Surface & Coatings Technology 200 (2006) 6368 6372 www.elsevier.

com/locate/surfcoat

Optimization of spray pyrolysis zirconia coatings on aluminized steel

R. Lpez Ibez, F. Martn, J.R. Ramos-Barrado, D. Leinen
Laboratorio de Materiales y Superficie (Unidad Asociada al CSIC), Dpto. Fsica Aplicada I and Dpto. Ingeniera Qumica, Facultad de Ciencias, Universidad de Mlaga, Mlaga E-29071, Spain Available online 13 December 2005

Abstract An exhaustive study of spray conditions is presented for ZrO2 thin films on hot-dip aluminized steel to be used as protective coatings against degradation. A fractional factorial design (24-1) has been applied to reduce the number of experiments and find out main and secondary effects of IV spray pyrolysis variables on growth rate. Factors, as substrate temperature, solution concentration, air pressure and precursor flow rate have been studied at two levels. Results from this work will be discussed by X-ray photoelectron and UVVISNIR spectroscopy and by scanning electron microscopy, finding great changes in substrate coverage, as well as optical and morphological properties for the analysed layers. A statistically significant influence of substrate temperature and precursor flow rate has been found for film growth. The co-optimization of both variables has conducted to a considerable reduction of deposition time, as industrial applications and process improvements require, being that the first step for a later up-scaling of these barrier coatings. 2005 Elsevier B.V. All rights reserved.
Keywords: Spray pyrolysis; Zirconium oxide thin film; Galvanized steel; ANOVA; Factorial design

1. Introduction Galvanized steels are frequently used for outdoor structures because of their good behaviour in humid environments. Nevertheless, degradation occurs after several years [1], and the need to increase its corrosion resistance in aggressive atmospheres, as well as recent restrictions for using chromates are demanding the development of new alternative coatings to protect galvanized steels [2]. Zirconium oxide thin films find many applications in different fields of interest, such as thermal barrier coatings, hard overcoats, catalyst support, optic devices or corrosion protection of metals [36], because of its good chemical stability, high melting point (2860 C), high refractive index and dielectric constant [7]. Many techniques have been employed to deposit zirconia layers on different substrates [8,9]. With regard to that spray pyrolysis is a very interesting option since it is suitable to cover large areas at low cost. The challenge is to obtain dense and homogenous thin films, which will be the first step to introduce this coating technique for industrial applica Corresponding author. Tel.: +34 9521 31928; fax: +34 9521 32382. E-mail address: [email protected] (D. Leinen). 0257-8972/$ - see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.11.017

tions. However, the presence of surface cracking continues to be the main problem for these overcoats, requiring a specific study of deposition conditions to obtain continuous, compact and dense zirconia layers. In a previous paper [10], a preliminary study of deposition conditions by spray pyrolysis was made for ZrO2 onto three different types of galvanized steels. It was found that aluminiumsilicon hot-dip galvanized steel substrates are suitable in obtaining good optical and morphological properties of the zirconia coatings, as well as a high growth rate and good adherence. Some cracks were detected in scanning micrographs, probably due to thermal stresses during the cooling step, or maybe due to temperature changes across the film while spraying. In this work, the aim is to test a broad range of spray conditions. A factorial design for experiments is used as an objective planification at laboratory for best parameter selection. Factorial design is usually required in industry for product quality improvement, but it is a simple and powerful method which allows estimating a variable effect related to the rest of working parameter values. Results from the factorial design can be statistically analysed by analysis of variance (ANOVA), assessing whether an increase or a decrease in a factor value help to obtain the desired properties [11].

R. Lpez Ibez et al. / Surface & Coatings Technology 200 (2006) 63686372 Table 1 Sample names and deposition conditions Sample name Zr 0000 Zr 1001 Zr 0101 Zr 1100 Zr 0011 Zr 1010 Zr 0110 Zr 1111 T (C) 200 300 200 300 200 300 200 300 C (mol/l) 5 10 2 5 10 2 10 1 10 1 5 10 2 5 10 2 10 1 10 1 p (bar) 1 1 1 1 1.5 1.5 1.5 1.5 Qp (ml/h) 25 50 50 25 50 25 25 50

6369

A stand-free, steel-made solar absorber to be constructed will be the application for these coatings as top coat (EU Project: SOLABS1). To guarantee a long-life for the unglazed collector, a barrier feature against steel degradation should be ensured by the zirconia layer. As this coating will be used in optical devices, adequate optical properties should be obtained too. To optimize these features, the Zr/Al ratio and the solar absorptance must be increased along the coating. Furthermore, a study on corrosion resistance of zirconia-coated galvanized steel by electrochemical techniques has been started, as well as by outdoor exposure in different climates (EU Project: SOLABS1). Those results will be presented in a different paper. 2. Experimental Zirconia layers have been prepared onto 40 mm 40 mm 0.88 mm size aluminized steel, provided by ThyssenKrupp Stahl (TKS), in a chemical spray station described in previous papers [10]. Galvanized sheets consist of low carbon steel (commercial DX 56/TKS) hot-dip coated with a 13 m layer of aluminium, silicon and iron, in an 87/10/3 atomic relation (FAL/TKS). The sprayed solutions have been prepared from a zirconium acetyl acetonate precursor from Aldrich, mixed with acetic acid to get a stabilized compound before dissolving in bi-distilled water. A factorial design has been used to identify the effects of main spray pyrolysis variables in film growth. Substrate temperature, solution concentration, air pressure and precursor flow rate have been chosen as the most important factors to be investigated in a spray pyrolysis system, studying their effects at two different symbolic levels: high (1) and low (0), while air flow rate (20 l/min), nozzle height (15 cm), deposition time (10 min) and substrate velocity under the nozzle have been kept constant for all the experiments. As the number of samples to analyse would become important, 4-1 a fraction for this design (2IV ) was proposed to reduce in a half the samples to prepare and still obtain the same results and conclusions. Sample names were given according to Table 1, indicating their deposition conditions. Coatings were prepared in random order to minimize experimental errors.

As-deposited zirconium oxide coatings have been analysed by X-ray photoemission spectroscopy (XPS) with a PHI 5700, using a 15 kV, 300 W, Mg K (1253.6 eV) X-ray source. For their in-depth analyses, a 4 keV Ar+ beam has been used to raster a 3 mm 3 mm surface area, analysing at surface and after 1, 3 and 6 min of sputtering, assuming a penetration rate of 3 nm/min for the oxide layer, as for the Ta2O5 reference under the same sputter conditions. High resolution photoelectron peaks for the C1s, O1s, Zr3d, Al2p and Si2p regions were registered, referring their binding energies to the Zr3d peak at 182.2 eV [12]. A Shimadzu UV-3100 spectrophotometer has been used for optical measurements in the UVVISNIR wavelength range (1902500 nm). Spectral reflectance, R(), and incoming solar radiation, IAM 1.5(), were used to calculate the solar absorptance (AM 1.5) of the oxide layers according to midEuropean latitudes (AM 1.5 standard): R 2:5 aAM
1:5

0:3 IAM 1:5 kd 1 Rkd dk : R 2:5 0:3 IAM 1:5 kd dk

Surface morphology of the coatings has been studied with a Jeol JSM-5300 scanning electron microscope (SEM). It was
Table 2 XPS Atomic concentrations and Zr/Al ratio during depth profile for the ZrO2 samples C Zr 0000 50.1 11.6 12.6 14.8 40.1 4.1 3.5 2.6 41.1 12.7 13.0 16.0 39.3 5.6 4.8 3.0 43.2 10.3 11.9 15.5 37.0 4.5 2.5 3.1 40.4 10.0 11.6 13.4 36.7 10.7 10.6 12.1 O 38.4 57.7 54.3 50.5 41.0 56.2 34.9 16.3 44.8 57.3 54.5 49.9 42.5 55.2 41.4 20.6 44.9 58.4 55.0 51.6 42.0 54.6 29.5 14.6 45.7 58.3 54.0 48.4 47.3 58.2 55.5 51.4 Zr 11.5 30.1 31.9 32.8 14.6 15.2 8.7 5.9 14.1 29.6 31.0 32.1 15.5 19.3 10.6 7.6 11.9 30.0 32.0 31.1 13.1 9.5 7.3 5.1 13.9 26.9 27.6 28.7 16.0 30.0 32.4 34.2 Al 0.4 0.9 1.5 3.7 21.9 36.8 47.4 0.2 1.0 1.4 2.0 17.6 32.0 45.4 0.5 0.6 1.1 6.5 27.1 40.6 47.6 3.8 5.0 6.3 0.7 0.8 1.3 Si 0.2 0.3 0.4 0.6 2.6 16.1 27.8 0.2 0.5 0.6 0.7 2.3 11.2 23.4 0.8 0.5 0.7 1.4 4.3 20.1 29.6 1.0 1.8 3.2 0.4 0.7 1.0 Zr/Al 75 35 22 3.9 0.7 0.2 0.1 148 31 23 7.8 1.1 0.3 0.2 60 53 28 2.0 0.4 0.2 0.1 7.1 5.5 4.6 43 41 26 Depth Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm Surface 3 nm 9 nm 18 nm

Zr 1001

Zr 0101

Zr 1100

Zr 0011

Zr 1010

Zr 0110

Zr 1111

http://www.solabs.net.

6370
Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

R. Lpez Ibez et al. / Surface & Coatings Technology 200 (2006) 63686372
284.8 eV

20000

0,8

c/s

15000 10000

C 1s Zr 0000
0,6

288.8 eV

282.0 eV

5000 0 295 100000 80000 290

Surface 1 min Ar+ 3 min Ar+ 184.6 eV 6 min Ar+

Reflectance

0,4
Substrate Zr 0000 Zr 1001 Zr 0101 Zr 1100 Zr 0011 Zr 1010 Zr 0110 Zr 1111

285
182.2 eV

280

Zr 3d Zr 0000

0,2

c/s

60000 40000 20000 0 190 80000 60000

Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

0,0 200

300

400

500

600

700

Wavelength (nm)
185 180
530.2 eV

15

Fig. 3. Diffuse reflectance spectra of zirconia coatings on the galvanized steel substrate in comparison to the substrate itself.

O 1s Zr 0000

40000
531.4 eV

was employed, with a CuK X-ray radiation source ( = 1.5406 nm). 3. Results and discussion

c/s

20000 0 540

535

530

525

520

Binding Energy (eV)

Fig. 1. C1s, Zr3d and O1s photoelectron spectra of the Zr 0000 sample at its surface and after 1, 3 and 6 minutes of 4 keV Ar+ sputtering.

not necessary to overcoat the samples with a gold film by sputtering. For the crystalline structure analysis (XRD), a Siemens D-5000 BraggBrentano geometry diffractometer
20000 15000 10000 5000 0 100000 80000 295 290 285 280

The XPS analysis for the as-deposited thin films revealed a soft carbon film, present on all samples, due to atmospheric contamination because of the ex situ analysis. After 1 min of sputtering this contamination layer was removed, and the real amount of carbon in the coatings could be estimated. Table 2 shows the XPS atomic concentrations for layers at the surface and after 1, 3 and 6 min of Ar+ sputtering. As can be seen in Table 2, samples Zr 0000, Zr 0101, Zr 0011, Zr 0110 and Zr
Surface 1 min Ar+ 3 min Ar+ 6 min Ar+
20000

Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

c/s

C 1s Zr 1100

Al 2p Zr 1100 15000
10000 5000 0

80

75

70

c/s

60000 40000 20000 0 190 80000 60000

Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

9000 6000 3000 0

185

180

175

105

100

95

Binding Energy (eV)

Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

c/s

O 1s Zr 1100

40000 20000 0 540

535

530

525

520

Binding Energy (eV)

Fig. 2. C1s, Zr3d, O1s, Al2p and Si2p photoelectron spectra of the Zr 1100 sample at its surface and after 1, 3 and 6 min of 4 keV Ar+ sputtering.

c/s

Zr 3d Zr 1100

Surface 1 min Ar+ 3 min Ar+ 6 min Ar+

Si 2p 12000 Zr 1100

c/s

R. Lpez Ibez et al. / Surface & Coatings Technology 200 (2006) 63686372 Table 3 Solar absorption coefficient for substrate and as-deposited ZrO2 thin films AM Substrate Zr 0000 Zr 1001 Zr 0101 Zr 1100 Zr 0011 Zr 1010 Zr 0110 Zr 1111 0.40 0.48 0.39 0.46 0.39 0.49 0.40 0.41 0.48
1.5

6371

Table 4 Results of ANOVA test for Zr/Al ratio Variable Effect estimate T 33.3 C 3.9 p 5.9 25.9 Qp Residual Sum of squares 2212 31 69 1343 1734 Degrees of freedom 1 1 1 1 3 Mean square 2212 31 69 1343 578 F 3.8 0.05 0.1 2.3

1111 show the best substrate coverage, for these coatings a homogeneous oxide thin film with at least 20 nm thickness was obtained. Binding energies (BE) of photoelectron peaks Zr3d (182.2 eV) and O1s (530.2 eV) indicate a ZrO2 film growth during pyrolysis (Fig. 1). At the same time, some carbon from the precursor solution was trapped inside the layer, reacting with zirconium during pyrolysis, stabilizing a new metallic carbide phase (C1s BE at 282 eV). For samples Zr 1001, Zr 1100 and Zr 1010 a high amount of aluminium and silicon from the galvanized layer of the steel was detected. These coatings are non-uniform and extremely thin, less than 510 nm. The

chemical state of aluminium changed from oxide (Al2p BE at 74.3 eV) to metal (71.8 eV) at increasing depth (Fig. 2), while silicon was always found in metallic state throughout the coating (Si2p BE at 98.3 eV). The small amounts of oxidized zirconium (Zr3d BE at 182.2 eV) almost disappeared after 3 min of sputtering being reduced to the metallic form. The oxygen photoelectron peak includes contributions to both zirconium and aluminium oxides (O1s BE at 530.2 eV and 531.0 eV). The XPS results were in accordance with UVVISNIR reflectance spectroscopy measurements (Fig. 3). All the coatings were highly transparent, and the presence of thicker zirconium oxide layers can be observed as a deep decrease in the diffuse reflectance spectra at UV wavelengths. Galvanized steel absorptance (AM 1.5) was therefore increased by the oxide layer, which acts as an antireflection layer, as seen in Table 3. A non-stoichiometric zirconium oxide film was assumed since a doublet absorption band appeared which we relate to Zr4+ ( 230 nm) and Zr3+ ( 300 nm) states. The XRD patterns of the as-deposited films showed an amorphous structure in all cases. The measured atomic Zr/Al ratio and the solar absorptance of the samples were the responses to study by an ANOVA test. Considering film thickness (Zr/Al ratio), results showed a significant influence of substrate temperature. A negative effect of substrate temperature and a positive effect of precursor flow rate on growth rate were found (Table 4). The estimated effects of solution concentration and air pressure were less important, even when compared with estimated effects of two-factor interactions. Due to the design resolution, no main effects are confounded with any other main effect or two-factor interaction, but two-factor interactions are confounded with another one. Once a multi-variable regression is made, estimated the Zr/Al

Substrate

Zr 0011

Zr 0000

Zr 1100

Fig. 4. SEM images of substrate, Zr 0000, Zr 1100 and Zr 0011 samples. The black bar corresponds to 10 micrometers.

6372

R. Lpez Ibez et al. / Surface & Coatings Technology 200 (2006) 63686372

Table 5 Results of ANOVA test for solar absorptance Variable Effect estimate T 4.5 C 0.5 p 1.5 Qp 3.5 Residual Sum of squares 40.5 0.5 4.5 24.5 65.5 Degrees of freedom 1 1 1 1 3 Mean square 40.5 0.5 4.5 24.5 21.8 F

assumed that Zr 0011 sample conditions are the best to fulfill optical properties. 4. Conclusions
1.9 0.02 0.2 1.1

ratio follow the model given below, where each factor can take any value in [ 1,1]: Zr=Al 26:3 33:3dT 25:9dQp 14:8dT dCpdQp 24:3dCdpT dQp : Thus we can conclude that for a maximum thickness, substrate temperature should be kept low, while precursor flow rate should be high. In that case, precursor solution concentration should be also at low level and air pressure at high, being Zr 0011 best conditions for homogeneous film growth. Nevertheless, in SEM images (Fig. 4) some cracks still can be seen for the estimated best deposition condition. This may be due to a temperature gradient across the coating during deposition or to a fast cooling step. With respect to the solar absorptance measured for the asdeposited layers, ANOVA results from the fractional factorial design showed no significant influence of any factor. However, a negative effect of substrate temperature and a positive effect of precursor solution flow rate on solar absorptance were found (Table 5), following the model given below after a multivariable regression, where each factor can take any value in [ 1,1]: a % 43:8 2:3dT 1:8dQp 2:3dT dCpdQp 1:8dT dpdCdQp : As two-factor interactions are confounded with another one for this fractional design, low level of substrate temperature should be combined with low level of precursor solution concentration to reach highest absorptance. In the same way, for precursor solution flow rate at a high level, air pressure should be as well at its high level for the same result. It can be then

Zirconium oxide thin films have been obtained onto aluminized steel by spray pyrolysis. Optical properties have been improved by the oxide layers, while acting as a physical barrier coating for steel. Best spray conditions have been studied with help of an ANOVA analysis, setting substrate temperature as the main variable to control during deposition. Low substrate temperature and high precursor flow rate have been found the main effects on substrate coverage and solar absorptance. Acknowledgements Funds from the EU (project SOLABS: ENK6-CT200200679) are gratefully acknowledged. References
[1] ThyssenKrupp Stahl: Hot-dip coated sheet, Booklet, order no. 2050 (2001). [2] S. Dalbin, G. Maurin, R.P. Nogueira, J. Persello, N. Pommier, Surf. Coat. Technol. 194 (2004) 363. [3] X.Q. Cao, R. Vassen, D. Stoever, J. Eur. Ceram. Soc. 24 (2004) 1. [4] A.V. Emeline, G.N. Kuzmin, L.L. Basov, N. Serpone, J. Photochem. Photobiol., A 174 (2005) 214. [5] G. Xu, P. Jin, M. Tazawa, K. Yoshimura, Sol. Energy Mater. Sol. Cells 83 (2004) 29. [6] J.P. Holgado, M. Prez-Snchez, F. Yubero, J.P. Espins, A.R. GonzlezElipe, Surf. Coat. Technol. 151152 (2002) 449. [7] S. Venkataraj, O. Kappertz, Ch. Liesch, R. Detemple, R. Jayavel, M. Wuttig, Vacuum 75 (2004) 7. [8] P. Peshev, I. Stambolova, S. Vassilev, P. Stefanov, V. Blaskov, K. Starbova, N. Starbov, Mater. Sci. Eng., B 97 (2003) 106. [9] Z.W. Zhao, B.K. Tay, L. Huang, G.Q. Yu, J. Phys., D Appl. Phys. 37 (2004) 1701. [10] R. Lpez Ibez, J.R. Ramos Barrado, F. Martn, F. Brucker, D. Leinen, Surf. Coat. Technol. 188189 (2004) 675. [11] R.H. Myers, D.C. Montgomery, Response Surface Methodology, 2 ed., Wiley-Interscience, 2002. [12] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain, R.C. King (Eds.), Handbook of X-ray Photoelectron Spectroscopy, PHI, 1995.