G.Gusmano a,b ,G.Montesperelli c,⁎,M.Rapone a,b ,G.Padeletti d ,

A.Cusmàd ,S.Kaciulis d ,A.Mezzi d ,R.Di Maggio e

a

Consorzio INSTM -Unitàdi Ricerca di Roma Tor Vergata,Italy

b

Dipartimento di Scienze e Tecnologie Chimiche,Universitàdi Roma -“Tor Vergata ”,Rome,Italy

c

Dipartimento di Fisica ed Ingegneria dei Materiali e del Territorio,UniversitàPolitecnica delle Marche,Ancona,Italy

d

ISMN -CNR,Monterotondo Stazione,Rome,Italy

e

Dipartimento di Ingegneria dei Materiali,Universitàdi Trento,Trento,Italy

Received 14July 2006;accepted in revised form 19October 2006

Available online 8December 2006

Abstract

The recent restriction in the use of chromium VI gave a new impulse in the field of anti-corrosion treatment for an alternative solution.A promising route seems to be the deposition of a thin ceramic layer on the metallic surface.In this study,four zirconia primers have been deposited on 1050aluminium sheets by sol –gel process.Sol preparation was obtained starting from two different precursors and two curing temperatures were tested.

Surface microstructure and chemical composition were determined by AFM and XPS for all samples.Electrochemical Noise Analysis (ENA)was used in Harrison's solution to evaluate the corrosion resistant features of the samples.Tests have been also performed on samples protected by polyester topcoat deposited on the aforesaid primers.

AFM analysis revealed that,for all the samples,covering and structure of the coating were uniform.XPS depth profiling allowed concluding that film thickness was in the range from 18to 30nm and that chemical composition was constant through primer thickness.

Corrosion tests demonstrated that zirconia primers showed good performances in terms of corrosion resistance,comparable to chromate and fluotitanate layers industrially prepared.

Topcoated samples gave an optimal corrosion resistance.©2006Elsevier B.V .All rights reserved.

Keywords:Zirconia;Coating;Sol –gel;Corrosion;Aluminium alloy;AFM;XPS;Electrochemical Noise

1.Introduction

In recent years,the European Union promulgated directives concerning restrictions in the use of heavy metals such as chromium VI,lead,mercury and cadmium in vehicles [1]and in electronic devices [2].In particular the limitations in the use of chromium VI,led a great demand in the corrosion protection field for new inhibitors and new coating formulations,characterized by low environmental impact,with the same corrosion prevention and protection properties [3–5].

A suitable solution seems to be the use of inorganic primers based on silica,zirconia or titania,topcoated with different polimeric layers [3,6,7].The corrosion resistant features of such a system depend on a number of characteristics such as:metal/primer and primer/organic film adhesion,number of defects and their dimension,lack of primer degradation phenomena,such as hydrolysis,when immersed in an aggressive environment,primer ability in inhibiting local corrosion phenomenon occurring on metal.In particular,the presence of defects on the primer,allows the onset of corrosion events in correspon-dence of such defects [3,4,6–8].

A very versatile technique for primer preparation is the deposition by sol –gel route [9–11].This method allows a good reproducibility of coating performances even though some problems have been evidenced due to complex shape

substrate

Surface &Coatings Technology 201(2007)5822–

5828

⁎Corresponding author.Tel.:+390712204401.

E-mail address:g.montesperelli@univpm.it (G.Montesperelli).0257-72/$-see front matter ©2006Elsevier B.V .All rights reserved.doi:10.1016/j.surfcoat.2006.10.036

and thick film deposition.Moreover,in some case,coating cracking induced by the thermal treatment performed at the end of deposition,has been reported [8].Usually,a thermal treatment at a temperature around 600°C provides a good chemical homogeneity and good corrosion resistant features to primers.Nevertheless,some promising results,in terms of corrosion resistant,have been obtained also at lower temper-ature (400°C)[12,13].

This paper presents the results of corrosion tests carried out on 1050aluminium sheets coated with zirconia primers,prepared by sol –gel route followed by a low temperature heat treatment.Primers obtained from both organic and inorganic precursors were tested.As a comparison,some cromate and fluotitanate primers industrially prepared have been also tested.In a second stage,the same primers coated with a polyester top layer have been tested.

Corrosion tests have been performed by means of current and potential Electrochemical Noise Analysis (ENA)in Harrison's solution.

Surface characterizations by X-ray Photoelectron Spectros-copy (XPS)and Atomic Force Microscopy (AFM)have been also carried out for all samples.2.Experimental

AA 1050aluminium sheets were used as substrates.Before deposition,the sample surface was degreased with acetone.Sol preparation was obtained starting from two different precursors.A 0.1M solution of zirconium tetrabutoxide (Zr(OBu n )4),containing acetic acid as complexing agents was used for organic (ZO)series [14]and a 0.4M solution of zirconyl nitrate (ZrO(NO 3))was used for inorganic (ZI)series.The withdrawal rate was 1mm/s for all samples.After film deposition,the samples were treated in an oven at 150°C for 1h or at 250°C for 4min as reported in Table 1.The film thickness was increased by repeating deposition procedure as reported in Table 1.

Comparison tests have been carried out on AA 1050sheets coated with a chromate or a fluotitanate layers industrially performed by Chemetall.Chromate layer was obtained by using Gardobond C714.Fluotitanate samples were prepared by Gardobond X4705.On fluotitanate primer,a further passivation step was performed by Gardolene D6800/CC.

Tests have been also performed on samples protected by polyester topcoat deposited on the already mentioned primers.

2.1.Test environment

Corrosion tests were carried out into two different Harrison's solutions.A diluted solution,in the case of samples only protected by a primer [3.5g/l (NH 4)2SO 4+0.5g/l NaCl]and a concentrated one for topcoated samples [3.5g/l (NH 4)2SO 4+5g/l NaCl].2.2.Electrochemical Noise Analysis (ENA)

Two working electrodes of 7.5by 10cm plate,0.7mm in thickness were used.The immersed surface was 9.6cm 2.Electrochemical current noise was measured as the galvanic coupling current between two nominally identical working electrodes,while electrochemical potential noise was detected as potential fluctuation between working electrodes and an Ag/AgCl reference electrode.Potential and current noises were simultaneously recorded using a Solartron 1287electrochem-ical interface.Data sets of 1024readings were recorded at a sampling interval of 250ms (sampling frequency =4Hz).The sampling interval was optimised by means of preliminary tests.After data collection the DC linear trend was removed by linear regression removal method.Data were transformed in the frequency domain through the Maximum Entropy Method (MEM)and Fast Fourier Transform (FFT)algorithms per-formed by dedicated software.Some blank tests on “as received ”electrodes have been also performed.2.3.Atomic Force Microscopy Analysis (AFM)

The morphological characterization of the coatings has been carried out in air by a Digital Instruments Dimension D3100AFM,operating in tapping mode with scanning areas from 1×1to 10×10μm.

2.4.X-ray Photoelectron Spectroscopy (XPS)

The surface chemical characterization has been carried out by using an Escalab Mk II spectrometer (VG Scientific Ltd,U.

Table 1

Experimental conditions of samples preparation Label Sample Precursors Temperature (°C)Time (min)Number of layers C Chromates ––––F Fluotitanates –

–––ZO1Zirconates Organic 25043ZO2Zirconates Organic 150603ZO3Zirconates Organic 25042ZI

Zirconates

Inorganic

250

4

2

Polyesther topcoated samples have been labelled adding “TC ”at the primer label (for example

ZO2TC).

Fig.1.AFM image (5×5μm)of ZO1zirconia primer.

5823

G.Gusmano et al./Surface &Coatings Technology 201(2007)5822–5828

K.)with 5channeltron detection system.Monochromatic Al and Mg X-ray sources (h ν=1486.6eV ,h ν=1253.6eV ,respectively)have been used for the excitation of photoelectron spectra.The binding energy scale was calibrated by fixing the peak of adventitious carbon at 285.0eV.For selected-area XPS depth profiling,the samples were fixed on the standard Escalab holder stub by means of an Au foil mask with a window of 3mm in diameter.The Ar +ion beam of 2.0keVenergy,rastered over a sample area of 3×3mm 2,has been used for the sample sputtering.More details on XPS depth profiling have been published elsewhere [15].3.Results and discussions 3.1.AFM results

AFM analysis showed a uniform covering for all samples.AFM image for ZO1sample,shown in Fig.1,pointed out a double granular structure characterized by rounded small grains (40÷60nm in diameter)upon which other larger grains (60÷100nm)grew.Larger grains appear to be aligned along a preferential direction.ZO2sample showed a microstructure quite similar to ZO1with slightly larger grains than ZO1(diameter 60÷120nm).ZO3sample showed a granular structure with grains in the range from 60to 100nm.ZI sample showed a significant increase of grain size that reached

dimension in the range 120÷250nm,thus emphasizing an influence of the inorganic precursors on the grain growth.On the contrary,as a general remark,it may be concluded that the grain size was not affected by heating condition.

Chromates samples (C)showed rounded grains with diameter from 50to 100nm.As already discussed in the case of ZO1sample,a double granular structure was noted,with outer grains arranged in annular structures (Fig.2).

Fluotitanates samples (F)evidenced a very similar micro-structure with respect to C samples,with grains in the range 40–120nm in diameter.

AFM examination did not evidence a granular structure for topcoated samples but the polymeric layer was thin enough to reveal the granular structure of the substrate.AFM also permits evaluation of the roughness of sample surface,taking into account the distribution curve of the relative highness between the acquired points within the scanning area and thus calculating the average roughness value (R a )with regard to a central plane and its standard deviation (R q ).Fluotitanates and chromates show higher roughness parameters.As a general trend it may be pointed out that passing from sample ZO1to ZO2(i.e.decreasing the temperature and increasing the length of heat treatment)

the

Fig.2.AFM image (1.75×1.75μm)of chromates (C)samples.

Table 2

AFM results Label Grain size (nm)R a (nm)R q (nm)C 50–10060.2075.67F 40–12090.52110.96ZO140–10016.2521.00ZO260–12023.3737.68ZO360–10036.0945.55ZI

120–250

26.45

33.17

Fig.3.Comparison between the chemical compositions of the samples ZO1and

ZO2.

Fig.4.XPS depth profile for ZO2sample.

5824G.Gusmano et al./Surface &Coatings Technology 201(2007)5822–5828

average roughness increases.The results of the AFM analysis for all primers are summarized in Table 2.Topcoated samples showed very low roughness in the range from 13to 16nm.3.2.XPS results

The results of XPS measurements enabled us to establish that all the samples have a constant composition through their thickness without any significant alteration of the chemical species.Fig.3shows the comparison of XPS quantification for ZO1and ZO2samples.Both samples are characterized by the presence of main constituents of the primer (zirconium,oxygen)and of the substrate (aluminium,silicon).

The peak-fitting analysis of C 1s spectra defined that this signal is very similar for all the samples:C –C and C –H component at 285.0eV ,C –O at ∼286.6eV ,C=O and –COOH at ∼288.6eV .It is worth noting that no substantial differences were detected in C 1s of the sample produced by using metallorganic and inorganic precursors.Anyway,due to the complex shape of C 1s signal,the contribution of organic residual of sol –gel process to the total carbon content can not be excluded and distinguished from the environmental contamination.

XPS depth profiling allowed calculation of the thickness of all the samples.The results of the depth profiling of the sample ZO2are shown in Fig.4.Considering that the ion sputtering rate at 2.0keV energy is about 0.1nm/min,total analyzed thickness amounts to nearly 30nm,while the coating is about 23nm thick.Fig.5reports XPS depth profile for the

sample

Fig.5.XPS depth profile for ZO3

sample.

Fig.6.Potential noise evolution for ZO1samples after 14and 15days of

immersion.

Fig.7.Current noise evolution for ZO1samples after 14and 15days of

immersion.

Fig.8.σv and σI for ZO3sample as a function of immersion time.

5825

G.Gusmano et al./Surface &Coatings Technology 201(2007)5822–5828

ZO3,with the coating thickness of about 32nm.The coating thickness in the sample ZI was estimated to be about 18nm,thus evidencing a negative influence of the use of inorganic precursors on the film thickness.3.3.Electrochemical results

Fig.6shows the potential noise evolution for the ZO1samples after 14and 15days of immersion.Given the experimental configuration,potential noise coincides with the open circuit potential (OCP)fluctuation.The analysis of noise signals in the time domain,allowed us to find out the initiation time of corrosion attack.Data acquired up to 14days of immersion are always characterized by small fluctuations.The fast decreasing and the irregular trend of the OCP after 15days,suggests the onset of a localised attack.Current noise acquisitions gave consistent indications with small fluctuations up to 14days and spikes starting from 15th day (Fig.7).

According to the literature data,for oxide-coated metal sys-tems,the corrosion processes should be controlled by the pene-tration of the electrolyte through the coating defects,causing the onset of localised attack [3,4,6,8].More clear and interesting information were obtained by the study of the standard deviations as a function of time.In Fig.8potential and current standard deviations (σv and σI respectively)as a function of immersion

time for ZO3samples are shown.In the potential signal,it can be observed the occurrence of some spikes during the first days of immersion.These spikes are followed by a flat trend until a strong increase occurred after 80days.As a consequence of this be-haviour,the current signal does not show any appreciable varia-tion during the first 60days of immersion.This behaviour can be explained in terms of a metastable pitting taking place in the first part of tests until a stable pitting propagation occurred after 80days.

In the case of sample ZI,although the potential standard de-viation varied in a very narrow range,a clear trend may be detec-ted,as shown in Fig.9A.During the first 60days of immersion,an almost constant value of σv was detected and thereafter a slight increase was observed.σI evidenced a more clear trend (Fig.9B)consistent with σv variation and very similar with that showed by ZO3,confirming the occurrence of a localised attack after 80days.

The analysis of σv and σI allowed the determination of the pitting onset time for all samples and they are summarized in Table 3.

Fluotitanate samples do not show stable corrosion phenom-ena,although potential fluctuations of higher amplitude were sometimes observed,as in the case of the three acquisitions shown in Fig.10,recorded during the 116th day after 30min from each

other.

Fig.9.σv and σI for ZI sample as a function of immersion time.

Table 3ENA results Label R n (Ohm)Onset time (days)Comments Blank 3.58·10418Uniform corrosion C 3.·10454Small pits F 8.06·104119No corrosion ZO1 4.75·10415Several pits ZO2 5.83·10444One deep pit ZO3 5.65·10476Several pits ZI

3.35·10463Several pits ZO1TC 6.83·106NA No corrosion ZO2TC

4.57·106NA No corrosion CTC 4.54·106NA No corrosion FTC

5.49·106

NA

No

corrosion

Fig.10.Potential noise acquisitions performed during the 116th day for F sample.

5826G.Gusmano et al./Surface &Coatings Technology 201(2007)5822–5828

In this stage the system was characterized by the increase of the potential mean value from −562mV to −553mV and the contemporary attenuation of the potential fluctuations with the increasing of immersion time.This behaviour can be explained with the formation of metastable pits.The current standard deviation plot evidences the days in which metastable pits took place (Fig.11).

A very useful parameter from a practical point of view is the Noise Resistance (R n )defined as the ratio of σv and σI and formally equivalent to the Polarization Resistance (R p )[16].R n was calculated for all samples.In particular,Fig.12shows the R n plot of the ZO1samples topcoated with organic coating.The multilayer coatings showed very high R n ,that means very good corrosion resistance from the initial immersion time,which improved gradually.Similar trend was recorded for all topcoated primer samples,and no corrosion attack was observed during more than 80days immersion tests.

Noise data were transposed in the frequency domain by means of FFT and MEM algorithms.Both the elaborations were significant even though MEM was preferred to FFT,because it produces smoother plots and the parameter determination is more accurate.

Fig.13reports a selection of current PSD plots by MEM algorithm.

It is known that,the power level is proportional to the kinetic rate of the involved reactions [17].The analysis in the frequency domain confirms the findings in the time domain.The power level increased during initiation and propagation of the attack.Moreover,the slope of PSD plot,estimated after linear trend removal of measured signals,was recognized in the main lite-rature as the most significant parameter to distinguish between different form of corrosion [18,19].From Fig.13is clear that,when the corrosive attack occurred,the current PSD slope dramatically increased and,as a consequence,the power level at the low frequency limit,increased.In particular for the ZO1sample the low frequency plateau increased from 3·10−16A 2/Hz to 2·10−15A 2/Hz and the slope increased more than 3dB/dec in the period between 14and 15days of test.

The increase was also greater for the samples ZO3.In fact,in the period between 75and 76days of immersion,the low fre-quency plateau raised from 2.5·10−17A 2/Hz to 2·10−15A 2/Hz with an increasing of the PSD slope more than 8dB/dec.

The results of the corrosion tests are summarized in Table 3.Amongst zirconia primers,ZO3sample showed the best corrosion resistance properties,even better than chromate reference samples.ZI samples also gave good results,comparable to the

chromate

Fig.12.R n for ZO1TC samples as a function of immersion

time.Fig.13.MEM power spectra of corrosive attack a)ZO1sample,b)ZO3

sample.

Fig.11.σI for F samples as a function of immersion time.

5827

G.Gusmano et al./Surface &Coatings Technology 201(2007)5822–5828

primers.Due to the additional passivation treatment performed after conversion,fluotitanate primers exhibited the best behaviour with no corrosion at all.As expected,blank tests showed the worst behaviour,with a uniform attack after18days of immersion. Topcoated specimens had an optimal corrosion resistance and no blisters were observed even for long term immersion.

4.Conclusions

The AFM analysis implies that,for all the samples,covering and structure of the coating were uniform.The grain size ranged between40and120nm for the organic precursors and between 120and250nm for the inorganic one.No clear correlation was found between roughness and deposition parameters.

From the XPS depth profiles,was determined the thickness of the coatings:23nm in the samples ZO1and ZO2,32nm in the sample ZO3and only18nm in the sample ZI.The chemical composition was always constant through the coating thickness.

Sol–gel primers increased the corrosion resistance of the alumi-nium AA1050.Amongst zirconia primers prepared by organic precursors,sample with two layers deposited and treated at250°C for4min had the best corrosion resistance.Zirconia primers prepared by inorganic precursors showed lower corrosion resistance feature.Both exhibited corrosion resistances comparable to chromates,but inferior to fluotinates.Corrosion resistance feature of fluotitanate primers was enhanced by a passivation stage at the end of preparation.The topcoated specimens gave an optimal corrosion resistance.References

[1]Directive2000/53/EC of18September2000,Official Journal of the

European Communities,21.10.2000,L269/34.

[2]Directive2002/95/EC of27January2003,Official Journal of the European

Union,13.2.2003,L37/19.

[3]X.F.Yang,D.E.Tallman,V.J.Gelling,G.P.Bierwagen,L.S.Kasten,J.Berg,

Surf.Coat.Technol.140(2001)44.

[4]E.Almeina,L.Fedrizzi,T.C.Diamantino,Surf.Coat.Technol.105(1998)

97.

[5]S.M.Cohen,Corrosion51(1995)71.

[6]L.Fedrizzi, F.J.Rodriguez,S.Rossi, F.Deflorian,R.Di Maggio,

Electochim.Acta46(2001)3715.

[7]S.Hirai,K.Shimakage,K.Wada,J.Am.Ceram.Soc12(1998)3087.

[8]M.Guglielmi,J.Sol–Gel Sci.Technol.8(1997)443.

[9]S.Sakka,J.Sol–Gel Sci Technol.37(2006)135.

[10]R.Di Maggio,L.Fedrizzi,S.Rossi,J.Adhes.Sci.Technol.15(2001)793.

[11]A.Conde,J.Damborenea,A.Durán,M.Menning,J.Sol–Gel Sci Technol.

37(2006)79.

[12]P.L.Neto,M.Atik,L.A.Avaca,M.A.Aegerter,J.Sol–Gel Sci.Technol1

(1994)177.

[13]Haibin Li,Kaiming Liang,Lefu Mei,Shouren Gu,Shuangxi Wang,J.Mat.

Sci.Letters20(2001)1081.

[14]F.Andreatta,P.Aldighieri,L.Paussa,R.Di Maggio,S.Rossi,L.Fedrizzi,

“Electrochemical behaviour of ZrO2pre-treatments on AA6060alloy”.

Proceedings of EMCR2006,9th Symposium Electrochemical Methods in Corrosion Research,Dourdan(Francia),18–20June2006.

[15]S.Kaciulis,G.Mattogno,Surf.Interface Anal.30(2000)502.

[16]Y.J.F.Chen,W.F.Bogaerts,Corros.Sci.37(1995)1839.

[17]G.Gusmano,G.Montesperelli,S.Pacetti, A.Petitti, A.D'Amico,

Corrosion53(1997)860.

[18]T.Fukuoka,T.Mizuno,Corros.Sci.38(1996)1085.

[19]K.Hladky,J.L.Dawson,Corros.Sci.22(1982)231.

5828G.Gusmano et al./Surface&Coatings Technology201(2007)5822–5828下载本文