle. Considering $D\approx L$, one can estimate from eqs. (1), (2), (3) the value of heating. Using the typical known literature data for the surface charge density [[51], [52], [53]], ${\rho }_q=0.1\frac{C}{m^2}\text{,}$ the material properties of silicon dioxide: $C=1\frac{kJ}{K\bullet kg},\rho =5\bullet {10}^3\frac{kg}{m^3},\gamma =5\frac{J}{m^2}$, and using the difference of potentials across a VGB as $U=\left(100-200\right)V\text{,}$ for the NP of diameter D = 50 nm, yields $\Delta T=\left(120-360\right)K$. This rather low heating explains inert incorporation of silica NPs in the oxide layer on Mg-alloy. On the other hand, the near-surface incorporation of NPs may lead to a local decrease of the breakdown voltage due to accumulation of positive surface charge during particle/surface friction. It therefore may lead to the appearance of additional microarcs and their branching (sparking) which were observed on the samples during PEO with the silica NPs addition. This, in turn, leads to an increase in the volume of the molted metal, and then the oxide layer volumes (increase of the coating thickness), similarly to the case of $n‐$ TiC using in PEO of Al–Si alloys [28].

4. Conclusion

The addition of silica NPs into the electrolyte led to a complex positive effect on the mechanical and anticorrosive properties of oxide layers obtained by PEO on Mg-LPSO alloy. The effect of hereditary chemical inhomogeneity caused by LPSO phase was partially reduced, the average coating thickness and adhesion strength were increased by ∼28 % and ∼36 %, respectively, the porosity was reduced by ∼2.7 times and the corrosion current density by ∼5.5 times. All these positive effects are associated with the change of number and power of microarc discharges. A certain decrease in the breakdown voltages caused by impact penetration of silica NPs into the oxide layer during their collisional interaction with the oxide layer during PEO treatment was observed and explained in the present work. The negatively charged and accelerated in electric field SiO₂ NPs collide with the forming coating and are embedded in the near-surface region almost without deformation being substantially harder than the oxide layer. It is suggested that this colliding interaction results in local accumulation of positive charges at the surface due to particle/surface friction. This, in turn, leads to appearance of additional local breakdowns and microarc branching responsible for increasing efficiency of the PEO process.

Data availability statement

Data associated with this study has not been deposited into a publicly available repository but it will be made available on request.

CRediT authorship contribution statement

M.M. Krishtal: Writing – review & editing, Validation, Supervision, Resources, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. A.V. Katsman: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis, Data curation. A.V. Polunin: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. A.O. Cheretaeva: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.