j ourna l homepage: www.e lsev ie r .com/ locate /sur fcoat
Effect of surface composition and microstructure of aluminised steel onthe formation of a titanium-based conversion layer
Ine Schoukens a, Isabelle Vandendael a, Joost De Strycker b, Adli A. Saleh c, Herman Terryn a, Iris De Graeve a,⁎a Vrije Universiteit Brussel (VUB), Electrochemical and Surface Engineering (SURF), Pleinlaan 2, 1050 Brussels, Belgiumb OCAS, Pres. J. F. Kennedylaan 3, 9060 Zelzate, Belgiumc Arab American University, Physics Department, 240 Jenin, State of Palestine
The presented work aims at determining the influence of surface composition and microstructure of hot dipaluminium–silicon coated steel on the formation characteristics and mechanism of a titanium-based conversionlayer. Varying the amount of silicon in the molten aluminium bath changes the surface nature which in turnaffects the subsequent pre-treatment process. Different phases in the layer formation process were identifiedin the open circuit potential evolution as a function of immersion time in the conversion solution. The homoge-neity and the thickness of the conversion layer in each of these regions were evaluated using secondary electronimages, Auger elemental surface maps and depth profiles. While intermetallic precipitates (FeAl3) act asnucleation centres in pure Al coatings, eutectic Si is responsible for the deposition process in a Si-rich coating.A continuous surface conversion layer is formed after 45 s of immersion in the conversion bath. Also, thepresence of silicon in the coating leads to the deposition of a thicker conversion layer. A model for build-up ofthe conversion layer is proposed based on these observations.
The use of steel is widespread throughout several industries includ-ing transport and general engineering & construction. Often galvanisedsteel is used for improved corrosion protection, as the zinc coating actsas a sacrificial anode in case a defect exposes the steel to the surround-ing environment [1]. However, due to volatile zinc prices, high levels ofleached zinc in the environmental waters and scarcity issues, steel com-panies have developed alternative products like hot dip aluminiumcoatings on steel. Different types of aluminised steel with variable prop-erties are commercially available [2]. In the present work two types ofcoatings were studied: a relatively pure aluminium coating (containingsome iron) and aluminium–silicon coatings (with a variable Si content).Their good thermal resistivity leads to such coatings being used inpower generation plants, exhaust systems and heating equipment [2].Aluminium–silicon coated steel (with a high silicon content) is also in-vestigated as an alternative for rolled electrical steel used, for example,as core metal in electrical motors [3,4].
During the aluminising hot dipping process, several parameters canbe varied: dipping time, composition of the aluminium bath, cooling-down procedure, etc. All these parameters as well as the characteristicsof the steel substrate have an influence on the composition and mor-phology of the coating [5,6]. Certain studies focused on the formationand characteristics of the interdiffusion layer that is formed at the
26293200.
ghts reserved.
, Surf. Coat. Technol. (2013), h
metal/coating interface underneath the so-called free aluminium layer[7–11]. This interdiffusion layer is a brittle intermetallic alloy compoundconsisting of FeAl3 in the upper part and Fe2Al5 in the lower part of thelayer [12]. Mechanical properties require the interdiffusion layer to beas thin as possible, which can be achieved by adding a certain amountof silicon to the molten aluminium-bath [13]. The addition of siliconalso changes the composition of the sub-surface layers by the formationof Al–Fe–Si ternary compounds [14], or a eutectic silicon structure in thefree aluminium layer, as was recently reported by the authors [15].
Many applications need the aluminised steel to be treated further,e.g. when a higher corrosion protection and/or a coloured finish arerequired. To achieve this, the metal surface is usually pretreated witha conversion layer that improves the adherence of thefinal organic coat-ing. Studies of conversion processes for different aluminium alloys, haveobserved that the cathodic behaviour of the precipitates versus the sur-rounding matrix is essential for the deposition of the conversion layer.Lunder et al. [16] and Nordlien et al. [17] studied the titanium deposi-tion mechanism on aluminium, showing that the essential processesfor the film formation were the oxygen reduction and hydrogen gasevolution taking place at cathodic precipitates. These processes lead toa local pH increase at themetal surface causing the precipitation of spe-cies from the conversion solution to form a thin solid oxide conversionlayer. Nordlien et al. [17] also reported that thedeposition of the conver-sion film around the precipitates in turn reduces the cathodic activity ofthe precipitates and as a consequence slows down further deposition ofthe layer. Andreatta et al. [18] studied the influence of the precipitates inan AA6016 alloy on the (Zr/Ti)-conversion process by investigating the
2 I. Schoukens et al. / Surface & Coatings Technology xxx (2013) xxx–xxx
Volta potential difference between the precipitates and the surroundingaluminiummatrix after different immersion times in the conversion so-lution. The Volta potential difference between the matrix and the pre-cipitates was seen to decrease at the start of the conversion layerformation reaching a stable and uniform value over the entire surfacewhen it is completely covered.
While the conversion processes of aluminium, galvanised steel andbare steel have been studied in detail, this is not the case for aluminisedsteel, where the complex metal surface composition and structure areexpected to have a strong influence on the ongoing electrochemicalreactions. This work studies the influence of the presence of silicon inan aluminium coating on a titanium based conversion layer formationprocess. Particular attention is paid to the influence of the silicon pres-ence in the aluminium coatings on the time taken for the surface ofthe aluminium coating to be completely covered by the conversionlayer and its final thickness. Finally, a model is proposed to explain thestructural and compositional nature of the conversion layer.
2. Experimental
2.1. Hot dip aluminising process
The steel substrate used was a 0.8 mm thick, vapour degreased,interstitial free DC06 steel (composition indicated in Table 1) with anelectron beam textured surface provided by OCAS-ArcelorMittal. Thesteel was aluminised on a laboratory scale using a RHESCA hot dip sim-ulator to imitate the industrial process. During the whole dipping pro-cess a reducing N2 (95%)–H2 (5%) atmosphere was used to preventthe oxidation of the steel substrate during the annealing step. The sam-ple was first heated up to 800 °C at a rate of 4 °C/s and was maintainedat this temperature for 75 s. This annealing step was necessary to im-prove the homogeneity of the coating. Subsequently the steel wascooled down to 700 °C at a rate of 10 °C/s and was kept at this temper-ature for 10 s. The substratewas then dipped for 3 s in amolten alumin-ium bath at 660 °C. Two different types of aluminium based coatingswere applied, starting from an aluminium bath containing 2 wt.% Feand from aluminium–silicon baths containing 2 wt.% Fe and 1, 3 or7 wt.% Si. The corresponding samples were designated as silicon-free,Si-1 wt.%, Si-3 wt.% and Si-7 wt.% respectively. 2 wt.% of iron wasadded to the bath to saturate it with Fe, in order to avoid the dissolutionof iron from the steel into the bath that could change the bath composi-tion when dipping a series of plates. No rinsing was applied when thesample was removed from the bath to control the coating thickness.Samples were put back in the sample chamber and cooled down toapproximately 50 °C at a rate of 10 °C/s using a pure nitrogen flow(400 l/min).
2.2. The conversion treatment
Prior to the conversion treatments, all samples were ultrasonicallycleaned for 10 min in ethanol. The titanium conversion stock solutionused was a Granodine 1445 T solution provided by Henkel, containingdihydrogen hexafluorotitanate, phosphoric acid and organic compounds.The conversion procedure consisted of immersing the samples in a 3 wt.%Granodine 1445 T solution to which 0.250 g of ammonium nitrate per500 ml solution was added as an accelerator. Various samples weredipped into the non-stirred solution at room temperature for 45, 105and 300 s. The immersion of the samples was pneumatically controlledby the homemade set-up illustrated in Fig. 1. After the treatment, the
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
samples were rinsed with distilled water and dried with pressurised air.The OCP of the samples was followed during the whole conversion treat-ment using a Ag/AgCl reference electrode and a measuring interval of0.1 s.
The Si-3 wt.% and Si-7 wt.% coatings had a eutectic silicon structureat the surface (see Section 3.1), which could be visualised better usingan extra pretreatment step consisting of a potentiostatic polarisationat −1.6 V versus SCE during 5 min in a 1 M NaCl-solution. After thepotentiostatic polarisation, the sample was rinsed with distilled waterand dried using compressed air. This pretreatment exposed the eutecticstructuremore clearly at the surface. Thiswas done to study the effect ofthe eutectic structure on the initiation of the conversion layer. Specifi-cally, this pretreatment was only applied to study the initiation step ofthe conversion process with FE-SEM-EDX.
2.3. Characterisation of the samples
A Jeol JSM-7000F FE-SEM in combination with the Jeol Hyper Mini-cup EDX detector was used to characterise the surface before and afterthe conversion treatment. The characterisation of the surface, prior tothe application of the conversion layer, consisted of identifying the in-termetallics present at the surface by obtaining Backscattered ElectronImages (BEI), EDX-spectra and X-ray compositional maps of Al, Fe andSi using an electron beam energy of 15 keV. After the conversion treat-ment, Secondary Electron Images (SEI) of the surfaces were recorded atamagnification of 5000 times andwith an accelerating voltage of 10 kV.These images were combined with elemental analysis by EDX.
Additionally, Auger elemental surface maps and depth profiles wereacquired after the conversion treatment with the Jeol JAMP-9500Fscanning Auger microprobe. The surface maps were obtained with aprimary electron beam of 10 keV and 5 nA and a resolution of 512pixels by 512 pixels. The value used to establish the colour scale in themapping is the {peak-background/background} intensity for eachelement. Prior to the acquisition of the maps, the surface was sputteredfor 2 min with argon ions at a rate of around 4.3 nm/min, where therate was calibrated on SiO2. This step was applied to remove the surfacecarbon contamination.
For the elemental depth profiles, a primary electron beam of 10 keV&5 nAand argon ion sputteringwere used. The analysed areawas 24 by24 μm2 and a sputter rate of 4.3 nm/min was used, calibrated on SiO2.The sputtering time between two adjacent points of the depth profilewas 30 s for conversion times of 45 and 105 s. For the conversiontime of 300 s a sputter interval of 60 s was used. All Auger spectrawere evaluated in the direct mode.
3. Results
3.1. The surface microstructure
The BEI of the planar surfaces of the silicon-free, Si-1 wt.% and Si-7 wt.% coatings are shown in Fig. 2. A comparison between the images,whichwere taken at the samemagnification of 2000, clearly shows thatthe amount of silicon in the molten bath has an important influence onthe surface structure.
At the surface of the silicon-free coating, a dense distribution of pre-cipitates (about 2 μm in size) was observed as shown in Fig. 2(a). EDXanalysis revealed that these precipitates contain approximately27 wt.% Fe and are embedded in the pure aluminium matrix. Based onthe Fe–Al phase diagram, the precipitates were identified as FeAl3[14]. The addition of 1 wt.% of Si to the bath, resulted in precipitates(1.5 to 3.5 μm in size) embedded in the Al-matrix (see Fig. 2(b)). EDXanalysis showed that these particles contained around 30 wt.% Fe and3 wt.% Si, corresponding to the θ-phase in the Fe–Al–Si phase diagram.These results were in good agreement with the results obtained byEBSD on cross sections of the coatings (measurements shown in [19]).When more silicon was added to the molten bath (3 wt.% and 7 wt.%)
the surface composition was completely transformed as shown inFig. 2(c). X-ray maps of the coating constituents revealed the presenceof silicon in a eutectic structure with aluminium as shown in Fig. 3 forthe Si-7 wt.% coating. Besides the eutectic structure, some FeAl3 precip-itates could be identified at the surface of the Si-3 wt.% coating, butwere no longer visible at the surface of the Si-7 wt.% coating.
3.2. The evolution of the OCP during the conversion treatment
Fig. 4 shows the OCP as a function of time during the immersion ofthe different aluminium coatings in the conversion solution for 300 s.For the silicon-free coating, the OCP decreases until a short plateau isreached, as indicated by a slight change in the slope after a few seconds.After this short plateau, the OCP decreases further until a minimum isreached after approximately 80 s. Beyond this minimum, the OCP in-creases to a stable value of around −0.84 V vs. Ag/AgCl reached afterapproximately 210 s.
Fig. 2. BEI of the surface of the silicon-free (a)
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
For aluminium–silicon coatings, these different regions could still beidentified, but the stable OCP was reached 1 min earlier than for thecoating without silicon, and the OCP values shifted towards nobler po-tentials (between −0.78 and −0.79 V vs. Ag/AgCl).
3.3. The homogeneity of the conversion layer
Literature on the conversion mechanism suggests that the depositionof the titanium layer initiates around the cathodic FeAl3-precipitates pres-ent on the surface of the silicon-free and Si-1 wt.% coatings. The SEI of thesurface of the silicon-free coating after different immersion times in theconversion solution are shown in Fig. 5(a). These images combinedwith the EDX analysis performed after 45 s of immersion show that tita-nium is only present around the precipitates. After 105 s, the depositionof titanium is seen to spread out further over the surface with the precip-itates being growth centres. After 300 s, the surface, including theprecipitates is completely covered. Similar results have been obtainedfor the Si-1 wt.% aluminised steel coating.
Fig. 3. X-ray maps of Al (b), Fe (c) and Si (d) at the surface of the Si-7 wt.% coating together with the SEI of the scanned area (a).
4 I. Schoukens et al. / Surface & Coatings Technology xxx (2013) xxx–xxx
Fig. 6 shows Auger surface maps of the various atomic constituentsof the Si-1 wt.% coating surface after 45 s of immersion. These mapsclearly show that a homogeneous layer of titanium was deposited onthe surface of the coating, including the precipitates, after 45 s of im-mersion. This was not seen in the EDX analyses, suggesting that thelayer was too thin after that short immersion time to be seen withEDX. Also oxygen and phosphorus were homogeneously distributedover the surface, while very low levels of carbon were detected.
An Al–Si eutectic structure is present at the surface for the Si-3 wt.%and Si-7 wt.% coatings. Based on the SKPFMmeasurements, whichwerecarried out to study the electrochemical behaviour of the coatings as afunction of depth [19], the deposition of the conversion layer isexpected to start around the eutectic silicon. The SEI shown in
Fig. 4. The OCP versus time recorded for the different coatings during immersion in theconversion solution.
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
Fig. 5(b) and EDX measurements taken after 45 s of immersion in theconversion solution, indeed show that the titanium deposition initiatesaround the cathodic eutectic silicon and then spreads out from thereover the surface until complete coverage is attained. Auger maps ofthe surfacewere also obtained after 45 s of immersion in the conversionsolution. Both coatings were completely covered by the titanium layer.
3.4. The thickness of the conversion layer
Fig. 7 shows the AES elemental depth profiles of aluminium, oxygen,titanium, phosphorus and carbon recorded for the conversion layers onthe silicon-free coating after 45, 105 and 300 s of immersion in the con-version solution. For the sample that had undergone 45 s of immersion,the depth profile exhibits titanium, oxygen and phosphorus intensitiesthat diminish after a few nanometers indicating the formation of athin conversion layer. The aluminium intensity shows the opposite evo-lution: at the beginning of the profile no aluminium is present, but theintensity of aluminium increases until a maximum plateau is reached.Finally, carbon has a high intensity near the surface, but then decreaseswith probed depth. It is assumed that contamination influences the in-tensity of carbon at the outer surface, but carbon is also clearly detectedinside the conversion layerwhichmeans that the organic compounds inthe solution are incorporated into the conversion layer. Longer immer-sion times in the conversion solution lead to increased layer thickness asshown in Fig. 7.
To get a better overview of the thickness of the conversion layersafter the different immersion times in the conversion bath, the thick-ness of the conversion layer and the virtual interface width, which is ameasure of the depth resolution of the profile, were calculated basedon the Auger depth profiles of aluminium and the method of De Laetet al. [21] (Table 2). The table shows that after 45 s, a conversion layerof approximately 15 nm forms on the silicon-free coating, while onthe Si-1 wt.% and Si-7 wt.% coatings, the layers are around 30 nmthick. After 105 s of immersion the thickness of the conversion layeron the different coatings increases to 35 nm, 50 nm and 95 nm,
Fig. 6. Auger surface maps of titanium (b), phosphorus (c), oxygen (d) and carbon (e) recorded on the Si-1 wt.% coating after 45 s of immersion together with the SEI of the scanned area (a).
6 I. Schoukens et al. / Surface & Coatings Technology xxx (2013) xxx–xxx
respectively. Finally, after 300 s of immersion, the thickness reaches50 nm, 125 nm and 115 nm, respectively.
Based on these results, a clear influence of the presence of siliconin the aluminium coating is visible. When comparing the thicknessof the conversion layer formed on the silicon-free coating and theSi-1 wt.% coating, it is seen that the thickness of the conversionlayer is doubled when silicon is present in the coating. The presenceof more silicon (i.e. for the Si-7 wt.% coating) in the coating influ-ences the thickness of the deposited conversion layer only after animmersion time of 105 s. At this immersion time, the thickness ofthe conversion layer is doubled when more silicon is present in
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
the coating. For the other 2 immersion times (45 s and 300 s), thethickness values of the conversion layers formed on the Si-1 wt.%and Si-7 wt.% coatings are comparable.
Table 2 observes that the virtual interface width increases withlayer thickness. This is in good agreement with the findings ofDe Laet et al. [21]. Thicker conversion layers require longersputter times, causing loss of resolution due to sputter-inducedroughness of the plane of analysis. Even for the thinnest conversionlayer (15 nm), the virtual interface width is quite large (10 nm).This is due to the significant initial roughness of the analysedsamples.
Fig. 7. Auger electron depth profiles of the conversion layers deposited on the silicon-free coating after 45 s (a), 105 s (b) and 300 s (c) of immersion.
3.5. In-depth analysis of the conversion layer composition
Fig. 8 shows the evolution of the Auger peaks of the different ele-ments as a function of depth acquired for the conversion layer depositedon the Si-1 wt.% coating after 300 s of immersion in the solution. Thepeak positions have been calibrated based on the position of the peakof metallic aluminium [22]. The changes in the chemical state of the el-ements can be derived by comparing the peaks at different depths.
For titanium, two characteristic Auger electron transitions are visibleat depths of approximately 65 nm and 115 nm: one at a kinetic energyof 381 eV and a second at a kinetic energy of 416 eV. Although the peakat 381 eV has a higher intensity than the peak at 416 eV, the chemicalstate of titanium is studied based on the peak occurring at 416 eV.This peak is chosen since it has been shown in literature that it is uniquefor titanium,while the peak at 381 eV can be influencedby thepresenceof nitrogen, whichwas present in the conversion solution as an acceler-ator [23]. For oxygen, two characteristic Auger peaks are visible at kinet-ic energies of 489 eV and 511 eV. The latter peak is used to study thechemical state of oxygen as a function of depth, since it has the highestintensity.
4. Discussion
4.1. OCP evolution during the conversion treatment
Andreatta et al. observed a similar OCP-evolution for an AA6016alloy during a zirconium–titanium-based conversion process; they de-scribed the different stages of the film formation as follows [18]. Thefirst regionwhere theOCPdecreased rapidly to aminimum is associatedwith the activation of the aluminium and the initiation of the film depo-sition. Unlike the observations in the present study, Andreatta et al. didnot report a (short) plateau region in this first stage of the film forma-tion. In the second region where the OCP increased again, the conver-sion layer was seen to develop and spread out over the surface. Lastly,the OCP reached a final stable plateau and the surface was completely
Table 2Thickness values and virtual interface widths after 45 s (a), 105 s (b) and 300 s (c)of immersion in the conversion solution.
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
covered. Campestrini et al. on the other hand also observed a short pla-teau in the first region of the OCP-curvewhen studying a chromate con-version process on AA2024 [20]. They found that the activation of thesurface occurred in the period before the short plateau, while the firstdeposition of the conversion layer started after the short plateau wasreached.
4.2. Influence of the surface structure on the conversion mechanism
The results obtained with FE-SEM and FE-AES clearly suggest thatthe amount of silicon present in the aluminium coating influences thegrowth rate of the conversion layer and consequently its thickness.Combining the EDX and Auger analysis results concludes that a contin-uous Ti conversion layer is formed on the silicon-free and the Si–Al coat-ings after 45 s of immersion in the conversion solution, although thislayer is thicker around the FeAl3-precipitates. Literature [16–18,27,28]shows that the deposition of the conversion layer is strongly dependenton the rate of the cathodic reactions, i.e. hydrogen evolution and oxygenreduction, taking place at precipitates which are nobler than thematrix.In the case of the aluminium coatings, these reactions take place at theFeAl3-precipitates in the silicon-free, the Si-1 wt.% and Si-3 wt.% coat-ings, and at the eutectic silicon in the Si-3 wt.% and Si-7 wt.% coatings.Although the OCP data (see Fig. 4) indicate complete surface coverageafter approximately 200 s for the silicon-free coating and 150 s for theAl–Si coatings, the Auger maps reveal a continuous Ti layer on all stud-ied surfaces after only 45 s of immersion. The identification of the differ-ent stages in the OCP-evolution found in literature, as discussed above,is typically based on SEM-EDX data. Based on the Auger resultspresented here for the first time, the link with the different regions inthe OCP-curve, defined in literature, needs to be reconsidered.
Besides, a clear influence of the presence of silicon in the aluminiumcoating was seen on the thickness of the layer (Table 2). For every im-mersion time, it was observed that the thickness values of the conver-sion layers doubled when silicon was added to the aluminium coating.For the aluminium coatings with silicon, the thickness values after45 s and 300 s of immersion were comparable. After an immersiontimeof 105 s the thickness of the conversion layer on the Si-7 wt.% coat-ing was approximately double that on the Si-1 wt.% coating.
Nordlien et al. [1] showed that intermetallics covered by the conver-sion layer hinder further deposition of titanium since the rate of the es-sential reduction processes rapidly decreases and as a result theconversion process is self-extinguishing. This would mean that, if thecathodic processes on the silicon-free coating take place at a higherrate, the conversion layer covers the surface faster, including the precip-itates. As a consequence the rate of the titanium deposition would de-crease earlier than on the silicon-containing coatings. This couldexplain the difference in thickness of the conversion layer on the alu-minium coatingswith andwithout silicon. This theory should be furtherinvestigated by recording sputter depth profiles both on a precipitateand/or the eutectic structure and next to the precipitates and/or the eu-tectic structure for all coatings.
Fig. 8. Auger peaks for Ti (a), Al (b), C (c), O (d) and P (e) acquired at depths of 65 nm, 115 nm and 380 nm for the Si-1 wt.% coating after 300 s of immersion in the conversion solution.
8 I. Schoukens et al. / Surface & Coatings Technology xxx (2013) xxx–xxx
4.3. Model for the conversion layer build-up
Based on the presented AES results, the literature on titanium con-version layers and the composition of the conversion solution, a firstsuggestion for the conversion layer build-up can be made. At a depthof 65 nm, titanium, oxygen, phosphorus and carbon were detected.The peak of titanium at a kinetic energy of 416 eV suggests that titani-um is present in its metallic form [23,24], but the intensity profiles oftitanium and phosphorus are very similar, suggesting that titaniumand phosphorus are present as one compound. Besides, phosphoruswas present in the solution as phosphoric acid. Wilson et al. [25] useda similar kind of conversion solution containing phosphoric acid as aconversion treatment for a zinc alloy and observed that phosphoruswas present in the conversion layer as a phosphate. This suggests thattitanium, oxygen and phosphorus are present as a titanium-phosphatecomplex.
Finally, carbon was present in the solution as an organic compoundthat is often added to a conversion solution to enhance the corrosionresistance or the adhesion abilities of the conversion coating. As the
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
evolution of the carbon intensity was not comparable with the evolu-tion of the other elements it is expected that carbon was still presentas an organic compound.
At the second depth, i.e. around 115 nm, the intensity of the peaks oftitanium, phosphorus and carbon decreased but did not shift. On theother hand, the intensity of oxygen remained high but its peak shiftedto a kinetic energy of 507 eV. Additionally, an aluminiumpeak appearedat a kinetic energy of 1387 eV. Based on literature this wouldmean thataluminium is present as an oxide [22,23,26]. The titanium-phosphatecomplex and organic compound that were already found at a depth of65 nmwere still present at the depth of 115 nm.
Finally, no titanium, phosphorus, oxygen and carbon were detectedat a depth of approximately 380 nm. This means that the compoundsfound in the two first regions are not present any more. Nevertheless,two Auger peaks of aluminium were observed: one at a kinetic energyof 1378 eV and one at a kinetic energy of 1392 eV. Literature hasshown that the first peak corresponds to the first bulk plasmon losspeak of aluminium and the second one to the KL2,3L2,3 line of metallicaluminium [22].
Based on the various observations, the conversion layer can be dividedinto two regions, as illustrated in Fig. 9. In the first outer region an organiccompound is present together with a titanium-phosphate complex. Thesecond region can be seen as an interface region between the conversionlayer and the aluminium coating; here aluminium oxide is present to-getherwith the titanium-phosphate complex and the organic compound.At a depth of 380 nm, the free aluminium layer of the coating is alreadyseen to appear. The microstructure of the aluminium coating did not in-fluence the build-up of this layer: for the different aluminium coatingsthe same compounds were found in the conversion layer.
5. Conclusions
The influence of surface composition and microstructure of hot dipaluminium coated steel on a titanium based conversion treatment wasstudied. The OCP of the aluminised steel was recorded as a function ofthe immersion time to study the formation of the conversion layer. Sec-ondary electron images and Auger elemental surface maps and depthprofiles were also recorded after immersion times of 45, 105 and 300 s.
The recordedAuger titaniummaps showed a continuous titaniumdis-tribution all over the surface of all studied coatings after already 45 s ofimmersion in the conversion bath. SEI combined with EDX analysis indi-cated that the conversion layer is thicker at the precipitates for the silicon-free, the Si-1 wt.% and the Si-3 wt.% and at the eutectic Si structure for theSi-3 wt.% and Si-7 wt.% coatings. This confirms the formation of the con-version layer initiates at the precipitates and/or eutectic structure, butalso that complete surface coverage is attained much faster than predict-ed by the OCP curves. It can be concluded that the commonly used inter-pretation of the OCP curves needs to be reconsidered.
A clear influence of the amount of silicon in the coating was also seenon the thickness of the deposited conversion layer. It was observed thatthe conversion layers on the silicon-containing aluminium coatingswere twice as thick as that on the silicon-free aluminium coating. This ob-servation was explained by two phenomena: the decreasing rate of thecathodic reactions due to an increasing amount of silicon and the factthat the conversion depositionmechanism is a self-extinguishing process.
Finally, a first suggestion of the build-up of the conversion layer as afunction of depthwas proposed based on the peak shifts observed in theAuger depth profiles in combination with the knowledge about thecomposition of the conversion solution and literature. Three differentregions could be identified. At the outer surface a titanium-phosphatecomplex along with an organic compound was present. Underneaththis outer surface layer a region composed of the same compounds to-gether with aluminium oxide was detected. In the third region thefree coating layer of the aluminium coating was reached.
Please cite this article as: I. Schoukens, et al., Surf. Coat. Technol. (2013), h
Towards the application, it is expected that all samples will showsimilar adhesion properties since all samples could be converted tofull coverage, although this needs to be experimentally confirmed.
Acknowledgements
Ine Schoukens was financed by the Flemish Institute for the Promo-tion of Scientific and Technological Research in Industry (IWT). Herwork was supported by ArcelorMittal Global R&D Gent/OCAS NV.
References
[1] F. Porter, in: L. Shreir, R. Jarman, G. Burstein (Eds.), Corrosion — corrosion control,Butterworth-Heinemann, 1994, pp. 12:51–12:63.
[2] Arcelor, “User manual, metallic coated steel”, 2010.[3] Y. Houbaert, T. Ros Yañez, M. Prado, Mater. Sci. Forum 373–376 (2001) 733–736.[4] I. Infante Danzo, K. Verbeken, Y. Houbaert, Defect Diff. Forum 297–301 (2010)
370–375.[5] C.M. Cottel, J.A. Sprague, F.A. Smidt, On Surface Engineering, Metals Handbook, vol.
5, ASM international, 1994, p. 346.[6] S.H. Hwang, J.H. Song, Y.S. Kim, Mater. Sci. Eng. A Struct. Mater. 390 (2005) 437–443.[7] G.H. Awan, F. Ul Hasan, Mater. Sci. Eng. A Struct. Mater. 472 (2008) 157–165.[8] K. Kee-Hyun, B. Van-Daele, G. Van-Tendeloo, Y. Jong-Kyu, Aluminium Alloys Pts 1
And 2 (519–521) (2006) 1871–1875(2006).[9] F. Garcia, A. Salinas, E. Nava, Mater. Lett. 60 (2006) 775–778.
[10] D. Phelan, B.J. Xu, R. Dippenaar, Mater. Sci. Eng. A Struct. Mater. 420 (2006)144–149.
[11] G. Eggeler, W. Auer, H. Kaesche, J. Mater. Sci. 21 (1986) 3348–3350.[12] S. Kobayashi, T. Yakou, Mater. Sci. Eng. A Struct. Mater. 338 (2002) 44–53.[13] J.E. Hatch, Aluminium Properties and Physical Metallurgy, American Society for
Metals, 1984.[14] J. Davis, Aluminium and Aluminium Alloys, ASM international, 1994.[15] I. Schoukens, I. De Graeve, J. De Strycker, H. Terryn, Corr. 68 (2012) 982–993.[16] O. Lunder, C. Simensen, Y. Yu, K. Nisancioglu, Surf. Coat. Technol. 184 (2004)
278–290.[17] J.H. Nordlien, J.C. Walmsley, H. Østerberg, K. Nisancioglu, Surf. Coat. Technol. 153
(2002) 72–78.[18] F. Andreatta, A. Turco, I. De Graeve, H. Terryn, J.H.W. De Wit, L. Fedrizzi, Surf. Coat.
Technol. 201 (2007) 7668–7685.[19] I. De Graeve, I. Schoukens, A. Lanzutti, F. Andreatta, A. Alvarez-Pampliega, J. De
Strycker, L. Fedrizzi, H. Terryn, Corros. Sci. 76 (2013) 325–336.[20] P. Campestrini, H. Terryn, J. Vereecken, J. De Wit, J. Electrochem. Soc. 151 (2004)
B359–B369.[21] J. De Laet, K. De Boeck, H. Terryn, J. Vereecken, Surf. Interface Anal. 22 (1994)
175–180.[22] I. Vandendael, T. Hauffman, Y. Van Ingelgem, A. Hubin, H. Terryn, Surf. Interface
Anal. 42 (2010) 897–901.[23] C. Powell, J. Electron Spectrosc. Relat. Phenom. 182 (2010) 11–18.[24] Y. Sakai, Y. Iijima, K. Hiraoka, J. Surf. Anal. 15 (2008) 172–176.[25] B. Wilson, N. Fink, G. Grundmeier, Electrochim. Acta 51 (2006) 3066–3075.[26] B. Timmermans, N. Vaeck, A. Hubin, F. Reniers, Surf. Interface Anal. 34 (2002)
356–359.[27] P. Campestrini, G. Goeminne, H. Terryn, J. Vereecken, J. De Wit, J. Electrochem. Soc.
151 (2004) B59–B70.[28] G. Goeminne, Study of the Mechanism of the Chemical Conversion of Aluminium,
