UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Formation of zinc phosphate coatings on aluminum alloys Kok, Wai-Hoong 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2001-0048.pdf [ 4.57MB ]
JSON: 831-1.0061474.json
JSON-LD: 831-1.0061474-ld.json
RDF/XML (Pretty): 831-1.0061474-rdf.xml
RDF/JSON: 831-1.0061474-rdf.json
Turtle: 831-1.0061474-turtle.txt
N-Triples: 831-1.0061474-rdf-ntriples.txt
Original Record: 831-1.0061474-source.json
Full Text

Full Text

Formation of Zinc Phosphate Coatings on Aluminum Alloys by Wai-Hoong Kok B.Sc., University of Malaya, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Chemistry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 2000 © Wai-Hoong Kok. Aooo In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract This thesis describes research to determine some of the factors that affect zinc phosphate coatings of aluminum alloys. The basic parameters for a coating bath to be used for dipping AA6061-T6 alloy are P-to-Zn atomic ratio, pH and F~ ion concentration, but the effect of acid etching in the pre-treatment was studied for both the AA6061-T6 and AA2024-T3 alloys. The surface characterization methods used are X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and energy-dispersive X-ray (EDX) spectroscopy, and the corrosion protection provided by the coatings formed are assessed by electrochemical polarization measurements. Three main zinc phosphate coating recipes used in this study are referred to as the high-ratio normal-zinc (h-Zn), normal-zinc (n-Zn) and low-zinc (/-Zn) formulations. It is shown that the h-Zn formulation (modified from n-Zn), with a P-to-Zn atomic ratio of 15, provides the best coating condition at the original pH as compared with the n-Zn and 1-Zn formulations. The coating baths have optimal pH values of 4 for the main types of coating recipe, and this is judged in terms of the quality of the coatings formed and by the observed corrosion resistance. The best coating recipes determined here are for the /-Zn process at pH 4 with 600-1000 ppm of added F~ and for the n-Zn process at pH 4 with 200-400 ppm F~. The inclusion of acid etching in the pre-treatment of AA6061-T6 aluminum alloy yields better coatings than when mechanical polishing alone is used; it appears that the acid etching yields a rougher surface on which the coating nucleates more favorably. However, acid etching of the AA2024-T3 alloy results in the presence of enriched copper at the alloy/film interface, and the amount of copper detected by XPS increases with the time of the acid etching. A study was made of the effect of this copper enrichment on the zinc phosphate coating formed on the AA2024-T3 alloy. For the conditions studied, a 5 min acid etch leads to zinc phosphate coatings ii that have a uniform distribution of small crystallites and improved corrosion resistance. However, when the copper is in excess, for example from a 10 min acid etch, the coating formation is strongly inhibited. 111 Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures viii Acknowledgements xi 1. Introduction 1 1.1 General 1 1.2 Aluminum and its alloys 2 1.2.1 Properties of aluminum 2 1.2.2 Classification of aluminum alloys 3 1.2.3 Corrosion and corrosion protection 7 1.3 Phosphate conversion coating 10 1.3.1 Theoretical consideration 12 1.3.2 Factors affecting zinc phosphate coating 15 1.3.3 Recent developments 16 1.4 Outline and aim of thesis 17 2. Experimental methods 19 2.1 X-ray photoelectron spectroscopy (XPS) 19 2.1.1 Introduction 19 2.1.2 History 19 2.1.3 Basic Principles 21 2.1.4 Instrumentation 23 iv 2.1.5 Spectral features 30 2.1.6 Quantitative Analysis 33 2.1.7 Data processing 36 2.2 Scanning electron microscopy (SEM) 38 2.2.1 Introduction 38 2.2.2 Instrumentation 39 2.3 Energy-dispersive X-ray (EDX) spectroscopy 43 2.3.1 Introduction 43 2.3.2 Instrumentation 43 2.4 Electrochemical Polarization Curve Measurement 45 2.4.1 Introduction 45 2.4.2 Instrumentation 50 3. Factors affecting zinc phosphate coating on AA6061-T6 aluminum alloy 53 3.1 Introduction 53 3.2 Experimental 54 3.3 Effect of different P-to-Zn atomic ratio 55 3.3.1 SEM studies 57 3.3.2 XPS study 62 3.4 Effect of pre-treatment 68 3.5 Effect of pH 69 3.5.1 Normal-zinc solution 69 3.5.2 Low-zinc solution 73 3.5.3 High-ratio normal-zinc solution 73 3.6 Fluoride ion concentration 74 3.6.1 Normal-zinc solution 74 v 3.6.2 Low-zinc solution 77 3.7 Corrosion studies 77 4. Effect of copper enrichment on zinc phosphating of AA2024-T3 alloy 85 4.1 Introduction 85 4.2 Principles associated with surface enrichment of aluminum alloy 86 4.3 Pre-treatments of AA2024-T3 aluminum alloy 89 4.4 Effect of copper enrichment on zinc phosphate coating 92 4.4.1 General observation 92 4.4.2 Corrosion behavior 94 4.4.3 Comparison of zinc phosphate coatings 96 5. Concluding remarks and future direction 99 5.1 Concluding remarks 99 5.2 Future directions 101 5.2.1 Different bath modifications 101 5.2.2 Pre-treatment and post-treatment 101 5.2.3 Evaluation of coatings 102 5.2.4 Study of elemental enrichment on aluminum alloys 102 References 103 vi List of Tables Table 1.1 Aluminum alloy designation according to A A designation 5 Table 1.2 The main temper designations for aluminum alloys 6 Table 1.3 Characteristics of different phosphate coatings 11 Table 3.1 Specification of treatments used for AA6061-T6 samples 56 Table 3.2 XPS analysis for coatings formed from different phosphating baths 65 Table 3.3 Current density values measured for coated AA6061-T6 samples 81 V l l List of Figures Figure 1.1 Pourbaix diagram for the A l - F b O system 8 Figure 1.2 Diagrammatic representation reactions occurring during phosphating 13 Figure 2.1 Schematic diagram for (a) photoelectric effect, (b) X-ray 20 fluorescence and (c) Auger emission Figure 2.2 A schematic diagram of the M A X 2 0 0 system viewed from the top 24 Figure 2.3 Schematic representation of the pumping system for the M A X 2 0 0 25 Figure 2.4 Dual anode X-ray source 27 Figure 2.5 Schematic diagram for the C H A and lens system in the M A X 2 0 0 29 Figure 2.6 Schematic diagram of energy levels for the binding energy measured in 31 XPS Figure 2.7 XPS survey and narrow scan spectrum of Zn 32 Figure 2.8 Inelastic mean free path of electrons as a function of kinetic energy 35 Figure 2.9 Shirley non-linear background subtraction applied to a O Is spectrum 37 Figure 2.10 Sectional view of the Hitachi S4100 S E M 40 Figure 2.11 Schematic of the pumping system for Hitachi S4100 S E M 41 Figure 2.12 The illuminating/imaging system of a S E M unit 42 Figure 2.13 Construction of lithium-drifted silicon detector for E D X spectrometer 44 Figure 2.14 E D X spectra of sample after coating with zinc phosphate solution 46 Figure 2.15 Electrochemical polarization curve for a corrosion system 47 Figure 2.16 Schematic of the potentiostatic polarization circuit 48 Figure 2.17 Electrode kinetic behavior of pure zinc in acid solution 51 viii Figure 3.1 SEM images of AA6061-T6 alloy after treatment in different normal- 58 zinc formulations using H 3 P0 4 to adjust the P-to-Zn atomic ratio Figure 3.2 SEM images after treatment in nomal-zinc formulations using fixed 59 H,P0 4 and varying NaH 2P0 4 to adjust the P-to-Zn atomic ratio Figure 3.3 SEM images after treatment with different sources of phosphate 61 Figure 3.4 SEM images for coatings formed by different reference solutions 63 Figure 3.5 XPS from AA6061-T6 alloy samples: (a) after mechanical polishing 64 and (b) with the n-Zn reference solution Figure 3.6 EDX spectrum measured from area marked by X in Fig. 3.4 67 Figure 3.7 Phosphate species present at different pHs and associated effects on 70 normal-zinc coatings Figure 3.8 SEM images samples coated by n-Zn, h-Zn and /-Zn reference 71 solutions at their original pH values and at pH 4 Figure 3.9 SEM images of AA6061-T6 alloy after coating by some noral-zinc 75 processes at pH 4 where the only variation is in the F concentration Figure 3.10 XPS from AA6061-T6 alloy coated at pH 4, 800 ppm F" by: 76 (a) normal-zinc and (b) low-zinc formulation Figure 3.11 SEM images of AA6061-T6 alloy after coating by some low-zinc 78 processes at pH 4 where the only variation is in the F" concentration Figure 3.12 Polarization curves measured from AA6061-T6 alloy after coating with 80 n-Zn, /-Zn and h-Zn reference solution Figure 3.13 Polarization curves measured from AA6061-T6 alloy after coating with 82 modified n-Zn, /-Zn process and h-Zn process, all at pH 4 Figure 3.14 SEM images before and after corrosion test for samples coated by n-Zn 83 process at different pH values ix Figure 4.1 Illustration of the development of an enriched alloy layer during oxide 87 film growth on aluminum Figure 4.2 Gibbs free energy of formation per equivalent for different oxides 88 compared with enrichments observed Figure 4.3 XPS spectra of AA6061-T6 aluminum alloy; (a) after mechanical 90 polishing and (b) acid etching Figure 4.4 Copper enrichment by XPS obtained for different acid etching times 91 Figure 4.5 SEM images of zinc phosphate coatings formed on AA2024-T3 93 samples after different pre-treatments Figure 4.6 Polarization curves for AA2024-T3 aluminum alloy coated by h-Zn 95 reference solution after different pre-treatments Figure 4.7 SEM images of samples coated by different zinc phosphate coating 97 x Acknowledgements I would like to thank my supervisor, Professor K.A.R. Mitchell, for accepting me to his research group and introducing me to the area of surface science. I am grateful for his sponsoring this work as well as his invaluable advice and guidance. I am especially grateful to Dr. X. Sun for many valuable discussions and Dr. K.C. Wong for helping with the XPS measurements and interpretations. I have greatly appreciated interactions with other members of the group, including Ms. L. Shi, Ms. R. Li , Ms. D. Susac, Ms. M . Kono, Dr. M . Saidy, Mr. H. Fourtier, Ms. J. Jing, Mr. J.T.K. Kim, Mr. M.Y.C. Teo and Mr. M.C.C. Lee. I am thankful to Professor D. Tromans and Mr. T .M. Ahmed for their assistance with the electrochemical polarization measurements and interpretation. I also acknowledge the guidance from J. Mackenzie for the operation of SEM, and M . Mager for the operation and interpretation of the E D X spectrometer. I am grateful to my friends for their support during my stay in Canada, especially Daryl Rubinof for proof reading my thesis. Last, but not least, I would like to thank my family for their support and encouragement during my studies in Canada. xi Chapter 1 Introduction 1.1 General Aluminum is the most important non-ferrous metal, second only to iron in annual consumption. This is mainly because of its favorable properties such as low density, reasonably high strength, good corrosion resistance, and relatively low cost [1]. The good corrosion resistance is mainly due to an inert oxide layer, which forms almost instantaneously when bare aluminum is exposed to air. However, the oxide layer has limited thickness, and evidence shows that flaws pre-exist in the oxide film and can act as nucleation sites for film breakdown [2]. Conversion coating may be defined as the chemical treatment of a metal's surface so as to form a corrosion-resistant layer that more easily accepts and bonds to applied coatings. Chromate conversion coatings are generally considered to give the best corrosion resistance and are specified for use in military applications [2-4]. However, in recent years, chromium, especially in the hexavalent state, has been found to cause irritation of the respiratory tract, produce ulceration and perforations of the nasal septum, and lead to lung cancer in workers employed in chromium manufacturing plants in West Germany and the United States [5]. Thus, there is a strong incentive to explore and evaluate alternative non-chromate treatments of metal surfaces. Recent studies [5-11] have shown that chemical conversion treatment in chromate-free phosphate baths is a promising alternative for the surface modification of iron-aluminum components. It has long been recognized that the phosphated metal exhibits substantially greater corrosion resistance and paint adhesion [12-14]. More importantly, this technique has advantages over chromate-phosphate treatment in terms of environmental effects. 1 In the last few years, some extensive studies has been made in our research group in developing a phosphating recipe for AA7075-T6 aluminum alloy [9, 10, 15-18] and most recently by Shi for AA6061-T6 [19]. The studies undertaken here aim to further develop this recipe, building on the experience already obtained in this laboratory. X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and energy-dispersive X-ray analysis (EDX) were used to characterize the coatings formed. The corrosion protection was assessed by electrochemical techniques. 1.2 Aluminum and its alloys The element "aluminum" is very widely distributed in the form of aluminum oxide or 'alumina'. Alumina (A1203) comprises 15% of the earth's crust and is second only to silica in abundance. Ore reserves are therefore virtually unlimited [20]. Aluminum was discovered in 1809 by Sir Humphrey Davy and first isolated in 1825 by H. C. Oersted but it was not until 1886 that an economic industrial reduction process was invented. In that year, two scientists working independently, Charles Martin Hall in the United States and Paul Heroult in France, each invented the same electrolytic process for reducing alumina to aluminum. The Hall-Heroult process is widely used today for extracting aluminum from alumina [20]. I. 2.1 Properties of Aluminum The three main properties on which applications of aluminum are based are its low density (-2.7 g/cm3), high mechanical strength which increases by suitable alloying and heat treatments, and the relatively high corrosion resistance of the pure metal. The combination of low density and high strength gives aluminum and its alloys a unique strength-to-weight ratio advantage which 2 makes them prized materials for construction where lightness is at a premium, such as in the aerospace and automobile industries. In many applications, the high electrical and thermal conductivities, coupled with excellent strength-to-weight ratios, make aluminum and its alloys suitable replacements for copper and stainless steel, both of which are more expensive. In its pure state, aluminum is a relatively soft metal with a yield strength of only 34.5 N/mm 2 and a tensile strength of 90 N/mm 2 . Its strength and hardness can be improved by alloying, cold working and thermal treatment, and in practice a combination of all three is often used [1]. Many metals and combinations of metals are used to alloy aluminum. The range of alloys obtained offers to industry a remarkably wide variety of combinations of properties, such as mechanical strength, ductility, electrical conductivity, and corrosion-resistance. However, the optimal properties are not available for every end-use, or for any one alloy, and therefore compromises are needed in general [20]. 1.2.2 Classification of aluminum alloys The major alloying elements are copper, magnesium, silicon, manganese, nickel and zinc. A l l these are used to increase the strength of pure aluminum. The aluminum alloys can be classified into two classes. The first are the 'cast alloys' which are cast directly into their desired forms by one of three methods (i.e. sand casting, gravity die casting or pressure die casting). The second class, the 'wrought alloys', are cast in ingots or billets, and are later hot and cold worked mechanically into extrusions, forging, sheet, foil, tube and wire [2]. The specifications and properties of aluminum and its alloys are covered by standard codes, which facilitate selection and procurement of the right grade for any particular application. This usually also covers temper designations to signify the cold working or heat treatment applied to the alloy. These codes are governed by two main bodies. In the United 3 States, the Aluminum Association (AA) designation numbers are used quite widely. The Unified Numbering System (UNS), developed jointly by the Society of Automotive Engineers, "and American Society for Testing and Materials (ASTM) , is a six-character/space alpha-numeric designation system; the first character on the left usually being a letter of the English alphabet, followed by a recognized and well accepted designation system. The U N S designation system for aluminum alloys essentially uses the A A designations in the U N S format; with the letter A and number 9, followed by the four-digit A A designation. Designation codes used in the U K , Canada, France and Spain use combinations of one or two letters and the numerals. The A A codes for cast and wrought alloys are separate and slightly different from each other as shown in Table 1.1. For the wrought alloys, the first digit denotes the major alloying element and the second digit indicates alloy modifications; 0 for original alloy while 1 through 9 are assigned consecutively to indicate the alloy modifications in sequence. For cast alloys, the second and third digits identify the alloy or aluminum purity, and the last digit, separated by a decimal point, indicates the product form, such as casting or ingot. The temper designation system for both wrought and cast alloys, to indicate their cold-working and heat treatment histories, consists of a letter, separated by a hyphen, following the numeric alloy designation. This is followed where required by one or two digits indicating the specific sequence of basic treatments which significantly influence the characteristics of the product. The temper designations are indicated in Table 1.2 [1]. The main classes of aluminum alloys are the high-strength A l - C u alloys (2000 series) used mainly in the aircraft industry, the A l - M n alloys (3000 series) used mainly in the canning industry, the A l - M g alloys (5000 series) used unprotected for structural and architectural applications, the most common extrusion Al -Mg-Si alloys (6000 series) used particularly in the 4 Table 1.1 Aluminum alloy designation according to the Aluminum Association (AA) designation numbers [1]. Wrought Alloy Designations: aluminum, 99.00% minimum or greater 1xxx aluminum alloys grouped by major alloying elements copper 2xxx manganese 3xxx silicon 4xxx magnesium 5xxx magnesium and silicon 6xxx zinc 7xxx other elements 8xxx unused series 9xxx Cast Alloy Designations: aluminum, 99.00% minimum or greater 1xx.x aluminum alloys grouped by major alloying elements copper 2xx.x silicon with copper and/or magnesium 3xx.x silicon 4xx.x magnesium 5xx.x zinc 7xx.x tin 8xx.x other elements 9xx.x unused series 6xx.x 5 Table 1.2 The main temper designations for aluminum alloys [1 ]. F as fabricated 0 annealed H strain-hardening (wrought products only) W solution heat treated T heat treated to produce stable tempers other than F, O or H 6 building industry, and the high-strength Al-Zn-Mg alloys (7000 series) for aircraft and military vehicle applications [2]. 1.2.3 Corrosion and corrosion protection The word corrosion comes from the Latin word 'corrodere', which means 'gnaw away'. Corrosion may be defined as the deterioration of a metal by reaction with its environment to result in metal oxidation. Aluminum is a base metal having a strong tendency to react with its surroundings. However, the good corrosion-resistance of aluminum in natural water is due to the presence on the surface of a very thin protective film of aluminum oxide, which has strong self-healing properties when damaged. The film begins to form immediately when the surface of bare metal is exposed to air, and grows rapidly for several days, then slowly for a month, until it reaches a thickness in the 0.005-0.015 um range [2, 20] The corrosion characteristics of aluminum are shown in the potential-pH (E-pH or Pourbaix) diagram for the A1-H,0 system (Figure 1.1) [21]. It is a thermodynamic diagram which shows the stability of a species as a function of the variables plotted, and it is used to predict the spontaneous direction of a reaction. As for a thermodynamic phase diagram, it does not provide any information on the kinetics of reaction. Lines (a) and (b) are the hydrogen and oxygen stability lines for water by itself at 25°C, 1 atm. Below line (a), water reduces to form hydrogen gas, and above line (b), water oxidizes to form oxygen gas. The stability region of water in the absence of aluminum is therefore between lines (a) and (b). Figure 1.1 shows that the whole domain of aluminum stability lies below that of water. In the presence of sufficiently acidic solutions (corrosion region I), it decomposes water with the evolution of hydrogen, and dissolves as trivalent A l 3 + ions. In sufficiently alkaline solutions 7 Corrosion 1 Passivation 2 H 1 E H (volt) 0 - H -2-f Al 3+ A l 2 0 3 - 3 H 2 0 Al Immunity 6 7 p H Corrosion II A10 2 (b) — (a) Figure 1.1 Pourbaix diagram for the A1-H 20 system at 25°C, 1 atm pressure [21 8 (corrosion region II), aluminum decomposes water with the evolution of hydrogen, and dissolves as the aluminate ions, AlO,". In non-complexing solutions, with the pH roughly between 4 and 9, aluminum tends to become covered with a film of oxide and this defines the passivation region. This oxide can act to protect the metal from a surrounding corroding medium, and so provide a kinetic stability. The molecular volume of the oxide film is stoichiometrically 1.5 times that of the metal used up in oxidation, and therefore the oxide film formed on the surface is under compressive stress. This oxide cannot only cover the metal continuously, but it can also cope with a certain amount of substrate deformation without rupturing [2]. Since the aluminum alloys encompass a whole group of metallic properties, it is understandable that some perform better than others under corrosive conditions. Super-pure aluminum has better corrosion-resistance than any aluminum alloy, but this material is not normally suitable for commercial application because of its less favorable mechanical properties [2, 20]. Breakdown of the oxide film can result from mechanical rupture or from chemical attack by anions such as chloride ions. Corrosion of aluminum can occur when the film is damaged or removed and conditions prevent its reformation. The information gathered from both electrochemical studies [2, 22, 23] and transmission electron microscopy (TEM) [24], shows that flaws pre-existing in the oxide film can act as nucleation sites for film breakdown. It is proposed that, in solution, flaws are continually being developed and repassivated, and that the presence of aggressive ions hinders the repassivation step so allowing pits to nucleate [2]. The methods of corrosion prevention are varied and they may be classified as: a) modification of the metal by alloying and/or surface modification; b) modification of the environment by the use of inhibitors; and c) change of metal/environment potential by cathodic or anodic protection. Surface modification is generally used for corrosion protection because it 9 is more economical than the other methods. This process involves forming a physical barrier to protect the metal against its corrosive environment [2, 25, 26]. Some of the more conventional techniques for surface modifications are: i) painting; ii) anodizing and iii) chemical conversion coatings. 1.3 Phosphate conversion coating Phosphate processes originated with the Romans in the third century A.D, but the idea of using phosphate coatings to impart corrosion resistance goes back over 100 years to applications on iron surfaces [2]. Phosphating of a steel part by immersion in dilute phosphoric acid solution results in the metal being coated with a ferrous phosphate layer, but the basis of modern methods of phosphating developed from the first coating bath compositions of Coslett [27-30], Richards [31] and Allen [32-34]. All conventional phosphating solutions contain dilute phosphoric acid with the addition one or two metal ions (e.g. Fe2 +, Zn 2 + , Mn 2 +), and the processes are often classified as zinc, iron or heavy (manganese) phosphating. Table 1.3 shows the characteristics of the coatings obtained from these baths [35]. Zinc phosphate is found to be suitable for use as a paint base in highly corrosive environments. The main application for a conversion coating on aluminum is as a pre-treatment before painting, however their usefulness has been demonstrated for metal treatment prior to adhesive bonding. The most commonly used conversion coatings for adhesion are based on chromate or chromate-phosphate chemistry. Only more recently, the phosphating processes developed for steels have been adapted for use on aluminum as an alternative to using the chromating process [2, 6-8, 10, 15-18, 36-48]. 10 Table 1.3 Characteristics of different phosphate coatings [35]. Iron Phosphate Zinc Phosphate Heavy Phosphate Coating weight Types Primary use 0.16-0.80 gm" Organic phosphate Paint base for low-corrosion environments 1.4-4.0 gm" Nickel-modified Low-zinc Calcium-modified Manganese-modified Paint base for high-corrosion environments 7.5-30 gm" Manganese phosphate Zinc phosphate Ferrous phosphate Unpainted applications Limitations Application method Low corrosion resistance (both painted and unpainted) Spray and immersion Poor unpainted corrosion resistance Expensive, long processing times Spray and immersion Immersion only 1.3.1 Theoretical considerations Orthophosphoric acid, H 3 P 0 4 is a tribasic acid, which gives rise to three series of salts with zinc: Zn(H 2P0 4) 2 primary ZnHP0 4 secondary Zn 3(P0 4) 2 tertiary Modifications occur in the crystalline form according to the degree of hydration. The primary zinc phosphate is soluble in water, whereas secondary forms are only sparingly soluble, and tertiary zinc" phosphate is insoluble [12, 49]. The tertiary phosphate tetrahydrate readily precipitates on neutralization of dilute primary zinc phosphate solution [12], and Wustefeld [34, 50] indicated that the presence of tertiary phosphate in colloidal form is a prerequisite to the formation of phosphate coating at a metal surface. Phosphating is essentially an electrochemical phenomenon in which dissolution of metal and discharge of hydrogen occurs at micro-anodes, and this is accompanied by hydrolysis and precipitation of insoluble phosphates at micro-cathodes (Figure 1.2) [12, 51]. Even for high-purity (99.99%) samples, the aluminum surface can be inhomogeneous because of structural defects and impurities, but with alloying the different anodic and cathodic regions become even more important in coating formation [52]. The first reaction that occurs when an aluminum surface is exposed to acid is the pickling reaction, which results in metal dissolution at anodic areas A1 20 3 + 6H + - 4 2A13 + + 3H 2 0 (1.1) A1 + 3FT ^ A l 3 + + 3/2H2(g)T (1.2) The pickling reaction is an essential step for coating formation because it acts to chemically clean the surface, so enabling subsequent coatings to adhere to the base metal. The initial metal 12 Figure 1.2 Diagrammatic representation of the micro-electrode reaction occuring during the phosphating of aluminum (modified from Freeman [12]). 13 attack is frequently localized at 'active centers' (e.g. grain boundaries or other defects) due to their high energy [12, 13,53]. The initial solution containing Zn 2 + ions in H,P0 4 is subject to the equilibria H 3 P0 4 ^ FLPO; + H + (1.3) H 2 P O ; — HP0 4 2 +H + (1.4) HP0 4 2- ^ P 0 4 3 + H + (1.5) and as the local pH increases following removal of Ff according to Equation (1.1) and (1.2), the equilibria in (1.3) to (1.5) become driven to the right hand side by the solubility product of Zn3(P04), being exceeded with precipitation of the latter as Zn,(P04)2-4H20 (hopeite). Some other insoluble salts may precipitate as sludge and also deposit on the metal surface. The experience is that a certain amount free phosphoric acid must always be present in the phosphating solution in order to repress hydrolysis and keep the bath stable for effective deposition of phosphate at the microcathodic sites (Equations (1.3) to (1.5)) [12, 35]. In addition, the A l , + ions in solution can react with phosphate ions to form aluminum phosphate (Equation (1.6)), and this can have a deleterious effect on the coating formation. To avoid the aluminum ions from poisoning the phosphating bath, fluoride ions are added (Equation (1.7)). A l 1 + + PO/" -+ A1P04 (1.6) A13+ + 6F" -+ A1F 6 V (1.7) The presence of fluoride ions is generally considered essential to the coating bath, but a balance is required since an excess of fluoride ion can induce sludge formation through the reaction [8] A l , + +6F"+3Na+ -+ Na 3 AlF 6 4 (1.8) In contrast to conversion treatments based on chromic acid, where amorphous films are formed, the zinc phosphating treatments produce crystalline coatings [54]. A surface activation is normally needed in order to refine the crystal size, and especially to obtain a dense zinc 14 phosphate layer with small crystals. One activation step involves dipping or spraying with a titanium colloidal suspension [9, 38, 55, 56]. The structure of a phosphate coating can vary with the particular metal. On steel, the zinc phosphating process can yield the presence of Zn,Fe(P04)2-4H20 (phosphophyllite) and Zn,(P04)2-4H20 (hopeite) [12, 14, 35, 57]. However, Ishii [8] reported X-ray diffraction (XRD) studies to show that only the hopeite form was obtained on aluminum. Sankara Narayanan [58] in studying the stabilities of phosphate coatings concluded that the phosphophyllite has greater alkaline stability and lower tendency for rehydration. Therefore, the phosphophyllite-rich coating is best suited to withstand the environments that occur during painting, paint curing and under service conditions. 1.3.2 Factors affecting zinc phosphate coatings The zinc phosphating treatment on aluminum alloy sheets can be significantly affected by the F" concentration in the coating solution, and it has been reported that there is an optimum value for the free fluoride ion concentration [8, 48]. When this concentration is below the optimum value, the proper weight of phosphate coating could not be obtained and the corrosion resistance was reduced accordingly. However, for F~ concentrations in excess of the optimum value, the adhesion deteriorates as a result of cryolite (Na,AlF6) deposition. Cavanagh et al. [59] in studying the role of pH in the phosphating of steel reported that phosphate coatings cannot be produced in the presence of excess H,P0 4. Baths in that study, with and without phosphoric acid accumulation show satisfactory coatings. However, if the accumulation continues, a point is reached where the acid increases the rate of attack on the steel and this results in the formation of a sludge. The type of alloying elements is also important. A highly modified alloy layer immediately beneath the oxide film may have a significant influence on a wide range of 15 properties of the alloy/film interface. Pits have been shown to initiate near microsegregated iron, copper and silicon impurities in high purity aluminum [60]. It has been reported that enrichments of alloying elements in aluminum due to various surface pre-treatments [61-63] only occur for elements with oxides that have Gibbs free energy per equivalent similar to, or less than, that of alumina [64]. Such surface enrichments can have a significant effect on the phosphating process [65]. 1.3.3 Recent Developments Different phosphating processes have been developed to address corrosion resistance and adhesion for different metals, as well as the demand from industry to treat different metals with the same coating solution. The traditionally used zinc phosphating processes, often referred to as normalxzinc processes, have been gradually replaced with the use of the low zinc processes, which use lower amounts of zinc in the coating solutions. This new process shows outstanding performance in all corrosion testing environments for steel [54]. The difference between these processes involves the process chemistry. In normal zinc phosphating baths, approximately 2000-4000 ppm (mg/L) zinc and 5000-10000 ppm phosphorus pentoxide (P2Os) are present in the bath, whereas approximately 400-1700 ppm zinc and 12000-16000 ppm P,0 5 are present in low zinc phosphating baths. Further, the low zinc phosphate technology has been refined by incorporating a third metal, namely manganese in the process bath to enhance the coating properties [54]. According to Donofrio, the low zinc coating technology shows a better coating performance because of several factors. First, a longer pickling reaction results in a better chemical cleaning of the metal surface; second the slower deposition reaction gives a denser phosphate structure and increased amounts of zinc-iron phosphate (phosphophyllite) on steel surfaces [54]. 16 A most recent development is of the trication (zinc, nickel and manganese) process [58, 66-68]. This solution includes the addition of Zn 2 + , Ni 2 + and Mn 2 + ions as well as small amounts of fluoride, and the coating process results in the formation of Zn2Ni(P04)2-4H,0 (phosphonicollite) and Zn2Mn(P04),-4H,0 (manganese phosphophyllite) which codeposit along with hopeite. These codeposited crystals are considered as modified hopeite crystals, and they appear to resemble phosphophyllite (the iron-zinc phosphate formed on steel) in chemical behavior [58]. 1.4 Outline and Aim of Thesis The motivation for this work originates from a grant provided by the Department of National Defence and the Natural Sciences and Engineering Research Council (NSERC) to establish home-based corrosion-protective phosphate-coating recipes for different aluminum alloys. Satisfactory phosphating processes for AA7075-T6 [10] and AA6061-T6 [19] have been developed by our group, although our emphasis is more to establish principles and mechanistic understanding. The work done in this thesis aims to study the effect of some of the basic parameters that have to be optimized in order to develop effective phosphating processes for AA6061-T6 aluminum alloy. The basic parameters include Zn-to-P atomic ratio, pH and F" concentration. In order to develop an optimum phosphating process, some understanding of the basic parameters affecting it is required, and that is one emphasis of this study. The second objective is to determine the effect of pre-treatment, especially insofar as it is associated with copper enrichment, on the zinc phosphate coating process for AA2024-T3 aluminum alloy. Enriched layers developed during such surface pre-treatments of aluminum alloys, by chemical polishing [61], electropolishing [62] or alkaline etching [63], have been reported to be important on the surface treatment of aluminum, although no previous study has 17 assessed the effect of copper enrichment on the formation of a zinc phosphate conversion coating. 18 Chapter 2 Experimental Methods 2.1 X-ray photoelectron spectroscopy (XPS) 2.1.1 Introduction X-ray photoelectron spectroscopy (XPS, also called electron spectroscopy for chemical analysis, ESCA) is one of the most widely used and powerful surface characterization methods. Its popularity as a surface analysis technique is attributed to its high information content, its flexibility in addressing a wide variety of samples, and its sound theoretical basis [69]. This method depends on X-rays to eject electrons by the photoelectron effect from bound energy levels within the material. The kinetic energies of these photoelectrons are measured, but a spectrum is usually presented as a plot of intensity versus binding energy for the electron inside the solid. XPS analysis of a surface can provide qualitative and quantitative information for all elements (except H and He) in the topmost 20-100 A for any material that is vacuum stable, or can be made vacuum stable by cooling. By careful analysis, detailed information about the chemistry, organization, and morphology of a surface can often be obtained [69, 70]. The basic XPS experiment is illustrated in Figure 2.1(a). The surface to be analyzed is first placed in a vacuum environment and then irradiated with photons from the soft X-ray range. Common sources are those for MgKa (1253.6 eV) and AlKa (1486.6 eV) radiation [71]. 2.7.2 History The discovery of the photoelectric effect, its explanation, and the development of the XPS method are entwined with the revolution in physics that took place in the early years of the 19 Photoemission Photoelectron • A \ Valence band X-ray fluorescence Valence band _Q • hv (b) (a) L2.3 K L23 Li K Auger emission KL]L2,3 Auger electron Valence band _Q • L23 L, K (c) Figure 2.1 Schematic diagram for (a) photoelectric effect, (b) X-ray fluorescence and Auger emission. 20 twentieth century. This revolution led from classical physics based upon observational mechanics to quantum physics, whose impact is most clearly appreciated at the atomic scale. Hertz [72], in the 1880's, noticed that metal contacts in electrical systems, when exposed to light, exhibit an enhanced ability to spark. Hallwachs, in 1888, observed that a negatively charged zinc plate lost its charge when exposed to ultraviolet (UV) light, but that positively charged zinc plates were not affected. In 1889, J.J. Thompson [73] found that subatomic particles (electrons) were emitted from the zinc plate exposed to light. Finally, in 1905, Einstein [74], using Planck's 1900 quantization of energy concept, correctly explained all of these observations - photons of light directly transferred their energy to electrons within an atom, resulting in the emission of the electrons consistent with the conservation of energy. Einstein received the Nobel Prize for Physics in 1921, in part for this work. As an analytical method, a more straightforward history can be presented. In 1914, Robinson and Rawlinson [75] studied photoemission from X-ray irradiated gold and, using photographic detection, observed the energy distribution of electrons produced. Although they were hampered by poor vacuum systems and inhomogeneous X-ray sources, they were still able to publish a recognizable gold photoemission spectrum. In 1951, Steinhardt and Serfass first applied photoemission as an analytical tool. Throughout the 1950's and 1960's, Kai Siegbahn [76] developed the instrumentation and theory of XPS to give us the method we use today, and he shared the Nobel Prize for Physics in 1981 [69] primarily for his work in XPS. 2.1.3 Basic Principles An understanding of the photoelectron effect and photoemission is essential in order to appreciate the surface analytical method, XPS. The basic process in XPS is due to emission of photoelectrons from a solid as a result of the total transfer of energy from impinging X-rays to 21 electrons inside the material. The physics of this process can be described by the Einstein equation, simply stated: E B = hv-EK (2.1) where E B is the binding energy of the electron in the material, hv is the photon energy, and E K is the kinetic energy of the photoelectron that is measured in XPS [69, 71, 77]. To a good approximation, the binding energy is characteristic of the specific atom, and for any given electron shell, the electron binding energy increases with increasing atomic number. Usually, information on the binding energies observed in a particular sample allows direct identification of which atoms are present (i.e. qualitative analysis). Since the energy levels occupied by electrons are quantized, the photoelectrons have a kinetic energy distribution N(E), consisting of a series of discrete bands that reflects the shell form of the electronic structure of the sample. Small variations in binding energy of up to 5 eV or so, may arise from the changes in valence electronic structure when a particular atom experiences different chemical environments. Observations of such effects may provide information on chemical bonding for the atoms concerned [77]. The ion in an excited state, produced by the photoelectric process, can dissipate the excess energy through one of two relaxation processes: (i) photon emission (e.g. X-ray fluorescence); or (ii) ejection of another electron (an Auger electron), as illustrated in Figure 2.1 (b,c). These two relaxation processes compete and their relative probabilities vary with atomic number (Z) and binding energy associated with the core vacancy, but Auger emission dominates when the initial binding energy is 2 keV or less [78]. An XPS spectrum generally also includes electrons generated by the Auger process. Figure 2.1 (c) shows schematically the production of a K L T , , Auger electron whose kinetic energy, in a first approximation, is EKL,L 2 , = E K - E L , - E L 2 3 (2.2) 22 where E K , E L , , E L , 3 are the binding energies for the levels involved. The kinetic energies of Auger electrons are independent of the excitation source, and thus they can be distinguished from photoelectrons. Direct elemental identification can also be made from observed peaks due to Auger electrons. Each element has its own set of Auger electron kinetic energies, determined basically by Equation (2.2). 2.1.4 Instrumentation A plan of the XPS facility used for the research in this thesis is presented in Figure 2.2. This is based on a MAX200 instrument (Leybold), and the important components of the facility consist of: (i) sample manipulator and transfer system, (ii) X-ray source, (iii) energy analyzer and (iv) pumping system. The XPS experiment is done under vacuum for three reasons. First, the emitted photoelectrons must be able to travel from the sample through the analyzer to the detector without colliding with gas phase particles. Second, some components such as the X-ray source, electron gun and electron multiplier require vacuum conditions for their effective operation. Third, the surface composition of the sample under investigation must not change during the analysis. Only a modest vacuum (10"6-107 Torr) is necessary to meet the first two requirements, but more stringent vacuum conditions are necessary to avoid contamination of the s^ample [69]. The high surface sensitivity of XPS and Auger electron spectroscopy necessitates the use of an ultra-high vacuum (UHV) system with typical pressure range of 10s-10"'° mbar to ensure the sample surface does not change during the time taken to make the measurements. Contaminants can build up at the rate of one monolayer per second at a pressure of 10"6 Torr [77]. The MAX200 pumping system is indicated schematically in Figure 2.3. The main chamber is pumped by a combination of pumps, specifically rotary and turbomolecular pumps with an 23 • Leak valve Gas lines, Transfer rod #1 / Preparation chamber #2 Preparation chamber #1 CHA analyzer Analysis chamber Transfer rod #2 Monochromatized X-ray source Manipulator Automatic 1 X-ray source transfer rod L Transfer chamber Figure 2.2 A schematic diagram of the MAX200 system viewed from the top 24 Gas manifold Manual transfer rods Automatic transfer rod Manipulator Ion pump Ti sublimation pump by) Turbo pump Rotary pump Figure 2.3 Schematic representation of the pumping system for the MAX200 25 auxiliary titanium sublimation pump, and an ion pump for the X-ray source. The base pressure of 2x10""' mbar in the analysis chamber can be achieved by baking the TJHV system from time to time at around 100-120"C for up to 24 hours. The pressure in the transfer chamber can go down to around 2x10s mbar. Samples for analysis are mounted on standard sample holders by using double-sided sticking copper tape where the holders (up to seven) are placed in the transfer chamber. Once the sample is placed in this chamber, the pressure is reduced by the combination of rotary and turbomolecular pumps to the 107-108 mbar range. The sample is then transferred into the analysis chamber via the automatic transfer rod (x-direction) where the sample holders lock to the manipulator (PTM 60) dock. The PTM 60 manipulator allows three linear (x,y,z) and two rotational degrees of movement. These motions enable the sample to be appropriately positioned for the XPS analysis, including angle dependent measurements. X-rays are generated by bombarding an anode material with electrons accelerated through a potential difference of about 10 kV from the filament (at ground potential). The core holes are created in the target material where the relaxation process favors photon emission. The choice of anode depends on three main considerations: (i) the spectral width should not limit the energy resolution required, (ii) the X-ray intensity should be high, and (iii) the photon energy should be large enough that a reasonable range of photoelectron kinetic energies is available for analysis. The MAX200 is equipped with a dual anode X-ray source (Figure 2.4) which gives K a radiation of either Al (1486.6 eV) or Mg (1253.6 eV) with approximate line widths of 0.7 and 0.85 eV respectively. Two separate anode filaments (Al and Mg) can be connected to the external circuit so that by switching power from one to the other enables an easy change in the X-rays produced. 26 cooling water Filament 1 Focusing shields Filament 2 Al anode face Mg anode face Al window hv Figure 2.4 Dual anode X-ray source 27 Energy analysis is done with an EA200 concentric hemispherical analyzer (CHA). Figure 2.5 schematically indicates the two metal hemispheres, as well as the collection lens and multichannel plate detector. The collection and focussing of photoelectrons towards the entry slit is done in two stages by ramping voltages on the lens elements. The first stage controls the analysis area (spot size), and collection angle for the electrons coming from the sample surface, while the second stage acts to retard the electrons to the particular pass energy required and it also controls the angle (a) at which the electrons enter the analyzer. A potential difference AV is applied between the concentric hemispheres of radii R, (inner) and R 2 (outer), and electrons can only travel on the central circular trajectory through the energy analyzer and reach the detector if their kinetic energy, E 0 inside the analyzer (the pass energy) satisfies [79] eAV = E 0(R 2/R, - R,/R2) (2.3) The MAX200 is set for particular pass energies, namely 24, 48, 96 and 192 eV. The relative analyzer resolution (AE a n a l /E 0 ) is a function of the aperture size, the electron entrance angle, and the electron trajectory radius (R0), but this is constant for a given analyzer setting. Therefore, a better energy resolution is obtained with lower pass energy, but this is also accompanied by a reduction in the signal intensity. To obtain an optimum balance between resolution and intensity (counts per second), an appropriate pass energy should be chosen for each measurement. In this work, a pass energy of 192 eV is used for survey scans, while 96 eV is used for the higher-resolution measurements. The energy resolution for a peak in a measured spectrum is expressed by the full-width-at-half-maximum (FWHM) height. This measurement contains contributions from the analyzer (AF 7u.), the natural line width of the X-ray source (AEw i i r c c), as well as the inherent line width of the atomic level involved (AE l inc). The observed peak width can be expressed by: 28 CHA Second lens stage First lens stage Emitted Photoelectrons Multichannel Detector 4r < X-ray source sample Figure 2.5 Schematic diagram for the concentric hemispherical analyzer (CHA) and lens system in the MAX200. 29 A E p c ; , k = ( A E \ maly/cr + A E2 + A E 2 .) source line' t/2 (2.4) provided all contributions have the Gaussian form [79]. The kinetic energy of a photoelectron measured in the spectrometer (E'k) is conventionally referenced to the spectrometer's vacuum level, while the binding energy of the electron inside the sample (Eh) is referenced to the Fermi energy of the sample (Figure 2.6). For a conducting sample in electrical contact with the spectrometer, so that the Fermi energies are equal, the energy balance requires and this represents a modification of Equation (2.1). The spectrometer work function (W ) remains constant while the analyzer is held under TJHV; its value is routinely checked by calibrating with a standard gold sample. The spectra shown in this thesis have been calibrated against the gold 4f7/2 peak, whose binding energy is fixed at 84.0 eV. 2.1.5 Spectral features Measured kinetic energies are converted to electron binding energies with Equation (2.5), and an XPS spectrum is then normally displayed as a plot of electron binding energy versus intensity (i.e. electron counts per second). An XPS analysis is typically performed by first taking a wide scan or survey scan spectrum, often covering a range of 1000 eV. That helps identify the elements present, and then specific features can be studied in more detail over smaller energy ranges (e.g. 10-20 eV) [69]. Figure 2.7 (a) shows a series of peaks from a survey scan spectrum from a zinc sample [71]. The number of counts attributed to the background increases with increasing binding energy (decreasing kinetic energy) due to inelastic scattering. The spectral peaks can be grouped into three types: peaks due to photoemission from core-levels, peaks from valence-levels, and peaks due to the emission of Auger electrons. The low-intensity features E \ = hv - E. - W k b sp (2.5) 30 i E \ 1 I \ \ • v ^ N \ 1 \ V sp Spectrometer Sample hv = energy of photon \V = work function of sample (i.e. energy difference between Fermi level and vacuum level) E k = kinetic energy of photoelectron with respect to vacuum level of sample E h= binding energy of electron in solid with respect to Fermi level W = work function of spectrometer E ' k = kinetic energy of photoelectron measured by spectrometer Figure 2.6 Schematic diagram of the relevant energy levels for the binding energy measured by spectrometer in XPS for a conducting sample. 31 Binding energy, eV (a) 1065 1015 1045 1035 Binding energy, eV (b) Figure 2.7 (a) XPS survey scan spectrum of Zn, and (b) narrow scan spectrum for Zn 2p excited by MgKoc radiation [71]. 32 seen between 0 and 30 eV are due to photoemission of valence electrons. The interpretation of features from the outer shell electrons may be complex, and this work entirely emphasizes structure from core levels [69]. The core-level peaks often directly reveal the electron orbital structure; in the example for zinc, the 2p, 3s, 3p and 3d levels are seen in Figure 2.7 (a). At higher resolution, doublet structure may be observed for photoelectrons from states with non-zero orbital angular momentum (e.g. p and d core levels) due to spin-orbital coupling. Figure 2.7 (b) shows a high-resolution (narrow scan) spectrum for the Zn 2p level. The component of the doublet are characterized by j values (i.e. 1/2, 3/2 for p orbitals; 3/2, 5/2 for d orbitals, and 5/2, 7/2 for f orbitals); the relative areas in each case are determined by the respective (2j+l) number of states. Figure 2.7 (b) shows the 2p1/2 and 2p3/, contributions from zinc with the peak area ratio of 1:2; the energy separation between peaks is 23.1 eV in this case. 2.7.6' Quantitative Analysis Quantitative analysis in XPS is done by making measurements of the peak areas in a spectrum, although it is first necessary to subtract out the background contributions. Factors that contribute to peak intensities include X-ray flux (f), photoelectron cross section (a) , the number of atoms per unit volume (n), the area of the sample from which the photoelectrons are collected (A), the instrumental transmission function (T) and the inelastic mean free path (k) [79]. Although during an XPS measurement, the incident photons can penetrate deeply into the solid (e.g. into the um range), not all the photoelectrons emitted from this region are readily detected. This is because many of these electrons lose energy via inelastic scattering with the solid. A relevant parameter is the mean free path (IMFP, X), which is defined as the mean distance traveled in the solid by an electron before it undergoes some inelastic scattering. 33 Empirical values of IMFP (or X) have been reported by Seah and Dench [80], and Figure 2.8 shows the dependence of this parameter on energy for different materials [81]. It is seen that for electron kinetic energies in the range 100-1000 eV, as in typical XPS and AES measurements, the IMFP is generally around 6 to 20 A. These values are sufficiently short that the XPS and AES techniques are highly surface sensitive. The sampling depth for XPS is conventionally taken to be 3X, and this gauges the depth from which 95% of a measured signal is contributed. The intensity contribution from an incremental thickness dx in the sample is: dl = f a n A T exp (-x/X) dx (2.6) Simple integration from x=0 to x=°° for a semi-finite homogeneous sample gives: I = f o n A T A (2.7) For comparing relative surface compositions of two elements in a sample, a commonly used approach is to group f, A, T and a into a sensitivity factor, S, whose value is determined by the particular instrumental settings and the particular peak under study. These factors, derived relatively to the F Is peak, are available for the MAX200 spectrometer, where the transmission function is corrected for the particular instrumental conditions used for each measurement [82]. The composition ratio for two elements in a sample can then be expressed as: n /n^KI /S .Vd/S , ) ]^ ) /^ , ) ] (2.8) In principle, the atomic ratio (n/n2) can be determined by using tabulated values of Xt and X, for the appropriate photoelectrons and sample [71], but for semi-quantitative work, the XJX^ ratio is often taken as constant and equal to unity. The I/I, ratio, from the measured peak intensities, and the ratio of sensitivity factors are used to estimate the n/n, ratio within the depth probed. This approach was used in this work for estimating relative compositions by XPS. 34 T — i i i i n 11 1—i i i i n 11 Electron energy (eV) Figure 2.8 Inelastic mean free path of electrons as a function of kinetic energy inside a solid [81]. 35 2.7.7 Data processing Some processing of raw XPS data is generally needed before doing a detailed spectral analysis to identify contributions from the surface region. This starts with a background correction, and the non-linear subtraction approach proposed by Shirley [83] (Figure 2.9) is used for the work in this thesis. The dominant contributions to the background come from inelastically scattered photoelectrons, and this method assumes that, at any point in a spectrum, the background signal is proportional to the number of electrons elastically scattered at higher kinetic energy. This correction for inelastic contributions over an energy range E, to E 2 (lower and upper kinetic energies chosen by the operator) is determined by the iterative algorithm: where N(E) is the measured count rate and the N'k(E) identify count rates after subtraction of background contributions (the index k indicates the k'h iteration). The reference background level is provided by N(E2) and C is fixed by the requirement that N\(E,) = 0. The process starts with N',(E) = 0 and continues until N' k + I ~ N' k ; usually the functional form of the Shirley background correction converges after three or four iterations. Then the background can be subtracted from a measured spectrum. High resolution spectra after a background correction often show structure arising from overlapping chemical species, and a curve synthesis procedure is then needed to identify the individual components with regard to peak position, intensity and peak width (FWHM). The mathematical approach chooses a functional form for each component profile, and simulates a measured spectrum (after subtraction of background) by a set of individual components. A mixed Gaussian/Lorentzian functional form [79] is available in the computer processing program for the MAX200 spectrometer: (2.9) E f(E) = peak height/ [l+M(E-E0)2/(32]exp{(l-M)[ln2(E-E0)2](32} (2.10) 36 Figure 2.9 Shirley non-linear background subtraction applied to a O Is narrow scan spectrum. 37 where E 0 is the energy for the maximum of an individual component peak, (3 fixes the FWHM, and M is a mixing ratio (1 for pure Lorentzian; 0 for pure Gaussian). After an initial estimation of these parameters, for all components needed to fit a spectrum, the program multiply iterates to optimize the fit between the simulated spectrum and the measured spectrum. This involves minimizing the least-squares function (%): X = { l / N r ) Z . ( Y -Y r,.)7Y }2 j=l,...N (2.11) A, t free-7 | v mca,| ln,j-' mea,j > J 1 v 1 where Y m c i j is the measured count rate at the j l h data point, Y f i l j is the corresponding value of the simulated function expressed as a sum of functions of the type in Equation (2.10); N is the number of data points and N l r a . equals N-N f i l, where N ( 1 1 is the number of parameters to be fitted through the minimizing process. It is important that all components added meet reasonable chemical criteria for the particular system being studied, and a visual comparison of the measured and simulated curves is also advised in order to guard against any weaknesses in the numerical fitting routine (e.g. insufficient components added). 2.2 Scanning electron microscopy (SEM) 2.2.1 Introduction The scanning electron microscope (SEM) is commonly used for imaging topographical features on surfaces of materials. The specimen is scanned with a highly-focussed electron beam (e.g. energy 20-50 keV) [84], and detection is made of the low-energy (e.g. < 50 eV) secondary electrons emitted. The incident beam is rastered and the secondary-electron images are collected from each specific area probed. When the whole desired area of the surface has been scanned, the secondary electron images are collected and displayed together so each region in the secondary electron image matches in a one-to-one fashion to the corresponding area probed by the incident beam. 38 The number of secondary electrons emitted per incident high-energy electron depends on the shape, chemical composition, and crystal orientation of each local area on the surface [85]. Commonly a large number of secondary electrons will be detected from elevated regions, with fewer electrons detected from depressed regions; such factors influence the contrast in the combined secondary electron images. In modern SEM instruments, the higher spatial resolution is achieved by using a field-emission electron gun for the incident beam. Details can be resolved to better than 10 nm with the most advanced equipment [84]. The first commercial scanning electron microscope was introduced in 1965 [85]. SEM shows some analogy with the reflected light microscope, but use of an incident electron beam can provide a thousand-fold increase in resolving power compared with the use of light. Also by detecting low-energy secondary electrons, the three-dimensional appearance of textured surfaces can be revealed with excellent detail. Such factors have made the SEM an indispensable tool in materials science research and development. 2.2.2 Instrumentation The Hitachi S4100 SEM (Figure 2.10) in Prof. Tiedje's laboratory in A M P E L is used for this study; it operates under a vacuum of 106 mbar, obtained by a diffusion pump backed by a rotary pump (Figure 2.11). The SEM optical column and specimen chamber is operated under high vacuum (>104 mbar) to avoid serious scattering of the electron beam by gas molecules [86]. The main components of the SEM are illustrated in Figure 2.12. The illuminating system consists of an electron gun and condenser lens assembly to focus the beam on the sample. The electron gun has three main components: a filament, a shield and an anode. Electrons produced by field emission are accelerated to about 30 keV. Scanning of the beam across the sample is accomplished by deflection coils, and the detector counts the secondary electrons emitted from 39 40 Electron gun chamber 1st intermediate chamber 2nd intermediats chamber MV-5 IP-1 ~ IP-3 Ion pumps RP-1. RP-2 Rotary pumps DP Oil diffusion pump PI-1,Pi-2 Pirani gauges Pe Penning gauge LV-1 ~LV-3 Leak valves V-1 ~ V-4 Vacuum valves MV-1 Manual airlock valve for specimen exclianae chamber MV-2 ~ MV-5 Valves for pra-evacuationi-of electron gun chamber V-1 MV-1 S.C. AIR LOCK VALVE S.E.C. VALVE Figure 2.11 Schematic representation of the pumping system for Hitachi S4100 SEM. 41 i 7-- l-'V,,.- ZZZD Aperture I Sample Figure 2.12 The illuminating/imaging system of a SEM unit 42 each region probed on the surface. Detection is by scintillator and photomultiplier, and the final amplified electron signal can be displayed on the cathode-ray tube (CRT), and photographs taken for permanent record. 2.3 Energy-Dispersive X-ray (EDX) Spectroscopy 2.3.1 Introduction An energy-dispersive X-ray (EDX) spectrometer is a common addition for providing localized elemental analysis of samples in conjunction with SEM. EDX depends on the sample being irradiated by a finely-focussed energetic electron beam so that X-rays can be produced according to the process in Figure 2.1 (b). These X-rays can come from deeper in the material (e.g. from depths up to the um range), but they have characteristic energies and wavelengths according to the nature of the atom present. The elemental composition can be revealed by measuring either property, but in general EDX is more common than wavelength-dispersive spectroscopy for combining with SEM [86]. 2.3.2 Instrumentation The EDX spectrometer on the Hitachi S2300 SEM in the Department of Metals and Materials Engineering was used in studies reported here. Detection of X-rays is made with a lithium-drifted silicon diode (Figure 2.13); this detector is composed of silicon doped with lithium. The drift zone contains equal numbers of L i + donors and B" acceptors, and they are evenly distributed in the active area of the diode. The lithium atoms diffuse into the crystal to electrically compensate for the effect of the boron impurities present [87, 88]. The front of the diode is covered by a thin gold film, which during operation is reverse biased by 1000 V to establish a depletion zone. Absorption of X-rays in this region can produce electron hole pairs that are 43 Preamplifier -4-Linear amplifier • ADC • FET Cold finger 200 A Gold Li-drifted < zone • holes electrons < X-rays -H k-0.1 um dead layer Ultra thin window M C A ADC: analog-to-digital converter MCA: multichannel analyzer Figure 2.13 Construction of lithium-drifted silicon detector for EDX spectrometer [86] 44 swept out by the bias to manifest charge pulses on opposite sides of the diode. The X-ray photons that reach the detector produce pulses that are converted to digital form by the analog-to-digital converter (ADC); they are then sorted, stored, and displayed in a multichannel analyzer (MCA). Data acquisition is a real-time event, in which each pulse is sorted by energy and entered into the appropriate channel. The spectrum is displayed as counts per second on the vertical axis and energy in keV on the horizontal axis (Figure 2.14) [86]. EDX is normally most efficient for the analysis of X-rays with energies in the range 0.7-20 keV; except for elements with low atomic number (Z < 10), all elements will release at least one X-ray within this range of energies. The spectrometer resolution at below 0.7 keV may be degraded by X-ray absorption, but that is minimized in the EDX system used in this study since it uses an ultra thin window (UTW) instead of the regular beryllium window. This allows detection of lower energy X-rays (> -0.2 keV) for elements with low atomic number (C and up) [88]. For X-ray energies greater than 20 keV, the spectrometer resolution is diminished because these energetic X-rays possess significant probabilities for passing completely through the Si(Li) detector without producing the electron hole pairs [86]. 2.4 Electrochemical Polarization Curve Measurement for Corrosion Assessment 2.4.1 Introduction Figure 2.15 shows a polarization curve, which is presented as a plot of potential versus logarithm of current (or current density). Polarization is described as the extent of change in electrode potential from the equilibrium value associated with a net current flow at the electrode [89]. The polarization curve is measured by a potentiostat as shown schematically in Figure 2.16. The potential is measured between the working electrode and the reference electrode, but the measured current is between working electrode and counter electrode (the potentiostat ensures 45 Relative Counts A! Zn p I ' I 1 I r -Zn 6 8 9 Energy/ keV Figure 2.14 EDX spectra of AA6061-T6 aluminum alloy sample after coating with zinc phosphate solution. 46 Figure 2.15 Electrochemical polarization curve for a corrosion system. 47 Figure 2.16 Schematic of potentiostatic polarization circuit: A identifies working electrode; B reference electrode; G counter electrode; HL Haber-Luggin capillary; MV amplifier; U voltmeter and I ammeter 48 current in the working-electrode-to-reference-electrode circuit is zero). Because the area of the working electrode (i.e. that for which corrosion effects are being studied) is constant, the plot in Figure 2.15 is frequently given in terms of current density. The measured polarization curves often show two linear regions which correspond to Tafel plots (i.e. the Tafel regions in Figure 2.15). The Tafel equation is derived from the Butler-Volmer equation [90] for the kinetics of irreversible electron transfer at an electrode in a well stirred solution. The general Tafel equation [91] is ti = p log (i/io) (2.12) where (3 has the form pc = -2.303RT/aF (2.13) for the cathodic current, and (3a = 2.303RT/(l-a)F (2.14) for the anodic current. In Equation (2.12), i is the measured current density, r\ is the overpotential (i.e. potential change from the equilibrium value), io is the exchange current density which corresponds to the anodic and cathodic currents flowing when the overpotential is zero (i.e. at equilibrium where these two currents are equal and opposite). In Equations (2.13) and (2.14), R is the gas constant, F is the Faraday constant, T is temperature, and a is the transfer coefficient (lies in the range 0 to 1 and depends on how the applied potential modifies the activation energy in the Butler-Volmer treatment [90, 92]). The parameters of a Tafel analysis, io and a, can be obtained from the straight-line slopes of measured polarization curves for the anodic and cathodic Tafel regions. The current flowing in a polarization curve measurement is a direct indication of the rate of the associated corrosion reaction. In practice, the current density io from a Tafel analysis indicates stability to corrosion insofar as the smaller io, the better the corrosion protection. 49 However, in general, for an actual metal sample the situation may be more complicated with the electrode being in contact with two (or more) oxidation-reduction systems. Then as the metal corrodes, both the anodic and cathodic half-cell reactions occur simultaneously on the surface. Each has its own half-cell electrode potential and exchange current density, but since the two half-cell electrode potentials cannot coexist separately on an electrically conductive surface, they change to a common intermediate value, E a i n , which is called the corrosion potential. An example of such a mixed potential system is shown in Figure 2.17 for zinc corroding in hydrochloric acid. At E a i r i . , the anodic and cathodic reaction rates are equal, and the current density, i c o i r , still provides a direct measure of the corrosion rate [25, 93]. 2.4.2 Instrumentation The polarization measurements were made with a Solartron 1286 Electrochemical Interface (Schlumberger Technologies) using the polarization circuit shown schematically in Figure 2.16 [94]. The potential between working electrode and the reference electrode is stepped linearly and the associated current is recorded for each setting. Solartron Corrware and Corrview software were used for data acquisition and for plotting polarization curves. In this work, the corrosion current density, imn is usually estimated using the Solartron Corrview software from the intercept of the extrapolated cathodic plot (from two appropriate points on the Tafel-like curve) with the observed corrosion potential. The reference electrode is in contact with the solution close to the working electrode through a Haber-Luggin capillary. This helps eliminate any ohmic potential may otherwise act to obscure Tafel behavior [94, 95]. Measurements of polarization curves directly give electron transfer rates, but without necessarily identifying the actual chemical reaction. The polarization 50 E(V) i (A/cm2) Figure 2.17 Electrode kinetic behavior of pure zinc in acid solution [25]. 51 curves measured in this work are entirely used to follow trends in behavior, shown by measured values of iam and E a i r r , as changes are made in the coatings applied to the aluminum surfaces. 52 Chapter 3 Factors affecting Zinc Phosphate Coatings on AA6061-T6 Aluminum Alloy 3.1 Introduction Zinc phosphate can form a crystalline conversion coating in the form of hopeite, Zn3(P04)2-4H20 by precipitating on aluminum from solutions containing Zn 2 + , H2P04~ and some F" [54]. The phosphating process can be sensitive to many factors, but the basic reactions are: Pickling reaction: A1203 + 6H + - » 2A13+ + 3H 20 (3.1) Al + 3FT - » A f + + 3/2H2(g)T (3.2) Coating formation: 3Zn2 + + 2H 2 PO; + 4FLO ^ Zn3(P04)2-4H20 + 4H + (3.3) Complex formation: A13+ + 6F" - » A1F 6 U (3.4) Sankara Narayanan [96] has reviewed various factors affecting the phosphating process on steel, but not many studies have been done on the aluminum alloys. Different phosphating baths react differently depending on specific values of basic parameters like pH and fluoride ion concentration. Factors affecting the formulation of a phosphating bath currently used in our laboratory have been studied mainly for the AA7075-T6 aluminum alloy [10, 15-18], although some initial observations have been made recently for AA6061-T6 aluminum alloy [19]. Nevertheless, further studies are required in order to find the best phosphating recipe for the AA6061-T6 alloy, and some basic parameters under investigation here include the effects of the P-to-Zn atomic ratio, pH and different fluoride ion concentration [97]. 53 3.2 Experimental The compositions in wt% for AA6061-T6 aluminum alloy (also called Al-Mg-Si alloy) are 0.8-1.2 Mg, 0.25 Zn, 0.15-0.4 Cu, 0.7 Fe, 0.4-0.8 Si, 0.15 Mn, 0.15 Ti and 0.04-0.35 Cr. It is the most common extrusion alloy, and mainly used in the building industry [2]. AA6061-T6 aluminum alloy samples were cut into square panels with dimensions 1x1x0.12 cm3 for characterization of phosphate coatings, while samples with dimensions 1x3x0.12 cm' were used for studying the electrochemical corrosion properties. All samples were polished prior to treatment using a mechanical polisher (Metaserv Universal Polish); 400-grit aluminum oxide sandpaper was used first, and this is followed by 1200-grit sandpaper. The polishing was done in a water flow in order to wash out residues from the metal surface. The samples were then degreased in an ultrasonic bath, first in acetone and then in methanol, for 1 min each. Some samples were subsequently given an acid etching pre-treatment by immersing for 2 min in 50/50-vol% H 2 S0 4 -H 2 0 solution at 40"C. Three key reference solutions were used for the phosphating baths. Their designations and compositions are: n-Zn reference solution which contains 5.3 g ZnO, 16 mL (85%) H,P0 4 and 0.59 g NaF in 1 L solution; h-Zn reference solution which has 5.3 g ZnO, 16 mL (85%) H 3 P0 4 , 89.25 g NaH 2 P0 4 .H 2 0 and 0.59 g NaF in 1 L solution; and /-Zn reference solution with 1.0 g ZnO, 8 mL H,P0 4 (85%), 6.3 g NaH 2 P0 4 H 2 0 and 0.59 g NaF in 1 L of solution. These reference solutions all have 219 ppm F"; n-Zn has pH 2 and P-to-Zn atomic ratio of 4; h-Zn has pH 3 and P-to-Zn atomic ratio of 15; and /-Zn has pH 2 and P-to-Zn atomic ratio of 15. Various modifications of these reference solutions are used through this research. Modifications of n-Zn define the normal-zinc formulations, similarly modifications of h-Zn and /-Zn define the high-ratio normal-zinc and low-zinc formulations respectively. The pH adjustments for these 54 modified phosphating solutions were made by adding sodium hydroxide or acetic acid; measurements were made with an Orion pH meter (model 420A). All coatings were done by dipping the pre-treated alloy samples into the various phosphating solutions for 3 min at 60°C in a water bath (Microprocessor 280 Series). The strength of adhesion of coating to the metal was assessed by the Scotch-tape test just after its formation. That is done by simply sticking Scotch-tape to the coated sample and then pulling away from the surface. The Scotch-tape was then examined to see if there are any coating particles present. Samples for studying corrosion resistance were covered with epoxy paint except for a 1 cm2 area which was exposed to 3.5% NaCl solution. The electrochemical corrosion properties of the alloy were measured via polarization curves, which were obtained using a Solartron model 1286 Electrochemical Interface (Schlumberger Technologies) with a scan rate of 1 mV/s. Other details were summarized in Section 2.4.2. A set of important coating conditions for the work in this chapter is specified in Table 3.1. The notation for the different coatings uses an N if it was formed from a variant on the n-Zn reference solution; similarly H and L are used as codes for the coatings when formed from h-Zn and 1-Zn type solutions. 3.3 Effect of different P-to-Zn atomic ratio Both normal-zinc and low-zinc phosphating are used in commercial applications, although the low-zinc phosphating approach is more popular. The differences depend not only on the P-to-Zn atomic ratio (for which the low-zinc formulations have higher P-to-Zn values), but also on the total amount of zinc used. Normal-zinc solutions typically have Zn concentrations of 2-4 g/L, while the P 20, concentrations are in the range 5-10 g/L. By contrast, low-zinc solutions have Zn 55 Table 3.1 Specification of initial pre-treatment and coating solutions used for AA6061-T6 samples discussed in text Sample code Treatment NI Mech. polished, n-Zn ref. soln.* N2 Acid etched, n-Zn ref. soln. N3 Mech. polished, modified n-Zn soln (pH 4 and 219 ppm F~) N4 Mech. polished, modified n-Zn soln (pH 4 and 800 ppm F~) HI Mech. polished, h-Zn ref. soln.* H2 Acid etched, h-Zn ref. soln. H3 Mech. polished, modified h-Zn soln (pH 4 and 219 ppm F") LI Mech. polished, /-Zn ref. soln.* L2 Acid etched, /-Zn ref. soln. L3 Mech. polished, modified /-Zn soln. (pH 4 and 219 ppm F~) L4 Mech. polished, modified /-Zn soln. (pH 4 and 800 ppm F~) *The three reference solutions: n-Zn ref. soln. (pH 2, 219 ppm F), h-Zn ref. soln. (pH 3, 219 ppm F) and /-Zn ref. sol. (pH 2, 219 ppm F) refer respectively to the normal-zinc, high-ratio normal-zinc and low-zinc reference solutions specified in Section 3.2. The other solutions are modified from the reference solutions only by changes noted in parenthesis. 56 concentrations in the range 0.4-1.7 g/L, and P 2 0 5 concentrations of 11-16 g/L [14, 40, 41, 54]. Donofrio [54] in his review of zinc phosphate coating on steel reported that low-zinc coating baths generally provide better quality coatings because of the associated longer pickling times and slower deposition rates. Studies are done here to determine the effect of different P-to-Zn atomic ratios on zinc phosphate coatings formed on AA6061-T6 aluminum alloy surfaces. 3.3.1 SEM studies Different P-to-Zn atomic ratios were used initially by adjusting the amount of phosphoric acid (H3P04) in the normal-zinc formulation. SEM micrographs for the coatings formed for P-to-Zn atomic ratio values of 3, 4 (sample NI), 5 and 6 are shown in Figure 3.1. Generally crystal growth occurs along lines or scratches created by the mechanical polishing. However, for coating solutions with P-to-Zn atomic ratio of 5 or higher (pH value is less than 2), it can be observed that no coating occurs and some second phase particles are revealed. The pH of the coating solutions with P-to-Zn atomic ratio equal to 3 and 4 are both very close to 2. The presence of excess phosphoric acid will also affect Equation (3.3) and prevent the formation of zinc phosphate coating. These observations are consistent with the conclusion that phosphate coatings cannot be produced with excess H 3 P0 4 [59], which will act to emphasize the aluminum dissolution. In order to study the effects of higher P-to-Zn atomic ratios in the coating solution, but without having excess H 3 P0 4 , NaH 2 P0 4 H 2 0 was added as an alternative source of phosphate. Figure 3.2 shows SEM micrographs for AA6061-T6 alloy after coating with n-Zn reference solution (NI), with P-to-Zn atomic ratio 4 and modified n-Zn solutions with the P-to-Zn atomic ratio adjusted to 5, 10, 15, 30 and 40 (no NaH 2 P0 4 H 2 0 is added for sample NI and so it provides a reference for assessing the effect of the additions). The same amount of phosphoric acid (16 57 Figure 3.1 S E M images (2500x) of AA6061-T6 alloy after treatment in different normal-zinc formulations using H3PO4 to adjust the P-to-Zn atomic ratio: (a) 3, (b) 4 (sample N I ) , ( c )5and (d) 6. 58 (e) (0 Figure 3.2 S E M images (2500x) of AA6061-T6 alloy after treatment in different normal-zinc formulations using fixed H3PO4 and varying N a H 2 P 0 4 to adjust the P-to-Zn atomic ratio: (a) 4 (sample NI ) , (b) 5, (c) 10, (d) 15 (sample HI ) , (e) 30 and (f)40. 59 ml/L 85% H 3 P0 4 ) was added as in the treatment for forming sample NI but the P-to-Zn atomic ratio was increased by adding the N a H 2 P 0 4 H 2 0 . It appears that the P-to-Zn atomic ratio of 15 provides the best coating morphology insofar as the surface is covered with a higher density of small crystalline particles (dimension -500 nm) as compared to the other ratios. Generally, the coatings are composed of clusters (~ 15 um dimension) of crystallites. Leblanc [98] reported that smaller particle size both enhances corrosion protection and improves the adhesion property. It is therefore concluded that the P-to-Zn atomic ratio of 15 (HI) is favored for the coatings compared in Figure 3.2. Figure 3.3 shows S E M micrographs for coatings formed from normal-zinc formulations with different phosphate species used. A l l the phosphating baths used to obtain coatings in Figure 3.3 have the P-to-Zn atomic ratio equal to 15, obtained by different combinations of phosphate species. It is observed that little coating is obtained when the secondary phosphate salt Na,HP0 4-7H,0 is used with H 3 P0 4 (Figure 3.3b). Probably phosphate is precipitating out as ZnHP0 4-3H,0 [49], so preventing availability of H,P0 4" to form the tertiary phosphate Zn,P0 4-4H 20 (hopeite) according to Equation (3.3). When only the primary phosphate salt, NaH 2 P0 4 H,0 , is used, with pH adjustment to 3 by addition of acetic acid, the coating is observed to contain slightly larger coating crystallites at low coverage (Figure 3.3c). This observation is consistent with the conclusion by Freeman [12] and Sankara Narayanan [35] that a small amount of free phosphoric acid must always be present in the phosphating bath. When only H,P0 4 was used, and the pH was adjusted to 3 by addition of NaOH, the coating formed consists of large clusters with little evidence for the small crystallites (Figure 3.3d). Therefore, in order to form favorable coatings at a high P-to-Zn atomic ratio, a combination of free phosphoric acid, H,P0 4 , and primary phosphate salt, NaH 2 P0 4 H,0 , is preferred. Formulations of coating baths with P-to-Zn atomic ratio 15 are referred to here as being of the high-ratio 60 (a) (b) Figure 3.3 SEM images (2500x) of AA6061-T6 alloy after treatment with normal-zinc formulations at pH 3, P-to-Zn atomic ratio equals 15, using different sources of phosphate: (a) NaH 2 P0 4 and H 3 P 0 4 (sample HI), (b) Na 2 HP0 4 and H 3 P 0 4 , (c) N a H 2 P 0 4 and (d) H 3 P 0 4 . 61 normal-zinc (h-Zn) type, since they do not correspond within either the conventional normal-zinc or low-zinc categories. The effect of low-zinc solutions (/-Zn) is also under investigation here. Figure 3.4 shows S E M images for AA6061-T6 aluminum alloy samples when coated with the three key reference solutions: n-Zn, h-Zn and /-Zn reference solutions. Samples NI (n-Zn), HI (h-Zn) and LI (/-Zn) were all obtained on alloy surfaces which had just been mechanically polished. Table 3.1 specifies modifications to these processes. From the S E M images, sample LI has a similar coating coverage as sample N I , but the coated particles are smaller in L I . By contrast, sample HI shows a better distribution of small crystallites, and the clusters are smaller than those in samples NI and L I . It is concluded that HI has the best coating morphology for the parameters under investigation here. 3.3.2 XPS study XPS was used to assess the chemical composition of the mechanically polished (blank) AA6061-T6 alloy sample and of coatings formed in it. Figure 3.5a shows an XPS survey scan of the blank sample, and the presence of A l , O, C and M g is detected. Oxygen is present mainly from aluminum oxide; this may include both oxide which is incompletely removed by polishing, as well as some re-growth after the polishing. Some oxygen, as well as carbon, is also expected from air-borne contamination and residue from the degreasing procedure. The M g is present as an intrinsic element of the AA6061-T6 aluminum alloy, although this element is not detected by XPS after the coatings have been applied. The XPS survey scan of AA6061-T6 alloy after coating with the n-Zn reference solution (sample NI) shows the presence of Zn, P, O, C and A l (Figure 3.5b). Table 3.2 gives percentage compositions measured by XPS for this alloy after treatment with different zinc phosphate 62 Mechanically polished Acid etched X N I H I N2 L I L2 Figure 3.4 Comparison of SEM images (lOOOx) for coatings formed by the different reference solutions (Table 3.1) on AA6061-T6 alloy surfaces that have been mechanical polished or acid etched according to the pre-treatments detailed in Section 3.2. 63 Intensity (a.u) 1200 1 0 0 0 800 600 400 2 0 0 Binding energy Figure 3.5 XPS survey scans from AA6061-T6 alloy samples: (a) after mechanical polishing and (b) after a subsequent coating with the n-Zn reference solution (sample N1) 64 Table 3.2 Percentage compositions and Zn/Al and P/Al ratios from XPS for coatings formed from different phosphating baths specified in Table 3.1. Sample Zn 0 P Al(3+) Al(0) Zn/Al P/AI N1 12.3 66.6 7.9 6.3 6.8 0.9 0.6 N2 13.4 67.7 6.3 6.4 6.1 1.1 0.5 N3 11.5 66.0 15.2 5.9 1.4 1.6 2.1 N4 7.7 67.3 12.1 13.0 0.0 0.6 0.9 H1 10.8 59.8 15.8 5.6 7.9 0.8 1.2 H2 12.0 57.8 12.4 8.5 9.3 0.7 0.7 H3 12.6 58.2 22.4 9.5 7.9 0.7 1.3 L1 9.1 67.2 4.5 9.5 9.8 0.5 0.3 L2 11.8 66.9 3.7 10.0 7.5 0.7 0.2 L3 14.9 71.0 10.3 3.1 0.8 3.8 2.6 L4 19.8 70.5 6.1 3.7 0.0 5.4 1.7 65 coating procedures defined in Table 3.1. In all cases, the values apply within the probing depth of XPS, and these values usually represent averages over several independent measurements. All indications are that the C is concentrated at the surface, and arises from air-borne contamination, and therefore it is not considered to be an inherent component of the coating. Accordingly, this element has not been included in the relative compositions reported in Table 3.2. High-resolution Al 2p spectra show that this element is present in both metallic and 3+ oxidized forms, and these spectra were used for the allocations made in preparing Table 3.2. The high-resolution P 2p spectra are consistent with the 5+ oxidation state of P [18]. The EDX spectrum shown in Figure 3.6 is obtained from the particular crystallite cluster in sample NI, which is marked by an X in Figure 3.4. The large Al signal is obtained from the underlying substrate, and the presence of Zn, P and O is consistent with a thin coating of zinc phosphate crystals, as indicated in a similar situation by Ishii et al. [8]. The XPS measurements provide complementary information insofar as they assess an average composition from across the surface region of a sample to the depth probed (typically -50 A). The Zn-to-Al and P-to-Al atomic ratios from XPS analysis (Table 3.2) show that samples NI and HI have similar Zn contents, but NI has a smaller relative amount of P. SEM images (Figure 3.4) indicate that the coating coverage is similar for both samples, and therefore it is concluded that the higher amount of P in sample HI is probably due to the formation of A1P04. As compared with coating by the /-Zn reference solution (sample LI), sample NI coated by the n-Zn reference solution contains higher relative amounts of Zn and P, but less Al was detected. This indicates that the coating coverage for sample NI is higher than that of sample LI. Together, these observations are consistent with sample NI having a more extensive coating 66 Relative i Al I Counts 4 i I I I i I i l i I i — 2 3 4 5 6 7 8 9 Energy/ keV Figure 3.6 EDX spectrum measured from area marked by X in sample NI (Fig. 3.4) 67 (i.e. net increase in thickness and coverage) compared with sample LI for coatings at pH 2. However, other parameters like pH and fluoride concentration have yet to be optimized. 3.4 Effect of pre-treatment The effect of different pre-treatments was assessed in this work especially by comparing the coatings formed, after dipping in a common reference solution (3 min, 60"C), on surfaces which had first been given either a mechanical polish or an acid etch (according to procedures noted in Section 3.2). Significantly different morphologies were observed for all the coating formulations tested (Figure 3.4). Generally, the coatings formed after the acid etching pre-treatment are more uniform, with a denser structure of small crystallites, compared with the coatings formed on surfaces that had just received a mechanical polish. The relative compositions of coatings indicated by XPS analysis are given in Table 3.2, and the differences are small between the coating by the n-Zn reference solution with only a mechanical polish (NI) and one that is formed from the same solution after the acid etch in the pre-treatment (N2). A similar observation was made for coatings LI and L2 formed by the /-Zn reference solution. For samples HI and H2 formed by the h-Zn reference solution, the amount of Zn detected is similar in each case, although there is less P for sample H2 (with acid etching pre-treatment). In general, for a given coating formulation, the overall compositions are similar for all coatings formed with and without acid etching in the pre-treatment. However, coatings on samples that received the acid etching pre-treatment appear to have better morphologies, both because of more a uniform coating distribution and slightly smaller crystallites. In comparing the different pre-treatments, it is noted that mechanical polishing has the basic role of scraping off layers, whereas the acid etching process is more active at highlighting 68 heterogeneities on the surface. For example, the acid reactions involve anodic dissolution of aluminum and hydrogen evolution at local cathodic sites, including grain boundaries and second-phase particles, leading to increase in local pH [99, 100]. Such processes open up more nucleation sites for the phosphating coatings and thereby influence the ultimate morphology. 3.5 Effect of pH 3.5.1 Normal-zinc solution A range of different pH values was used in order to optimize the coating formation on AA6061-T6 alloy. Figure 3.7 shows the different species of phosphate ions present at different pHs, and SEM images show the effect of this variable on coatings formed with the normal-zinc formulation. From this figure, it can be observed that the dominant species are H 3 P0 4 for pH less than 2, H2P04" for pH between 2 and 7, HP04 2~ for pH between 7 and 12 and P043" for pH greater than 12 [101]. Test studies were done here with the pH of coating bath ranging from 2 to 13. The clear indications are that the phosphating process works best in acidic environments. Figure 3.7 shows that little coating was observed for a coating solution with pH greater than 5, where there is no free H 3 P0 4 species. This result supports the conclusion by Freeman [12] and Sankara Narayanan [35] that some H 3 P0 4 species is required for the formation of zinc phosphate coating. It seems that free phosphoric acid, H,P0 4, must always be present to repress hydrolysis and keep the bath stable for effective deposition of phosphates at the microcathodic sites. Figure 3.8 shows SEM images of coatings formed on AA6061-T6 alloy after treatment with the different reference solutions at two specific pH values. The pH of normal-zinc formulation under study here is at pH 2 (sample NI) and at pH 4 (sample N3). Sample NI appears to consist of relatively large clusters of crystallites. However, sample N3 has a more uniform coating and consists of finer crystallites (dimension -500 nm) with higher coating 69 Figure 3.7 Phosphate species present at different pHs and associated effects on normal-zinc coatings shown by S E M images (2500x). The samples at pH 2 and 4 correspond to N I and N3 respectively. 70 Original p H p H 4 NI N3 H I H3 L I L3 Figure 3.8 S E M images (lOOOx) of AA6061-T6 alloy samples coated by n-Zn, h-Zn and /-Zn reference solutions at their original pH values (NI , H I , L I respectively) and by modified n-Zn, h-Zn and /-Zn solutions at pH 4 (N3, H3, L3). 71 coverage. Such characteristics are expected to be beneficial for improved adhesion bonding and corrosion resistance [98]. The XPS signal detected from metallic Al for sample N3 is considerably reduced over that of sample NI (Table 3.2), and that is consistent with a more effective coverage by the conversion layer. This depends in part on the coating being uniform, but it also appears to be thicker. The amount of P has increased for sample N3 compared with N1, although the Zn content has reduced somewhat. That is possibly indicative of some A1P04 being incorporated into the coating. Various factors will influence the comparison between coating processes at pH values of 2 and 4. One is that the Al etching rate should reduce at the higher pH value, but the associated longer pickling time may increase the number of nucleation sites for formation of a subsequent coating. Also, the primary phosphate ion (H,P04~) concentration is greater at pH 4 compared with 2, and that could promote faster phosphate deposition [59]. This is supported by the observation made by Machu [51] that supersaturation is reached earlier at higher pH values of the solution, and that may encourage earlier crystallization. Increase in the number of nuclei at higher pH may be expected to lead to finer grains of crystallites being precipitated. The phosphating process is an electrochemical phenomenon where aluminum dissolution occurs at anodic sites, while the H+-to-H, conversion at cathodic sites helps drive the precipitation of zinc phosphate [12]. A consequence is that the details of this process depend on local pH values at micro-electrodes, as well as on the general solution pH. Machu [51] proposed that zinc phosphate deposition is initiated at micro-cathodic sites. 72 3.5.2 Low-zinc solution The effect of pH on the low-zinc formulation is similar to that just noted for the normal-zinc formulation. SEM images (Figure 3.8) show that AA6061-T6 alloy coated with low-zinc formulation at pH 4 (sample L3) has uniform small crystallites with high coating coverage compared with the coating formed at pH 2 (sample LI), which has relatively large crystals. Therefore, the morphology of coating is better at pH 4 (sample L3), and XPS analysis (Table 3.2) also indicates higher Zn-to-Al and P-to-Al ratios for sample L3. Therefore, it is concluded that a better zinc phosphate coating is present in sample L3 as compared with sample LI. At pH 4, the low-zinc formulation (sample L3) appears more effective than the normal-zinc formulation (sample N3). This can be observed from the SEM images (Figure 3.8) that sample L3 has better coverage compared with sample N3. The XPS analysis (Table 3.2) also shows higher Zn and P contents (i.e. from Zn-to-Al and P-to-Al ratios) for sample L3 as compared with sample N3. These observations together indicate that a more favorable zinc phosphate coating is present in sample L3 compared with that in sample N3. The denser phosphate coatings formed with the low-zinc formulation at pH 4 appear associated with a slower deposition reaction (Equation. 3.3) and longer pickling reactions (Equations 3.1 and 3.2). The latter may give better chemical cleaning of the surface and in turn produce more nucleation sites for the coating deposition. 3.5.3 High-ratio normal-zinc solution The high-ratio normal-zinc formulations show different behaviors with varying pH compared with the normal-zinc and low-zinc formulations. Samples coated with the high-ratio normal-zinc formulation at pH 3 (sample HI) and at pH 4 (sample H3) have similar coverages and size of crystallites (Figure 3.8). This is further supported by XPS analysis (Table 3.2), where both 73 samples HI and H3 have similar amounts of Zn and P. Therefore, unlike the situation for the normal-zinc and low-zinc formulations, the coating formed with the high-ratio normal-zinc formulation at pH 4 does not improve as compared with the situation for the coating at pH 3. 3.6 Fluoride ion concentration Fluoride ions are added to phosphating baths in order to complex A l 1 + and so prevent the formation of A1P04, which has the effect of poisoning the phosphating process. However, an excess of F~ will induce the formation of cryolite (Na,AlF6) which is believed to weaken adhesion of the coating [8]. 3.6.1 Normal-zinc solution Figure 3.9 show SEM images of AA6061-T6 alloy after coating with normal-zinc formulations through a range of F" concentration from 0 to 1000 ppm (mg/L). Little coating was observed from the bath when no F~ is added (Figure 3.9a). This observation is consistent with the conclusion by Ishii [8] that F~ must be present to inhibit the formation of A1P04. Significant coatings are observed for samples formed with the normal-zinc formulation containing between 200 and 1000 ppm F (Figure 3.9 (b) to (f)), but the coverages do not change much in this range of F~ concentration. Figure 3.10 shows XPS survey scans of AA6061-T6 alloy after coating with a normal-zinc formulation (sample N4) and with a low-zinc formulation (sample L4), both at pH 4 and 800 ppm F". The first spectrum (N4) indicates the presence of Na and a significant amount of F for sample N4, and that observation is consistent with the formation of cryolite (Na,AlF6). Therefore, it appears that the range of 200-400 ppm F" is suitable for the normal-zinc formulation at pH 4. 74 Figure 3.9 S E M images (lOOOx) of AA6061-T6 alloy after coating by some normal-zinc processes at pH 4 where the only variation is in the F~ concentration: (a) 0 ppm, (b) 219 ppm (N3), (c) 400 ppm, (d) 600 ppm, (e) 800 ppm (N4) and (f) 1000 ppm. 75 Intensity (a.u) 1200 1000 800 600 400 200 Binding energy (eV) Figure 3.10 XPS survey scans from AA6061T6 alloy samples coated at pH 4, 800 ppm F" by: (a) normal-zinc formulation (N4) and (b) low-zinc formulation (L4). 76 3.6.2 Low-zinc solution Initial tests showed with low-zinc formulations that little coating is formed when no F" is added to the phosphating solution. Figure 3.11 shows SEM images of AA6061-T6 alloy after coating with low-zinc formulations at pH 4 with varying F" concentrations from about 200 to 1000 ppm. Reasonably good coverages are observed when F" is present in the 200-400 ppm range, but the coating coverage appears complete for the 600-1000 ppm range (this is consistent with the range given by Rossio [40, 41]). XPS data (Table 3.2) shows that the increase in F" yields more Zn, but less P (from the Zn-to-Al and P-to-Al ratios) for sample L4 as compared with sample L3. This suggests that more zinc phosphate (and less A1P04) is formed in sample L4. The XPS survey scan spectrum (Figure 3.10) of AA6061-T6 aluminum alloy after coating with the low-zinc formulation at pH 4 with 800 ppm F~ (sample L4) does not indicate the presence of Na. Therefore, it is presumed that cryolite (Na3AlF6) is not significantly present in sample L4. It is concluded that effective F~ concentrations for low-zinc formulations at pH 4 are in the range 600 to 1000 ppm F . The coating appears loose (Scotch tape test) for samples coated with solutions containing more than 1500 ppm F". Overall, it is apparent that the optimal amount of F" for this coating process is dependent on other parameters of the coating bath, including P-to-Zn atomic ratio and pH. 3.7 Corrosion studies The electrochemical polarization method was used for gauging corrosion rates in selected samples. The main objective is to compare the effectiveness of coatings formed from different types of phosphating bath, including variation in their pH values. 77 Figure 3.11 S E M images (lOOOx) of AA6061T6 alloy after coating by some low-zinc processes at pH 4 where the only variation is in the F~ concentration: (a) 219 ppm (L3), (b) 400 ppm, (c) 600 ppm, (d) 800 ppm (L4) and (e) 1000 ppm. 78 Figure 3.12 shows polarization curves measured for AA6061-T6 aluminum alloy after coating with three zinc phosphate reference solutions: n-Zn to form sample NI, h-Zn to form sample HI and /-Zn to form sample LI; all these coatings were made at the original pH for the reference solutions (i.e. 2, 3 and 2 respectively). The current densities deduced from the extrapolated Tafel plots are tabulated in Table 3.3. It is seen that samples NI and LI have similar current density values, and hence it is concluded that they have similar corrosion rates. By contrast, sample HI has the smallest current density, and that is taken to indicate that sample HI provides the best corrosion protection. This supports the view that the presence of smaller crystallites in the coating helps enhance the corrosion protection. Figure 3.13 shows the corresponding polarization curves measured after forming the coatings at pH 4. The current densities reported in Table 3.3 suggest that all coatings produced at pH 4 enhance the corrosion resistance compared with values measured with coatings formed at the original pH values for the three reference solutions. The improvement for sample H3, compared with HI, is only small, but the improvements for N3 compared with NI, and for L4 compared with LI , are more significant. The latter changes appear as a consequence of the more uniform coverages with small crystallite coatings when formed at pH 4 compared with the lower pH values. Figure 3.14 shows SEM images of AA6061-T6 alloy samples when coated with the normal-zinc formulation at pH 2 (sample NI) and at pH 4 (sample N3) before and after the electrochemical polarization corrosion test. These measurements necessarily drive the samples to corrode, and therefore substantial changes in the coatings necessarily occur after the corrosion test. This is quite different from the simple immersion or salt spray tests, which depend on visual observation to evaluate the effectiveness of coatings formed. It is observed that after the measurements of polarization curves, no coatings are left on samples NI and N3. These surfaces 79 E (Volts) -1 0 • • • yf" -1 1 _==- f^-^ -1 2 = ^ = 5 ^ 7 N %. X \ %- \ \ \ \ N I -1 .3 \ \ \ * \ * \ \ \\A LI • \ \ -1 4 \<\ ^ HI \ \ \ \ \ 1 t V -1 5 • i i i i 1 1 1 J i i i i i i I I 11 i i i i 11111 i i i i 1 1 1 n i i i i 1 1 1 ii l 1 l l 1 l II 1CT" IO"1 10° io 1 10" 10" io 4 I (uA/cm2) Figure 3.12 Polarization curves measured from AA6061-T6 alloy after coating with n-Zn reference solution (NI), /-Zn reference solution (LI) and h-Zn reference solution (HI). 80 Table 3.3 Current density values (uA/crrr) measured for AA6061-T6 samples after coating with some procedures specified in Table 3.1. Sample Current density N1 9.3 H1 5.9 L1 8.8 N3 1.6 H3 3.0 L4 1.6 81 E (Volts) -0.9 , N3 -1 0 L4 -1 1 -^ ^xS^*** ^ H3 -1 -1 2 3 - \\ • \ * V -1 4 - \ -1 5 i i i i i n 11 1 1 1 1 1 M l 1 1 1 1 1 1 M l 1 1 1 1 1 1 M l 1 ' 1 1 1 11 1 1 i i i 1 1 1 1 1 IO"" 10"" 10 _ 1 10° 10 1 10" 10 J I (uA/cm2) ;ure 3.13 Polarization curves measured from AA6061-T6 alloy after coating with modified n-Zn process (N3), modified /-Zn process (L4) and modified h-Zn process (H3), all at pH 4. 82 10 um (a) (b) Figure 3.14 S E M images (lOOOx) before and after dipping in 3.5% NaCl solution for AA6061-T6 alloy samples coated by normal-zinc formulation at different pH values: (a) pH 2 (sample N I ) and (b) pH 4 (sample N3). 83 end up in a rougher state; also pitting holes may form as is usually the case for aluminum corrosion in chloride media. The low-zinc formulation at pH 4 with 800 ppm F" results in a coating that shows a promising corrosion protection ability, and the performance of this coating appears comparable with that formed by the normal-zinc formulation at pH 4, 219 ppm F". This study supports the concept that coatings with good coverages of relatively small particles are helpful to improve corrosion resistance [98]. 84 Chapter 4 Effect of copper enrichment on zinc phosphating of AA2024-T3 aluminum alloy 4.1 Introduction In order to improve the properties of aluminum (e.g. mechanical strength, ductility, electrical conductivity, etc.) for practical applications, this element is alloyed with other metals. The major alloying elements, either singly or in combination, are copper, magnesium, silicon, manganese, nickel and zinc [20]; Table 1.1 shows a classification of aluminum alloys according to the different added elements. An alloy may have chemically heterogeneous regions in addition to the main alloy matrix, and in particular there may be a modified alloy layer immediately beneath the outermost oxide film, which can also significantly influence properties of an aluminum alloy. Bond et al. [60] showed that pits initiate near microsegregated iron, copper and silicon impurities in high purity aluminum, and knowledge of the enriched alloy layers developed during surface pre-treatments (e.g. chemical polishing [61], electropolishing [62] and alkaline etching [63]), is needed to understand what is happening during the surface treatment of aluminum alloys. The enrichment of copper during anodizing has been studied [61, 64, 102-105], and Habazaki et al. presented an overview of alloying elements segregating to a surface during this process [106]. Elemental enrichment at an interface can, in certain alloys, assist the formation of flaws within the oxide film so weakening the corrosion resistance [103], and it may also influence the electrochemical behavior of the alloy [107] as well as resistance to pitting [108]. Such enrichments are not just limited to the elements that have been deliberately added, but they can also involve accumulation of impurities at the interface [63], which may have adverse effects on the corrosion behavior. 85 The effect of copper enrichment on zinc phosphate conversion coatings has not been reported previously, but that provides the subject for the current chapter of this thesis. The emphasis is on the effect of copper enrichment due to acid etching in pre-treatments. 4.2 Principles associated with surface enrichment of aluminum alloys The surface enrichment of an alloying element at aluminum has been revealed to proceed in two stages (Figure 4.1) [109, 110]. The first stage of the oxidation of aluminum involves 0~ ions migrating into the solid while A l 3 + ions migrate upwards to form the A1,0, region. At the metal interface, there will be atoms of the alloying element M , and basically there is a competition at this point as to whether the Al or M preferentially oxidizes. This is determined by which oxide has the lower Gibbs free energy of formation on an equivalent basis. If the oxide of M has a lower Gibbs free energy of formation per equivalent than A1,0, then M m + ions are expected to migrate into the oxide layer along with the A l 3 + ions (Al is in excess in the whole alloy). Alternatively if the oxide of M has a higher Gibbs free energy of formation per equivalent than A120,, M is expected not to oxidize initially but to enrich at the metal-oxide interface (since the Al will continue to oxidize). In the later case, an excess of M can build up at the interface and eventually will oxidize as the dominant element exposed to the oxygen attack. That corresponds to the second stage of anodic oxidation indicated schematically in Figure 4.1 (c). As the oxidized alloying element moves outwards, the steady-state distribution in the oxide layer will depend on the relative migration rate of M n , + ions compared with A l 3 + [61, 104]. Habazaki et al. [64, 106] measured enrichments for binary aluminum alloys containing chromium, copper, gold, molybdenum, niobium, tantalum, titanium, tungsten and zinc when subjected to a particular anodizing procedure. Figure 4.2 plots the degree of enrichment measured against the Gibbs free energy of formation per equivalent of the alloying element 86 electrolyte electrolyte electrolyte marker T Aluminum alloy enriched alloy ^ layer A l 3+ Aluminum alloy Aluminum alloy (a) (b) (c) Figure 4.1 Illustration of the development of an enriched alloy layer during oxide film growth on aluminum: (a) initial condition, (b) first-stage formation of alumina film, (c) second-stage oxidation of alloying element, M . 87 Enrichment (IO 1 5 atoms cm"2) | Z r 0 2 BaO AG7eq (kJ mol"1) j !!Nd 2Oj BcO j Th02 Y 2 0 3 , CaO, Sc 2 Oj Figure 4.2 Gibbs free energy of formation per equivalent for different oxides compared with enrichments observed when aluminum alloys are anodized at 5 mA cm 2 (the bold entries corresponds to actual measurements) [64, 106]. 88 oxide, and this correlation is extended to other oxides. Elements whose oxides are more positive on the x-axis of this plot than A12Q, have a thermodynamic tendency to enrich at the metal-oxide interface, whereas those that are more negative (e.g. Mg) are not expected to enrich at the interface. An investigation of this was made by studying samples of AA6061-T6 aluminum alloy (with Mg as the major alloying element). Figure 4.3 compares XPS spectra before and after 2 min of acid etching. In this case, there is a thermodynamic driving force for MgO formation; it simply migrates into the oxide layer along with A1203, and does not show any special Mg interfacial enrichment. 4.3 Pre-treatments of AA2024-T3 aluminum alloy AA2024-T3 aluminum alloy (composition wt%: 1.2-1.8 Mg, 0.25 Zn, 3.8-4.9 Cu, 0.5 Fe, 0.5 Si, 0.30-0.9 Mn, 0.10-0.15 Ti and 0.10 Cr), also known as Al-Cu alloy, is used in the studies of this chapter. All samples were initially polished using a mechanical polisher (Metaseiv Universal Polish) with aluminum oxide sandpaper (first 400-grit, then 1200-grit). A water flow is maintained during polishing in order to wash out debris from the metal surface. The samples were then sequentially degreased in acetone and methanol in an ultrasonic bath for 1 min each. These treatments were followed by an acid etching step, which was done with 50/50-vol% H 2 S0 4 -H 2 0 at 40"C. The copper contents of the surfaces resulting after these pre-treatments were analyzed using XPS, and Figure 4.4 reports measurements of Cu 2p spectra and peak areas for different times of acid etching. The sample that had the mechanical polishing, but no acid etching, did not reveal any significant copper content in the surface region. It is observed that the increase in etching time to 5 min only slightly increases the copper content. However, a substantial increase in copper was observed when the time of etching was increased from 5 to 10 min. The 2p 89 Figure 4.3 XPS spectra of AA6061-T6 aluminum alloy; (a) after mechanical polishing and (b) after subsequent acid etching. 90 2024 etch with H 2 S 0 4 120000 -i— 1 r — — . 1 1 T • r o min _ i , , , i i i i i i i i . i i . i i i i 960 950 940 930 920 Binding Energy (eV) Figure 4.4 Copper enrichment by XPS obtained for different acid etching times applied to AA2024-T3 aluminum alloy. 91 photoelectron and Auger spectra of copper show that this segregated element is in the metallic form, and therefore it is concluded that the enriched copper layer is located just beneath the oxide film [65]. This result corresponds to the independent conclusion made by Habazaki et al. with a study using transmission electron microscopy (TEM) [61, 64]. The interfacial enrichment noted for the Al-Cu alloy clearly modifies the local alloy composition. It must therefore be expected to have a significant effect on the behavior of the alloy/film interface, for example as a result of the interfacial stress created by the changes in lattice parameter and structure with changing composition of the enriched layer. This opens the question of whether such enrichments can be used to optimize the materials in relation to their physical, chemical, mechanical, electrical and structural properties, and especially whether the amount of enrichment can be controlled at the pre-treatment stage for such purposes. The next section develops this idea by considering the effect of copper enrichment on the zinc phosphate coating process. 4.4 Effect of copper enrichment on zinc phosphate coating 4.4.1 General observations In order to study the effects of copper enrichment on the zinc phosphate coating process, different times were selected for the acid etching pre-treatment applied to the AA2024-T3 aluminum alloy. The zinc phosphate solution used in this study is the high-ratio normal-zinc (h-Zn) reference solution; its composition is described in Section 3.2, and the same coating procedures are used here as summarized in the previous chapter. Figure 4.5 shows SEM micrographs of zinc phosphate coatings formed on AA2024-T3 aluminum alloy for different acid etching pre-treatment times (0-10 min). It is clear that the 92 S E M images (2500x) of zinc phosphate coatings formed on AA2024-T3 samples after different pre-treatments: (a) mechanical polishing only, (b) 2 min acid etch, (c) 5 min acid etch, and (d) 10 min acid etch. 93 different etching times have appreciable effects on the coverages and morphologies of the coatings formed [65]. For the sample pre-treated with only a mechanical polish, the zinc phosphate coating shows local clusters of 10 um dimension, which correspond to individual grains (Figure 4.5 a). In addition, it is clear that the coating shows a tendency to form along the scratches produced by the mechanical polishing. A more uniform distribution of crystal particles with smaller clusters was observed for coatings formed after 2 min of acid etching in the pre-treatment (Figure 4.5 b). When the time of acid etching is increased to 5 min, the coatings are observed to consist of very fine crystallites (Figure 4.5 c), whereas little coating is observed for the sample which received a 10 min acid etch (Figure 4.5 d). It is clear from the observations in Figure 4.5 that the degree of copper enrichment has a significant influence on the form of the zinc phosphate coatings. It has been reported that Cu 2 + ions in solution can be used as an accelerator for this coating process [12, 111], but the copper introduced in the present work is due to interfacial enrichment and it is directly present in the metallic form. It is considered that small amounts of copper may create more cathodic sites, on which the zinc phosphate coating can be initiated. Ideally this can result in relatively fine grained particles distributed uniformly on the surface (Figure 4.5 c). By contrast, if the amount of copper increases to approximate formation of a uniform layer (e.g. after 10 min acid etching), it appears that the coating formation becomes strongly inhibited by the excess copper (Figure 4.5 d). 4.4.2 Corrosion behavior Figure 4.6 shows electrochemical polarization curves measured (scan rates 0.5 mVs"1) for zinc phosphate coatings after immersing the samples in 3.5% NaCl solution. The corrosion rates, 94 E(mV) -250 I (viA/cm2) Figure 4.6 Polarization curves for AA2024-T3 aluminum alloy coated by h-Zn reference solution after different pre-treatments: (A) mechanical polishing only, (B) 2 min acid etching, and (C) 5 min acid etching. 95 expressed in terms of current densities, for samples prepared with different times for the acid etch in the pre-treatment were estimated by extrapolating the cathodic Tafel curves back to the intercept with corrosion potential. These rates decrease as the acid etch time increases to 5 min. The estimated current densities at the corrosion potential are 13.8, 3.6 and 0.76 pA cm 2 for the samples treated respectively by 0, 2 and 5 min acid etch, followed by the zinc phosphating treatment. Higher potentials are observed for coatings formed on the acid etched samples, compared with that which just received a mechanical polish in the pre-treatment. The decrease in corrosion rate, and increase in potential, correlate with the reduced grain size in the zinc phosphate coating (and perhaps also with some integration of copper into the zinc phosphate coating layer). 4.4.3 Comparison of zinc phosphate coatings Figure 4.7 shows SEM images from AA2024-T3 alloy samples after coating by different zinc phosphate formulations following either a mechanical polish or an acid etch pre-treatment. All samples were mechanically polished with up to 1200-grit aluminum oxide sandpaper prior to the coating treatment, and some samples were subsequently acid etched by immersing into a 50/50 vol% H 2 S0 4 -H 2 0 solution for 2 min at 40°C. After each pre-treatment, a coating was applied by one of the reference solutions described by the notation high-ratio normal-zinc (h-Zn), normal-zinc (n-Zn) and low-zinc (/-Zn) in Section 3.2. The treatments discussed here are specified in Table 3.1. For the samples with only mechanical polishing in the pre-treatment, it is observed that the coating produced by the h-Zn process (sample HI) has a better coverage than those prepared by the n-Zn (sample NI) and /-Zn (sample LI) processes. The addition of acid etching in the pre-treatment has a different effect for the different coating formulations. With acid etching in 96 Mechanical polish pre-treatment Acid etch pretreatment Figure 4.7 S E M images (lOOOx) from AA2024-T3 samples after different zinc phosphate coating procedures specified in Table 3.1. 97 the pre-treatment, the samples coated by the h-Zn and n-Zn processes (samples H2 and N2 respectively) have coatings composed of finer grains compared with the corresponding coatings formed after a pre-treatment that just has a mechanical polishing (samples HI and NI). This shows that the coating improved with the acid etching pre-treatment for both the h-Zn and n-Zn processes. However, less coating is observed for the sample coated by the /-Zn process after an acid etch in the pre-treatment (sample L2) compared with that which only had a mechanical polish at the pre-treatment stage (sample LI). From the XPS analyses, copper is detected for samples LI and L2 but not for samples NI, N2, HI and H2. The relative compositions of copper in sample L2 (2.0% Cu(2+), 7.1% Cu(0)) are higher than in sample LI (1.1 % Cu(2+), no Cu(0)). This is consistent with the acid etching increasing the copper enrichment, but the metallic Cu(0) detected in sample L2 probably suggests that redeposition of elemental copper has occurred on the sample surface during the coating process. While it is concluded that the /-Zn process increases copper enrichment, correlation with the SEM images (Figure 4.7) shows that this has a negative effect on formation of the coating. This is consistent with excess copper inhibiting coating formation as proposed by Freeman [12]. Therefore, the study here indicates that the copper enrichment may both enhance a zinc phosphate coating and inhibit it, depending on the degree of enrichment. 98 Chapter 5 Concluding remarks and future directions 5.1 Concluding remarks The studies that were done in this thesis show that different coating recipes have substantial effects on the form of the zinc phosphate coatings deposited on aluminum alloy surfaces. The three main zinc phosphating recipes under investigation are referred to as the high-ratio normal-zinc (h-Zn), normal-zinc (n-Zn) and low-zinc (/-Zn) formulations, as defined in Section 3.2. The h-Zn formulation, with a P-to-Zn atomic ratio of 15, is indicated to provide better coatings compared with the n-Zn and /-Zn formulations at the original pH values (3, 2 and 2 respectively) applied to AA6061-T6 aluminum alloy samples. The h-Zn formulation gives a coating composed of small crystallites, and smaller clusters, compared with the n-Zn and /-Zn formulations for comparable conditions of application. Acidic coating baths are required for zinc phosphate deposition, and appreciable variations are observed to occur with changing pH. It can be concluded that pH 4 is optimal for the three reference coating solutions emphasized in this research. However, the h-Zn process did not provide significant improvement in terms of coating morphology when the coating is applied at pH 4 as compared with samples coated at the original pH values. Samples coated by the n-Zn and /-Zn formulations at pH 4 have morphologies that consist of uniformly distributed small crystallites, and good corrosion resistance is shown. The addition of F" ions to the phosphating bath in general has a large effect on the subsequently formed coatings. Higher F concentrations (600-1000 ppm) are suitable for the /-Zn process but lower concentrations (200-400 ppm) are optimal for the n-Zn process on AA6061-T6 aluminum alloy [97]. 99 The acid etching in the pre-treatment on the AA6061-T6 alloy helps the zinc phosphate coating process by reducing the size of the clusters formed. Acid etching therefore provides a more effective pre-treatment than mechanical polishing alone, with the former being a better generator of nucleation sites for the subsequent coating. No magnesium enrichment is observed after an acid etching pre-treatment of AA6061-T6 aluminum alloy, but copper enrichment is observed for the AA2024-T3 aluminum alloy. This fits the correlation with the Gibbs free energy of formation of oxide per equivalent as discussed by Habazaki et al. [64, 106]. The interfacial enrichment of copper increases with acid etching time applied to AA2024-T3 alloy. For the conditions of application used here, etching times up to 5 min help the zinc phosphate coating, but an etching time of 10 min results in a clear deterioration of the coating. In addition, different phosphate coating recipes will have different effects on the copper enrichment. For the special reference formulations considered here, the /-Zn process increases copper enrichment compared to the n-Zn and h-Zn processes. The presence of elemental copper by the enrichment process has been revealed to have a significant role in affecting the morphology, composition and properties of zinc phosphate coatings. An optimal amount of copper in the alloy subsurface (e.g. 5 min acid etching of AA2024-T3 alloy) leads to fine grains in the phosphate coating and to increased corrosion resistance. A new idea broached in this work is that the interfacial enrichment of particular alloying elements may provide a potential means for optimizing the composition and properties of zinc phosphate coatings [65]. 100 5.2 Future directions Future research is suggested within the following approaches: 5.2.7 Different bath modifications The study of basic parameters for coating baths for other aluminum alloys will also be valuable in order to study the effect of these parameters on each type of alloy. For all alloys, the zinc phosphating recipe could be further improved by adding additives and accelerators like Ca 2 + , Mn 2 + , Fe2 +, Ni 2 + , and Cu 2 + [12-14, 35]. Mechanistic studies of coating formation using additives and the trication approach (i.e. phosphates containing Zn 2 + , Mn 2 + and Ni 2 + ions) would also be useful for designing the most favorable coatings [37, 58]. 5.2.2 Pre-treatment and post-treatment A titanium pre-treatment is used in industry prior to applying zinc phosphate coatings on aluminum alloys and steel [9, 38, 55, 56]. Study is needed to determine the effects of copper enrichment on the titanium pre-treatment. Post-treatments other than the water rinse done in this study would be useful to improve the coating properties. Normally, this is applied by dipping a coated sample into a chromate solution [12]. However, alternative post-treatments need developing in order to eliminate the use of Cr(VI), particularly given the environmental concerns associated with the use of chromate. Some of the methods being studied include treatments with Ce 3 + solutions [112] and organosilane coupling agents [19]. 5.2.3 Evaluation of coatings Other corrosion evaluation tests like the simple immersion and salt spray test should be used as a comparison to the electrochemical polarization measurement in order to evaluate more 101 completely the performances of the coatings formed. Also more rigorous adhesion tests could be used to evaluate better the adhesion properties of the coatings formed. 5.2.4 Study of elemental enrichment on aluminum alloys Quantitative determination of the amounts of enriched copper seems important in order to determine the optimal amount of copper for enhancing the zinc phosphate coating formation. Mechanistic information should also be obtained for the formation of zinc phosphate coatings in the presence of enriched copper layers. Techniques applicable for this include Rutherford backscattering spectroscopy (RBS) and depth profiling in secondary ion mass spectrometry (SIMS) to determine the distribution of copper. Transmission electron microscopy (TEM) can lso be used to study the thickness of the copper layer. In addition, studies of the effects of the presence of other elements at aluminum alloy surfaces on the formation of zinc phosphate coatings would also be useful for establishing the bigger picture of how to modify surfaces in order to optimize properties of zinc phosphate coatings. 102 References 1 S. L. Chawla and R. K. Gupta, "Materials Selection for Corrosion Control" (ASM International, Materials Park, Ohio, 1993). 2 S. Wernick, R. Pinner and P. G. Sheasby, "The Surface Treatment and Finishing of Aluminium and Its Alloys" (Finishing Publications Ltd., Teddington, Middlesex, 1987). 3 J. W. Bibber, Metal Finishing, 91 (1993), 46-47. 4 Z. Yu, H. Ni, G. Zhang, Y. Wang, Q. Dong and G. Zhao, App. Surf. Sci., 62 (1992), 217-221. 5 T. Foster, G. N. Blenkinsop, P. Blattler and M . Szandorowski, Journal of Coatings Technology, 63 (1991), 91-99. 6 B. Cheng, S. Ramamurthy and N. S. Mclntyre, JMEPEG, 6 (1997), 405-412. 7 W. F. Heung, M.Sc thesis, University of British Columbia (1993). 8 H. Ishii, O. Furuyama and S. Tanaka, Metal Finishing, 91 (1993), 7-10. 9 J. F. Ying, M . Y. Zhou, B. J. Flinn, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci., 31 (1996), 565-571. 10 J. F. Ying, B. J. Flinn, M. Y. Zhou, P. C. Wong, K. A. R. Mitchell and T. Foster, Prog. Surf. Sci., 50 (1995), 259-267. 11 S. M . Cohen, Corrosion, 51 (1995), 71-78. 12 D. B. Freeman, "Phosphating and Metal Pre-treatment: A Guide to Modern Processes and Practice" (Industrial Press Inc., New York, 1986). 13 G. Lorin, "Phosphating of Metals: Constitution, Physical Chemistry and Technical Applications of Phosphating Solutions" (Finishing Publications, Hampton Hill, 1974). 14 W. Rausch, "The Phosphating of Metals" (ASM International, Ohio, 1990). 15 W. F. Heung, Y. P. Yang, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci., 29 (1994), 1368-1373. 16 W. F. Heung, Y. P. Yang, M . Y. Zhou, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci., 29 (1994), 3653-3657. 17 W. F. Heung, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci. Lett., 14 (1995), 1461-1463. 103 18 J. Fang, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci., 33 (1998), 1541-1547. 19 L. Shi, M.Sc thesis, University of British Columbia (2000). 20 Aluminium Company of Canada, "Handbook of Aluminum" (Montreal, 1961). 21 M. Pourbaix, "Atlas of Electrochemical Equilibria in Aqueous Solutions" (NACE, Houston, 1974). 22 G. C. Wood, W. H. Sutton, J. A. Richardson, T. N. K. Riley and A. G. Malherbe, "Proceeding Conference on Localised Corrosion", (Houston, 1974). 23 J.R. Galvele, J. Corr. Sci., 28 (1981), 551. 24 G . E . Thompson, P. E. Doherty and G. C. Wood, J. Electrochem. Soc., 129 (1982), 1515-1517. 25 M . G. Fontana, "Corrosion Engineering" (McGraw-Hill Inc., New York, 1986). 26 E. Mattsson, "Basic Corrosion Technology for Scientists and Engineers" (The Institute of Materials, London, 1996). 27 T. W. Coslett, British Patent No. 8667 (1906) 28 T. W. Coslett, British Patent No. 15628 (1908) 29 T. W. Coslett, British Patent No. 22743 (1909) 30 T. W. Coslett, British Patent No. 28131 (1909) 31 R. G. Richards, British Patent No. 37563 (1911) 32 W. H. Allen, U.S Patent No. 1206075 (1914) 33 W. H. Allen, U.S Patent No. 1311726(1914) 34 T. B. Biestek and J. Weber, "Conversion Coatings", ed. C. R. Draper (Portcullis Press Limited, Redhill, 1976). 35 T. S. N. Sankara Narayanan, Corrosion Reviews, 12 (1994), 201-238. 36 J. Fang, M.Sc thesis, University of British Columbia (1997). 37 I. Van Roy, H. Terryn and G. Goeminne, Trans IMF, 76 (1998), 19-23. 38 I. Van Roy, H. Terryn and G. Goeminne, Colloids and Surfaces A, 136 (1998), 89-96. 39 I. Van Roy, H. Terryn and G. Goeminne, ATB Metallurgie, 37 (1997), 161-164. 104 40 C. E. Rossio, US Patent No. 5797987 (1997) 41 C. E. Rossio, US Patent No. 5868874 (1999) 42 A. Turono, K. Toyose and H. Fujimoto, Kobelco Technology Review, 11 (1991), 14-18. 43 H. Gehmecker, Society of Automotive Engineers Inc., (1997), 45-52. 44 L. F. Vega, E. L. Colvin, J. Moran, M . Egbert, T. Nakayama, K. Ikeda, H. Shige and T. Ueda, Society of Automotive Engineers Inc., (1997), 21-36. 45 G. Bierwagen, R. Twite, G. Chen and D. Tallman, Progress in Organic Coatings, 32 (1997), 25-30. 46 M . Handke, A. Stoch, S. Sulima, P. L. Bonora, G. Busca and V. Lorenzelli, Materials Chemistry, 7 (1982), 7-18. 47 H. Gehmecker, ATB Metallurgie, 37 (1997), 130-136. 48 W. A. Roland and P. Droniou, Metal Finishing, 91 (1993), 57-60. 49 J. E. Salmon and H. Terrey, J. Chem. Soc, (1950), 2813-2824. 50 H. Wustefeld, Fortschritte, B III, (1950), 67. 51 W. Machu, "Interface Conversion for Polymer Coatings: Proceedings of the Symposium on Interface Conversion for Polymer Coatings, General Motors Research Laboratories", (Michigan, 1967), p. 128-149. 52 K. Shimizu, G. Brown, K. Kobayashi, P. Skeldon, G. Thompson and G. Wood, Corr. Sci., 40 (1998), 1049-1072. 53 W. Machu, "Die phosphatierung" (Verlag Chemie, Weinnharin, 1950). 54 J. Donofrio, Metal Finishing, 97 (1999), 71-86. 55 P. E. Tegehall, Colloids and Surfaces, 42 (1989), 155-164. 56 P. E. Tegehall, Colloids and Surfaces, 49 (1990), 373-383. 57 A. Neuhaus and M . Gebhardt, Werkstojfund Korrosion, 17 (1966), 567. 58 T. S. N. Sankara Narayanan, Metal Finishing, 91 (1993), 51-52. 59 W. R. Cavanagh and R. R. Cavanagh, Plating and Surface Finishing, 78(1991), 40-43. 60 A. P. Bond, G. F. Boiling and H. A. Domian, J. Electrochem. Soc, 113 (1966), 773-778. 105 61 H. Habazaki, K. Shimizu, M . A. Paez, P. Skeldon, G. E. Thompson, G. C. Wood and X. Xhou, Surf. Interface Anal., 23 (1995), 892-898. 62 X. Zhou, G. E. Thompson, H. Habazaki, K. Shimizu, P. Skeldon and G. C. Wood, Thin Solid Films, 293 (1997), 327-332. 63 X. L. Wu and K. Hebert, J. Electrochem. Soc., 143 (1996), 83-91. 64 H. Habazaki, K. Shimizu, P. Skeldon, G. E. Thompson, G. C. Wood and X. Zhou, Corr. Sci., 39 (1997), 731-737. 65 X. Sun, R. Li , W. H. Kok, K. C. Wong, K. A. R. Mitchell and T. Foster, ATB Metallurgie, (2000), in press. 66 T. W. Cape and H. R. Charles, US Patent No. 5238506 (1993) 67 J. Riesop, K. Hosemann and K. Gottwald, US patent No. 5312492 (1994) 68 M . L. Sienkowski and G. J. Cormier, US Patent No. 6019858 (2000) 69 B. D. Ratner and D. G. Castner, "Electron Spectroscopy for Chemical Analysis", ed. J. C. Vickerman, (John Wiley & Sons Ltd., West Sussex, 1997) 70 J. D. Andrade, "Surface and Interfacial Aspects of Biomedical Polymers", (Plenum Press, New York, 1985) 71 C D . Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder and G. E. Muilenberg, "Handbook of X-Ray Photoelectron Spectroscopy" (Perkin-Elmer Corporation, Physical Electronics Division, 1979). 72 H. Hertz, Ann. Physik, 31 (1887), 983. 73 J. J. Thompson, Phil. Mag., 48 (1899), 547. 74 A. Einstein, Ann. Physik, 17 (1905), 132. 75 H. Robinson and W. F. Rawlinson, Phil. Mag., 28 (1941), 277. 76 K. Sieghbahn, C. N. Nordling, A. Fahlman, R. Nordberg, K. Hamrin, J. Hedman, G. Johansson, T. Bergmark, S. E. Karlsson, I. Lingren and B. Lindberg, "ESCA: Atomic Molecular and Solid State Studies by Means of Electron Spectroscopy" (Uppsala, 1967). 77 J. M . Walls and A. B. Christie, "Surface Analysis by Auger and X-ray Photoelectron Spectroscopy", ed. D. M. Brewis, (Applied Science Publishers, London, 1982) 78 E. Bauer, Vacuum 22, (1972), 539. 79 D. Briggs and M . P. Seah, "Practical Surface Analysis" (John Wiley & Sons Ltd., West Sussex, 1990). 106 80 M . P. Seah and W. A. Dench, Surf. Interface Anal., 1 (1979), 2. 81 G. A. Somorjai, "Chemistry in Two Dimensions: Surfaces" (Cornell University, Ithaca, New York, 1981). 82 "MAX200 User Manual" (Leybold, Koln, Germany). 83 D . A . Shirley, Phys. Rev., B5 (1972), 4709. 84 J. M . Cowley, "Electron Microscopy", ed. A. T. Hubbard, (CRC Press Inc., Boca Raton, 1995) 85 O. C. Wells, "Scanning Electron Microscopy" (McGraw-Hill Book Company, New York, 1974). 86 B. L. Gabriel, "SEM: A User's Manual for Materials Science" (American Society for Metals, Ohio, 1985). 87 J. Ruste, "X-Ray spectrometry", ed. F.Maurice, L. Meny and R. Tixier, (Les Editions De Physique, St-Martin-d'Heres, 1978) 88 J. C. Russ, "Fundamentals of Energy Dispersive X-ray Analysis" (Butterworths & Co Ltd., Raleigh, 1984). 89 O. W. Siebert, "Laboratory Electrochemical Test Methods", ed. G. S. Haynes and R. Baboian, (American Society for Testing Materials (ASTM), Philadelphia, 1985) 90 J. A. V. Butler, Trans. Faraday Soc. ,19 (1924), 734. 91 J. Tafel, Z. Physik. Chem., 50 (1905), 641. 92 A. C. Fisher, "Electrode Dynamics" (Oxford University Press, New York, 1996). 93 D. A. Jones, "Principles and Prevention of Corrosion" (Prentice Hall, New Jersey, 1996). 94 E. Heitz, R. Henkhaus and A. Rahmel, "Corrosion Science; an Experimental Approach", ed. D. R. Holmes (Ellis Horwood Ltd., West Sussex, 1992). 95 P. W. Atkins, "Physical Chemistry" (ELBS, Oxford, 1994). 96 T. S. N. Sankara Narayanan, Metal Finishing, 96 (1996), 86-90. 97 W. H. Kok, X. Sun, L. Shi, K. C. Wong, K. A. R. Mitchell and T. Foster, J. Mater. Sci., (submitted for publication). 98 O. Leblanc, J. Oil & Colour Chemists' Assoc., 73 (1990), 231-251. 99 M . A. Alodan and W. H. Smyrl, J. Electrochem. Soc, 144 (1997), L282-L284. 107 100 M . A. Alodan and W. H. Smyrl, J. Electrochem. Soc, 145 (1998), 1571-1577. 101 W. L. Masterton, E. J. Slowinski and C. L. Stanitski, "Chemical Principles with Qualitative Analysis" (Saunders College Publishing, Philadelphia, 1986). 102 M. A. Paez, T. M . Foong, C. T. Ni, G. E. Thompson, K. Shimizu, H. Habazaki, P. Skeldon and G. C. Wood, Corr. Sci, 38 (1996), 59-72. 103 H. Habazaki, M . A. Paez, K. Shimizu, P. Skeldon, G. E. Thompson, G. C. Wood and X. Zhou, Corr. Sci., 38 (1996), 1033-1042. 104 H. Habazaki, X. Zhou, K. Shimizu, P. Skeldon, G. E. Thompson and G. C. Wood, Electrochimica Acta, 42 (1997), 2627-2635. 105 X. Zhou, G. E. Thompson, P. Skeldon, G. C. Wood, K. Shimizu and H. Habazaki, Corr. Sci., 41 (1999), 1089-1094. 106 H. Habazaki, K. Shimizu, P. Skeldon, G. E. Thompson, G. C. Wood and X. Zhou, Trans IMF, 75 (1997), 18-23. 107 M . Amstutz and M . Textor, Surf. Interface Anal. ,19(1992), 595-600. 108 G. D. Davis, B. A. Shaw, B. J. Rees and M . Ferry, J. Electrochem. Soc, 140 (1993), 951-959. 109 H. Habazaki, K. Shimizu, P. Skeldon, G. E. Thompson and G. C. Wood, J. Electrochem. Soc, 143 (1996), 2465-2470. 110 H. Habazaki, K. Shimizu, P. Skeldon, G. E. Thompson and G. C. Wood, Phil. Mag. B, 73 (1996), 445-460. 111 W. Wimmer, J. Gottschlich, B. Motorenwerke, A. G. Dingolfing, E. Busch and H. Gehmecker, Metal Finishing, 96 (1998), 18-19. 112 F. Mansfeld, C. Chen, C. B. Breslin and D. Dull, J. Electrochem. Soc, 145 (1998), 2792-2798. 108 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items