Applying ANOVA to Study the Effect of Weight Factions of the Nano-Hafnia Dispersed in Spin Coated Gelatin on the Corrosion Behaviour of AISI 316L Steel

This study aims to examine the potential improvement of uniform corrosion resistance and critical pitting potential of AISI 316L stainless steel (SS) in a simulated marine environment containing 3.5 wt. % NaCl by applying layer of gelatin containing HfO2 nanoparticles using spin coating. Also, to examine the statistical significance of using different weight fractions of HfO2 in the applied layer on the corrosion protection performance. The corrosion performance was assessed from potentiodynamic polarization. X-ray diffraction (XRD) was used to check the presence of HfO2 in gelatin matrix. Scanning electron microscopy (SEM), and energy dispersive X-ray analysis (EDX) were used to examine the surface morphology and elemental composition of corrosion products. The results showed that the applied nanocomposite coatings enhanced both uniform and localized corrosion resistance of the underlying AISI 316L SS, also Analysis of Variance (ANOVA) test approved that the weight fractions of dispersed HfO2 nanoparticles have a statistical significant effect on the corrosion performance. Nanocomposites coating having 2 wt.% of HfO2 showed the best performance as compared to other examined coatings.


Introduction
Nanocomposite coatings of polymer reinforced with inorganic ceramic nanoparticles have been used recently to enhance the corrosion resistance of the underlying metal substrate.They provide barriers for corrosive ions diffusion and penetration [1][2][3][4][5][6] .
Gelatin is considered a promising candidate material in medical applications because of its high biocompatibility [7][8][9] .Therefore, it was used by several researchers as a matrix in many nanocomposite coatings containing ceramic powder to protect the medical grade steel AISI 316L against corrosion 3,5,6 .These coatings show good protection capability against corrosion in corrosive environments containing chlorine 10 .
The high dielectric constant of hafnium dioxide (HfO 2 ) makes it a good choice in corrosion protection coatings [11][12][13] .HfO 2 films give superior protection performance against corrosion as compared to ZrO 2 , TiO 2 , Al 2 O 3 , ZnO, ZrO 2 atomic layer deposited films from the potential reactive etchants and solvents 14,15 .Plasma reactive sputter deposited HfO 2 nanolayer was used to protect 316L SS against corrosion.This layer improves the uniform corrosion resistance of the underlying 316L SS substrate and increases the pitting potential to a more noble value 16 .Over the last few years, HfO 2 thin films were successfully deposited using thermal and plasma-enhanced atomic layer methods to provide protection against corrosion over different kinds of metal substrates, i.e., copper and AZ23 magnesium alloy 14,17 .
Corrosion rate (r) and critical pitting potential (CPP) represent two important electrochemical parameters measured from a polarization experiment.The former one highlights to the materials' protection against uniform corrosion, whereas the latter is related to localized corrosion resistance and refers to the minimum potential level at which the pits become stable and autocatalytic growth begins 18,19 .
ANOVA is a simple, widely known statistical approach that is used to examine the statistical significance difference between the means across multiple groups 20 .This method has become very popular in materials engineering to assess variation in the material properties with either elemental composition 21 or processing parameters 22,23 .
In this paper, ANOVA approach is conducted with the aim to examine the statistical significance of weight fractions of HfO 2 nanoparticles incorporated into gelatin matrix nanocomposite coatings on both corrosion rate and pitting potential of 316L SS substrate in a 3.5 wt.% NaC solution.

Preparation of AISI 316L stainless steel substrates
Test coupons of AISI 316L SS with a square shape and size of 55 mm x 55 mm were cut from 1 mm thick sheet.The supplier provided the following chemical compositional analysis for 316L SS as listed in Table 1, and it is not based on a chemical analysis of each coupon used in this case.These coupons were used as substrate for coatings.Prior to coating, the substrates were chemically etched, ultrasonic cleaned, and then dried.The chemical etching was conducted using a 70:30 solution of nitric acid (HNO 3 ) and hydrogen peroxide *e-mail: maalmomani7@just.edu.jo(H 2 O 2 ).Hence, substrate surface roughness increased which enhanced the coatings adhesion to the underlying substrate.Thereafter, substrates were ultrasonic cleaned in two stages using acetone and deionized water to remove surface stains.Lastly, the hot air stream was used to dry off the cleaned substrates.Figure 1 presents a flow chart of the sequential steps that were used to prepare the substrates for coating.

Preparation of gelatin-HfO 2 coatings
In this research, HfO 2 nanoparticles (size of 61-80 nm, purity 99.99%, US Research Nanomaterials Company) were dispersed at different wt.%. in gelatin (type A porcine skin, Mw= 50,000-100,000, Sigma).All four types of coatings were applied at the same spinning speed (500-2500 rpm), and the spinning time was chosen to be 15 seconds, in order to prepare the sol gel coating material.A crosslinking agent called 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide (EDC) is used to join carboxyl groups to primary amines.It creates an amine-reactive intermediate when it combines with a carboxyl.Figure 2 presents a schematic diagram for the gelatin-HfO 2 nanocomposites sol gel preparation process, and then spin coatings on an AISI 316L substrate.Those steps were similar to those used for other coatings and published elsewhere [1][2][3][4] .

Characterization of the nanocomposite coated substrates
Due to the strong dielectric properties of the produced nanocomposite coatings, the surface was sprayed with a thin layer of gold.Then, SEM was used to examine the surface morphology of the coated samples before the corrosion test to investigate the homogeneity and dispersion of the nanoparticles within the gelatin matrix, and after the corrosion test to examine and observe carefully the surface morphological changes occurred during polarization experiment.EDX was used to study the elemental composition of the substrate surface and corrosion products.SEM and EDX were conducted using SEM model FEI Quanta FEG 450 at Nanotechnology Institute of Jordan University of Science and Technology.To confirm the presence of HfO 2 nanoparticles within the gelatin coating on the AISI 316L SS substrate, Diffractometer Rigaku Ultima IV model was used to conduct XRD at the Pharmaceutical Research Center of Jordan University of Science and Technology.In order to develop an understanding of the roughness of the coatings, image of the surface topography of the created nanocomposites' surfaces was collected using an AISI-NT atomic force microscope (AFM) at the Nanotechnology Institute of the Jordan University of Science and Technology.To evaluate the performance of the produced coating in protecting underlying substrate against uniform and localized pitting corrosion, potentiodynamic polarization test was used.This test was conducted with the aid of Gamry potentiostat Ref 600 corrosion testing instrument that was connected with three electrodes flat cell.The cell used graphite as a counter and saturated calomel (SCE) as reference electrodes.For each coating type, three replicates were used in this test.

Statistical analysis
The weight fraction of HfO 2 nanoparticles dispersed in the gelatin matrix in the produced nanocomposites determines the coating type.Four different types of spin coating layers were applied on AISI 316L SS substrate.The coatings were given the symbols A, B, C, and D as listed in Table 2; hereafter; these symbols will be used in the manuscript to refer to the examined nanocomposite coatings.
The statistical significance of the examined samples coating type was measured using one-way ANOVA analysis by Minitab software (Version 18).Differences were chosen to be significant for P-value < 0.05.

Electrochemical corrosion investigation
Figure 3 shows the potentiodynamic polarization curves of AISI 316L coated with different weight percentages of HfO 2 nanoparticles dispersed in a gelatin matrix at 500 -2500 rpm spinning speeds.In order to build confidence in the scientific validity of the results, three replicates were examined at each weight fraction of HfO 2 .The variation of the corrosion rate and critical pitting potential between the samples examined at each level was shown in the bar chart with different colors in Figures 4 and 5, respectively.The mean value and the standard deviation of the three replicates at each level of weight fractions for each one of the measured parameters are listed in Table 3.Here, the difference in color just denotes an independent experiment; the use of a certain color in the illustration does not imply that all trials that used that color are similar, since every experiment was run independently and at random.The corrosion parameters that were obtained from the Tafel region of the examined samples include: corrosion potentials (E corr ), corrosion current densities (I corr ), and critical pitting potentials (E pit ) 24 .The sample that has the least corrosion current density experienced the best corrosion resistance 25 .Thus, coating AISI 316L with gelatin reduces its' corrosion current density and thus its' corrosion rate.Also, dispersing HfO 2 nanoparticles in the gelatin reduced the corrosion    current density of the underlying AISI 316L substrate.The increase of pitting potential and shift to a higher noble value is a sign that the metastable pits that form during the anodic Tafel scan need higher potential to transform to stable pits.The curves show that no significant improvement on its' critical pitting potential was observed by coating AISI 316L SS with gelatin only.Nevertheless, dispersing HfO 2 nanoparticles in the gelatin increased the critical pitting potential significantly to more positive values.Where HfO 2 coating film provides a significant electrochemical corrosion protection and decreases the corrosion current density due its strong dielectric properties 14,26 .
Figure 4 shows the effects of dispersing different weight fractions of HfO 2 nanoparticles on the corrosion current density of AISI 316L SS, dispersing 2 wt.%HfO 2 nanoparticles in gelatin outperform better than dispersing 3 wt.% of HfO 2 in gelatin, such behaviour can be attributed to the increase of tendency of nanoparticles to agglomerate in gelatin as the weight fraction percentage of nanoparticles increased, which increased the surface roughness.This will affect the corrosion current density adversely; these results are in agreement with previous research works 6 .
Figure 5 shows the critical pitting potential behaviour of AISI 316L coated samples with different weigh percentages of HfO 2 nanoparticles (0,1, 2 and 3).It is shown that HfO 2 nanoparticles coating improved the critical pitting potential of the AISI 316L significantly, this behaviour is due to the formation of the stable coating layers on AISI 316L surface compared to the unstable oxide film on the uncoated AISI 316L, which could not sufficiently protect against pitting corrosion 6 .The change of critical pitting potential with weight fractions of HfO 2 are compatible with the corrosion current density results; where the highest electrochemical enhancement occurs at 2 wt.% HfO 2 nanoparticles coating due to the negative effect of agglomeration on surface roughness and critical pitting potential as well.

Statistical analysis
One-way ANOVA analysis with the significance level 95% was used to test the statistical significance of corrosion current density and critical pitting potential on the coating type.Minitab software (version 18) was utilized to conduct the test.Five levels of the examined coating type were used: uncoated AISI 316L SS, and samples of coatings (A, B, C, and D), each level has three replicates.Statistical procedures based on analysis of variance (ANOVA) methods were used to analyse the data.ANOVA test results for the effect of coating type on corrosion rate and critical pitting potential were summarized in Tables 4 and 5, respectively.The results confirmed that corrosion current density and the critical pitting potential are dependent on coating type.Therefore, the use of different coating type will affect significantly on both corrosion current density and critical pitting potential.It is shown that all coating types can improve the critical pitting potential of the uncoated AISI 316L significantly, this behaviour might be due to the formation of the stable coating layers on AISI 316L surface compared to the unstable oxide film on the uncoated AISI 316L, and the best electrochemical enhancement occurs at 2 wt.% HfO 2 nanocomposite coating.Hence, the statistical analysis is fully supported the electrochemical and morphology results.The individual value plot of the corrosion current Density (nA/cm 3 ) and the critical pitting potential (mV vs. SCE) of AISI 316L are graphically presented in Figure 6 and Figure 7.

Surface morphology and elemental composition analysis 3.3.1. Uncoated and coated AISI 316L SS with gelatin
The surface morphology of uncoated and coated AISI 316L SS with gelatin was examined previously and reported in our published work [3][4][5] .Prior to potentiodynamic polarization tests, only polishing lines appeared on the uncoated substrate surface; whereas autocatalytic stable pits were formed after polarization testing; the mechanism of these pits formation was described elsewhere 16 .Spin coating of AISI 316L SS with gelatin resulted in a smooth, uniform surface layer free of cracks and defects; however; after polarization this layer was cracked and failed to protect the underlying substrate.These results were in agreement with electrochemical polarization results that didn't show any improvement in the critical pitting potential of AISI 316L SS.

AISI 316L SS coated with HfO 2 nanoparticles dispersed in gelatin
Figure 8 shows SEM image of the sample of coating type C before conducting the polarization test.Nanoparticles were uniformly and homogeneously dispersed throughout the gelatin matrix, EDX spectrum for these dispersed particles is shown in Figure 9 and has peaks that belong to hafnium and oxygen (in addition to the substrate elements peaks), supporting the existence of HfO 2 nanoparticles.
On the other hand, the dispersion of HfO 2 nanoparticles in gelatin improves the coating's performance in protecting AISI 316L against corrosion.Figure 10 shows SEM image for the sample of coating type C after conducting the polarization, the surface was free of cracks and pits; the coating had nanoparticles similar to those observed in the coating prior to polarization experiments with small precipitates of salt.This is a decent indicator of the stability and functionality of this coating.Figure 11 shows an EDX spectrum of this sample; the spectrum had the substrate AISI 316L SS elements peaks.Also, hafnium, and oxygen peaks were observed, indicating that HfO 2 particles were not leached out.Sodium and chlorine peaks belong to precipitated salt on the surface.

XRD analysis
X-ray diffraction was used to confirm the existence of HfO 2 dispersed particle in the gelatin matrix of the produced nanocomposite coating.The XRD spectra for the uncoated 316L SS, and AISI 316L SS spin coated with HfO 2 -gelatin nanocomposites are shown in Figure 12.XRD spectrum for AISI 316L SS substrate had three peaks belonging to     the austenitic phase and assigned to crystallographic planes (100), (200), and (220).The XRD spectrum of the coated sample was compared with the standard one and it confirmed the presence of m-HfO 2 .All peaks are indexed accordingly.

AFM measurement
In order to develop an understanding of the roughness of the coatings, image of the surface topography of the sample of coating type C was collected using an AISI-NT atomic force microscope (AFM), and it is presented in Figure 13, showing that the average height of the coating is 10.57µm.

Conclusions
This study aims to examine the potential improvement in the corrosion behaviour of AISI 316L SS achieved by applying layer of gelatin containing HfO 2 nanoparticles using spin coating technique in a simulated marine environment containing 3.5 wt.% NaCl.Based on the results of the study, the following points can be concluded: • Coating AISI 316L with gelatin reduces its' corrosion current density and thus its' corrosion rate.However, no significant.improvement on it's critical pitting potential.• Dispersing HfO 2 nanoparticles in the gelatin reduced the corrosion current density of the underlying AISI 316L substrate, and increase its pitting potential to more noble value.

•
The gelatin coating containing 2 wt.%HfO 2 nanoparticles outperforms better than that containing 3 wt.%HfO 2 due to the high tendency of particles agglomeration at higher weight fraction, which increase the coating roughness.• ANOVA test shows that the coating type (weight fraction of HfO 2 ) has a statistical significance effect on the performance of the applied coating in protecting the underlying AISI 316L SS substrate against uniform corrosion and localized pitting corrosion.

Figure 3 .
Figure 3.The potentiodynamic polarization curves of uncoated AISI 316L SS and samples with coating's type A, B, C, and D.

Figure 5 .
Figure 5. Critical pitting potential (mV vs. SCE) of uncoated AISI 316L SS and samples with coating's type A, B, C, and D.

Figure 4 .
Figure 4. Corrosion Current Density (nA/cm 3 ) of uncoated AISI 316L SS and samples with coating's type A, B, C, and D.
of Freedom; (b) Adj SS = Adjusted Sum of Squares; (c) Adj MS = Adjusted Means of Squares; (d) F-Value = Value on the F distribution; (e) P-Value = Value of the level of marginal significance

Figure 6 .
Figure 6.The individual value plot of the corrosion current Density (nA/cm 3 ) of uncoated AISI 316L SS and samples with coating's type A, B, C, and D.

Figure 7 .
Figure 7.The individual value plot of the critical pitting potential (mV vs. SCE) of uncoated AISI 316L SS and samples with coating's type A, B, C, and D.

Figure 8 .
Figure 8. SEM image of the sample of coating type C before corrosion test.

Figure 9 .
Figure 9. EDX spectrum of the sample of coating type C before corrosion test.

Figure 10 .
Figure 10.SEM image of the sample of coating type C after corrosion test.

Figure 11 .
Figure 11.EDX spectrum of the sample of coating type C after corrosion test.

Figure 13 .
Figure 13.AFM image of the surface topography of the sample of coating type C.

Table 1 .
Chemical composition of AISI 316L SS as provided by the supplier.
Figure 1.Flow chart of sequential steps used to prepare the substrates for coating.

Table 3 .
Electrochemical parameters for the uncoated AISI 316L SS and samples with coating's type A, B, C, and D.

Table 4 .
ANOVA test results of the samples corrosion current density (I corr ) by the coating type.

Table 5 .
ANOVA test results of the critical pitting potential (E pit ) by the coating type.Applying ANOVA to Study the Effect of Weight Factions of the Nano-Hafnia Dispersed in Spin Coated Gelatin on the Corrosion Behaviour of AISI 316L Steel