Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Surface science studies of organosilane and zinc phosphate coatings on 2024-T3 aluminum alloy Susac, Darija 2003

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

Item Metadata

Download

Media
831-ubc_2004-902722.pdf [ 22.08MB ]
Metadata
JSON: 831-1.0061188.json
JSON-LD: 831-1.0061188-ld.json
RDF/XML (Pretty): 831-1.0061188-rdf.xml
RDF/JSON: 831-1.0061188-rdf.json
Turtle: 831-1.0061188-turtle.txt
N-Triples: 831-1.0061188-rdf-ntriples.txt
Original Record: 831-1.0061188-source.json
Full Text
831-1.0061188-fulltext.txt
Citation
831-1.0061188.ris

Full Text

Surface Science Studies of Organosilane and Zinc Phosphate Coatings on 2024-T3 Aluminum Alloy by D A R I J A S U S A C B . S c , U n i v e r s i t y o f Belgrade , Facu l ty o f P h y s i c a l Chemis t ry , 1992 M . S c , Un ive r s i t y o f B r i t i s h C o l u m b i a , 1999 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y I N T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department o f Chemis t ry) W e accept this thesis as confo rming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A O C T O B E R 2003 © Dar i j a Susac, 2003 Abstract The effects o f 2024-T3 a l l oy microstructure o n the format ion o f organosilane and z inc phosphate ( Z P O ) coatings have been studied by X - r a y photoelectron spectroscopy ( X P S ) , scanning electron mic roscopy ( S E M ) , energy dispersive X - r a y ( E D X ) spectroscopy and scanning A u g e r mic roscopy ( S A M ) . Some studies o f cor ros ion protect ion were also made us ing e lect rochemical polar iza t ion measurements, especial ly for compar ing different post-treatment rinses appl ied to the Z P O coatings. T h i s thesis describes n e w informat ion for the different chemistries occur r ing at m i c r o -regions o f 2024-A1 a l loy , especia l ly those associated w i t h intermetal l ic compounds (second-phase particles) o f composi t ions A l - C u - M g and A l - C u - F e - M n . These particles have been shown to strongly affect in i t ia t ion o f the Z P O coatings. Coa t i ng crystall i tes start to fo rm at A l - C u - M g particles, w h i l e the coverage o n the A l - C u - F e - M n particles is more amorphous. D u r i n g the etching stage o f the coat ing process, the natures o f the A l - C u - M g particles appear to change f rom anodic to cathodic character, and this helps dr ive the precipi ta t ion o f Z P O crystals. A d d i t i o n a l l y , a C u enrichment is found at the interface between the A l - C u - M g particles and the Z P O coat ing. Af te r comple t ion , the coat ing thickness is least o n the matr ix regions closest to the A l - C u - M g particles, where the compet i t ion between substrate etching and Z P O deposi t ion is strongest. A lower ing i n the p H o f the coat ing solu t ion results i n thinner and more amorphous Z P O coatings, and a greater C u enrichment. The adsorption o f b is- l ,2-( t r ie thoxysi ly l )e thane ( B T S E ) si lane is also s trongly affected by the dis t r ibut ion o f second-phase particles and by the amount o f ox ide at the 2024-A1 a l loy surface. F o r a mechanica l ly -po l i shed sample, B T S E adsorbs o n the matr ix regions away f rom the particles, and o n the particles themselves, but less adsorbs near the par t ic le-matr ix interface ii areas. F o r an a i r -ox id ized sample, the amount o f adsorbed B T S E increases markedly , but w i t h a h igh degree o f non-uniformity . It is postulated that the B T S E deposi t ion depends o n the precise natures o f ox ide layer formed at the var ious micro-areas, and o n the compet i t ion between silane deposi t ion and oxide etching. In contrast, y-aminopropyl t r ie thoxysi lane ( y - A P S ) forms a film o f re la t ively even thickness over a large sample area, apparently due to the format ion o f hydrogen bonds through the amino groups. iii Table of Contents Abstract i i Table o f Contents i v L i s t o f A c r o n y m s v i i i L i s t o f Tables i x L i s t o f F igures x i i Acknowledgemen t s x x Chapter 1: A l u m i n u m , its A l l o y s and Protect ive Coat ings 1 1.1 A l u m i n u m and its A l l o y s 1 1.1.1 Introduction 1 1.1.2 A l l o y Micros t ruc ture 4 1.2 C o r r o s i o n Protect ion Layers 8 1.2.1 N a t i v e O x i d e 8 1.2.2 Spec ia l Approaches 14 1.3 Organosi lanes 15 1.3.1 Introduction 15 1.3.2 Si lane B o n d i n g to M e t a l s 16 1.3.3 C o r r o s i o n Protect ion by Silanes 18 1.4 Phosphate C o n v e r s i o n Coat ings 20 1.4.1 Introduction 20 1.4.2 M e c h a n i s m o f Coa t i ng 25 1.5 Object ives o f Research 27 iv Chapter 2: Exper imenta l Me thods 30 2.1 Introduction X P S and A E S 30 2.1.1 B a s i c Pr inc ip les 3 0 2.1.2 Spectral Features ' 34 2.1.3 Quanti tat ive A n a l y s i s 37 2.1.4 U s e o f H i g h Ene rgy E lec t ron B e a m 3 8 2.2 X P S Instrumentation 40 2.2.1 M A X 2 0 0 41 2.2.2 X - r a y Source 43 2.2.3 Ene rgy A n a l y z e r 44 2.2.4 Da ta Process ing 48 2.3 S A M and S E M Instrumentation ^ 50 2.3.1 The rmo V G M i c r o l a b 350 50 2.3.2 Other Instruments for S E M 56 2.4 Energy Dispers ive X - r a y Spectroscopy 57 2.4.1 Introduction 57 2.4.2 E D X Instrumentation 59 2.5 E lec t rochemica l Po la r i za t ion C u r v e Measurement 61 2.5.1 B a c k g r o u n d 61 2.5.2 Instrumentation 63 Chapter 3: A d s o r p t i o n o f B T S E and y - A P S Organosi lanes o n Different Micros t ruc tura l Reg ions o f 2024-T3 A l u m i n u m A l l o y 68 3.1 Introduction 68 3.2 Exper imenta l 69 V 3.3 Resul ts and D i s c u s s i o n 70 3.3.1 B T S E A d s o r p t i o n o n M e c h a n i c a l l y - p o l i s h e d 2 0 2 4 - A l 70 3.3.2 B T S E A d s o r p t i o n o n O x i d i z e d 2024-A1 79 3.3.3 y - A P S A d s o r p t i o n 89 3.4 C o n c l u d i n g Remarks 90 Chapter 4: C o m p a r i s o n o f C h r o m i c A c i d and a B T S E F i n a l R inse A p p l i e d to Phosphated 2024-T3 A l u m i n u m A l l o y 92 4.1 Introduction 92 4.2 Exper imenta l 93 4.3 Resul ts and D i s c u s s i o n 94 4.3.1 C h r o m i c A c i d Post-rinse 94 4.3.2 B T S E Post-treatment 102 4.3.3 Protect ion at M i c r o - r e g i o n s 106 4.4 C o n c l u d i n g Remarks 111 Chapter 5: Micros t ruc tura l Effects o n the Ini t ia t ion o f Z i n c Phosphate Coat ings o n 2024-T3 A l u m i n u m A l l o y 113 5.1 Introduction 113 5.2 Exper imenta l 114 5.3 Results 115 5.3.1 P re l iminary Observat ions 115 5.3.2 Ini t ia l Coat ings o n M i r r o r - p o l i s h e d Samples 118 5.3.3 F u l l Coverage o n a M i r r o r - p o l i s h e d Sample 125 5.3.4 Sputtering and Re -ox ida t i on 12 8 5.3.5 Coa t i ng at p H 3 133 5.3.6 Z P O Coa t i ng Af te r N a O H / N a F E t c h i n g 136 vi 5.4 5.5 Chapter 6: 6.1 6.2 References D i s c u s s i o n C o n c l u d i n g Remarks C o n c l u d i n g Remarks N e w Resul ts Future Di rec t ions List of Acronyms Y - A P S y-aminopropyl t r ie thoxysi lane A E S A u g e r electron spectroscopy B S E backscattered electrons B T S E bis-1,2-(tr iethoxysilyl)ethane C H A concentric hemispher ica l analyzer C R R constant retardation ratio C F E c o l d field emiss ion E B S D electron backscattering diffract ion E D X energy dispersive X - r a y spectroscopy E I S e lect rochemical impedance spectroscopy F W H M fu l l -wid th-a t -ha l f -maximum I E P isoelectr ic point M C P mul t ichannel plate M P mir ror -pol i shed M P E mir ror -pol i shed and etched M P E Z P O mir ror -pol i shed, etched and z inc phosphate coated M P Z P O mir ror -pol i shed and z inc phosphate coated S A M scanning A u g e r mic roscopy S C E saturated ca lomel electrode S E M scanning electron mic roscopy S F E Schot tky field emiss ion S I M S secondary i o n mass spectrometry T O F t ime o f fl ight U H V ul t rahigh v a c u u m X P S X - r a y photoelectron spectroscopy Z P O z inc phosphate viii List of Tables Table 1.1 M a j o r classes o f wrought A l a l loys . Table 1.2 Some examples o f silanes. 15 Table 1.3 Phosphate phases formed o n var ious metals f rom coat ing baths conta ining different metal ions [79]. 21 Table 3.1 E lementa l ratios measured by E D X from selected matr ix regions (see F i g . 3.1) o f the po l i shed 2 0 2 4 - A l a l l oy sample before and after coat ing w i t h B T S E . 75 Table 3.2 Compos i t i ons measured by E D X from selected second-phase particles (see F i g . 3.3) i n the 2 0 2 4 - A l a l loy sample w h i c h had been coated by B T S E immedia te ly after po l i sh ing . 76 Table 3.3 E lementa l ratios measured by E D X from selected mat r ix regions (see F i g . 3.5(b)) o f the o x i d i z e d 2 0 2 4 - A l a l l oy sample before and after coat ing w i t h B T S E . 80 Table 3.4 E lementa l ratios measured by E D X from selected second-phase particles (see F i g . 3.6) o f the o x i d i z e d 2024-A1 a l loy sample before arid after coat ing w i t h B T S E . 82 Table 3.5 Compos i t i ons measured by E D X from selected second-phase particles (see F i g . 3.9) and from the a l l oy mat r ix i n a 2 0 2 4 - A l sample w h i c h had been coated b y y - A P S immedia te ly after po l i sh ing . 89 Table 4 .1 . Re la t ive a tomic composi t ions (mole%) from X P S for phosphated 2 0 2 4 - A l samples before and after a post-treatment rinse i n ch romic ac id (0.5 g / L C1O3) at different temperatures. 99 IX Table 4.2. Re la t ive a tomic composi t ions (mole%) from X P S for phosphated 2 0 2 4 - A l samples post-treated i n B T S E solutions o f different concentrations before and after ul trasonic test. 105 Table 4.3. Percentage composi t ions (wt%) f rom E D X measured from selected micro-regions o f mi r ror -po l i shed and phosphated 2024-A1 samples post-treated i n either ch romic ac id (0.5 g / L C r 0 3 at 7 0 ° C ) (particle P I and matr ix nearby) or 4 % B T S E (particle P 2 and matr ix nearby) and measured before the immers ion test. 108 Table 4.4. A t o m i c composi t ions (mole%) from X P S for mi r ror -po l i shed and phosphated 2024-A1 samples post-treated i n either ch romic ac id (0.5 g / L C r 0 3 at 7 0 ° C ) or 4 % B T S E and measured before and after the immers ion test. 109 Table 5.1 E lementa l composi t ions (a tomic%) measured b y S A M point analysis from A l - C u - M g (P1) and A l - C u - F e - M n (P2) particles after r emova l o f Z P O coat ing by sputtering. 122 Table 5.2 E lementa l composi t ions (a tomic%) measured by S A M point analysis from the coated surface s h o w n i n F i g . 5.7(c): A l - C u - M g part icle P3 at locations 1,2; nearby matr ix at locat ions 3-5. 127 Table 5.3 E lementa l composi t ions (a tomic%) measured by S A M point analysis f rom different regions o f A l - C u - M g particle P 4 showing Z P O coatings w i t h crystal l ine ( C ) and amorphous ( A ) characters. 128 Table 5.4 E lementa l composi t ions (a tomic%) measured by S A M point analysis from different regions o f the coated surface shown i n F i g . 5.11 obtained after us ing the coat ing solu t ion at p H 3.0. 133 Table 5.5 E lementa l weight ratios determined by E D X from different regions o f the coated surface s h o w n i n F i g . 5.14. 139 x Table 5.6 E lementa l composi t ions (a tomic%) measured by X P S for mir ror -po l i shed a l loy ( M P ) , for mi r ror -po l i shed and coated a l loy ( M P Z P O ) , for mir ror -po l i shed and etched a l l oy ( M P E ) , and for mir ror -pol i shed, etched and coated a l l oy ( M P E Z P O ) samples. 141 xi List of Figures Figure 1.1 P r i n c i p a l a l u m i n u m a l loys [4]. 3 F igure 1.2 Micros t ruc ture o f 2024-T3 A l a l loy : (a) second-phase particles and (b) grain structure. 6 F igure 1.3 Phase diagrams for: (a) A l - C u a l loys and (b) A l - C u - M g a l loys [9]. 7 F igure 1.4 Schematic ind ica t ion o f ox ide o n A l surfaces: (a) pores i n the ox ide and (b) cor ros ion in i t ia t ion at l oca l sites i n a l loy [1]. 9 F igure 1.5 Pourba ix D i a g r a m for the AI -H2O system at 2 5 ° C , 1 a tm pressure [13]. 11 F igure 1.6 (a) S E M micrograph o f a pi t i n 2 0 2 4 - A l a l l oy and (b) schematic representation o f its cross section. 13 F igure 1.7 A d s o r p t i o n o f B T S E o n a l u m i n u m substrate. 19 Figure 1.8 S E M micrographs for phosphated surfaces o f A l and its a l loys produced under different condi t ions: (a) 99 .99% pur i ty A l w i t h laboratory based coat ing solu t ion conta ining 1 p p m o f C u ; (b) same as (a) but w i t h 10 p p m o f added C u ; (c) 6061 A l a l l oy treated w i t h commerc i a l coat ing solut ion, and (d) 2024 A l a l l oy treated under same condi t ions as (c). 23 F igure 1.9 S E M micrographs f rom 2 0 2 4 - A l a l l oy samples after treating w i t h laboratory based Z P O coat ing solutions: (a) no st irr ing; (b) same condi t ions as (a) but w i t h st irr ing and (c) as (b) but at h i g h magnif ica t ion . 24 Figure 1.10 Schematic representation o f the mechan i sm o f z i nc phosphate coat ing format ion o n A l a l loy . 26 xii Figure 2.1 Schematic d iagram for (a) photoelectron emiss ion ; (b) X - r a y emiss ion and (c) A u g e r electron emiss ion . 31 Figure 2.2 Inelastic mean free path o f electrons as a funct ion o f k inet ic energy inside the so l id [126]. 33 F igure 2.3 X P S spectra from z inc phosphated 2024-A1 a l loy : (a) survey spectrum, and higher resolut ion spectra for (b) Z n 2p and (c) A l 2p. 35 F igure 2.4 A u g e r spectra measured from 2024-A1 a l loy showing regions o f (a) meta l l ic A l and (b) o x i d i z e d A l . 36 F igure 2.5 Energy spectrum o f electrons emitted from a so l id upon interaction w i t h incident electron beam [132]. 39 F igure 2.6 M A X 2 0 0 X - r a y photoelectron spectrometer. 41 Figure 2.7 Schematic representation o f the p u m p i n g system for the M A X 2 0 0 . 42 F igure 2.8 Schematic d iagram for the concentr ic hemispher ica l analyzer ( C H A ) and lens system i n the M A X 2 0 0 . 45 Figure 2.9 Schematic d iagram o f the relevant energy levels for the b ind ing energy measured i n the spectrometer from a conduct ing sample (redrawn f rom Ref . [115]). 47 Figure 2.10 Data processing for a C Is spectrum: (a) Sh i r l ey non-l inear background subtraction and (b) curve fitted spectrum. 49 Figure 2.11 Thermo V G Scient i f ic M i c r o l a b 350. 51 Figure 2.12 V a c u u m system for M i c r o l a b 350. 52 xiii Figure 2.13 F i e l d emiss ion electron source i n M i c r o l a b 350 and its power supply [134]. 54 Figure 2.14 I l lumina t ing system i n M i c r o l a b 350. 55 F igure 2.15 E D X spectrum measured f rom an A I - C u - M g part icle. 58 Figure 2.16 S i l i c o n diode X - r a y detector for E D X (redrawn f rom Ref . [137]). 60 F igure 2.17 Schematic d iagram o f po lar iza t ion curve for O + e~ <-*• R showing anodic and cathodic Tafe l regions. 64 F igure 2.18 Po la r i za t ion curves for z i nc i n ac id so lu t ion [ 16] showing two separate electron transfer processes. 65 F igure 2.19 Schematic d iagram o f po la r iza t ion circui t : A identifies w o r k i n g electrode; B reference electrode; G counter electrode; H L H a b e r - L u g g i n capi l la ry ; M V ampli f ier ; U voltmeter and I ammeter. 66 F igure 3.1 S E M micrograph measured f rom a 2 0 2 4 - A l a l l oy sample that was coated w i t h B T S E immedia te ly after po l i sh ing (magnif ica t ion 300x) . 71 F igure 3.2 E D X spectra measured f rom: (a) dark areas and (b) l ight areas s h o w n i n F i g . 3.1. 72 F igure 3.3 S E M micrograph showing microstructure for a 2 0 2 4 - A l a l l oy sample that had been coated w i t h B T S E immedia te ly after po l i sh ing (magnif icat ion 800x) . Second-phase particles o f composi t ions A l - C u - M g ( A 1 - A 5 ) and A l - F e - C u - M n - S i ( B ) are apparent, a long w i t h zones o f influence around particles (l ighter appearance) and a l l oy matr ix regions, w h i c h are further f rom the particles (darker appearance). 73 Figure 3.4 E D X spectra measured f rom: (a) part icle A l and (b) part icle B ( locat ion 1) shown i n F i g 3.3. 74 xiv Figure 3.5 S E M micrographs measured for (a) o x i d i z e d 2024-A1 a l loy sample before and (b) after B T S E coat ing (magnif ica t ion 200x) . Cor responding areas are shown i n each case, a l though the image for (b) is rotated by 180° w i t h respect to that i n (a). 81 Figure 3.6 S E M micrographs showing different A l - F e - C u - M n - S i particles (P1-P3) o n an o x i d i z e d 2 0 2 4 - A l a l l oy sample. The images o n the left-hand side apply to before exposure to B T S E , w h i l e the corresponding images o n the r ight-hand side app ly to after the B T S E exposure. 83 Figure 3.7 X P S spectra measured f rom the o x i d i z e d 2 0 2 4 - A l sample after coat ing w i t h B T S E : (a) survey scan and (b) A l 2p h igh resolut ion spectrum. 85 Figure 3.8 G r a i n structure for the mechanica l ly -po l i shed 2024-A1 a l loy sample: (a) S E M image after e tching i n HNO3 and H 3 P 0 4 ; (b) E B S D image w i t h grain index ing , co lo r coded according to the plane (hkl) para l le l to the r o l l i n g plane and d i rec t ion [uvw] paral le l to the r o l l i n g di rect ion: p ink (110) [100], red-orange (112) [111], y e l l o w (100) [100], blue (110) [211], v io le t (123) [634]; the green areas cannot be indexed and are referred to i n the text as deformed. 87 F igure 3.9 S E M micrographs measured f rom a 2 0 2 4 - A l a l l oy sample w h i c h had been coated w i t h y - A P S immedia te ly after po l i sh ing : (a) larger-area v i e w (magnif ica t ion 400x) and (b) higher magni f ica t ion ( l , 0 0 0 x ) v i e w o f another reg ion w h i c h shows second-phase particles o f A l - F e - C u - M n - S i (S4 arid S5) and A l - C u - M g (T1 and T2) types. 88 Figure 4.1 Po la r iza t ion curves measured f rom phosphated 2 0 2 4 - A l samples (a): (1) no post-treatment, (2) post-treated i n 0.1 g / L C r 0 3 , (3) 0.5 g / L C r 0 3 and (4) 1.0 g / L C r 0 3 a l l at 2 5 ° C ; (b) post-treated i n 0.5 g / L C r 0 3 at 2 5 ° C and at 7 0 ° C . 95 X V Figure 4.2 S E M micrographs o f phosphated 2024-A1 samples: (a) no post-treatment rinse (x5,000); (b) no rinse (x50,000); (c) rinse i n 0.5 g / L C r 0 3 at 2 5 ° C (x5,000); (d) rinse i n 0.5 g / L C r 0 3 at 2 5 ° C (x50,000) ; (e) rinse i n 0.5 g / L C r 0 3 at 7 0 ° C (x5,000) , and (f) rinse i n 0.5 g / L C r 0 3 at 7 0 ° C (x50,000). 97 F igure 4.3 (a) X P S survey spectra measured f rom phosphated 2024-A1 samples (1) before and (2) after post-treatment i n 0.5 g / L C r 0 3 so lu t ion at 7 0 ° C ; (b) corresponding spectra for a l ower b ind ing energy range. ' 98 Figure 4.4 C r 2p3/2 spectra measured from phosphated 2024-A1 samples post-treated i n : (a) 0.5 g / L C r 0 3 at 2 5 ° C ; (b) same so lu t ion at 7 0 ° C ; and (c) as (b) but after i m m e r s i o n test. 101 Figure 4.5 Po la r i za t ion curves measured from phosphated 2024-A1 samples: (a) no post-treatment; (b) after post-treatment i n B T S E concentrations 1 % , (c) 4 % and (d) 10%. 103 Figure 4.6 X P S from phosphated 2 0 2 4 - A l sample after post-treatment i n 1% B T S E so lu t ion at 2 5 ° C : (a) survey spectrum; and (b) C Is spectrum. 104 Figure 4.7 S E M micrographs o f mir ror -po l i shed and phosphated 2024-A1 samples: (a) post-treated i n 0.5 g / L C r 0 3 solut ion at 7 0 ° C (x3,000); (b) post-treated i n 4 % B T S E at 2 5 ° C (x5,000); (c) as (a) but after immers ion i n 3 .5% N a C I solu t ion for 12 h ; (d) as (b) but after 12 h immers ion test. 107 F igure 4.8 S E M micrographs o f mir ror -po l i shed and phosphated 2024-A1 samples: (a) post-treated i n 0.5 g / L C r 0 3 so lu t ion at 7 0 ° C and after i m m e r s i o n i n 3 .5% N a C I solu t ion for 21 h (x3,000); (b) post-treated i n 4 % B T S E at 2 5 ° C and after 21 h i m m e r s i o n test (x3,000). 110 xvi Figure 5.1 S E M micrographs o f mechanica l ly -po l i shed (1200 grit) 2 0 2 4 - A l a l loy after treating w i t h the Z P O coat ing so lu t ion at 5 0 ° C for different immers ion t imes: (a) 3 s, (b) 15 s, (c) 30 s, (d) 1 m i n (e) 2 m i n and ( f ) 3 m i n . 116 F igure 5.2 X P S survey spectrum o f mechanica l ly -po l i shed 2 0 2 4 - A l a l loy after treating w i t h Z P O coat ing so lu t ion at 5 0 ° C for 3 m i n . 117 F igure 5.3 A l - C u - M g particle ( P I ) f rom mir ror -pol i shed 2024-A1 sample after exposure to Z P O coat ing solu t ion at 2 5 ° C for 1 m i n : (a) S E M micrograph; and (b) A u g e r spectra measured f rom different locations: (1) par t ic le ' s center, (2) near edge o f part icle, and (3) matr ix near to P I . 119 F igure 5.4 A l - C u - F e - M n particle (P2) f rom mir ror -po l i shed 2024-A1 sample after exposure to the Z P O solut ion at 2 5 ° C for 1 m i n : (a) S E M micrograph , and (b) A u g e r spectra measured f rom different locat ions: (1) and (2) par t ic le ' s surface; (3) and (4) mat r ix near to P 2 . 120 Figure 5.5 A u g e r spectra measured f rom different regions o n mir ror -po l i shed 2 0 2 4 - A l sample after exposure to Z P O coat ing solu t ion at 25 ° C for 1 m i n : (a) (1) A l - C u - M g particle ( P I ) after coat ing, (2) after subsequent 6 m i n A r + sputtering, (3) A l - C u - F e - M n particle (P2) after coat ing, (4) after subsequent 3 m i n A r + sputtering; (b) A l K L L spectra measured after A r + sputtering for t imes to remove coat ing: (1) part icle P I , (2) particle P 2 , and (3) mat r ix . 123 F igure 5.6 Images f rom A l - C u - M g particle o n mir ror -po l i shed 2024-A1 sample after exposure to Z P O coat ing solu t ion at 2 5 ° C for 1 m i n : (a) S E M micrograph; and S A M elemental maps for: (b) Z n and O ; (c) A l and P ; and (d) C u map after r emova l o f coat ing. 124 xvii Figure 5.7 M i r r o r - p o l i s h e d 2 0 2 4 - A l sample after exposure to Z P O coat ing solu t ion at 5 0 ° C for 3 m i n : (a) w i d e - v i e w S E M micrograph ; (b) coated A l - C u - M g particle (P3) ; (c) A u g e r spectra measured f rom different locat ions: (1-2) par t ic le ' s surface, and (3-5) posi t ions o n matr ix close to P 3 . 126 Figure 5.8 S A M depth profi les measured from Z P O - c o a t e d A l - C u - M g part icle o n mir ror -pol i shed 2 0 2 4 - A l sample after so lu t ion appl ied at 5 0 ° C for 3 m i n . 129 Figure 5.9 A l - C u - M g particle (P4) o n mir ror -po l i shed 2 0 2 4 - A l sample after exposure to Z P O coat ing solu t ion at 5 0 ° C for 3 m i n : (a) S E M micrograph; (b) A u g e r spectra taken from (1) crystal l ine region, and (2) flat amorphous region. 130 Figure 5.10 A u g e r spectra measured f rom: (a) A l - C u - M g part icle before and after ox ida t ion (spectra 1 and 2 respect ively); mat r ix before and after ox ida t ion (spectra 3 and 4) , and (b) A l K L L spectra for matr ix before and after ox ida t ion (spectra 1 and 2) and A l - C u - M g part icle before and after ox ida t ion (spectra 3 and 4). 132 F igure 5.11 S E M micrographs o f Z P O coat ing obtained o n a l l oy sample us ing coat ing bath w i t h p H 3.0: (a) selected area at l ower magnif ica t ion; (b) region w i t h A l - C u - M g part icle; and (c) region w i t h A l - C u - F e - M n part icle. 134 F igure 5.12 A u g e r spectra measured f rom different locations for 2 0 2 4 - A l sample coated at p H 3: (a) w i d e scan for matr ix (spectrum 1), A l - C u - M g particle (spectrum 2), A l - C u - F e - M n particle (spectrum 3); and (b) same as (a) but shorter energy range. 135 F igure 5.13 S E M micrographs o f (a) mir ror -po l i shed and N a O H / N a F etched 2 0 2 4 - A l a l loy sample and (b) after Z P O coat ing. 137 xviii Figure 5.14 (a) C r a c k e d Z P O coat ing o n grains w i t h different orientation, and selected locat ions for E D X measurements; (b) gra in boundary before coat ing; and (c) same reg ion as i n (b) but after coat ing. 138 Figure 5.15 A l 2p and M g 2p spectra measured f rom mir ror -po l i shed and etched sample, M P E (a) and (b) respectively; and mir ror -pol i shed , etched and coated sample, M P E Z P O (c) and (d) respectively. 140 Figure 5.16 Schematic ind ica t ion o f mechanism for the in i t ia t ion o f Z P O coat ing o n A l - C u - M g second-phase part icle: (1) before coat ing the mat r ix and part icle are covered w i t h different oxides . (2) Af te r immers ion i n solut ion, e tching gives C u enrichment especia l ly d r iven by galvanic coup l ing between particle and matr ix . (3) E x p o s e d C u - r i c h surface regions act as cathodic sites for coat ing precipi tat ion. (4) O n comple t ion the particle surface is covered w i t h Z P O w h i l e a C u - r i c h interface exists between coat ing and particle. 145 xix Acknowledgements I w o u l d l i ke to thank m y supervisor, Professor K . A . R . M i t c h e l l , for p r o v i d i n g me the opportunity to enjoy research i n the area o f surface science. I a m grateful for his teaching and guidance dur ing m y studies, cont inuing support and encouragement, invaluable discussions, and for his careful reading o f this thesis. Spec ia l thanks are directed at a former member o f our research group, D r . X . S u n for in t roducing me to the area o f microstructural properties and characterizat ion o f a l u m i n u m a l loys , for his assistance i n experimental work , and many useful discussions and suggestions. I a m especia l ly grateful to D r . P . C . W o n g for he lp ing m e w i t h S A M measurements, and to D r . K . C . W o n g for he lp ing me w i t h X P S measurements. T h e y taught me the experimental aspects o f these techniques, helped w i t h data interpretations and part icipated i n numerous useful discussions. I thank Prof . D . Tromans for assistance w i t h the e lec t rochemical polar iza t ion measurements, M r . J . M a c k e n z i e for help i n S E M operation, M s . T . B i g g s for E B S D analysis and M s . M . M a g e r for assistance i n E D X measurements and interpretations. I have greatly appreciated interactions w i t h other colleagues (present and former) i n our research group inc lud ing M s . R . L i , Ms ." A . A k h t a r , M r . W - H . K o k , M r . M . Sa idy , M r . J . K i m and M r . M . Teo . I w o u l d also l ike to acknowledge N S E R C and D N D for p rov id ing the f inancia l support for this research. Las t but not least, I w o u l d l ike to thank m y fami ly for their support and encouragement. XX Chapter 1 Aluminum, its Alloys and Protective Coatings 1.1 Aluminum and its Alloys 1.1.1 Introduction M e t a l s have had a k e y role i n the development o f human c i v i l i z a t i o n . A l u m i n u m is second to i r o n as the most plent i ful meta l l ic element o n earth, but its pract ical uses started o n l y i n the 1 9 t h century because o f diff icul t ies associated w i t h extract ion f rom its ores [1-3]. The electrolyt ic reduct ion o f a l umina (AI2O3) d i s so lved i n mol t en cryol i te was developed independently i n 1886 by Char les H a l l i n O h i o and P a u l Heroul t i n France [1-3]. In 1892, the A u s t r i a n K a r l Joseph B a y e r developed an efficient process that uses caustic soda to extract a lumina f rom bauxite, a natural ore conta in ing 35 -50% o f a lumina [1]. M o d e r n methods for a l u m i n u m extract ion use essential ly the same process p r inc ip les invented by Baye r , H a l l and Heroul t . The first commerc i a l applicat ions o f a l u m i n u m were novel ty items such as mi r ro r frames, serving trays and c o o k i n g utensils. In t ime, the divers i ty o f applicat ions o f a l u m i n u m and its a l loys g rew to the extent that n o w their use affects v i r tua l ly every aspect o f modern l i fe , f rom food packaging to the automotive and aircraft industries. The properties o f a l u m i n u m that contribute to such a broad range o f applicat ions inc lude: l ight weight (density 2.7 g /cm is m u c h less than for example steel, copper or brass), excellent cor ros ion resistance under no rma l condi t ions , ab i l i ty to conduct electr ici ty and heat near ly as w e l l as copper, non- toxic i ty , h i g h elasticity, toughness at l o w temperature, workab i l i t y and formabi l i ty . Besides these favorable characteristics a l u m i n u m is nonferromagnetic, w h i c h is important for electr ical and electronics applicat ions, and it has h i g h ref lect ivi ty w h i c h is useful for decorating purposes [3-5]. 1 Pure alurninum is a relatively soft material and that limits its usefulness in construction, but other elements can be deliberately added to form alloys with modifications to the physical and mechanical properties. For example, Zn, Ag, Mg and Li have solid solubilities greater than 10%; Ga, Ge, Cu and Si are in the 1-10% range, with other elements being less soluble [4]. Although alloying can produce materials with some enhanced properties, other properties may be less favorably affected, and it is often the case that aluminum alloys have increased susceptibility towards corrosion. Figure 1.1 summarizes the principal types of aluminum alloys based on the elements added [4]. Aluminum alloys are divided in two major categories: cast and wrought. The further division for each category is based on the primary approach to property development: heat treatable and non-heat treatable. In heat treatable alloys, considerable strengthening can be achieved by heating and cooling, while this is not the case for non-heat treatable alloys, which depend primarily on cold-working to increase strength [2-4]. The nomenclature for aluminum alloys has been developed to help identification of alloy composition and the treatment applied. For wrought alloys a four digit system is used while a separate designation scheme is used for cast alloys [2-4]. Table 1.1 shows the major classes of wrought aluminum alloys, its nomenclature, principal alloying element and area of application. After the four digits a basic temper designation (a letter and possibly one to three numbers) is often added to indicate the mechanical and/or heat treatment to which the alloy has been subjected [2-4]. The addition of Cu significantly increases the strength of aluminum alloys to the level that can be comparable to, or even exceed, those of carbon steels. However, copper reduces the corrosion resistance of aluminum more than any other alloying element [2-6]. Therefore alloys containing Cu are often clad with high-purity aluminum or magnesium-silicon alloys of the 6xxx group to maintain corrosion resistance. The addition of Mn imparts moderate strength without adversely affecting the corrosion resistance, Si lowers the melting point without causing 2 Figure 1.1 P r i n c i p a l a l i ra i inum a l loys [4]. brittleness, M g improves strength and weldab i l i ty , and Z n increases strength, and i n combina t ion w i t h M g , provides good duct i l i ty , formabi l i ty and we ldab i l i t y [4]. Table 1.1 Major classes of wrought Al alloys. Nomencla ture P r i n c i p a l a l l o y i n g elements A p p l i c a t i o n s 1 x x x e.g. 1100 S m a l l amounts ( - 9 9 . 0 0 % puri ty) E lec t r i ca l , chemica l industry 2 x x x e.g. 2024 C u , M g Airc ra f t 3 x x x e.g. 3003 M n Archi tec ture 4 x x x e.g. 4032 S i W e l d i n g rods, b raz ing sheet 5 x x x e.g. 5052 M g M a r i n e industry 6 x x x e.g. 6061 M g , S i Arch i tec tura l extrusions 7 x x x e.g. 7075 Z n , C u Ai rc ra f t industry 8 x x x e.g. 8017 Other elements (e.g. B , L i ) 9 x x x Reserved for future use In the w o r k presented i n this thesis, the 2024-T3 a l u m i n u m a l loy has been used. Its overa l l compos i t i on is w i t h i n the ranges: 3.8-4.9 C u , 1.2-1.8 M g , 0.5 F e , 0.3-0.9 M n , 0.5 S i , 0.1 C r , 0.25 Z n o n a weight percent (wt%) basis [4], The temper designat ion (T3) reveals that the a l loy has been heat treated i n solut ion, c o l d w o r k e d after so l id i f ica t ion , and natural ly aged (at normal ambient temperature) to a substantially stable cond i t ion [2]. 1.1.2 Alloy Microstructure The strengthening and hardening i n a l loys is general ly achieved b y a method ca l led precipi ta t ion hardening, w h i c h is accompl i shed by appropriate heat treatments [1,3,4,7]. A g e hardening is another frequently used term to designate this procedure, because the strength can increase as the a l l oy ages. W h e n the amount o f an a l l o y i n g element exceeds the so lub i l i ty l i m i t i n the s o l i d state, other phases m a y precipitate out after c o o l i n g from l i q u i d so lu t ion [1,3,4,7]. 4 Often one or more types o f second-phase particles, o f separate chemica l compos i t ion , m a y be dispersed w i t h i n the m a i n s o l i d so lu t ion (a l loy matr ix) . T h e second-phase particles as intermetall ic compounds represent microstructural heterogeneities, and together w i t h the gra in structure, represent the microstructure o f a mater ial [4,7]. F igure 1.2 shows second-phase particles at the surface o f a mir ror -po l i shed 2024-T3 a l u m i n u m a l loy and the gra in structure obtained after etching i n K e l l e r ' s reagent [8]. The phase diagrams for the A l - C u and A l - C u - M g systems are shown i n F i g . 1.3 [8,9]. In the first, w i t h a C u concentrat ion o f about 4 % , the C u A f z phase (referred to as the 6 phase) exists i n equ i l i b r i um w i t h the so l id solut ion. H o w e v e r w i t h the addi t ion o f M g ( F i g . 1.3(b)), a further second-phase o f compos i t i on A ^ C u M g (the S phase) forms [4,8,9]. La te r i n this thesis, the S phase and variants o n it are often designated as an A l - C u - M g phase. W i t h addi t ion o f Fe , M n or S i , further phases may be formed such as: M g 2 S i , A l n S i ( M n , F e ) , A F j ( M n , F e ) and A l - C u - M n - F e -S i [4]. The usual procedure for fo rming the A l - C u a l loys invo lves so lu t ion treatment at h i g h temperature to m a x i m i z e so lub i l i ty o f a l l the elements, and this is f o l l o w e d by rap id c o o l i n g or quenching to 1 0 0 ° C or sometimes r o o m temperature [3,7]. T h i s leads to a nonequ i l ib r ium situation w i t h a dominant phase ( so l id solu t ion a) supersaturated w i t h another phase. The A l - C u a l loys exhibi t natural aging after be ing solu t ion heat treated and quenched, w h i c h means that w h e n the a l l oy is coo led and kept at a l o w temperature (e.g. 2 5 ° C ) for a pe r iod o f t ime (e.g. one week or longer) , other phases fo rm as f ine ly dispersed particles that g r o w over t ime [4]. The strength and hardness o f the material can increase w i t h the t ime o f natural aging and w i t h the C u content. In general, the microstructure is unique for each a l l oy and depends o n compos i t ion and process treatments appl ied. The mechan i sm by w h i c h precipi ta t ion hardening increases strength and hardness i n a l loys depends o n the presence o f second phase particles, and spec i f ica l ly o n the 5 Figure 1.3 Phase diagrams for: (a) A l - C u a l loys and (b) A l - C u - M g a l loys [9]. 7 strain fields in the matrix surrounding them that obstruct and retard the movement of dislocations [1,4,7]. Knowledge of the microstructure of a material is essential for a deeper understanding of its properties, particularly those associated with corrosion resistance and the various mechanical (fracture, fatigue, plastic) and magnetic properties [4]. The presence of uneven distributions of the alloying elements creates local galvanic cells on the surface when the material is exposed to an aqueous environment. This results from differences in electrochemical activities between regions on the alloy surface rich in the added elements (second-phase particles) and those where its concentration is less (matrix) [1,4]. As a consequence, aluminum alloys are prone to localized rather than uniform corrosion, and that is particularly dangerous in weight-bearing structures since sudden failure can take place without early warning of deterioration in the physical structure of the material. 1.2 Corrosion Protection Layers 1.2.1 Native Oxide The corrosion resistance of pure aluminum under common conditions is due to the oxide layer that forms directly upon exposure to the atmosphere. If damaged, this layer can quickly re-form, thereby maintaining the metal protection. When formed in the atmosphere at room temperature the oxide film is amorphous and composed of a base compact barrier layer and a hydrated upper layer with thickness around 5-10 nm [1-4]. Crystalline Y-AI2O3 forms at around 450°C, and this transforms to CI-AI2O3 at around 1000°C when A l is already in the molten state [3,10]. The native oxide films on A l and its alloys have pores and flaws (Fig. 1.4(a)) associated with scratches, surface irregularities and microscopic heterogeneities such as grain boundaries, sub-grain boundaries and second-phase particles (in alloys) [11-14]. These flaws can act as 8 (a) Pores in the oxide layer Barrier layer Aluminum 99.99% Figure 1.4 Schematic ind ica t ion o f ox ide o n A I surfaces: (a) pores i n the ox ide and (b) cor ros ion in i t ia t ion at loca l sites i n a l l oy [1]. nucleat ion sites for f i l m breakdown, a l though i n a s l igh t ly aggressive m e d i u m , the cor ros ion products m a y fo rm and seal the pores, so protecting the surface f rom further damage [1]. H o w e v e r , i n stronger corros ive environments, the presence o f aggressive ions hinders the repassivation step and provides the onset o f corrosion. D u e to the presence o f the added elements, surfaces o f a l u m i n u m al loys are e lec t rochemica l ly heterogeneous. In a corros ive environment the anodic and cathodic regions g ive meta l d i sso lu t ion and hydrogen evo lu t ion respect ively A l —> A l 3 + + 3 e _ (anodic reaction) (1.1) 2 H + + 2e "-> H 2 (cathodic reaction). (1.2) W i t h oxygen , another cathodic react ion can take place as w e l l G-2 + 2 H 2 0 + 4e 4 0 H (1.3) The cor ros ion characteristics o f an AI -H2O system are convenient ly characterized by a Pourba ix d iagram (Figure 1.5) [15]. T h i s is a phase d iagram w h i c h shows the ranges o f stabil i ty for phases i n the wate r -a luminum system as a funct ion o f p H and electr ical potential . S u c h diagrams predict the spontaneous d i rec t ion o f reactions, and hence help estimate the compos i t ion o f cor ros ion products; further they can identify environmental changes that can lead to prevent ion or reduct ion i n the meta l l ic cor ros ion [16,17]. The stabil i ty reg ion o f water i n F i g . 1.5 corresponds to the reg ion between the l ines marked as (a) and (b). Thus , b e l o w l ine (a), water reduces to f o r m hydrogen 2 H 2 0 + 2 e - - + H 2 + 2 0 H " (1.4) and above l ine (b), water ox id izes to fo rm o x y g e n 2 H 2 0 - » 0 2 + 4H+ + 4e.~ (1.5) The doma in o f a l u m i n u m stabil i ty l ies b e l o w that o f water. In suff icient ly ac id ic solutions ( p H < 4), a l u m i n u m dissolves to produce A l 3 + , w h i c h is accompanied b y hydrogen evo lu t ion (corrosion 10 Figure 1.5 Pourba ix D i a g r a m for the AI-H2O system at 2 5 ° C , 1 a tm pressure [13]. 11 region I), w h i l e i n sufficiently a lka l ine solutions ( p H > 9), AIO2 ions are formed (corrosion region II). The intermediate area is the passive region, where an A l surface tends to be covered b y ox ide film, w h i c h can impart cor ros ion protect ion b y a k ine t ic stabil i ty. B u t , as for other phase diagrams, a Pourba ix d iagram cannot p rov ide any informat ion o n the kinet ics o f a reaction. F o r a l u m i n u m a l loys , the cor ros ion resistance shows a general decrease w i t h increase i n the amounts o f the a l l o y i n g elements. T h e presence o f second-phase precipitates plays an important role i n the beginning o f a cor ros ion process. F igure 1.4(b) schemat ica l ly indicates cor ros ion o f an a l u m i n u m a l loy (99 .5% puri ty) that is covered w i t h a natural ox ide layer, and has second-phase particles conta ining F e and S i , elements that are nobler than A l [1]. The kinet ics o f ox ide format ion at temperatures b e l o w 4 0 0 ° C is governed especia l ly by the di f fus ion o f 0 2 ~ ions [18]. The ox ide film grows into the metal , and i n case o f an a l l oy exhibi ts a different chemica l compos i t ion at points conta ining inclus ions . These different ox ide regions can have different electr ical characteristics, for example be semiconduct ing. Therefore, the under ly ing a l l o y microstructure can affect the permeabi l i ty o f the ox ide layer to electrons, and hence its barrier properties [1]. Once a corrosive species penetrates to the actual metal surface, l oca l i zed cor ros ion starts usual ly i n the fo rm o f pi t t ing, crevice or stress cor ros ion c rack ing . Pi ts can be created by the facil i tated d isso lu t ion o f A l f rom the matr ix due to the close presence o f second-phase particles conta ining metals nobler than A l , or b y the d isso lu t ion o f the particles themselves, as w h e n they conta in metals less noble than A l [19-22]. F igure 1.6(a) shows a mic rograph f rom scanning electron mic roscopy ( S E M ) o f a p i t formed o n a 2024-A1 a l loy sample upon exposure to 3 .5% N a C I solut ion. The cor ros ion product, A 1 0 x ( O H ) y is deposited at the mouth o f the pit , as represented schemat ical ly i n F i g . 1.6(b) [23]. 12 alloy Figure 1.6 (a) S E M micrograph o f a pi t i n 2 0 2 4 - A l a l l oy and (b) schematic representation o f its cross section. 13 1.2.2 Special Approaches In order to improve the corrosion resistance of Al and its alloys, a variety of treatment processes are used. Most commonly in industrial practice, corrosion protection is provided by anodizing, during which the native oxide film is thickened and its pores sealed [3]. However, this can be expensive when dealing with large Al sheets, and it is not applicable for parts that need further mechanical treatment. For such cases conversion coatings are frequently applied in order to provide a chemically stable, nonconductive and inexpensive base for paint application to further the corrosion protection. The term conversion coating refers to a change in a metal surface, by a chemical treatment that leads to the surface being converted into an insoluble inorganic compound [24]. The most widely used approaches include chromating, phosphating and a mixed treatment which involves phosphating followed by a chromic acid post-treatment [3, 24]. Chromating provides excellent corrosion resistance and paint adhesion for most metals including aluminum. The coatings are composed of Cr(III) and Cr(VI) oxides/hydroxides [3,24-28]. It has been proposed that the exceptional corrosion protection afforded by these coatings is due to the ability of Cr(VI) to dissolve, when the protected metal is exposed to a corrosive environment, and migrate to the damaged site and repassivate it by reduction to Cr(III) [26,29-31]. However, in recent years there have been strong pressures to reduce the use of chromate conversion coatings now that the toxicity of the Cr(VI) compounds is becoming more recognized [32]. The alternative methods are reviewed in recent publications [33-35], and for aluminum and its alloys one approach is the application of the phosphate coatings originally developed for protection of iron and steel. Another promising possibility involves application of organosilanes, although this is still at the research stage [33-35]. 14 1.3 Organosilanes 1.3.1 Introduction Silane coup l ing agents were developed i n the late 1940s to pretreat glass fibers for applications i n composites [36,37], and for 40 years or more they have been used i n industry to promote adhesion between organic po lymers and a w ide variety o f substrates [37,38]. The organosilanes o f interest are h y b r i d organic- inorganic compounds w i t h general fo rmula X3Si (CH2 ) n Y, where Y is an organofunctional group, X is a hydrolysable group and n is t yp i ca l ly i n the range 0 to 3. The organofunctional groups are chosen to be compat ible w i t h a subsequently appl ied po lymer or paint, w h i l e the hydrolysable groups are present to fo rm s i lanol groups for bond ing to the substrate [37]. Table 1.2 lists some o f more frequently used silane coup l ing agents. Table 1.2 Some examples of silanes. C o m p o u n d name A c r o n y m F o r m u l a y - A m i n o p r o p y l si lane y - A P S H 2 N - ( C H 2 ) 3 - S i ( O C 2 H 5 ) 3 y - U r e i d o p r o p y l si lane y - U P S H 2 N - C O - N H - ( C H 2 ) 3 - S i ( O C H 3 ) 3 Bis-1,2-( t r ie thoxysi lyl)ethane B T S E ( C 2 H 5 0 ) 3 - S i - ( C H 2 ) 2 - S i ( O C 2 H 5 ) 3 V i n y l silane V S H 2 C = C H - S i ( O C H 3 ) 3 Methyl t r imethoxys i lane M M S C H 3 S i ( O C H 3 ) 3 Methyl t r ie thoxys i lane M E S C H 3 S i ( O C 2 H 5 ) 3 In recent years, attention has been g iven to invest igat ing the appl ica t ion o f organosilanes for cor ros ion- inhib i t ing pre-treatments o f metals. The in i t i a l w o r k indicates that silane treatment before paint ing can give improved cor ros ion resistance o n a variety o f substrates i nc lud ing steel, ga lvanized steel, G a l v a l u m e and a l u m i n u m a l loys [39-42]. In some cases, an improved protect ion was achieved even wi thout an appl ica t ion o f paint [39]. It appears that the l eve l o f 15 protection provided depends on specific details of the silane structure, and on the form of the metal substrate, particularly insofar as they influence the nature of the interfacial bonding, as well as on processing details like silane concentration, hydrolysis time and solution pH [39]. 1.3.2 Silane Bonding to Metals Various theories have been proposed for the mechanism by which silane coupling agents enhance the adhesion. One approach suggests that the silane adhesion occurs as a result of penetration into surface irregularities, which then leads to the so-called mechanical keying or interlocking of the adhesive [43-46]. The basic physical interactions like London dispersion forces and electrostatic interactions (e.g. hydrogen bonds) [37,44-50] must then be responsible for at least part of the silane-to-substrate adhesion. Another theory emphasizes strong primary interfacial bonds between the silane and substrate. These bonds are characterized as mainly covalent in nature, but with some degree of ionic character due to the differing electronegativities of the atoms involved. This approach was developed in the mechanism of Plueddemann [37] for the silane adsorption on Si02 surfaces, and was extended to metals since their surfaces are generally oxidized and contain -OH groups. According to this model, hydrolysis of the Si(OR)3 groups in aqueous solution occurs with production of silanol groups (i.e. Si(OH)3), which may then condense with -OH groups on the oxidized metal surface, so leading to the formation of direct S i -O-M chemical bonds OH H20 - H 2 0 I Y(CH2)„Si(OR)3 Y(CH2)„Si(OH)3 + HO-M -» Y(CH2)„ ^ i -O-M. (1.6) OH The pH of the silane solution is important since it affects both the silane hydrolysis and the properties of the oxidized surface. At a pH of 4, the hydrolysis reaction for many silanes is fast, while the condensation reaction is slow [39,51], and this pH is often helpful for ensuring 16 that as many silanol groups as possible are available for the condensation reaction with the surface -OH groups. However, the situation is different for the aminosilanes (such as y-APS) since, if deposited from acidic solution, they form hydrogen bonds between the amino group and the surface hydroxyls. That is referred to as the "upside-down" silane orientation [52,53]; it does still improve adhesion, but the interactions are not sufficiently strong to passivate the metal surface and enhance the corrosion protection. A convenient characterization of a metal oxide surface in water is provided by the isoelectric point (IEP), which corresponds to the pH at which the net surface charge is zero [37,54,55]. Values of the IEP can vary over a relatively broad range: 0.5-12.5 (a value determined for oxidized films on A l is 9.5 [56]), reflecting the composition and chemical nature of the oxide surface, including presence of impurities, degree of hydration and oxidation state [55]. A metal oxide surface in aqueous solution at pH below its IEP value will have a net positive charge, due to an excess of protonated hydroxyl groups (i.e. -MOH2 + ). Correspondingly, in solutions where the pH is larger than IEP, the surface is negatively charged due to the presence of - M O - [55]. A hydrolized silane molecule, RaSiOH, will have a net positive or negative charge dependent on the relation [Ra-Si-OHjI /pts-Si-Ol = K ^ ^ / K a K w (1.7) where K a and K b are the acid and base equilibrium constants, and K w is the ionic product of water. A common view has been that the solution conditions should be set so that there is an electrostatic attraction between the oxidized surface and the silane, one being positively charged and the other negatively charged. This may help to form the direct interfacial bonding, although desolvation and the condensation process must occur before the Si-O-Al bonding can be established. 17 Plueddemann's model has received considerable attention in a variety of studies that have explored the interfacial bonding between organosilanes and metal substrates, using such techniques as secondary-ion mass spectrometry (SIMS) [57-60], including in its time-of flight (TOF) form [61,62], as well as other techniques such as vibrational spectroscopy [63-66] and X-ray photoelectron spectroscopy (XPS) [53,67]. These efforts to date have been directed at detecting the presence of S i -O -M interfacial bonding, rather than identifying conditions that optimize their presence. 1.3.3 Corrosion Protection by Silanes The procedure for silane deposition involves the steps of: metal cleaning, rinsing with water, dipping into the silane solution and drying [39]. The commercially available alkaline etching solutions are usually used for metal cleaning, since the acid or neutral cleaners produce surfaces with insufficient -OH groups for the subsequent condensation reaction with silanol groups [39]. A most promising organosilane for the corrosion protection of A l and its alloys is the non-functional compound bis-l,2-(triethoxysilyl)ethane, or BTSE (Table 1.2), that according to van Ooij and Child [39] can outperform the chromating pre-treatment on the 2024-T3 A l alloy. BTSE has several advantages in its ability to form high-quality protective films compared to the other widely used silanes. Its bis-silyl function allows as many as six silanol groups to be formed upon hydrolysis, and this increases the probability for the formation of S i -O-A l bonds (Fig. 1.7). Upon drying, a dense, interpenetrated and highly adhered network of BTSE can form on top of a metal and provide the protective properties. The thickness of the BTSE film (as for other silanes) appears to be governed mainly by solution concentration, with the temperature or dipping time having little or no effect [39]. However, films that are too thick become less desirable since they may become brittle. 18 CH3CH20 OCH2CH3 \ / CH,CH o 0-Si-C-C-Si-0CH,CH, / CH3CH20 H2 H 2 \ OCH2CH3 BTSE: bis-1,2-(triethoxysilyl)ethane H20 HO OH \ / HO-Si -C-C-Si -OH / H 2 H 2 \ Q O H r r H OH ?H -Al I -AJ-Al substrate H20 X H O - / S i V - C - C H, H. 0 ' Al substrate OH Si I O OH Figure 1.7 A d s o r p t i o n o f B T S E o n a l u m i n u m substrate. 19 The lack o f an organofunctional group l imi t s the capabi l i ty o f B T S E to bond w i t h a subsequent po lymer coat ing, but a second organofunctional si lane (e.g. y - A P S ) can be deposited o n a f i l m before it is comple te ly cross- l inked, so that i t reacts w i t h the B T S E to fo rm a double layer film. S u c h structures can thereby be attached to the substrate and have a h i g h degree o f organofunctionali ty for bond ing to paint. Indeed examples have been reported to show p romis ing cor ros ion resistance properties w h e n appl ied to co ld - ro l l ed steel and to the 3003 and 2024-T3 A l a l loys [39,42]. Bu t , knowledge o f the nature o f the si lane-metal ( inc lud ing S i - O - A l ) bond ing is s t i l l incomplete . In order to strengthen pract ical applicat ions, more research is needed for the pre-treatments that m a y op t imize such interfacial adhesion and its effects o n cor ros ion protection. 1.4 Phosphate Conversion Coatings 1.4.1 Introduction Phosphate convers ion coatings are used i n industry as metal pre-treatments for the purpose o f cor ros ion protection. In pr inc ip le they fo rm a chemica l l y stable and wel l -adhered phys ica l barrier o n a metal surface, both to prevent penetration o f a corros ive species, and to improve the adhesion o f any coat ing subsequently appl ied. The appl ica t ion o f phosphate convers ion coatings started around a hundred years ago w i t h the p ioneer ing w o r k o f T . W . Coslet t , w h o patented the treatment o f i ron and steel w i t h di lute phosphoric ac id [68]. H i s w o r k was soon f o l l o w e d b y French , U . S . and G e r m a n patents [69-71]. H o w e v e r , the actual idea o f m a k i n g use o f a phosphate coat ing for cor ros ion protect ion dates back to the late eighteenth century [72-74], a l though some archaeological excavat ions suggest that the process m a y have been used b y the R o m a n s i n the th i rd century A . D . [75]. O v e r the years, the phosphating process developed w i t h improvements i n the coat ing recipes, and the so-cal led "Cos l e t t i s i ng" process evo lved into the "Pa rke r i z ing process" and the 20 "Bonder i te process", among others [75]. M o d e r n convent ional phosphat ing solutions (also k n o w n as phosphating baths) are supersaturated solut ions that conta in di lute phosphoric ac id and divalent metal ions such as F e 2 + , Z n 2 + or M n 2 + [75-81]. Other addit ives i n c l u d i n g N i 2 + , C u 2 + , hydrogen peroxide, nitrates and chlorates are k n o w n as accelerators, since they increase the speed o f the coat ing deposi t ion [75-83]. Table 1.3 [79] shows the m a i n phases i n phosphate coatings formed o n different metals f rom var ious baths. Based o n the dominant metal i o n i n the phosphating solut ion, the processes are c lass i f ied as i ron , z i nc or manganese phosphating, a l though there are mod i f i ed variat ions [79]. The emphasis i n this w o r k is o n the z inc phosphate ( Z P O ) process. Table 1.3 Phosphate phases formed on various metals from coating baths containing different metal ions [79]. M e t a l i o n i n bath Fe substrate Z n substrate A l substrate F e 2 + F e 3 H 2 ( P 0 4 ) 4 • 4 H 2 0 F e P 0 4 • 2 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 2 F e ( P 0 4 ) 2 • 4 H 2 0 F e 5 H 2 ( P 0 4 ) 4 • 4 H 2 0 M n 2 + ( M n F e ) 5 H 2 ( P 0 4 ) 4 • 4 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 M n 5 H 2 ( P 0 4 ) 4 • 4 H 2 0 M n 5 H 2 ( P 0 4 ) 4 • 4 H 2 0 Z n 2 + Z n 2 F e ( P 0 4 ) 2 • 4 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 2 + , C a 2 + Z n 2 F e ( P 0 4 ) 2 • 4 H 2 0 Z n 2 C a ( P 0 4 ) 2 • 2 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 Z n 2 C a ( P Q 4 ) 2 • 2 H 2 0 Z n 2 C a ( P 0 4 ) 2 • 2 H 2 0 Z n 3 ( P 0 4 ) 2 • 4 H 2 0 The phosphat ing process i n the metal finishing industry i n general invo lves several stages: c leaning o f the metal surface, r ins ing , surface condi t ion ing , format ion o f a phosphate convers ion coat ing, r ins ing , seal ing (f inal rinse) and d r y i n g [75-79,82]. T h e surface c leaning m a y be done mechan ica l ly (e.g. b y sand blasting), by appl ica t ion o f organic solvents, b y 21 etching (either alkaline or acid), or by some combination of them. In principle, the initial rinsing procedure removes chemicals left over from the cleaning step. Such contamination would be expected to significantly affect the subsequent phosphating process. Surface conditioning refers to the immersion of the metal to be coated into an activating solution (titanium colloid solution) for a short time prior to phosphating. It is believed that this part of the procedure creates nucleation sites for the formation of the phosphate coating [75-77,84-88]. A phosphate conversion coating is formed either by immersion in the phosphate bath (typical bath temperature 30-99°C) or by spraying, with the processing time in either approach ranging from a few seconds to several rninutes depending on the nature of the metal to be coated, the thickness and weight of coating required, and the bath parameters [75,79]. For example, zinc phosphate (ZPO) coatings on aluminum can be obtained by immersion in the coating bath for 1-4 min at 50-60°C and pH ~4. Equilibria in the coating bath for pH ~4 encourage precipitation of the insoluble crystalline tertiary zinc phosphate, Zn3(P04)2'4Ff20 (hopeite) on a metal surface, rather than the primary phosphate (soluble in water) or the secondary form (less soluble), which are unfavorable as coatings [77]. In addition, fluoride is added to the coating bath in order to complex any free A l 3 + and prevent formation of A I P O 4 . However, the F~ concentration has to be monitored, since any excess can cause precipitation of cryolite, with contamination of both the phosphating bath and the conversion coating [89]. The nature of the phosphate coatings depends strongly on the associated details. For example, Fig. 1.8 shows SEM micrographs of some ZPO coatings on pure aluminum and its alloys (6061 and 2024) obtained from both a laboratory-based coating bath and a commercial bath. Comparisons are shown for changing just a single parameter in the coating bath. Those shown emphasize that variations in the size of the coating crystals, as well as coverage, that occur depending on the nature of the substrate and the exact composition of the coating bath. In addition, solution agitation can considerably improve the coverage (Fig. 1.9). 22 »**>>. ' \S •* ' W - * • • • • Figure 1.8 SEM micrographs for phosphated surfaces of A l and its alloys produced under different conditions: (a) 99.99% purity Al with laboratory based coating solution containing 1 ppm of Cu 2 + ; (b) same as (a) but with 10 ppm of added Cu 2 + ; (c) 6061 Al alloy treated with commercial coating solution, and (d) 2024 Al alloy treated under same conditions as (c). 23 Figure 1.9 S E M micrographs from 2024-A1 alloy samples after treating with laboratory based ZPO coating solutions: (a) no stirring; (b) same conditions as (a) but with stirring and (c) as (b) but at high magnification. 24 The final step in a ZPO coating procedure is a post-treatment (also known as final rinse), which is usually done by application of dilute chromic acid [75-80]. The purpose of this step is to wash away unreacted species and unwanted residue, as well as to seal the pores and unprotected areas in the coating. Phosphate coatings are crystalline compared to chromate conversion coatings, which are amorphous, and that leads to the former having a more porous nature. The final rinse therefore represents an integral part of the metal-treatment process and, together with the preceding steps, contributes to the overall protection against corrosion. 1.4.2 Mechanism of Coating Machu [90] has proposed that the initiation of phosphating should be considered as an electrochemical process. When in contact with the acidic phosphating solution, the metal behaves as a "multi-electrode" due to the heterogeneity of its surface. This implies the existence of numerous short-circuited micro-cells comprising microanodic zones, where metal dissolution takes place, and microcathodic zones where hydrogen evolution occurs. While the majority of the metal surface exhibits anodic behavior, cathodic areas are confined to the heterogeneities with distinctive physical or electrochemical properties, such as grain boundaries or second-phase precipitates [90]. Figure 1.10 shows a schematic representation for the initiation of a zinc phosphate coating on aluminum, although the mechanism has been developed from observations related to iron and steel phosphating. The process starts with the pickling reaction that involves etching of the initially-present oxide layer, and the subsequent metal oxidation and dissolution A 1 2 0 3 + 61^ - » 2 A 1 3 + + 3 H 2 0 (1.8) Al° + SFf -* A l 3 + + 3/2 H2(g)T. (1.9) The acid attack removes contamination and provides a chemically clean surface which is favorable for the deposition of the coating. In addition, pickling roughens the metal surface and that can also act to improve the adhesion of the phosphate coating. The concentration of free 25 Phosphating solution: H + , P043% Zn 2 + , F" F Zn 2 + Figure 1.10 Schematic representation o f the mechan i sm o f z inc phosphate coat ing format ion o n A l a l loy . 26 ac id determines the extent o f the etching stage [75]. A t p H 4, the dominant an ion species i n solu t ion is H2PO4" [91], and the precipi ta t ion o f the tertiary z inc phosphate 3 Z n 3 + + 2H2PO4- Z n 3 ( P 0 4 ) 2 i + 4 H + (1.10) is d r iven by discharge o f H + ions at the cathodic sites [90]. Al te rna t ive theories for phosphate in i t ia t ion o n i ron and steel substrates suggest that anodic phenomena m a y be i n v o l v e d [92,93]. In the Z P O process for i ron and steel, it has been observed that an amorphous i ron- r ich layer exists under the crystal l ine hopeite coat ing. It is be l ieved that dur ing the metal d issolut ion, a rapid saturation by F e 2 + occurs at anodic areas, leading to the format ion o f the m i x e d i ron-z inc phosphate, w h i c h then becomes a basis for subsequent nucleat ion o f z inc phosphate crystals. K o z l o w s k i [94,95] assumed the existence o f so-cal led "quasi-anodic areas" o n a steel surface; these are sma l l areas w i t h posi t ive potential at w h i c h epi taxial format ion o f the subcrystal l ine layer composed o f very f ine-grained ferrous phosphates and amorphous i ron oxides takes place. H e proposed that this process facilitates the transformation o f quasi-anodes into microcathodic sites, and that a crystal l ine layer o f insoluble phosphate forms o n top o f these microcathodes. Regardless o f the details i n the mechan i sm at the beg inn ing o f the coat ing process, the in i t i a l z i nc phosphate crystals provide nucleat ion sites for further coat ing deposi t ion. The nature o f the f inal coat ing, especia l ly its effectiveness i n cor ros ion protection, depends i n part o n the in i t ia t ion stage. A gu id ing pr inc ip le is that the smaller the size o f the in i t i a l crystals, and the higher their coverage, the better for obta ining a more protective coat ing. 1.5 Objectives of Research In industry, the phosphat ing o f a l u m i n u m and its a l loys has not yet reached the same leve l o f performance as obtained for the phosphat ing o f steel. The search for improved coat ing recipes is s t i l l based o n "trial-and-error" testing without m u c h reference to an approach 27 dependent on scientific design. While some phosphate coatings can work for the A l alloys used in the automotive industry, their performance on other A l alloy types, such as used in aerospace applications, still requires improvement. The factors that determine the mechanism of coating in relation to the chemical compositions of the alloy surfaces are insufficiently known, and that in turn prevents the systematic development of more effective phosphate coatings. For a number of years, the research in our group has been dedicated to establish the principles and mechanistic understanding of the ZPO process on aluminum and its alloys, Laboratory-based coating recipes were developed for 7075-T6 and 6061-T6 alloys and a better understanding was obtained for the influence of basic parameters such as Zn/P atomic ratio, pH, and F~ in the coating bath [96-98]. Additionally work was reported for the effect of Cu enrichment on the development of improved coatings on 2024-T3 [99], and in another direction research was started on studies related to the adsorption of organosilanes on aluminum [100,101]. The work in this thesis aims to increase understanding for the interfacial chemistries involved in coating processes on aluminum alloys. This research builds on earlier work in this laboratory, but its emphasis is different insofar as it is focused on the role of alloy microstructure in coating processes. Other work recently has shown that A l alloy microstructure, especially involving second-phase particles through their type, size and distribution, markedly affects chemical processes that occur on surfaces upon exposure to corrosive environments [21,102-110]. However, to our knowledge, no corresponding research has been published in relation to the formation of protective coatings. For the work in this thesis, surface science methods have been used to study the effects of 2024-T3 A l alloy microstructure on silane adsorption and the initiation of zinc phosphate coatings. The specific objectives of this work include: • To characterize the adsorption of BTSE and y-APS silanes on different microstructural regions of 2024-Al alloy and identify factors involved in coating formation. 28 • T o characterize the natures o f z i n c phosphate coatings o n 2024-A1 a l loy after f inal rinses us ing either B T S E or di lute ch romic ac id . T h i s is directed at f ind ing alternatives to the use o f C r ( V I ) i n the post-rinse stage, but also to see h o w the subsequent cor ros ion protect ion is dependent o n the under ly ing a l l oy microstructure. • T o identify the factors i n v o l v e d i n the in i t ia t ion o f the z inc phosphate coat ing over different microstructural regions o f 2024-A1 a l loy , and to propose a mechan i sm for coat ing formation. In summary, a l l the w o r k discussed here a ims to help establish the pr inc ip les that provide the basis for deve lop ing new approaches for the cor ros ion protect ion o f A l a l loys , t ak ing account o f their dis t inct ive microstructures. T h i s thesis is organized as fo l lows . Chapter 2 introduces the techniques used. T h i s includes the methods o f X - r a y photoelectron spectroscopy ( X P S ) , scanning electron mic roscopy ( S E M ) , energy dispersive X - r a y spectroscopy ( E D X ) and scanning A u g e r mic roscopy ( S A M ) , used for character iz ing surfaces and coatings, as w e l l as the e lectrochemical po lar iza t ion measurements used for assessing cor ros ion stabil i ty and reactivi ty. Chapter 3 reports o n the adsorption o f B T S E and y - A P S o n different microstructural regions o f 2024-A1 a l loy , and Chapter 4 describes investigations o n the natures o f phosphated coatings and their cor ros ion resistance after app ly ing different post-rinse procedures. Chapter 5 reports o n the dependence o n the in i t ia t ion o f z i nc phosphate coatings o n a l l oy microstructure, and Chapter 6 summarizes some conclus ions and directions for future work . 29 Chapter 2 Experimental Methods 2.1 Introduction XPS and AES 2.1.1 Basic Principles X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES) belong to the category of analytical methods referred to as electron spectroscopy for chemical analysis. These are key techniques for the surface analysis and characterization of solid materials which are stable under vacuum conditions. They enable elemental analysis and provide information on chemical bonding (except for hydrogen and helium) for the topmost 50 A or so from the surface [111-115]. The kinetic energies of photoelectrons emitted from the sample surface upon irradiation with soft X-rays are measured in XPS, while the energies of Auger electrons are measured in AES. The Auger electrons can be excited by photon bombardment, but for studying local regions of a sample, an incident electron beam (energy ~10 keV) is preferred because it can be focused better. XPS has its origin in investigations of the photoelectric effect, which was discovered by Hertz in 1887 [116], and explained twenty-five years later by Einstein [117]. In 1914, Robinson and Rowlinson did the first experiments which used this phenomenon to perform an analytical function [118]. Throughout the 1950s and 1960s, Siegbahn and coworkers [119] refined the instrumentation, and this led to the first commercial XPS instruments being developed around 1970. The principle of XPS is illustrated in Fig. 2.1(a). The kinetic energy of the photoelectron is described by the Einstein equation: E k = h v - E b (2.1) 30 Photoemission photoelectron (Valence band (a) 2,3 K X-ray fluorescence V a l e n c e band (b) 2,3 K Auger emission K L j L 2 3 A u g e r electron 9 Valence baa t • (c) 2,3 K Figure 2.1 Schematic d iagram for (a) photoelectron emiss ion ; (b) X - r a y emiss ion and (c) A u g e r electron emiss ion . 31 where hv is the photon energy and Eb is the binding energy that corresponds to the amount of energy required to remove a particular electron from a solid [111-115]. The process that became known as the Auger effect was identified experimentally by Pierre Auger [120-122] in 1923, after its theoretical prediction by Rosseland [123]. Auger studied X-ray interactions in a cloud chamber and noticed the appearance of electrons with energies independent of the excitation source, which he explained in terms of the radiationless de-excitation of ionized atoms. Thus the ionized atom is generally in an excited state, and the excess energy can be dissipated through one of two relaxation processes: (i) photon emission, or (ii) emission of Auger electrons, as illustrated in Fig. 2.1(b) and (c) respectively. The kinetic energy of the emitted Auger electron depends on the three energy levels involved. For the example in Fig. 2.1(c), this kinetic energy can be expressed as: ERLL = E K - E L I - E u - U (2.2) in terms of the three binding energies involved, and U corrects for the hole-hole interaction energy and relaxation effects [111,124]. Each element has its characteristic set of Auger electron energies [111,113-115]. The potential of AES for surface analysis was recognized in 1953 by Lander [113] and commercial instruments were available by around 1970. The use of both XPS and AES for surface analysis depends on studying electrons emitted from solids with energies of 2 keV or less. Such electrons have a very restricted inelastic mean free path, X. This parameter represents the average distance that an electron can travel in the solid before it loses energy by inelastic scattering. Empirical values of X [125] are shown in Figure 2.2, which represents the so-called "universal curve" of Seah and Dench [126]. Variations with material are often small, and although there is an energy dependence, for the energy range 20-2000 eV, X is generally close to 10 A. The practical sampling depth corresponds to 3X, the depth from which 95% of a signal is contributed after background correction [115]. 32 Figure 2.2 Inelastic mean free path o f electrons as a funct ion o f k ine t ic energy inside the s o l i d [126]. 33 2.1.2 Spectral Features An XPS spectrum is displayed as a plot of 'intensity' or 'number of counts' (on the y-axis) against either kinetic energy or, more usually, binding energy on the x-axis. A measurement is typically performed by first taking a wide scan (or survey scan) spectrum covering a range of 1000 eV or so, and then looking in more detail at higher energy resolution over narrow energy ranges (e.g. 10-20 eV) [115]. Figure 2.3(a) shows a survey scan from a ZPO-coated 2024-T3 Al alloy sample after exciting with Mg Ka radiation. The peaks seen are associated with core-level photoemission events (Zn 2p, O Is, C Is, P 2s, Zn 3s, P 2p, Zn 3p), as well as with X-ray induced Auger electron emission (O KLL and Zn LMM); each peak can be identified by comparing its position on the binding energy scale with tabulated data [127-129] and the use of either Equation (2.1) or (2.2). This provides the basis for qualitative analysis in both XPS and AES. An XPS spectrum at high resolution provides more detail for the core level peaks. For example, the narrow scan of the Zn 2p photoemission peak in Fig. 2.3(b) [127] shows a doublet structure. Such a splitting is observed for electrons coming from p, d, f,... orbitals, that is from orbitals with angular momentum quantum number fi greater then zero (i.e. 6 = 1,2,3...). The splitting originates from the spin-orbit coupling interaction, which results in the energy for a state corresponding to j = fi + s being different from that for j = 6 - s, where s (= 1/2) is the electron spin quantum number. The components of the doublet are distinguished by the quantum number j , for total angular momentum. Possible values for j are: 1/2 and 3/2 for p orbitals, 3/2 and 5/2 for d orbitals, and 5/2 and 7/2 for f orbitals. The intensities of these components are proportional to their respective degeneracies, 2j +1, after correction for the contribution of inelastically scattered electrons to the spectral background. Narrow scan spectra may also provide information about the specific chemical state (Fig. 2.3(c)). For example, when an electronegative species like O is bonded to Al , the latter atom assumes a partial positive charge. 34 Zn 2p (a) / 0 1s 0 (A) \ C l s I Zn (A) 1 , 1 , 1 ^~\k P2p Hi 1 C/3 • * — i I fl CD 1000 800 600 400 200 T~ * 1 • 1 1 r 2p3 2 Pl /2 1065 1055 1045 1035 1025 79 77 75 73 71 Binding Energy (eV) 0 1015 69 Figure 2.3 X P S spectra f rom z inc phosphated 2024-A1 a l loy : (a) survey spectrum, and higher resolut ion spectra for (b) Z n 2p and (c) A l 2p. 35 A 1 K L L u 1 1 1 1 1 1 1 1 1 1 400 500 600 700 800 900 1000 1100 1200 1300 1400 Kinetic Energy (eV) Figure 2.4 A u g e r spectra measured f rom 2 0 2 4 - A l a l l oy showing regions o f (a) meta l l ic A l and (b) o x i d i z e d A l . 36 In turn this results i n an increase i n the A l b ind ing energy, and a " chemica l shift" is seen between a meta l l ic A l atom and one bonded to O . Conven t iona l ly A u g e r spectra are labeled w i t h K , L , M for p r inc ipa l quantum number n = 1,2,3...., and this is often extended w i t h subscripts (e.g. L i for 2si/2, L 2 for 2pm, L3 for 2p3/2) [111]. A s w i t h X P S , A u g e r peak posi t ions and shapes can be inf luenced by loca l chemica l and electronic environments as i l lustrated i n F igure 2.4 for a A l K L L spectrum from a sample conta ining both meta l l ic and o x i d i z e d a l u m i n u m . T h i s shows that an A u g e r l ine shape can be used to extract chemical-state informat ion, even though i n general for A E S three different energy levels are i n v o l v e d , so m a k i n g the situation more complex than for X P S . 2.1.3 Quantitative Analysis Quanti tat ive analysis w i t h X P S or A E S depends o n m a k i n g measurements o f peak intensity; this can be defined b y the area o f the peak left after r emov ing any background contr ibut ion. Factors that contribute to the intensity measurement inc lude: the incident beam flux (f), the scattering cross section (0), the number o f atoms per unit v o l u m e (n) that g ive rise to the peak, the area o f the sample from w h i c h the electrons are col lec ted ( A ) , the instrumental t ransmiss ion funct ion (T) and the inelastic mean free path (X) [112]. The intensity contr ibut ion for electrons ex i t ing i n the normal d i rec t ion from an inf in i tes imal thickness dx at depth x i n a sample is [130]: d i = f o n A T exp (-xA.) dx (2.3) and integration for a semi-inf ini te homogeneous sample gives: I = f o n A T L (2.4) In pr inc ip le , ion iza t ion m a y also occur from backscattered electrons, and a correct ion to E q . (2.4) m a y be needed, par t icular ly for the higher energies used i n A E S and w h e n l ight atoms are present a long w i t h heavier atoms [131]. 37 For a particular peak under study, and particular instrumental settings, the parameters f, a, A and T in Eq. (2.4) can be conveniently grouped into a sensitivity factor (S), whose value is set for each particular instrument. Then the composition ratio for two elements in a sample can be expressed as: ni/na = [(I,/S,)/(I2/S2)] [h/hl (2.5) Specific values of X\ and A,2 may be used in a quantitative analysis, but for semi-quantitative work, the ratio X2A,i is commonly taken as constant and equal to unity. Then the elemental composition ratio ni/n2 within the depth probed may be estimated directly from measured peak intensities and sensitivity factors. This can be generalized for more than two components, by expressing the atomic percentages as: m atomic %={( Ii/Si )/SjIj/Sj}x 100 j = 1,2,3... (2.6) Information on the variation of composition with depth near the surface can be obtained by measuring peak intensity as a function of collection angle [112], but a more general approach is to depth profile [114]. However this is a destructive technique since the sample is bombarded by a beam of chemically inert ions (usually Ar) accelerated to several keV. The ion beam is rastered across the sample surface for a known time in order to remove a controlled amount of the sample, and spectra are recorded at chosen sputtering times in order to indicate the depth distribution of different species in the sample. In general, the sputtering rates depend on such factors as the ion energy, beam intensity and incident angle. 2.1.4 U s e o f H i g h E n e r g y E l e c t r o n B e a m Focused electron beams with energies of the order of 10 keV have a key role for three techniques used in the research discussed in this thesis, namely scanning Auger microscopy (SAM), scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) spectroscopy. At this energy, electrons from the incident beam can undergo various elastic and inelastic 38 Figure 2.5 Ene rgy spectrum o f electrons emitted f rom a s o l i d upon interaction w i t h incident electron beam [132]. 39 scattering events upon striking the sample. A schematic diagram of the energy distribution of electrons emitted from a solid is shown in Fig. 2.5 [132], where some regions are classified as secondary electrons (SE), backscattered electrons (BSE) and low-loss electrons (LLE). The LLE region is especially involved with plasmon excitations, and Auger electrons (marked AES) are also shown (in magnified form) in the slowly-varying background for the BSE region. The electrons in the later region are associated with multiple energy losses and scattering events. The SE electrons are especially used for imaging in SEM, while specific Auger electron peaks are studied in SAM. Beside electrons, X-rays are also emitted as a result of the electron-solid interaction, and their measurement is a basis for the EDX analysis (see Section 2.4). Whether electrons or photons are being studied, there is a broadening insofar as they come from a wider area then that impacted by the incident electron beam, which in practice may be of the order of 10 nm or less. In general, care is needed that measurements are not affected by electron beam damage. 2.2 XPS Instrumentation 2.2.1 MAX200 Fig. 2.6 illustrates the Leybold MAX200 spectrometer, which was used in this work. It consists of four interlinked chambers: the transfer chamber (for sample introduction), the analysis chamber (for performing the analytical measurements), and two preparation chambers for 'in situ' sample treatment. The X-ray source, energy analyzer with lens system and detector, as well as the ion gun for sample cleaning and electron gun for Auger analysis are accommodated in the analysis chamber. A manual transfer rod takes a sample from the transfer chamber to the analysis chamber, while two additional rods are available for sample transfer from the preparation chambers to the analysis chamber [130]. 40 Figure 2.6 M A X 2 0 0 X - r a y photoelectron spectrometer. Metal deposition chamber Plasma chamber f , ; » s « l B i Gate valve Analysis chamber Transfer chamber & (Q^ Rotary pump Sublimation pump Turbo pump (W) Ion pump Figure 2.7 Schematic representation o f the p u m p i n g system for the M A X 2 0 0 . 42 The spectrometer is designed to operate in the ultrahigh vacuum (UHV) pressure range (10"8 - 10"10 mbar). This is necessary for two reasons: (i) to prevent low energy electrons from being scattered by residual gas molecules and thus to limit noise in a measured spectrum; and (ii) to limit contamination of the sample surface by adsorption from the residual gas. UHV conditions are achieved with the pumping system shown schematically in Fig. 2.7. The main chamber is pumped by a turbomolecular pump (pumping speed: 360 L/s), with its associated rotary pump, and by an auxiliary titanium sublimation pump. The X-ray source is also pumped by an ion pump (20 L/s). UHV systems need to be baked from time to time in order to slow down outgassing by removing adsorbed molecules from the chamber walls; in practice for the MAX200 this is done at around 100-120°C for up to 24 h. After such a baking, the base pressure can be obtained at around 2 x 10'10 mbar for the analysis chamber, although during use the pressure is generally a little higher (e.g. 5 - 8 x 10"10 mbar); the typical pressure in the transfer chamber is about 2 x 10"8 mbar. 2.2.2 X - r a y S o u r c e X-rays are produced by bombarding the anode material with electrons accelerated through a potential difference of about 10 kV from the filament (at ground potential). The electron impact creates core holes in the target, and the relaxation process dominantly involves photon emission. Different anode materials have characteristic emission lines with different energies. The MAX200 spectrometer is equipped with a dual anode X ray source that gives the Ka lines of either Al (1486.6 eV) or Mg (1253.6 eV); either may be chosen by the setting of a simple switch that selects between the two separate filaments. The anode is water cooled to prevent failure due to heat generation during operation. The radiation generated passes through a thin A l (~2 urn) window, which protects the sample from bombardment by stray electrons. This window also helps reduce the intensity of additional X-ray lines (KP), which can add satellite 43 structure i n an X P S spectrum ( l o w intensity, broad peak at l ower b ind ing energy side o f a photoelectron peak). These addi t ional contr ibutions can be r emoved by us ing the monochromator w h i c h is avai lable i f higher energy resolut ion is needed. 2.2.3 E n e r g y A n a l y z e r F i g . 2.8 outlines the analyzer system o f the M A X 2 0 0 spectrometer; the m a i n components are the co l l ec t ion lens, the energy analyzer and the mul t ichannel plate ( M C P ) detector. The two-stage input lens system controls the co l l ec t ion and focusing o f photoelectrons toward the energy analyzer sl i t by ramping voltages o n the different lens elements. The first stage controls the analysis area (spot size) and co l l ec t ion angle for the electrons c o m i n g f rom the sample surface. The second stage acts to retard the electrons to a particular pass energy, and it also controls the angle (a) at w h i c h electrons enter the analyzer. The concentric hemispher ica l analyzer ( C H A ) is constructed f rom two concentric hemispheres o f radius R i (inner) and R 2 (outer) to w h i c h a potential difference A V is appl ied. Elect rons travel o n the central c i rcu lar trajectory through the energy analyzer and reach the detector on ly i f their k ine t ic energy E 0 inside the analyzer (the pass energy) satisfies e A V = E o ( R 2 / R i - R i / R 2 ) . (2.7) Therefore, for different A V appl ied, different pass energies are required. The relative analyzer resolut ion (AEanaiyzer/E0) is a funct ion o f the analyzer sl i t widths (entrance S\, exit S 2 ) , and o f the electron entrance angle a , and the electron trajectory radius (Ro): A E analyzer / E 0 = (5 , + S2) I 4R« + tt2/4 . (2.8) Since S\, S 2 , Ro and a are fixed by the spectrometer construction, the analyzer resolut ion necessari ly varies w i t h E 0 [111]. T h e energy analyzer i n the M A X 2 0 0 operates i n the constant pass energy mode. A c c o r d i n g to E q . (2.8), the lower the pass energy, the better the energy resolut ion. H o w e v e r , this i s offset b y a reduct ion i n the s ignal intensity. T o obta in an o p t i m u m 44 Concentric hemispherical analyzer (CHA) / \ ^ — \ 1 S ] i f Second lens stage A3 — Multichannel V///////A detector (MCP) NSSSSSSvl TTTT First lens stage A1 ^^^^/^K-xzy source t* sample Figure 2.8 Schematic diagram for the concentric hemispherical analyzer (CHA) and lens system in the MAX200. 45 balance between resolution and intensity, 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 48 eV is used for narrow scans. 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 (AE analyzer), the natural line width of the X-ray source (AE source), as well as the inherent line width of the atomic level involved (AE iine). The observed peak width can be expressed by: AE peak = (AE2 analyzer + AE 2 S0Urce + AE 2 \me)m (2.9) provided that all contributions have the Gaussian form [111]. The detector is constructed from multi-channel plates assembled in a back-to-back configuration to form a chevron array. Each MCP is an array of 18 capillary-type microchanels, which act as individual electron multipliers. The voltage across the plate is set to allow count rates at above 107 s*1. The measured kinetic energy of an electron emitted from a solid sample is referenced to the vacuum level of the spectrometer, while the binding energy of the electron inside the sample is referenced to the Fermi energy of the sample. These values are schematically illustrated in Fig. 2.9 [115]. For a conducting sample in electrical contact with the spectrometer, the Fermi levels of sample and spectrometer are equal, and the measured binding energy can be expressed as: E b=hv-E k-<D s p (2.10) where O s p is the spectrometer work function. In practice, O s p is determined by calibration with a measurement on a standard gold sample, given that the binding energy of the Au 4f7/2 line can be referenced to 84.0 eV [111,112]. Insulating samples require an internal reference, and the C Is peak associated with adventitious hydrocarbon contamination (C-C and C-H bonds) at 285.0 eV is commonly used. 46 T vacuum level vacuum level hv <t>s 1 sp Fermi level Fermi level sample spectrometer hv = energy of photon <|)s = work function of sample (i.e. energy difference between Fermi level and vacuum level) = kinetic energy of photoelectron with respect to vacuum level of sample Fi,, = binding energy of electron in solid with respect to Fermi level (|)sp = work function of spectrometer E^, = kinetic energy of photoelectron measured by spectrometer Figure 2.9 Schematic diagram of the relevant energy levels for the binding energy measured in the spectrometer from a conducting sample (redrawn from Ref. [115]). 47 2.2.4 D a t a P r o c e s s i n g T o m a x i m i z e the informat ion from high-resolut ion X P S and A E S spectra, some processing o f the r aw measured data is required. The first step is to remove the background contr ibut ion from the overa l l measured peak. T h i s w o r k used the nonl inear background correct ion introduced by Shi r ley [133], and is discussed i n relat ion to the photoelectron spectrum i n F i g 2.10(a). T h i s method assumes that the number o f inelas t ical ly scattered electrons at any point i n a spectrum is proport ional to the number o f e last ical ly scattered electrons at higher k inet ic energies. Th i s correct ion (over an energy range chosen b y the operator) is determined by the iterative a lgor i thm: f max N ' k (E) d E (2.11) where N ( E ) is the measured count rate, N ' k ( E ) represents the count rate after subtraction o f background contr ibut ion (k t h iteration), C is a constant and N ( E m a x ) is the reference background leve l . The process starts w i t h N ' i ( E ) = 0 and continues unt i l N V i ( E ) ~ N ' k ( E ) w h i c h general ly requires three or four iterations. Af te r the background subtraction, a curve fi t t ing process is used i n X P S to identify a l l i nd iv idua l components w i t h regard to peak pos i t ion , intensity and peak w i d t h ( F W H M ) . T o op t ima l ly apply the fi t t ing method, chemica l knowledge o f the system has to be connected to the mathematical approach. F r o m the chemistry, a realist ic estimate can be made for the number o f different chemica l components for the element o f interest. A n example o f the curve fi t t ing process is i l lustrated i n F i g 2.10(b) for a C Is spectrum. The mathematical approach is to simulate a measured spectrum (after subtraction o f background) by a set o f i nd iv idua l components, and for each o f these the processing program avai lable w i t h the M A X 2 0 0 uses the standard m i x e d Gauss ian-Loren tz ian funct ion: f(E) = peak height / [1+ M ( E - E o ) 2 / p 2 ] e x p { ( l - M ) [ln2 ( E - E o ) 2 ] / p 2 } (2.12) 48 Figure 2.10 Da ta processing for a C Is spectrum: (a) Sh i r l ey non-l inear background subtraction and (b) curve fitted spectrum. 49 w h i c h represents an approximat ion to the V o i g t funct ion (convolu t ion o f Loren tz i an w i t h Gaussian) [111]. In E q . (2.12), E o is the energy for the m a x i m u m o f the i n d i v i d u a l component peak, P is nearly 0 . 5 F W H M and M is a m i x i n g ratio (1 for pure Loren tz ian ; 0 for pure Gaussian) . Af te r an in i t i a l estimate o f the parameters, for a l l components needed to fit a spectrum, the program iterates to op t imize the fit between the s imulated spectrum and the measured spectrum. T h i s i nvo lves m i n i m i z i n g the least-squares funct ion (x): X= {(1/Nfree) Sj ( Y m e a j - Y f l t J ) 2 / Y m e a j } 1 / 2 j = 1,.... N (2.13) where Y m e a j is the measured count rate at the j - t h data point , Yf l t j is the corresponding value o f the s imulated function obtained as a s u m o f functions o f the type i n E q . (2.12); N is the number o f data points and Nfr e e equals N - Nf l t where Nf,t is the number o f parameters to be fitted through the m i n i m i z i n g process. V i s u a l compar i son o f the measured and s imulated curves is s t i l l important to guard against any artifacts introduced by the curve fi t t ing procedure. 2.3 SAM and SEM Instrumentation 2.3.1 Thermo VG Microlab 350 F i g . 2.11 illustrates the The rmo V G M i c r o l a b 350 system used for this work , especial ly for measur ing A u g e r spectra and S E M images. It consists o f four l i nked chambers: transfer chamber for sample introduct ion, preparation chamber for sample treatment, a chamber where sample can be fractured and the analysis chamber for the analyt ical measurements. Sample transfer between the chambers is accompl i shed w i t h wobb le st icks and a r ack -p in ion transfer rod. The analysis chamber is equipped w i t h the Schot tky f ie ld emiss ion ( S F E ) electron source, a hemispher ica l energy analyzer w i t h detector for A u g e r electrons, and i o n gun for sample c leaning and depth p ro f i l ing . In addi t ion, detectors are present ( F i g . 2.12) for measur ing the secondary and backscattered electrons, and X - r a y s , as required for the addi t ional characterizations. 50 Figure 2.11 The rmo V G Scient i f ic M i c r o l a b 350. Analysis chamber Rotary pump Turbo pump Sub l ima t ion pump Transfer chamber Ion pump B S D = electron backscattered detector S E D = secondary electron detector E D X = X - r a y detector F igure 2.12 V a c u u m system for M i c r o l a b 350. 52 The M i c r o l a b 350 operates under U H V condi t ions that are achieved w i t h a combina t ion o f pumps. A n i o n p u m p (180 L / s ) and a t i tan ium subl imat ion p u m p provide pressures i n the l o w l O - 1 0 mbar range for the analysis chamber. The preparation chamber also has an i o n pump (50 L / s ) , w h i l e the transfer chamber uses a turbomolecular p u m p (240 L / s ) backed by a rotary pump; p u m p i n g for the electron source ( S F E ) i s b y an i o n p u m p (20 L / s ) . The system bak ing fo l lows s imi la r procedures to those described i n Sec t ion 2.2.1. The S F E source ( F i g . 2.13) consists o f a fine z i r con i a (Z1O2) coated single crystal tungsten w i r e that is we lded o n to a tungsten ha i rp in fi lament, used to resis t ively heat the t ip (~1800 K ) . A p p l i c a t i o n o f an electr ic f i e ld ( 1 0 7 V c m " 1 ) to the t ip reduces the potential barrier to electron emiss ion by tunnel ing [132]. T h i s source is fragile and depends o n l o w pressure (<10*8 mbar) for stable operation, but it provides h i g h brightness (5x10 A / c m steradian) combines w i t h l o n g operating l i fet ime ( - 5 0 0 0 h), and an energy spread i n the range 0.3-1.0 e V [134,135]. Af t e r emiss ion , the beam is focused to <10 n m w i t h a system o f electromagnetic lenses and apertures schemat ica l ly s h o w n i n F i g . 2.14. T h i s enables the h i g h spatial resolut ion for the S A M technique, but i n practice care must be taken to l i m i t problems w i t h mechanica l vibrat ions, presence o f stray magnetic or electric f ields, and any inhomogenei ty i n the deflect ing magnet ic f ie ld and its ax ia l asymmetry [136]. The hemispher ica l analyzer (basic in t roduct ion i n Sec t ion 2.2.3) is operated i n the constant retardation ratio ( C R R ) mode for A u g e r analysis i n contrast to the constant pass energy mode used i n X P S [111]. In the C R R mode , the pass energy E 0 is var ied to main ta in a constant value for the E / E 0 ratio. The detection o f peaks at l o w kinet ic energy is easier i n this mode due to the reduced background, but the absolute resolut ion ( A E ) progress ively increases as the k inet ic energy increases [134]. The detection i n the M i c r o l a b 350 is by six-paral lel-connected channeltrons, and the analyzer can be set to detect a part icular element o f interest, or p rogrammed to detect several signals sequential ly so that resolved compos i t iona l informat ion 53 Suppressor Extractor Beam Defining Aperture 1st Electrode 2n d Electrode (focus) 3r d Electrode (Anode) Crossover Source FILAMENT SUPPLY O - J A (T>pledyZ4-Z7SA| SUPRESSORHV SUPPLY P-1kV OVpic«»y J«v) EXTRACTOR HV SUPPLY 0 - 1 0 k V . (Tjpte.Hy4.71iV) LENS HV SUPPLY 0-7.SKV ( I n o t w M S Vfth Bi ramHT) BEAM HT SUPPLY (27.5KVFor CondUoning) 100-200 mm Figure 2.13 F i e l d emiss ion electron source i n M i c r o l a b 350 and its p o w e r supply [134]. f i lament suppressor extractor lens gun aperture upper scanning co i l s lower scanning co i l s st igmator co i l s objective c o i l objective aperture Figure 2.14 I l lumina t ing system i n M i c r o l a b 350. 55 can be obtained. The electron beam can be focused to one spot for point analysis, moved along the line for line profile or rastered across the sample surface to generate two-dimensional Auger maps for the various elements detected. Further, a three-dimensional analysis of a sample surface can be obtained by combining the two-dimensional maps with sputter etching. An Avantage software is used in this thesis for SAM data analysis. In practice, the particular microareas of interest for Auger analysis are first identified from a SEM image, but the focused electron beam in the Microlab 350 can be used to generate these images as well. In this context, the incident electron beam is rastered across the sample surface and the emitted secondary electrons are collected by the scintilator/photomultiplier detector also included in the analysis chamber (Fig. 2.12). For this mode of operation, no voltage is applied to the lens system of the energy analyzer, so as not to deflect the electrons from the scintilator. The detector is interfaced to the PC monitor and the image is build up in synchronism with the scanned beam on the specimen. With detection of the slower secondary electrons, SEM images can have a three-dimensional appearance that arises from the differences in contrast between various structural features on the sample surface. This difference in yield of secondary electrons detected relates to the relative orientation of the local topography, as well as mass density and the chemical natures of the different regions. To a first approximation, elevated sample areas often appear brighter, but any charge accumulation can also affect the yield of secondary electrons [136]. 2.3.2 Other Instruments for SEM The Microlab 350 is recently purchased equipment that only became available for research during the winter 2002-3. Accordingly a fair amount of the SEM imaging reported in this thesis was done using other instruments. Thus, images shown in Chapters 3 and 4 were generated by the Hitachi S4100, Hitachi S2300 and Hitachi S3000N instruments, while those 56 shown in Chapter 5 were produced by the Microlab 350. The Hitachi S4100 SEM has a cold field emission (CFE) electron source that operates at room temperature and hence has a lower energy spread (0.2-0.3 eV) compared to the SFE source in the Microlab 350. The CFE source has higher brightness (109 A/cm2steradian), but it is more susceptible to contamination. The Hitachi S2300 and Hitachi S3000N are both equipped with a tungsten hairpin filament source that produces electrons by thermionic emission with the filament operated at around 2700 K. This type of source can operate at higher pressure (10"5 mbar) than the others, and it is easier to exchange. However, those advantages are offset by lower brightness (106 A/cm2steradian), higher energy spread (1.0 eV) and shorter lifetime (e.g. 60-200 h) [132,135]. Besides secondary and backscattered electron detectors, the additional instruments are also equipped with X-ray detectors, and accordingly they were also used for EDX analysis. 2.4 Electron Dispersive X-ray Spectroscopy 2.4.1 Introduction Electron dispersive X-ray (EDX) spectroscopy has been used in this research to assess the chemical composition at different microareas of coated and uncoated 2024 alloy samples, prior to the arrival of the SAM. In EDX, X-rays are emitted from the sample through excitation by a focused electron beam following the mechanism shown in Fig. 2.1(b). For ionization from deeper core levels, X-ray emission dominates, and chemical analysis can in principle be obtained from the characteristic pattern of photon energies for all elements heavier than Be [136-139]. An example of an EDX spectrum is shown in Fig. 2.15, where peaks are labeled according to A l Ka, Mg Ka and Cu Ka (each peak involves a 2p —• Is transition). The example shown illustrates an application of EDX to qualitative analysis, where the peak energies observed are identified through comparison with an X-ray emission database. As for XPS, quantitative analysis depends on measuring the area under a peak after application of a non-linear 57 ure 2.15 E D X spectrum measured f rom an A l - C u - M g part icle. background correction. The composition ratio for two elements (a and b), on a weight percentage basis may be expressed as Ca/Q^kabla/Ib (2.14) where la, lb are the correspondent X-ray intensities, and kab is the appropriate proportionality factor [137]. The latter can be determined by analysis of standards under similar conditions, or it may be computed from the first principles taking account of the specific instrumental system and factors such as excitation efficiency and fluorescent yield [137]. The detection limit for EDX is at the 0.1 wt% level. In a sense EDX (electrons in, photons out) corresponds to the opposite process to that in XPS (i.e. photons in, electron out), but the probing depths, for the two techniques are very different. While information provided by EDX comes from depths in the um range [136-139], XPS is more surface sensitive by a factor of about 103. 2.4.2 E D X Instrumentation EDX measurements reported in this thesis were made using the Hitachi S2300 and Hitachi S3000N instruments, using a Quartz XONE software. The detection of X-rays is by a lithium-doped silicon diode (Fig.2.16) which is cooled by the liquid nitrogen to minimize electronic noise [137]. The front of the diode is covered by a thin gold film, which is biased to 1000 V to establish a depletion zone. Absorption of X-rays then produces electron-hole pairs that are swept out by the bias to manifest charge pulses on opposite sides of the diode. These pulses are converted into digital form by the analog-to-digital converter, sorted, stored and displayed in a multichannel analyzer. The Si(Li) crystal is protected against contamination by the Be window, but this in turn limits detection of the low-energy X-rays (e.g. elements with Z<10) [137]. 59 Cryostat coo led by l i q u i d N 2 C o l l i m a t o r B e w i n d o w \ X - r a y photons Au vt. contact layer S i ( L i ) r - l ^ e Crys t a l + e--e" signal Preampl i f ie r 0 Bias Voltage supply 0 C o l l i m a t o r A m p l i f i e r vol tage pulse V i d e o / C R T output F igure 2.16 S i l i c o n diode X - r a y detector for E D X (redrawn from Ref . [ 137]). 60 EDX spectra have the Gaussian peak shape and this reflects the performances of the detector and other components in the signal processing chain. The spectral resolution, expressed as the fAill-width-at-half-maximum (FWHM) height, is also a function of the measured X-ray energy. Different instruments are compared for a given X-ray line and operating conditions, for example the 5.9 keV Mn Ka line at 1000 counts s"1 and 8 us pulse processor time constant. On this basis, the spectral resolution for the Hitachi S2300 instrument is 149 eV, while that for the Hitachi S3000N is 131 eV. 2.5 Electrochemical Polarization Curve Measurement 2.5.1 Background Metallic corrosion in aqueous solution is an electrochemical process with electron transfer at the metal/solution interface. Considering an example of a reaction with single electron transfer between oxidized (O) and reduced (R) species in solution O(aq) + e"(m) <-»• R<aq) (2.15) the magnitudes of the cathodic current (ic) for the forward O + e~ —* R reaction, and of the anodic current (ta) for the backward R —»• O + e~ reaction are given as ic = FAk r e d[0] 0 (2.16) and i a = FAkoX[R]0 (2.17) where F is the Faraday constant, k r ed and koX are reaction rate constants, A is the electrode area, and [0]0 and [R]0 are concentrations of oxidized and reduced species at the electrode surface [16,140-142]. The net current usually expressed as a net current density (i = i/A), can then be written as i = (ia- ic)/A = F(koX[R]0 - k red[0]0) (2.18) 61 since the individual currents have opposite directions. At equilibrium, the net current flow is zero. The electron transfer reaction can be described in terms of transition state theory, which assumes that the reaction (2.15) depends on overcoming an energy barrier in order to yield products (Raq). According to this approach, the activation barriers for the forward and back reactions vary with the applied electrode potential, and this concept is developed in the formulation of the Butler-Volmer equation [142] i = i 0 {([R]o/[R]buik) exp[(l-a)Frl/RT] - ([0]o/[0]bu lk) exp(-aFn/RT)}. (2.19) In this equation, i 0 is exchange current density, which corresponds to the current density in each direction at equilibrium (i.e. when i c = ia); a is the transfer coefficient which lies in the range 0 to 1 and depends on how the applied potential modifies the energy of the transition state; and n is the overpotential, characterized as the difference between the applied and equilibrium potentials (E - Eeq). The Butler-Volmer treatment distinguishes the reactant concentrations at the electrode surface ([RJ and [00]) from those in the bulk ([Rbuik] and [Obuik])- However, when the solution is well stirred, the concentrations of the reactants at the electrode surface will remain equal to their concentrations in the bulk (i.e. [R]0 = [R]buik and [0]0 = [0]^) , and Eq. (2.19) simplifies to i = i0{exp[(l-a)Fn/RT] - exp[-aFn/RT}. (2.20) For applied potentials large enough to drive just the oxidative process, the corresponding reductive component becomes negligibly small and Eq. (2.20) effectively becomes tai = l n i 0 + (l-a)Fn/RT. (2.21) Equally for applied negative potentials, where the reductive process dominates, Eq. (2.20) effectively becomes ln i = lni 0-aFr]/RT. (2.22) Equations (2.21) and (2.22) represent the well-known Tafel equations, and they define the Tafel regions in polarization curves, such as that in Fig. 2.17, which is presented as a plot of measured current density versus applied potential. The exchange current density i 0, can be estimated from 62 a measured polarization curve by extrapolating these linear regions to the common intercept [16,140-142]. The situation described is for a simple half-cell process, but in an actual corroding system there will in general be two or more oxidation-reduction processes occurring simultaneously. For example, when Zn is corroding in acid, the anodic reaction Zn -+ Zn 2 + + 2e~ (2.23) is accompanied by the cathodic process, 2H + + 2e~-^H 2 (2.24) and their polarization curves become interconnected. This is shown in Fig. 2.18 [16], where the equilibrium potential for the two half-cell processes is marked as Ecorr, and the corresponding exchange current density is iCOn-Measurements of electrochemical polarization curves provide an easy and direct way of comparing trends in behavior. In this thesis, polarization curves are used to follow trends in measured icorr and Ecorr as different coating treatments are applied to 2024-A1 alloy surfaces. This follows the principle that the smaller the value of icorr, the lower is the corrosion rate, and hence the better the protective influence. However, it must be noted that these comparisons are made without necessarily identifying the actual chemical processes involved. 2.5.2 Instrumentation A schematic representation for the electrochemical polarization circuit is shown in Fig. 2.19. Three electrodes are immersed in 3.5% NaCI solution that is purged with N 2 to remove oxygen and stirred to ensure fast transport of reacting species. The potential is measured between the working electrode, which is the investigated metal sample, and the reference saturated calomel electrode (SCE), while current flows between the working electrode and a counter Pt electrode. A Haber-Luggin capillary is used to connect the reference electrode and 63 Figure 2.17 Schematic d iagram o f po lar iza t ion curve for O + e *-»• R showing anodic and cathodic Tafe l regions. 64 CURRENT DENSITY (A/crr>2) ;ure2.18 Polarization curves for zinc in acid solution [16] showing two separate electron transfer processes. 65 u constant Figure 2.19 Schematic d iagram o f po la r iza t ion circui t : A identifies w o r k i n g electrode; B reference electrode; G counter electrode; H L H a b e r - L u g g i n capi l la ry ; M V ampli f ier ; U voltmeter and I ammeter. 66 the solution close to the working electrode in order to substantially reduce any potential drop due to solution resistance. The potentiostat linearly steps the potential between the working electrode and the reference electrode, and the associated current is recorded for each setting. The polarization measurements are made with a Solartron 1286 Electrochemical Interface (Schlumberger Technologies), and Solartron Corrware and Corrview software are used for data acquisition and for plotting polarization curves. In this work, the corrosion current density i^rr is estimated from the intercept of the extrapolated cathodic plot with the observed corrosion potential. 67 Chapter 3 Adsorption of BTSE and y-APS Organosilanes on Different Microstructural Regions of 2024-T3 Aluminum Alloy 3.1 Introduction Recent developments reported i n Chapter 1, suggest that silane pre-treatments can help protect metals against corros ion, and investigations are be ing directed at deve lop ing this approach as an alternative to the currently-used chromat ing and phosphat ing processes [39-42]. In the m o v e to develop pract ical recipes for cor ros ion inh ib i t i on w i t h silane coatings, there has been an emphasis o n s tudying metal surfaces that appear to approach chemica l uni formi ty . Nevertheless , commerc i a l a l loys inevi tab ly conta in microstractural heterogeneities. F o r example , the second-phase particles i n an a l u m i n u m a l loy m a y encourage a desirable m i x o f mechanica l properties [2], but they also make the mater ial more susceptible to loca l i zed corros ion. Therefore, to help design n e w approaches to the use o f organosilanes for the cor ros ion protect ion o f a l loys , it becomes par t icular ly important to have basic knowledge for the w a y that the coat ing adsorbs o n the different regions o f an a l l oy surface. T o date knowledge i n this area appears re la t ively sparse. The w o r k presented i n this chapter uses the energy dispersive X - r a y ( E D X ) spectroscopy, scanning electron mic roscopy ( S E M ) and X - r a y photoelectron spectroscopy ( X P S ) techniques to study the adsorption o f organosilanes o n the different microstructural regions o f 2024-T3 a l u m i n u m a l loy . In part, the a i m is to explore h o w the silane adsorption varies over loca l regions, such as second-phase particles w i t h their dif fer ing chemica l composi t ions , compared w i t h the m a i n a l l oy matr ix . The organosilanes bis- l ,2-( t r ie thoxysi ly l )e thane ( B T S E ) and y -aminopropyl t r ie thoxysi lane ( y - A P S ) are used i n this w o r k because they have contrasting characteristics for interacting w i t h a metal . F o r example , B T S E has a b is s i l y l funct ion that m a y 68 help op t imize the direct M - O - S i interatomic bond ing and so contribute to an enhancement o f cor ros ion stabil i ty. B y contrast, y - A P S has a mono s i l y l funct ion p lus a terminat ing amino group, and bond ing to a metal may occur from either end. W h e n this occurs through the amino group, v i a hydrogen bonding , the cor ros ion protect ion effect appears to be less [143]. 3.2 E x p e r i m e n t a l The 2024-T3 a l u m i n u m a l loy samples (1 c m ) used i n this w o r k were mechan ica l ly po l i shed by d iamond paste to about 1 p m surface roughness and subsequently cleaned i n acetone and methanol solvents ( H P L C grade) for 10 m i n each. The samples used i n Sections 3.3.1 and 3.3.3 were then air-dr ied (5 min ) before d ipp ing into a freshly h y d r o l y z e d silane solu t ion for 2 m i n at r o o m temperature. Af te r r emov ing from the coat ing solut ion, the excess l i q u i d was dr ied o f f by compressed air. The samples used i n Sec t ion 3.3.2 were treated s imi l a r ly except each freshly po l i shed and dr ied sample was left i n air for 20 h pr ior to d ipp ing i n the coat ing solut ion. The silane solutions (4% concentration) were prepared b y m i x i n g a si lane-water-methanol mixture i n the respective v o l u m e ratios 4-6-90, and such solutions were hyd r o l yz e d for 7 h . The prepared solutions were used at the natural p H values: 3.9 for B T S E and 11.0 for y - A P S . The S E M and E D X characterizations were made w i t h a H i t a c h i S2300 spectrometer and some electron backscattered diffract ion ( E B S D ) measurements were done w i t h a H i t a c h i S-570 S E M instrument. Spec i f ica l ly , the E D X measurements used a 20 k e V incident electron beam, and elemental amounts were compared o n a weight-percentage basis. F o r example , for S i and A l , s ignals for K a radiat ion at 1.74 and 1.49 k e V respect ively were compared and adjusted for quantitative amounts after m a k i n g standard sensi t ivi ty factor corrections. Cor responding informat ion was obtained for the a l l o y i n g elements C u , M g , M n and Fe as detected w i t h i n the E D X prob ing depth. X P S spectra were measured w i t h a L e y b o l d M A X 2 0 0 spectrometer us ing the M g K a source (1253.6 e V ) operated at 10 k V and 20 m A ; the system pressure was 2 x 10" 9 69 mbar while measurements were being made. Survey spectra were obtained with the pass energy set at 192 eV, while the narrow scan spectra were measured with a 48 eV pass energy. Binding energies were referenced to the Au 4f7/2 peak at 84.0 eV for conducting samples, 3.3 Results and Discussion 3.3.1 B T S E Adsorption on Mechanically-polished 2024-A1 Figure 3.1 shows a SEM micrograph measured from a 2024-Al alloy surface that had been coated with BTSE immediately after polishing. Numerous irregularly-shaped second-phase particles, that appear white in the micrograph, are distributed over the sample surface, while the alloy matrix shows regions with different contrast. Specifically, the matrix regions that surround the second-phase particles have a lighter appearance in the SEM image, whereas the rest of the alloy surface is darker. EDX measurements (examples are shown in Figure 3.2) can help to assess the composition variation between the areas with different contrast. Comparisons are made for three alloy matrix regions (R1-R3) that occur within darker areas in the SEM micrograph and for two lighter areas (R4, R5) that are labeled in Fig. 3.1. Table 3.1 summarizes measured Si/Al ratios for these regions and provides information for the alloying elements. An indication of the composition of the alloy matrix (averaged over several positions) prior to the BTSE treatment is also included. It is clear that the Si detected by EDX from the matrix regions is to be associated with the BTSE adsorption. Indeed, the Si/Al ratios in Table 3.1 are consistent with the darker areas in the SEM micrographs corresponding to regions of higher silane adsorption, whereas the lighter matrix regions have lower Si concentrations. Figure 3.3 shows two types of second-phase particles that are present at the surface of the 2024-Al alloy. Those marked A1-A5 are Al-Cu-Mg particles; they have rounded sides with dimensions in the range 1-10 um. The other type, which occur with lower density, are Al-Fe-Cu-Mn-Si particles (e.g. that marked B) and they are larger (>10 um) with a more irregular shape. 70 Figure 3.1 S E M micrograph measured f rom a 2 0 2 4 - A l a l l oy sample that was coated w i t h B T S E immedia te ly after po l i sh ing (magnif ica t ion 300x) . 71 0 2 4 6 8 k e V 0 2 4 6 8 k e V Kinetic Energy Figure 3.2 E D X spectra measured f rom: (a) dark areas and (b) l ight areas shown i n F i g . Figure 3.3 S E M micrograph showing microstructure for a 2 0 2 4 - A l a l l oy sample that had been coated w i t h B T S E immedia te ly after po l i sh ing (magnif ica t ion 800x) . Second-phase particles o f composi t ions A l - C u - M g ( A 1 - A 5 ) and A l - F e - C u - M n - S i (B ) are apparent, a long w i t h zones o f influence around particles (lighter appearance) and a l l oy matr ix regions, w h i c h are further f rom the particles (darker appearance). 73 Counts 8000 J 6000 . 4000 ; 2000 . Al GU: Counts 4000 . 2000 . I Mn Mn Mn (a) Cu keV (b) 8 keV Kinetic Energy Figure 3.4 EDX spectra measured from: (a) particle A l and (b) particle B (locationl) shown in Fig 3.3. 74 Table 3.1 Elemental ratios measured by EDX from selected matrix regions (see Fig. 3.1) of the polished 2024-Al alloy sample before and after coating with BTSE. M a t r i x Loca t ions S i / A l C u / A l M g / A l M n / A l F e / A l N o coa t ing 1 * 0.055 0.019 0.008 * R l 0.014 0.055 0.019 0.008 * R 2 0.019 0.062 0.017 0.007 * R 3 0.014 0.056 0.019 0.009 * R 4 0.003 0.054 0.020 0.008 * R 5 0.001 0.101 0.018 0.025 0.016 * E lement i n numerator not detected. 1 V a l u e s represent averages from different regions o f a l l o y mat r ix before coat ing w i t h B T S E . The results i n Table 3.1 indicate that after the exposure to B T S E the A l - F e - C u - M n - S i particles are surrounded by zones o f uncoated (or w e a k l y coated) matr ix . Table 3.2 shows E D X measurements taken f rom central areas o f the particles that were used to assess whether B T S E adsorption occurred o n them and sample spectra taken from part icle A l and particle B ( loca t ion 1) are shown i n F igure 3.4. S ince no S i is detectable from the A l - C u - M g particles before coat ing, the detection o f this element after coat ing supports the presence o f a B T S E over layer o n these micro-regions . H o w e v e r , for the A l - F e - C u - M n - S i particles, the measured S i s ignal comes f rom both the particle compos i t i on and the silane f i l m . E D X analysis o f this part icle type i n uncoated samples indicates that the intr insic S i concentration does not exceed 0.40 w t%, but after coat ing the silane coverage, as w e l l as the par t ic le ' s chemica l compos i t ion , were found to va ry from the center o f the part icle ( locat ion 1) to its edge ( locat ion 2 , for part icle B i n F i g . 3.3). L o w e r amounts o f S i , M n and Fe were detected at the edge o f a A l - F e - C u - M n - S i particle, compared w i t h its central area, but the amount o f C u increased at the edge after coat ing. A d d i t i o n a l l y , the S E M micrographs show the presence o f substantial numbers o f microcavi t ies (e.g. locat ions C and D i n F i g . 3.1) w h i c h correspond to enhanced values o f the 7 5 Si/Al ratio (e.g. 0.034 or larger depending on cavity dimensions). These holes probably result from removal of small second-phase particles during the mechanical polishing in the sample preparation. Such sites apparently provide a favorable morphology for the silane adsorption, which appears especially strong in these localized regions. Table 3.2 Compositions measured by EDX from selected second-phase particles (see Fig. 3.3) in the 2024-A1 alloy sample which had been coated by BTSE immediately after polishing. Particles Percentage composition (wt %)1 Si A l Mg Mn Fe Cu A l 0.7 52.4 7.1 0.2 * 39.6 A2 0.9 35.0 10.1 0.2 * 53.8 A3 0.8 60.2 5.5 0.3 * 33.2 A4 0.8 66.3 4.6 0.6 * 27.7 A5 0.9 44.2 8.5 0.3 * 46.1 Biocation 1 0.7 46.6 * 15.6 13.5 23.6 Biocation 2 0.2 53.4 0.7 9.1 8.0 28.6 * Element not detected The evidence presented is consistent with the conclusion that the adsorption of BTSE on the freshly-polished 2024-A1 alloy is non-uniform. The silane coats all matrix regions except zones around the second-phase particles, and it is adsorbed on top of these particles as well. The average increase in the detected wt% for Si after the silane deposition is 1.5 for the coated matrix, 0.8 for Al-Cu-Mg particles and 0.25 for the Al-Fe-Cu-Mn-Si particles. Accordingly it is believed that the amounts of coating follow the order: coated matrix > Al-Cu-Mg particles > Al-Fe-Cu-Mn-Si particles > weakly or uncoated matrix in proximity of the particles. The non-uniform BTSE adsorption on a freshly polished 2024-Al alloy sample is strongly influenced by the microstructure at the alloy surface. The underlying chemistry is dependent on 76 several factors. Firs t , the format ion o f direct M - O - S i bond ing depends o n a condensation react ion between the silane molecu le and O H groups o n the th in o x i d i z e d f i l m at the metal surface [37]. The density and ac id i ty o f the h y d r o x y l groups depend o n a range o f factors i nc lud ing ox ide compos i t i on and structure, nature and l eve l o f impuri t ies , and the amount o f atmospheric exposure [55,144]. Second, it is noted that the p H o f the B T S E solut ion (3.9) is jus t outside the stabil i ty range for AI2O3 [2], and therefore the process for this silane adsorption is expected to start w i t h some etching o f the th in ox ide film present. T h i r d , the galvanic c o u p l i n g between the second-phase part icle and its surrounding matr ix is be l ieved to cause a l oca l p H change that w i l l affect the rate o f ox ide and matr ix etching i n a zone around the particle compared to more distant areas. T a k i n g these points together, it seems clear that variat ions i n the dis t r ibut ion and compos i t ion o f the surface microconst i tuents can have a marked effect o n the silane adsorption. A l u m i n u m - c o p p e r a l loys , such as 2 0 2 4 - T 3 , are susceptible to p i t t ing corros ion . T h i s is p r imar i l y associated w i t h the second-phase particles w i t h the degree o f damage be ing a funct ion o f part icle type as w e l l as their dis t r ibut ion. I f these l o c a l microregions are not w e l l protected b y the coat ing, the re la t ively anodic phases (particles or matr ix areas) w i l l d issolve w h e n exposed to a corrosive environment and provide the onset o f cor ros ion [21,22]. O u r observations indicate that B T S E coverages o n the A l - C u - M g and A l - F e - C u - M n - S i microconsti tuents are different f rom those o n ne ighbor ing mat r ix regions, as w e l l as o n mat r ix regions further away. T w o factors can be especia l ly i n v o l v e d . Firs t , the silane condensat ion react ion m a y occur differently w i t h the hydrated ox ide film o n the second-phase part icle, compared w i t h the si tuation for the a l l oy matr ix . Second, the etching m a y result i n some dea l loy ing for the second-phase particles w i t h further changes i n the l oca l morpho logy . These two factors are discussed further i n the f o l l o w i n g . 77 Studies o f ox ide g rowth o n a l u m i n u m a l loys at h igher temperatures show that the compos i t ion , morpho logy , thickness and cor ros ion suscept ibi l i ty o f the ox ide layer is dependent o n the type and concentrat ion o f a l l o y i n g elements [145-151]. The presence o f C u reduces the ox ida t ion rate o f the a l u m i n u m ox ide , and it m a y also produce a C u - r i c h layer at the oxide-meta l interface dur ing pre-treatments and anod iz ing o f A l - C u a l loys [145-147]; by contrast M g , for example , segregates to, and m a y enr ich at, the ox ide surface [152]. H o w e v e r , these observations relate to the properties o f ox ide f i lms formed over extensive surface areas, and they do not spec i f ica l ly i nvo lve the l oca l variat ions due to the presence o f the microheterogeneit ies, such as second-phase particles. F o r the latter regions, the concentrations o f a l l o y i n g elements can exceed those i n the mat r ix b y an order o f magnitude or more. Cor respond ing ly , the oxides formed at these loca l regions are l i k e l y to differ f rom the mat r ix ox ide w i t h regard to a number o f characteristics, i nc lud ing chemica l compos i t ion , thickness and degree o f hydrat ion. One useful parameter for assessing the ac id i ty o r bas ic i ty o f a surface is the isoelectr ic point ( IEP) , and this corresponds to the p H at w h i c h the surface (or a reg ion o f it) has a net charge o f zero [37,153]. O n this basis, an ox ide reg ion i n aqueous m e d i a w i l l be pos i t ive ly charged at p H values lower than the I E P (i.e net excess o f adsorbed H 4 ) , and negat ively charged at p H values greater than the I E P (i.e. net excess o f adsorbed O K T ) . Var ia t ions over a heterogeneous surface m a y be significant since the I E P values are quite var iable for metal oxides (e.g., for some cations relevant to this d iscuss ion, M g 2 + 12.2, F e 2 + 12.0, C u 2 + 9.1, A l 3 + 9.1-9.5, F e 3 + 8.5 and M n 3 + 4.2) [37,56,155]. W i t h m i x e d oxides , the overa l l I E P value m a y be approximated by a weighted average o f contr ibutions f rom the i nd iv idua l components [156]. Studies o f the adsorption o f propyl t r imethoxysi lane at a fixed p H o n different metal substrates suggest that the l ower the I E P value, the higher the surface h y d r o x y l concentrat ion that can lead to the silane adsorption [154]. F o r this w o r k , w e postulate that the second-phase particles o n a real a l l oy surface fo rm m i x e d oxides [1], and they represent micro-areas w i t h different I E P 7 8 values from the surface of the main alloy matrix. As a consequence, when the alloy is exposed to BTSE solution at pH 3.9, these regions develop different OH concentrations, which in turn lead to local variations in the silane coverage. The expectation is that the mixed oxide which builds up on a Al-Cu-Mg particle will have a higher IEP value than that for A10x(OH)y alone. But, during the BTSE coating, Mg most likely dissolves from the surface and edge of a particle (consequence of the low pH) leaving behind the Al and Cu oxides, which have similar IEP values. The preferential etching of Mg is expected to increase the surface porosity as well. Both these factors are likely to influence the BTSE bonding with the result that the coverage on the Al-Cu-Mg particles is similar to that on the coated matrix areas. This mechanism is consistent with recently-reported observations related to the preferential dissolution of Mg and Al from the edge of Al-Cu-Mg particles in a corrosive environment [110,158,159]. The situation is more complex for the mixed oxide that exists above Al-Fe-Cu-Mn-Si particles, but the A l is expected to dissolve preferentially, followed by Mn and Fe (at the edge) [110]. As the surface becomes richer in Fe and Cu , the IEP value is expected to increase with a lowering in the surface OH concentration. In turn, this can result in the decreased BTSE adsorption. 3.3.2 BTSE Adsorption on Oxidized 2024-A1 The adsorption process studied in the previous section was made directly after the mechanical polish, a process that suggests the silane exposure occurs on a thin oxide film. That opens the question of whether the distribution of adsorbed BTSE, over microregions of a 2024-Al alloy surface, is changed when the initial surface is modified with regard to the amount of oxide (or hydroxide) present. This point is examined by applying the BTSE treatment to a sample that was first mechanically polished (exactly as in Section 3.3.1) and then oxidized in air (room temperature) for 20 h. Figure 3.5 compares SEM micrographs from corresponding 79 regions o f such a sample before and after the B T S E adsorption. The latter case is marked ly different from the situation reported i n Sect ion 3.3.1, where the silane adsorption appears especia l ly inf luenced b y the p r o x i m i t y o f second-phase particles. T h e surface formed by silane o n the o x i d i z e d sample ( F i g . 3.5(b)) also shows l ighter and darker regions, but the scale o f this structure (~100 um) is considerably greater than that app ly ing after the si lane deposi t ion o n a freshly po l i shed sample ( F i g . 3.1). Tab le 3.3 reports E D X results from different locat ions w i t h i n the separate regions seen i n F i g . 3.5(b); the results relate to the observations i n Sec t ion 3.3.1 insofar as the darker areas i n the mic rograph correspond to more silane adsorption. Table 3.3 Elemental ratios measured by EDX from selected matrix regions (see Fig. 3.5(b)) of the oxidized 2024-Al alloy sample before and after coating with BTSE. M a t r i x Loca t ions S i / A l C u / A l M g / A l N o coa t ing 1 * 0.050 0.018 M l 0.064 0.018 M 2 0.002 0.056 0.018 M 3 * 0.063 0.018 M 4 0.005 0.060 0.018 M 5 0.011 0.061 0.017 M 6 0.122 0.068 0.015 M 7 0.154 0.070 0.015 M 8 0.287 0.072 0.011 * E lement i n numerator not detected. 1 V a l u e s represent averages from different regions o f a l l oy matr ix before coat ing w i t h B T S E . The larger value o f the S i / A l ratio for the darker regions i n F i g . 3.5(b) appears to be accompanied by a sl ight increase i n the C u / A l ratio, as w e l l as by a sma l l decrease i n the M g / A l ratio. T h i s suggests the poss ib i l i ty that etching m a y result i n some M g O d i s so lv ing from these regions, w h i l e the C u content is enhanced because o f that element 's segregation. F igure 3.6 80 Figure 3.5 S E M micrographs measured for (a) o x i d i z e d 2024-A1 a l l o y sample before and (b) after B T S E coat ing (magnif icat ion 200x) . Cor responding areas are s h o w n i n each case, a l though the image for (b) is rotated b y 1 8 0 ° w i t h respect to that i n (a). 81 shows S E M images i n the v i c i n i t y o f selected second-phase particles before and after the B T S E deposi t ion o n the o x i d i z e d sample, and Table 3.4 reports some corresponding E D X measurements. Table 3.4 Elemental ratios measured by EDX from selected second-phase particles (see Fig. 3.6) of the oxidized 2024-Al alloy sample before and after coating with BTSE. Part icles S i / A l C u / A l M n / A l F e / A l P I before 0.008 0.642 0.274 0.243 after 0.414 0.723 0.324 0.288 P 2 before 0.008 0.785 0.225 0.235 after 0.025 0.677 0.253 0.267 P3 before 0.009 0.552 0.314 0.292 after 0.009 0.586 0.336 0.310 The B T S E coverage o n the o x i d i z e d sample appears more diverse than o n the po l i shed sample. Whether a particle is coated or not for the o x i d i z e d sample depends especia l ly o n its loca t ion i n relat ion to the regions associated w i t h a higher or l ower silane f i l m thickness. F o r example , the A l - F e - C u - M n - S i particles P I ( F i g . 3.6(a) and (b)) and P 2 ( F i g . 3.6(c) and (d)) have s imi la r chemica l composi t ions , but the first is located i n a broad reg ion o f th ick silane deposit, w h i l e the second is i n a reg ion o f moderate si lane coverage. The regions w i t h thinner si lane coatings reveal the presence o f zones o f influence around the particles, m u c h l i ke the si tuation reported i n Sec t ion 3.3.1, and indeed the a l l oy mat r ix around part icle P 2 is o n l y w e a k l y coated at most. Par t ic le P3 ( F i g . 3.6(e) and (f)), w h i c h is pos i t ioned i n a generally uncoated area, is also uncovered, except for its edge region. The near-by matr ix shows an increased roughness after the silane exposure, and this is consistent w i t h etching occur r ing dur ing the silane exposure. T h e elemental composi t ions measured before and after the B T S E deposi t ion (Table 3.4) indicate increased M n / A l and F e / A l values after the coat ing, and that is consistent w i t h the preferential 82 Figure 3.6 S E M micrographs showing different A l - F e - C u - M n - S i particles (P1-P3) o n an o x i d i z e d 2024-A1 a l loy sample. The images o n the left-hand side apply to before exposure to B T S E , w h i l e the corresponding images o n the r ight-hand side apply to after the B T S E exposure. 83 dissolution of Al noted during the BTSE exposure described in Section 3.3.1. The Cu/Al ratio also increases on PI and P3 after the coating treatment, but not on P2, where some Cu-rich clusters may have been etched away [158]. Similar observations were made for the Al-Cu-Mg particles insofar as the silane coverage on them is dependent on their location within the broad regions of high or low coating. The dominant change in elemental composition for Al-Cu-Mg particles in low-coated regions is in the Mg/Al ratio, which decreases from 0.206 to 0.117 after the silane exposure (while other ratios remain unchanged). This supports the concept of Mg dissolution from the particle surface, as previously noted in Section 3.3.1. An XPS measurement has been taken from the sample prepared under similar conditions as the one shown in Figure 3.5(b). The relative atomic compositions obtained from a survey scan (Figure 3.7(a)) are found to be: O 46.7%, C 24.3%, Si 17.0% and Al 12.0%. The size of the A l signal suggests that the BTSE coverage is incomplete. In addition, the Al 2p narrow-scan spectrum shows two sub-components at binding energies of 71.5 eV and 74.6 eV that correspond to Al° and A l 3 + respectively (Figure 3.7(b)). The presence of a substantial metallic component suggests that the thickness of the oxide layer on top of regions that are not covered by BTSE must be considerably less than 100 A. The deposition of BTSE on the oxidized alloy sample shown in Fig. 3.5(b) corresponds to a high degree of non-uniformity, and the increased amount of oxide following the air exposure apparently allows more silane adsorption in certain regions. However, it is clear that the second-phase particles (with their associated zones of influence) now have no more than a second-order effect on the adsorption process. Figure 3.8 (a) shows the 2024-Al alloy grain structure revealed after etching in HNO3 and H3PO4. Apparently, the grain size does not exceed the dimensions 20 x 50 pm2, and that is confirmed by application of the electron backscattering diffraction (EBSD) technique to a mirror polished sample (Fig. 3.8(b)). An EBSD image corresponds to backscattered electrons that have been diffracted from different oriented grains, and in favorable 84 O Is (a) (Arb. Units) O ( A ) tensity C ( A ) C Is c i 1 i i , i i 1 1 S i 2 , ? i 2 p 1 1 | A 1 2 p 1 7**^ 1000 800 600 400 200 0 79 77 75 73 71 69 Binding Energy (eV) Figure 3.7 X P S spectra measured f rom the o x i d i z e d 2 0 2 4 - A l sample after coat ing w i t h B T S E : (a) survey scan and (b) A l 2p h i g h resolut ion spectrum. 85 cases the orientation can be indicated according to the recognized Kikuch i patterns [212]. That indexing is generally difficult for A l alloys but a partial identification is shown in Fig. 3.8(b)). A n important point for this work is that the grain dimensions are smaller than the regions covered by a thick coating of silane. The interpretation by E B S D indicates that some o f the underlying surface is textured, and some is deformed. The first is expected for alloys that have been cold-worked, and the second region could be due to some re-crystallization or growth of the oxide f i lm that prevents indexing o f grains [212]. The textured structure may perhaps influence the oxidation process, and in turn the silane adsorption. Indeed, there has been a recent report that the anodizing behavior of A l - C u alloy samples varies with grain orientation and Cu enrichment at the surface [148]. E D X in our work suggests that the silane-rich regions occur over areas with higher C u and lower M g contents. Such influences may contribute to the existence o f a thinner, more-compact oxide layer with different nature, and that could certainly affect the chemistry o f the silane deposition. But other factors can be involved. For example, the B T S E deposition is likely to depend on a competition between adsorption and etching, and the latter may be more prominent i f a higher B T S E solution concentration is applied. Further, the lateral distribution of silane over a surface may in general be influenced by the chain polymerization process, although that would not necessarily optimize the M - O - S i interfacial bonding required for corrosion protection. Finally, Quinton et al. have reported a dynamic adsorption-desorption mechanism for propyltrimethoxysilane ( P T M S ) on A l substrates [160,161], and a somewhat similar process could operate in the complex situation reported here. 86 Figure 3.8 G r a i n structure for the mechanica l ly -pol i shed 2 0 2 4 - A l a l l oy sample: (a) S E M image after etching i n HNO3 and H3PO4; (b) E B S D image w i t h gra in index ing , co lor coded accord ing to the plane (hkl) para l le l to the r o l l i n g plane and direct ion [uvw] paral le l to the ro l l i ng direct ion: p i n k (110) [100], red-orange (112) [111], y e l l o w (100) [100], blue (110) [211], v io le t (123) [634]; the green areas cannot be indexed and are referred to i n the text as deformed. 87 X= Measured matrix locations Figure 3.9 S E M micrographs measured from a 2 0 2 4 - A l a l loy sample w h i c h had been coated w i t h y - A P S immedia te ly after po l i sh ing : (a) larger-area v i e w (magnif ica t ion 400x) and (b) higher magnif ica t ion ( l , 0 0 0 x ) v i e w o f another region w h i c h shows second-phase particles o f A l - F e - C u - M n - S i (S4 and S5) and A l - C u - M g (T I and T2) types. 88 3.3.3 y-APS Adsorption Figure 3.9 shows two S E M micrographs from a freshly po l i shed 2 0 2 4 - A l a l loy sample that had been treated by y - A P S . That at l o w magni f ica t ion ( F i g . 3.9(a)) shows that the coated surface exhibi ts a un i fo rm contrast, except for the second-phase particles. E D X measurements at selected matr ix locations indicate that the S i / A l ratio is effect ively constant at 0.025, and it is therefore conc luded that the silane film has a re la t ively even thickness over a large area o f the surface. It is be l ieved that the y - A P S solut ion, w i t h p H o f 11.0, etches the surface too severely for zones o f influence to become established around the second-phase particles. The S E M images (especial ly that at h igher magni f ica t ion i n F i g . 3.9(b)) also show that more micro-holes are created by the y - A P S treatment, compared w i t h us ing B T S E , and this is consistent w i t h more particles d i s so lv ing i n the former case (correspondingly there is a smaller number o f A l - C u - M g particles avai lable for E D X measurement). Table 3.5 Compositions measured by E D X from selected second-phase particles (see Fig. 3.9) and from the alloy matrix in a 2024-A1 sample which had been coated by y-APS immediately after polishing. Part icles Percentage compos i t i on (wt %) S i A l M g M n Fe C u S I 2.3 50.0 0.2 20.6 8.5 18.4 S2 2.0 45.7 0.2 15.9 10.6 25.6 S3 2.3 44.2 0.2 13.6 13.8 25.9 S4 2.1 43.0 0.1 10.0 10.1 34.7 S5 2.3 47.8 0.4 7.9 8.5 33.1 T l 3.9 64.2 3.2 0.6 * 28.1 T 2 3.8 53.8 4.3 * * 38.1 M a t r i x 1 2.3 90.2 1.6 0.7 * 5.2 * E lement not detected 1 V a l u e s represent averages from different regions o f a l l oy mat r ix after coat ing. 89 Table 3.5 reports elemental composi t ions measured for the A l - F e - C u - M n - S i (S1-S5) and A l - C u - M g (T1-T2) particles. There is an increase o f about 1.8 w t % for S i o n the A l - F e - C u - M n -S i particles, and the evidence supports a relat ively constant thickness for the silane f i l m over the a l loy matr ix and the Fe-conta in ing particles. The y - A P S adsorption seems general ly less inf luenced (than for B T S E ) by the d ivers i ty o f I E P values o n the a l l oy surface, a l though the E D X signal for S i is c lear ly increased at the M g - c o n t a i n i n g particles. The latter appear par t icular ly favorable for silane adsorption. In the compar i son w i t h B T S E , it must be noted that y - A P S introduces a n e w feature: the latter can adsorb o n metal ox ide surfaces v i a either end, w i t h the orientation appearing especia l ly dependent o n solu t ion p H [37,52]. F o r example , i n ac id ic solut ion, the amino groups are protonated, and the NH3 + can bond to surface O H groups b y hydrogen bond ing ; this dr ives the so-cal led ' ups ide -down ' orientation [52,53]. O n a l u m i n u m , w i t h a p H o f 10.5, it is reported that the y - A P S orientation is m i x e d [53], and recent observations emphasize the dynamic nature o f its adsorption o n ox ide surfaces [162]. W i t h some hydrogen bond ing to the ox ide surface, the condensation react ion m a y occur more w i t h i n the silane layer; that m a y be a major contr ibut ing factor for observations i n this w o r k that a re la t ively even f i l m o f y - A P S is formed over a large surface area. 3.4 Concluding Remarks The situations covered i n this exploratory w o r k show contrasting behaviors. The thickness and coverage o f the B T S E f i l m formed o n different microstructural regions o f a mechanica l ly -po l i shed 2024-A1 , was found to be strongly affected by the dis t r ibut ion o f second-phase particles o n the a l l o y surface. Spec i f i ca l ly , a l though the adsorpt ion occurs o n the a l l o y matr ix away f rom the particles, and o n the particles themselves, the regions immedia te ly surrounding the part icles had less B T S E . S u c h a pattern i s be l ieved to be inf luenced by the 90 galvanic coupling between particles and matrix that affected the rate of the oxide and matrix etching around particles, and therefore the subsequent BTSE deposition. These differences in the BTSE adsorption on particles may occur because the nature of the oxide layer varies according to the variations in compositions of alloying elements in the different microregions. The amounts of BTSE adsorption on an air-oxidized sample is increased on average, although the coating is non-uniform with regions (-100 pm) of higher and lower coverages. The distribution of second-phase particles does not have a primary influence on silane coating in this case, but more important are variations in thickness and other characteristics of the oxide film. It is also believed that a competition between the silane deposition and the etching of the underlying oxide film at different areas still contributes to the overall coating coverage. The adsorption of y-APS is different insofar as films of relatively even thickness form over large areas of the polished sample; hydrogen bonding through the amino groups probably helps the distribution in this case. 91 Chapter 4 Comparison of a Chromic Acid and a BTSE Final Rinse Applied to Phosphated 2024-T3 Aluminum Alloy 4.1 Introduction In the metal finishing industry, the treatment of a metal surface prior to painting generally includes several stages such as cleaning, formation of a conversion coating and a final rinse [75-79,163]. Dilute chromic acid, or a mixture of Cr(VI) and Cr(III) solutions, have been used traditionally for the final rinse [75-79,163], but these approaches are less desirable because of concerns about the carcinogenic natures of the solutions, and their consequent safe handling and disposal. Attempts are being made to develop alternative rinse methods [164-166], but at the industrial level most of the chrome-free approaches are still inferior to the chromate post-treatment. Nevertheless, some investigations have been reported where new rinse methods can achieve comparable corrosion protection to that provided by chromate [167,168]. For example, particularly good results have been indicated by rinsing a phosphated cold-rolled steel with a solution that contains organosilanes in combination with group IV metal ions [168], and electrochemical impedance spectroscopy (EIS) has been used to study a methyltrimethoxy silane (MS)/Zr rinse applied to a zinc phosphate coating on cold-rolled and hot-dipped galvanized steel [169]. The application of organosilanes on metals was shown to improve their corrosion resistance properties [39-41], and among those studied, the non-functional silane bis-1,2-(triethoxysilyl)ethane (BTSE) appears notable. The work described in this chapter has the objective of characterizing, and comparing, the natures of zinc phosphate coatings on 2024-T3 aluminum alloy after final rinses using either BTSE or dilute chromic acid. This provides part of an attempt to understand the mechanisms involved. The characterization techniques used 92 emphasize polarization curve measurements and immersion tests for assessing susceptibility toward corrosion of the treated samples, while X-ray photoelectron spectroscopy (XPS) was used to obtain general information on surface chemical composition. The morphology of surface coatings was imaged using scanning electron microscopy (SEM), but electron dispersive X-ray (EDX) spectroscopy was also applied to assess chemical composition at local micro-regions. 4.2 Experimental All samples of 2024-T3 aluminum alloy ( lx l cm2) were polished with A1 2 0 3 sandpaper (up to 1200 grit), and a few were subsequently polished by a diamond suspension to give a mirror polish (1 pm roughness). After polishing, all samples were ultrasonically cleaned in acetone and methanol, and after a short air-drying they were immersed in Gardobond R2600 (Oakite) phosphating solution (pH 3.8-4.0) for 3 min at 50°C. Each phosphate coated sample was first rinsed in distilled water (15 s) and that was followed by a post-treatment. Two kinds of post-treatments were used: (a) Cr03 solution (concentrations 0.1, 0.5 or 1 g/L) for 30 s at 25°C or 70°C; (b) BTSE solution (1, 4 or 10 vol% of BTSE, with 6% water and methanol to bring to 100%) for 2 min at room temperature. The BTSE solutions were allowed to hydrolyze for 5 h before use. The concentration range for the CrG*3 rinsing solutions, as well as the specific temperatures, were chosen according to reports for the sealing of phosphate conversion coatings on iron and steel [75]; solutions with chromic acid concentrations higher than 1 g/L were not used since they are known to lead to paint blistering and poor adhesion [77,78]. For the coated samples that had been sealed in chromic acid, the excess liquid after post-treatment was blown out using compressed air and that was followed by drying at 150°C for 2 min. The BTSE-rinsed samples were dried by compressed air. XPS spectra were measured with a Leybold MAX200 spectrometer using the Mg K a source (1253.6 eV) operated at 10 kV and 20 mA; the system pressure was 2 x 10 - 9 mbar while 93 measurements were being made. Survey spectra were obtained with the pass energy set at 192 eV, while the narrow scan spectra were measured with a 48 eV pass energy. Binding energies were referenced to the A u Ain a peak at 84.0 eV for conducting samples, but for those coated samples where the metallic component could not be detected in the A l 2p spectrum, the C Is peak for adventitious hydrocarbon contamination was set at 285.0 eV as the alternative reference. Curve fittings used mixed Gaussian-Lorentzian functions, and relative compositions were estimated after background correction using sensitivity factors provided by the manufacturer. S E M micrographs from samples that received just the sandpaper polish were obtained by using a Hitachi S-4100 microscope with field emission source operated at 15 kV. For the mirror-polished samples, both before and after coating, S E M images and E D X spectra were measured by a Hitachi S-3000N variable pressure microscope (operated at 20 kV) with Quartz X O N E E D X system. Electrochemical corrosion measurements were conducted in 3.5% NaCI solution (after N 2 purging) at 25°C using a Solartron 1286 potentiostat with Corrware and Corrview software analysis. A conventional three-electrode cell was used over the potential range -1.5 to -0.5 V vs. SCE (scan rate 1 mVs - 1 ) . Immersion tests in 3.5% NaCI solution were applied to alloy samples that first received the mirror polish and were subsequently phosphated and post-treated. The morphology and composition of the coatings obtained were examined before and after the immersion test. 4.3 Results and Discussion 4.3.1 Chromic Acid Post-rinse Figure 4.1(a) shows measured polarization curves for zinc phosphated 2024-Al alloy post-rinsed in Cr03 solutions (concentration range 0.1 to 1 g/L) and Fig. 4.1(b) shows the effect for coatings sealed in 0.5 g/L CrC>3 solution on changing temperature from 25°C to 70°C. The icorr values from Tafel analyses provide convenient measures of the comparative corrosion 94 -o.50r -0.75 > w -l.ooi •1.25 (a) _ 1 ^Ql i i i 11 ml I I i i i ml i i i 111 ul i i i I I ml , i i i 111 ul i i i i 11 ul 10~8 IO - 7 IO"6 IO"5 IO - 4 IO"3 IO - 2 I (A/cm 2) Figure 4.1 Polarization curves measured from phosphated 2024-Al samples (a): (1) no post-treatment, (2) post-treated in 0.1 g/L Cr0 3 , (3) 0.5 g/L Cr0 3 and (4) 1.0 g/L Cr0 3 all at 25°C; (b) post-treated in 0.5 g/L Cr0 3 at 25°C and at 70°C. 95 stabilities for samples treated in different ways; specifically, a smaller icon- value indicates harder electron transfer and hence better corrosion protection [141]. For the phosphated alloy, without a sealing rinse, iCOn- was determined to equal 8.0 x 10"6 A/cm 2, but, after sealing with 0.5 g/L C1O3 solution at 25°C, its value reduced to 3.8 x 10"6 A/cm2. The value of icorr remained constant with increase in the C1O3 concentration, when the rinse was applied at 25°C, although icon- reduced to 7.2 x 10"7 A/cm 2 when the phosphated sample was rinsed in the 0.5 g/L Cr0 3 solution at 70°C. For that condition, the final rinse has lowered the icon- value by an order of magnitude compared with the unsealed coating. Observations for the anodic regions in Fig. 4.1(a,b) indicate decreases in the anodic current, for the potential range -1.02 V to -0.69 V, as the concentration of C1O3 solution is increased from 0.1 to 1 g/L, and when the temperature for the post-rinse is raised to 70°C. It is also noticed that Ecorr shifts to more cathodic values with increasing concentration and temperature of the C1O3 solution. This is believed to be associated with more effective barrier properties of the coating and therefore a decreased dissolution of the metal substrate. These observations support the view that the chromic acid post-treatment improves the corrosion resistance properties for phosphated 2024-A1 alloy samples. Further analysis in this work restricts to samples post-treated using the C1O3 solution of 0.5 g/L concentration. A low-magnification S E M micrograph for the phosphated alloy sample is shown in Fig. 4.2(a), and corresponding images after post-treatment in 0.5 g/L Cr03 solution at the two different temperatures are shown in Fig. 4.2(c,e). Crystalline zinc phosphate particles appear to be uniformly and densely distributed across the alloy surface, and at this magnification it seems that the coating morphology is unaffected by the post-treatment process. However, micrographs taken at higher magnification (Fig. 4.2(b,d,f)) show cracks in the coating after application of the post-treatment, and this is interpreted to indicate that some coating has dissolved during the final rinse. 96 Figure 4.2 S E M micrographs o f phosphated 2024-A1 samples: (a) no post-treatment rinse (x5,000); (b) no rinse (x50,000); (c) r inse i n 0.5 g / L C r 0 3 at 2 5 ° C (x5,000); (d) rinse i n 0.5 g / L C r 0 3 at 2 5 ° C (x50,000); (e) rinse i n 0.5 g / L C r 0 3 at 7 0 ° C (x5,000) , and (f) rinse i n 0.5 g / L C r 0 3 at 7 0 ° C (x50,000) . 97 Figure 4.3 (a) X P S survey spectra measured from phosphated 2024-A1 samples (1) before and (2) after post-treatment i n 0.5 g / L C r 0 3 so lu t ion at 7 0 ° C ; (b) corresponding spectra for a l ower b ind ing energy range. 98 XPS survey spectra were taken to assess the chemical composition of these coatings. For the phosphated sample with no rinse applied (Fig. 4.3(a), spectrum 1), the Zn 2p3/2 peak at 1022.9 eV binding energy, the P 2p signal at 134.2 eV, and O Is at 531.6 eV, are consistent with the presence of a zinc phosphate coating on the surface [170], although in addition there is a C Is peak at 285.0 eV from air borne contamination. After a chromic acid rinse (0.5 g/L C1O3 at 70°C), the sample shows the Cr 2p doublet structure at 578.5 eV (2p l /2) and 577.8 eV (2p3/2), but this post-rinse results in a reduction in the Zn and P signals (Fig. 4.3(a), spectrum 2). This is interpreted by some dissolution of the zinc phosphate coating [171], but the formation of a Cr-containing overlayer will also attenuate signals arising from deeper into the sample. Table 4.1 reports surface elemental composition information according to XPS spectra measured from samples before and after the post-rinse at different temperatures. Table 4.1. Relative atomic compositions (mole%) from XPS for phosphated 2024-A1 samples before and after a post-treatment rinse in chromic acid (0.5 g/L C1O3) at different temperatures. No rinse Zn Cu Cr O C P Al 14.9 0.0 0.0 58.4 18.2 8.5 0.0 Rinse (25°C) 7.3 0.0 5.4 57.3 24.1 5.9 0.0 Rinse (70°C) 3.5 0.0 4.4 57.9 24.4 4.1 5.7 Rinse (70°C) after immersion test 1.9 1.1 3.7 54.3 27.1 3.7 8.2 The A l signal is only detected for the sample that is post-treated in chromic acid solution at 70°C (Fig. 4.3(b), spectrum 2), so confirming that the thickness of the zinc phosphate coating has been significantly reduced by the higher-temperature rinse. The A l 2s signal at 119.2 eV binding energy is used for the composition estimates since there is an appreciable overlap between Cr 3 s and the A l 2p oxide component [172]. However, a higher-resolution A l 2p 99 spectrum did not show the metallic component at 72.6 eV, and that may indicate further oxidation of aluminum, for example where the surface etches more easily and/or the coating is less-well packed. This is consistent with the mechanism proposed for the reduction of CrtVI) on aluminum during the formation of chromium conversion coatings [28,173]. The Cr 2p3/2 spectra from samples sealed by the chromic acid solution (0.5 g/L Cr0 3) at 25°C and 70°C (Fig. 4.4(a) and (b)) are consistent with these surfaces involving trivalent chromium (i.e. Cr(III)), but with a smaller amount of hexavalent chromium. The Cr(III) part is interpreted to involve a Cr 2 0 3 component at 576.9 eV binding energy and a hydrated oxide CrOOH at 578.1 eV; the Cr(VI) component appears at 579.7 eV. This follows the interpretation from a chromate post-treatment applied to galvanized steel after phosphate coating [170]. For our 2024-A1 alloy sample, the ratio Cr(VI)/Cr(III) changes from 0.096 for the coating rinsed at room temperature to 0.295 for the coating post-treated at 70°C, and this is fully consistent with the coating sealed at the higher temperature having a higher amount of Cr(VI) at its surface. The CrOOH/Cr203 ratio equals 1.10 for the coating sealed at 25°C, and it remains essentially constant after rinsing at higher temperature. It is believed that different Cr species have differing roles in providing corrosion protection. Thus the CrOOH is thought to help improve paint adhesion, while the Cr(VI) component may be involved in a "self-healing" effect [170,174] which enables damaged areas to be repassivated during exposure to a corrosive environment [26,175]. In order to investigate the effect of a corrosive environment, a phosphated sample post-rinsed in 0.5 g/L Cr0 3 solution at 70°C was immersed in 3.5% NaCI solution for 12 h, and the surface composition measured (Table 4.1). XPS signals from Zn, Cr and P decreased, while those from A l increased, but in addition a small amount of Cu was detected. The 2024-A1 alloy contains ~4% Cu, but its presence is often not visible by XPS from surfaces of mechanically polished or phosphated samples when the outer oxide or coating is sufficiently thick. Various 100 583 580 577 574 Binding Energy (eV) Figure 4.4 Cr 2pm spectra measured from phosphated 2024-A1 samples post-treated in: (a) 0.5 g/L Cr0 3 at 25°C; (b) same solution at 70°C; and (c) as (b) but after immersion test. 101 pre-treatments of Al-Cu alloys lead to Cu enrichment at the surface [147], and that may also occur at the alloy-coating interface during the initial stage of phosphating. Then, as coating degradation occurs during the immersion in NaCI solution, XPS detection of Cu becomes easier. The Cr 2p3/2 spectrum (Fig. 4.4(c)) shows a much reduced concentration of the Cr(VI) component, compared to the situation prior to the immersion test; the Cr(VI)/Cr(III) ratio reduces to 0.04, while that for CrOOH/Cr 20 3 drops to 0.80. Inevitably, the nature of the Cr layer is affected strongly by exposure to a corrosive environment. The results presented indicate that an improved corrosion resistance is obtained for the zinc phosphate coatings on 2024-A1 alloy which have been post-treated with dilute chromic acid solution. The thickness of the phosphate coating reduces after the final rinse, especially when carried out at higher temperature, but the coating protection is still improved. The large reduction potential of Cr(VI) helps convert the phosphated surface into a more passive state, especially at porous regions in the coating (e.g. at intercrystalline zones where the initial protection of the metal is least). 4.3.2 B T S E Pos t - t r ea tmen t Figure 4.5 shows measured polarization curves for phosphated 2024-A1 alloy after post-treatment with BTSE solution (1, 4 and 10%). The i c o r r values decrease from 1.7 x 10"6 A/cm2 7 7 for the sample rinsed with 1% BTSE rinsed to 1.6 x 10" A/cm for the post-treatment with the 10% BTSE solution. In addition, the i C O rr value obtained from the sample with the 4% pre-treatment (4.0 x 10"7 A/cm2) is comparable to that obtained from the sample rinsed in Cr0 3 solution (0.5 g/L) at 70°C, indicating similar corrosion protection abilities of these coatings according to this measure. Although the value of the anodic current decreases after application of the post-treatment, E C O rr shifts to more cathodic values and it appears that the potential range for the passive behavior is reduced (compared with no post-treatment). According to SEM 102 Figure 4.5 Po la r i za t ion curves measured f rom phosphated 2024-A1 samples: (a) no post-treatment; (b) after post-treatment i n B T S E concentrations 1%, (c) 4 % and (d) 10%. 103 1000 800 600 400 200 0 Binding Energy (eV) 290 288 286 284 282 Binding Energy (eV) Figure 4.6 X P S f rom phosphated 2 0 2 4 - A l sample after post-treatment i n 1% B T S E solut ion at 2 5 ° C : (a) survey spectrum; and (b) C Is spectrum. 104 analysis, the morphology and coverage of phosphated samples has not changed after rinsing in BTSE solution. The XPS survey spectrum (Fig. 4.6(a)) from the sample post-treated in 1% BTSE shows the expected Si 2p peak at 102.3 eV, as well as the presence of Zn and P signals; the latter suggests that the thickness of the silane film is less than about 100 A, assuming a reasonably even distribution. The C i s spectrum (Fig. 4.6(b)) has two components, namely Ci at 285.0 eV and C 2 at 287.0 eV, with the ratio C1/C2 equal to 1.96. The component at the lower binding energy is associated with hydrocarbon-like bonding in the silane chain, as well as with contributions from air-borne contamination, while the component C2 at higher binding energy appears to indicate some C-O-Si bonding. This may originate from slow silane hydrolysis [51], or from the competing alcoholysis after dissolution in alcohol; both situations can result in the BTSE being incompletely hydrolyzed [58], and hence give rise to the component C 2 . Table 4.2. Relative atomic compositions (mole%) from XPS for phosphated 2024-A1 samples post-treated in BTSE solutions of different concentrations before and after ultrasonic test 1%, before Zn O C P Si 1.1 50.4 29.6 2.6 16.3 1%, after 1.7 50.5 29.1 2.5 16.2 4%, before 0.0 36.4 39.5 0.0 24.1 To examine the adhesive strength of coatings post-treated by the 1% BTSE solution, an ultrasonic rinsing test was performed in distilled water for 30 min, and the subsequent coating composition measured by XPS. Results in Table 4.2 show that after the sonication test, the surface chemical composition is not dramatically changed from the situation prior to the test. That observation contrasts with results obtained from the unsealed phosphated alloy, where the 105 detected A l is indicated to increase from zero to 24% Of the total measured. The latter is fully consistent with significant coating detachment during the adhesive test. We also conclude that the presence of some incompletely hydrolyzed BTSE does not appear to adversely affect the coating protection properties, to the level we have investigated. Coated samples rinsed in BTSE solutions of higher concentration (i.e. 4% and above) form silane films that are too thick to enable detection of signals from the underlying phosphate layer (Table 4.2). When used in a post-rinse, it is likely that BTSE molecules penetrate between the phosphate crystals and form Si-O-Si bonds between themselves; further, it is possible that they can reinforce unprotected areas on the substrate by forming Si -O-M bonds. In general, an interpenetrating silane network can strengthen adhesion to the metal, as well as provide a physical barrier to the onset of corrosion. Nevertheless, there is no evidence from this work to support direct chemical interaction between BTSE and the phosphate coating; other methods would be needed to investigate that possibility. 4.3.3 Protection at Micro-regions Since samples of phosphated 2024-A1 alloy post-treated in dilute chromic acid solution (0.5 g/L CrC>3 at 70°C) and in 4% BTSE (at 25°C) show similar corrosion resistance according to polarization curve measurements, further immersion tests were performed to follow more closely the changes happening at specific microregions on the coated surfaces. Figure 4.7 shows SEM micrographs taken from mirror-polished 2024-T3 alloy samples that have been phosphated and post-treated with either chromic acid or BTSE, before and after immersing in 3.5% NaCI solution for 12 h. First, it is noted that the zinc phosphate crystals appear to cover matrix regions of the alloy surfaces uniformly, although some enhanced coverage may be observed on particular microstructural regions, such as second-phase particles of Al-Cu-Mg type (PI in Fig. 4.7(a) and P2 in Fig. 4.7(b)). This qualitative observation from SEM is reinforced by EDX measurements 106 Figure 4.7 S E M micrographs o f mir ror -pol i shed and phosphated 2024-A1 samples: (a) post-treated i n 0.5 g / L C r 0 3 solut ion at 7 0 ° C (x3,000); (b) post-treated i n 4 % B T S E at 2 5 ° C (x5,000); (c) as (a) but after i m m e r s i o n i n 3 . 5 % N a C I so lu t ion for 12 h ; (d) as (b) but after 12 h immers ion test. 107 (Table 4.3) taken from part icular micro-regions w h i c h are representative o f behaviors across the who le surface. A subtle (and poss ib ly significant) point is that any C r o n the A l - C u - M g part icle ( P I ) after the ch romic a c i d rinse i s , at most , at the detection l i m i t (according to E D X ) , even though a smal l , but definite, s ignal is recognized for the ne ighbor ing matr ix region. Table 4.3. Percentage compositions (wt%) from EDX measured from selected micro-regions of mirror-polished and phosphated 2024-Al samples post-treated in either chromic acid (0.5 g/L C1O3 at 70°C) (particle PI and matrix nearby) or 4% BTSE (particle P2 and matrix nearby) and measured before the immersion test. P I M g C r A l S i P M n C u Z n 9.6 0.0 61.7 0.0 0.3 0.0 27.1 1.3 M a t r i x near P I 1.7 0.2 91.7 0.0 0.2 0.6 4.5 1.1 P 2 12.7 0.0 44.4 2.1 0.5 0.0 36.9 3.4 M a t r i x near P 2 1.8 0.0 89.6 1.8 0.2 0.6 4.2 1.8 Af te r immers ing i n N a C I solut ion, the sample post-treated w i t h ch romic a c i d shows the format ion o f a dark r i n g around the A l - C u - M g part icle ( P I ) ; that is be l i eved to be associated w i t h faster matr ix d isso lu t ion at the par t ic le-matr ix interface reg ion . In contrast, for the B T S E - r i n s e d sample, the matr ix adjacent to the A l - C u - M g part icle (P2) appears unaffected b y the immers ion test. The composi t ions o f the measured particles are not indicated to change as a result o f the immers ion , and that is consistent w i t h their cathodic natures relative to the matr ix . Af t e r the i m m e r s i o n tests, no changes have been observed for matr ix areas alongside the A l - C u - F e - M n particles, and that applies for both seal ing procedures investigated. X P S spectra measured from these samples, before and after the i m m e r s i o n test, also conf i rm that the coated surface w h i c h had been post-treated w i t h c h r o m i c ac id undergoes a greater change compared to the B T S E - r i n s e d sample (Table 4.4). The A l s ignal from the sample 108 rinsed w i t h ch romic ac id increases after the immers ion test, w h i l e the C r s ignal decreases, and both these observations are indica t ive o f destructive changes occur r ing i n the coat ing. Table 4.4. Atomic compositions (mole%) from XPS for mirror-polished and phosphated 2024-A1 samples post-treated in either chromic acid (0.5 g/L C1O3 at 70°C) or 4% BTSE and measured before and after the immersion test. C1O3, before Z n C u C r O C P S i A l 2.9 0.0 4.7 59.2 20.8 4.9 0.0 7.5 C r 0 3 , after 2.1 1.0 2.6 52.2 26.5 3.4 0.0 12.2 B T S E , before 0.0 0.0 0.0 39.2 38.9 0.0 21.9 0.0 B T S E , after 1.1 0.0 0.0 41.7 36.9 0.0 20.3 0.0 The appearance o f C u is be l ieved to be associated w i t h faster matr ix d isso lu t ion around the A l - C u - M g particles. In contrast, for the sample rinsed i n 4 % B T S E , the chemica l compos i t i on after the i m m e r s i o n test remains prac t ica l ly unchanged. S ince the phosphat ing has been performed o n a mir ror -po l i shed a l l o y surface, a m u c h thinner z i n c phosphate layer is formed. Cor responding ly a larger A l s ignal is measured from the sample r insed w i t h ch romic ac id , compared w i t h a s i m i l a r l y treated sample that i n i t i a l l y just had the 1200-grit a lumina paper po l i sh . D u e to differences i n a l l oy surface roughness, and thickness o f the z inc phosphate layers, the performances o f these samples after the immers ion test cannot be compared direct ly w i t h results obtained from the polar iza t ion curve measurements described i n Sections 4.3.1 and 4.3.2. H o w e v e r , the results presented i n this section are consistent w i t h the conc lus ion that the phosphate coatings sealed w i t h B T S E exhib i t comparable cor ros ion protect ion abil i t ies to those sealed w i t h di lute chromic ac id . Nevertheless, w h e n the coatings are assessed by immers ing i n the N a C I so lu t ion for 21 h , it is found that both post-treatment procedures fa i l ( F i g . 4.8). It also appears that the details o f the degradation differ. The coat ing sealed i n ch romic ac id shows cracks around second-phase particles. T h e damage is l oca l i zed i n regions w h i c h appear darker i n 109 Figure 4.8 S E M micrographs o f mir ror-pol ished and phosphated 2024-A1 samples: (a) post-treated i n 0.5 g / L C r 0 3 solut ion at 7 0 ° C and after i m m e r s i o n i n 3 .5% N a C I solu t ion for 21 h (x3,000); (b) post-treated i n 4 % B T S E at 2 5 ° C and after 21 h i m m e r s i o n test (x3,000). 110 the SEM micrographs; EDX identifies these regions as being associated with enhanced Cu contents compared with the brighter areas. By contrast, the sample post-rinsed in BTSE shows a more uniform degradation across the surface. 4.4 C o n c l u d i n g R e m a r k s Zinc phosphate conversion coatings are being used as an alternative to chromate coatings, the change being driven especially by the environmental concerns associated with the use of Cr(VI). However, the final rinse that is required to seal the phosphate coating is still based on application of chromic acid, and so this process does not fully mitigate the health concerns. Therefore the development of new post-treatments is needed in combination with the phosphating approach. The exploratory work reported here focuses on comparing two different post-treatment rinses. One involves the use of dilute chromic acid, and it forms a reference, while the second uses the organosilane BTSE, which in other contexts is believed to have potential as a protective coating for metals [39,40]. The corrosion stabilities of the post-treated phosphate coatings are characterized by polarization measurements and immersion tests, and the results are discussed in terms of observations for the surface morphologies and compositions of the coatings. With the chromic acid rinse, the effect on the phosphate coating depends in part on the etching involved and in part on the nature of the chromated layer formed. For example, higher concentrations of Cr(VI) in the rinse reduce the thickness of the phosphate coating, but also contribute to better corrosion resistance especially when the treatment is applied at 70°C rather than 25°C. That may be due to more mixed Cr-Al oxide formed at the higher temperature, as well as to a higher proportion of Cr(VI) compared with the Crflll) oxidation state. When the organosilane rinse is used, it is assumed that the BTSE molecules penetrate between the phosphate crystals and form Si-O-Si bonds among themselves, as well as Si -O-M bonds with the substrate, and thus help 111 passivate the surface. Such bonding details have not yet been confirmed in this context, but XPS analysis of the coatings before and after ultrasonic tests show no tendency for detachment. Therefore these coatings pass at least the first test for adhesive bonding. Initial observations were made in this work for the variation in chemical behaviors at different microstructural regions of the alloy surface. The etching of zinc phosphate coatings during the post-rinse in chromic acid creates sites, especially at borders between alloy matrix and Al-Cu-Mg particles, for the initiation of corrosion in NaCI solution. The application of the BTSE post-treatment appears to be more effective at protecting these micro-regions, but its overall performance is also highly dependent on the thickness of the silane overlayer. As corrosion proceeds, coatings from the two post-rinse methods develop different degradation chemistries, an observation that needs further study in order to establish the mechanisms of coating failure. Compared with conventional studies for corrosion protection, the tests made here are quite severe since the coatings remain unpainted. Also, in order to optimize information from the XPS technique in this work, it has been advantageous to use coatings that are thinner than those that maximize the corrosion protection. These points are accepted for the comparisons made, but further investigations are also needed for the natures of the interfacial bonding between post-treated coatings and a subsequently added paint layer. 112 Chapter 5 Microstructural Effects on the Initiation of Zinc Phosphate Coatings on 2024-T3 Aluminum Alloy 5.1 Introduction Z i n c phosphate convers ion coatings, as indicated i n Chapter 1, represent a p romis ing approach for the cor ros ion protect ion o f a l u m i n u m al loys . A number o f phosphating processes are avai lable through patents [176-182], and var ious studies have reported o n the effect o f changing solu t ion parameters, such as compos i t ion , p H , temperature, t ime o f coat ing, as w e l l as the influence o f pr ior treatments that m a y result i n surface enrichment o f the a l l o y i n g elements [96,98,183-187]. Howeve r , o n l y a f ew investigations have been reported that relate to the coat ing mechan i sm [86,188,189], and at this t ime it must be conc luded that the mechanis t ic details are insuff ic ient ly k n o w n to provide a basis for the des ign o f n e w processes. The z inc phosphat ing o f metals depends o n the precipi ta t ion o f Zn^^O^* 4H2O from supersaturated solutions, but the in i t ia t ion o f the phosphat ing process appears important to the nature o f the f ina l coat ing. The first part o f the phosphat ing process is governed by e lect rochemical reactions occur r ing at l oca l micro-anodes and micro-cathodes o n the heterogeneous metal surface [76,77,90]. E a r l y w o r k b y Cheever [190] showed that phosphate coatings o n i r o n and steel substrates p r i m a r i l y initiate at cathodic gra in boundaries, and more recently it has been reported that the base-metal microstructure i n i r on and steel also affects the thickness and loca l properties o f phosphate coatings [191]. A c c o r d i n g to the avai lable literature, the analogous w o r k has not been done o n a l u m i n u m al loys , and accord ing ly the research presented i n this chapter undertakes an in i t i a l explora t ion i n this area. The objective o f the present study is to explore the effect o f the different microstructural regions, that occur at a surface o f an 2024-A1 a l loy , o n the in i t ia t ion o f z i nc phosphate coatings. The pr ime 113 characterization techniques used are scanning electron microscopy (SEM), to assess coating morphology, and scanning Auger microscopy (SAM) to probe chemical compositions at local micro-regions; X-ray photoelectron spectroscopy (XPS) is also used to give more general information on chemical composition over broader regions of the coated surfaces. Electron dispersive X-ray (EDX) spectroscopy was additionally applied to get composition information from local regions, although this technique probes deeper than is the case with SAM. 5.2 E x p e r i m e n t a l Samples of 2024-T3 aluminum alloy (1x1x0.12 cm3) were mechanically polished with alumina paper up to 1200 grit; in addition, some samples were further polished with diamond paste to ltxm roughness, and such samples are referred to below as being "mirror-polished". These treatments were followed by ultrasonic cleaning in acetone and methanol, and drying in air. Zinc phosphate (ZPO) coatings were formed by immersing the polished alloy samples in Gardobond R 2600 (Oakite) phosphating solution (pH 4.0), held either at 25°C or at 50°C, while the immersions occurred for specific times within the range 3 s to 3 min. The coating at pH 3 was done after the original coating solution was acidified with HNO3. To observe the grain structure, a mirror-polished sample was etched in a NaOH/NaF solution (2 g NaOH, 5 g NaF and 93 ml H 20) at 25°C for 5 min [8]. SEM micrographs were measured with a Hitachi S-4100 microscope with field emission source operated at 15 kV. The scanning Auger electron microscopic analysis was performed by using a Microlab 350 spectrometer (Thermo VG Scientific) with field emission source and hemispherical energy analyzer. Point analysis was carried out using the primary electron beam set at 10 keV and 4 nA, but the respective values were 25 keV and 1.4 nA for generating Auger maps. In the latter case, the sample was tilted by 60° towards the analyzer in order to ensure 114 maximum sensitivity. To obtain a depth profile, an Ar + beam of 1 kV (current at sample 1.5 uA) was employed with a sputter period of 10 s for each analysis level. Supplementary XPS spectra were measured with a Leybold MAX200 spectrometer using the Mg Ka source (1253.6 eV) operated at 10 kV and 20 mA; the system pressure was 2xl0 - 9 mbar while measurements were being made. Survey spectra were obtained with the pass energy set at 192 eV. Binding energies were referenced to the Au Aim peak at 84.0 eV for conducting samples, but for coated samples for which the metallic component could not be detected (i.e. at 72.6 eV binding energy in the Al 2p spectrum), the C Is peak for adventitious hydrocarbon contamination was set at 285.0 eV as the alternative reference. EDX characterizations were done with a Hitachi S3000N spectrometer, using a 20 keV incident electron beam, and elemental amounts were compared on a weight-percentage basis. 5.3 Results 5.3.1 Preliminary Observations Figure 5.1 shows SEM micrographs of ZPO coatings formed on mechanically-polished (1200 grit) 2024-Al alloy samples after immersing in the coating solution at 50°C for different times (3 s to 3 min). The alloy surface is rough with polishing lines clearly detectable, and it is observed that the coating coverage increases as the immersion time increases to 1 min, although no change in coverage is apparent with a further increase to 3 min. For this coating time, XPS measurement indicates a complete coverage insofar as the A l 2p signal from the substrate is not detected (Fig. 5.2). The Zn 2p3/2, O Is and P 2p signals at binding energies 1022.9 eV, 531.6 eV and 134.2 eV respectively fit expectation for the presence of a ZPO coating [170], but the C Is signal signifies the presence of air-borne contamination. For the shortest immersion time (3 s), the coating crystallites are not observed on the alloy surface except at particular local sites associated with the second-phase particles. One such 115 Figure 5.1 S E M micrographs o f mechanica l ly -po l i shed (1200 grit) 2 0 2 4 - A l a l l oy after treating w i t h the ZPO coat ing solut ion at 5 0 ° C for different i m m e r s i o n t imes: (a) 3 s, (b) 15 s, (c) 30 s, (d) 1 m i n (e) 2 m i n and (f) 3 m i n . 116 Zn2p 1000 800 600 400 200 0 Binding Energy (eV) Figure 5.2 X P S survey spectrum o f mechanica l ly -po l i shed 2024-A1 a l l oy after treating w i t h Z P O coat ing so lu t ion at 5 0 ° C for 3 m i n . 117 micro reg ion is shown i n F i g . 5.1(a), and it is not iced that Z P O crystall i tes start to f o r m at the edge between that second-phase part icle and the surrounding matr ix . The two major types o f second-phase particles i n 2 0 2 4 - A l a l l oy are s h o w n i n F i g . 1.2(a) b y a S E M micrograph for a mir ror -po l i shed sample. The concentrat ion o f the a l l o y i n g elements w i t h i n these intermetal l ic precipitates exceeds their concentrat ion i n the matr ix by an order o f magnitude, and that results i n dis t inct ive phys ica l and chemica l properties at these loca l sites (as already seen, for example , i n F i g . 5.1(a)). 5.3.2 Initial Coatings on Mirror-polished Samples In order to obtain more informat ion o n the effect o f different microstructural regions o n coat ing chemistry, further studies were made o n a mir ror -po l i shed sample that was coated for 1 m i n i n the Z P O solut ion at 25 ° C . The intent ion was to s l o w d o w n the coat ing reaction, thereby enabl ing an ident i f ica t ion o f the in i t ia t ion sites. U n d e r such condi t ions , the Z P O coat ing coverage was found to be strongly dependent o n the part icular microstructure. The coat ing at an A l - C u - M g part icle (referred to b e l o w as particle P I ) is discussed i n re la t ion to F i g . 5.3. The S E M micrograph ( F i g . 5.3(a)) indicates that this part icle i s fu l ly covered by a dense crystal l ine coat ing, w h i l e the immedia te ly surrounding matr ix appears not to be covered. T h i s is substantiated by A u g e r point analysis ( F i g . 5.3(b)) where spectra taken f rom the center o f the particle (spectrum 1) and from its edge (spectrum 2) show the presence o f characteristic A u g e r peaks: Z n L M M at 987.9 e V , O K L L 509.6 e V and P K L L 1853.5 e V o n the k inet ic energy scale. The A l substrate s ignal is not detected indica t ing a complete and reasonably th ick coat ing, al though a C s ignal from air-borne contaminat ion is present at 266.9 e V . In contrast, a large A l K L L s ignal at 1387.5 e V , that corresponds to the A l ox ida t ion state [128], is measured a long w i t h O K L L from the surrounding matr ix , but o n l y trace amounts o f Z n and P are detected (spectrum 3). It is conc luded that the matr ix adjacent to the part icle is coated by a l u m i n u m oxide 118 -i 1 1 1 1 1 1 I I 200 400 600 800 1000 1200 1400 1600 1800 Kinetic Energy (eV) Figure 5.3 A l - C u - M g particle ( P I ) from mir ror -pol i shed 2 0 2 4 - A l sample after exposure to Z P O coat ing solu t ion at 25 ° C for 1 m i n : (a) S E M micrograph ; and (b) A u g e r spectra measured from different locat ions: ( l ) par t ic le ' s center, (2) near edge o f part icle, and (3) matr ix near to P I . 119 J 1 1 I L 200 400 600 800 1000 1200 1400 1600 1800 K i n e t i c E n e r g y ( e V ) Figure 5.4 A l - C u - F e - M n particle (P2) f rom mir ror -po l i shed 2 0 2 4 - A l sample after exposure to the Z P O solut ion at 2 5 ° C for 1 m i n : (a) S E M micrograph , and (b) A u g e r spectra measured from different locat ions: ( l ) and (2) par t ic le ' s surface; (3) and (4) matr ix near to P 2 . 120 (e.g. A10 x(0H)y) supplemented by small amounts of non-crystalline aluminum phosphate, AIPO4, or ZPO. A marked difference in coverage according to SEM is seen for the coating at the larger Al-Cu-Fe-Mn particle, referred to as particle P2 (Fig. 5.4(a)). Only a small number of coating crystallites are present on P2's surface, compared to PI. The surrounding matrix exhibits more crystallites at areas away from the particle compared to near its border. Figure 5.4(b) shows Auger spectra taken from selected spots on the particle and on the matrix close to it. Zn and P are both detected from the different locations considered on the particle's surface (Fig. 5.3(b), spectra 1 and 2), although no crystals were observed at location 1. However, the amount of Zn measured from the Al-Cu-Fe-Mn particle is lower compared to that from the Al-Cu-Mg particle, and an Al KLL signal at 1387.5 eV is observed for P2 (spectrum l), while that is not the case for PI. In addition, for P2, while a P L M M signal is detected at 120.0 eV, the higher kinetic energy P KLL signal appears lost in the background (Fig. 5.4(b), spectrum 1). Taking these points together, it seems clear that the ZPO coating on P2 is thinner and different in nature (e.g. more amorphous) compared with PL The matrix region closest to particle P2 appears lightly coated by phosphate, or possibly as a mixed oxide-phosphate (Fig. 5.4(b), spectrum 3), while areas further from the particle show the presence of ZPO crystals. When different second-phase particles are observed across the wider sample surface, the Al-Cu-Fe-Mn particles generally all exhibit similar levels of crystalline ZPO coverage, but variations are observed among the Al-Cu-Mg particles. For example, some in the latter group appear to be uncoated except for a few crystallites at the edge region. However, Auger measurements reveal this is misleading, and that a ZPO coating does exist at all measured particle locations. The coating is amorphous around the particle's central area, although crystallites nucleate at the edge regions (as already observed in Fig. 5.1(a)), often with a larger Al signal detected. While the crystalline areas generally give values of the Zn/P atomic ratio 121 (e.g. 1.3-1.6) which are fairly close to the value of 1.5 expected for stoichiometric Zn3(PC»4)2, lower values (e.g. 0.6) are indicated for the non-crystalline coated regions. The differences in the coatings formed on particles PI and P2 must depend on the natures of the individual surfaces, and some information on the relevant coating-particle interfaces can be provided by Ar + sputtering (Fig. 5.5). After 6 min sputtering, the coating from particle PI is completely removed, indicated by no Zn LMM, P KLL or C KLL signals being detectable. Figure 5.5(a) shows the Al KLL signal at 1393.2 eV, which is consistent with the A1(0) oxidation state (i.e. metallic Al); additionally there is a strong Cu L M M signal at 918.5 eV, and a Mg KLL signal is also present at 1185.6 eV. For the same conditions, the coating above the Al-Cu-Fe-Mn particle (P2) is removed after the 3 min sputtering (Fig. 5.5(a), spectrum 4), and Auger signals from Fe L M M and Mn L M M (occur at 703.1 eV and 588.1 eV) in addition to those from Cu L M M and A l KLL. Table 5.1 reports atomic percentages for the alloying elements detected at the surfaces of particles PI and P2 after removal of the ZPO coatings. Table 5.1 Elemental compositions (atomic%) measured by SAM point analysis from Al-Cu-Mg (PI) and Al-Cu-Fe-Mn (P2) particles after removal of ZPO coating by sputtering. PI Cu Mg Al Mn Fe 82.0 4.3 13.7 0.0 0.0 P2 10.2 0.0 23.3 51.0 15.5 It appears that the surface of PI is rich in Cu, while that for P2 is rich in Mn. This is in contrast to the usual compositions of these types of particles, as previously measured from other samples of the 2024-A1 alloy, by energy dispersive X-ray spectroscopy (EDX), where the average Cu content for the Al-Cu-Mg intermetallics is around 25% (atomic %), while the amount of Mn is usually similar to that of Fe (i.e. -10%) for the Al-Cu-Fe-Mn particles. 122 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 K i n e t i c E n e r g y ( e V ) 1220 1260 1300 1340 1380 1420 1460 K i n e t i c E n e r g y ( e V ) Figure 5.5 A u g e r spectra measured f rom different regions o n mir ror -pol i shed 2024-A1 sample after exposure to Z P O coat ing solu t ion at 2 5 ° C for 1 m i n : (a) ( l ) A l - C u -M g particle ( P I ) after coat ing, (2) after subsequent 6 m i n A r + sputtering, (3) A l -C u - F e - M n part icle (P2) after coat ing, (4) after subsequent 3 m i n A r + sputtering; (b) A l K L L spectra measured after A r + sputtering for t imes to remove coating: ( l ) part icle P I , (2) part icle P 2 , and (3) matr ix . 123 Figure 5.6 Images f rom A l - C u - M g particle o n mi r ro r -po l i shed 2 0 2 4 - A l sample after exposure to Z P O coat ing solut ion at 25 ° C for 1 m i n : (a) S E M micrograph; and S A M elemental maps for: (b) Z n and O ; (c) A l and P ; and (d) C u map after r emova l o f coating. 124 Figure 5.5(b) compares A l KLL Auger spectra taken from the matrix and from the particles PI and P2 after Ar + sputtering. The spectrum from the matrix after removal of the coating consists of a number of peaks in the kinetic energy range 1270 to 1400 eV, with the highest intensity component at 1393.2 eV. The shape and fine-structure appears consistent with expectation for A l in metallic form [128]. However, the corresponding spectra measured from the two particles differ in shape, compared to that from the matrix, although the peak positions are essentially unchanged. The intensities of the fine structure peaks (kinetic energy range 1270-1390 eV) for the Al-Cu-Fe-Mn particle are lowered compared with those from the matrix, and they are decreased even further for the Al-Cu-Mg particle. Although the A l is basically metallic in these regions, the detailed electronic properties are different when the Al is in the matrix and when it is bonded to the alloying elements within the intermetallic compounds. Auger maps shown in Fig. 5.6 provide more-detailed views of changes in chemical composition across areas containing an Al-Cu-Mg particle. The ZPO coating covers the particle, and a trench appears to be present at the border between the particle and the matrix. It is concluded that after immersion in the acidic ZPO solution, etching is increased in this area as a consequence of galvanic coupling between A l in the matrix and Cu in the particle. 5.3.3 Ful l Coverage on a Mirror-polished Sample Results from Section 5.3.1 indicate that saturation in the ZPO coating on the mechanically-polished samples occurs when the alloy samples are immersed for 3 min in the coating solution at 50°C. These conditions have been applied to the mirror-polished sample, and the coating coverage has been studied at local micro-regions, both across the surface and with depth below. The SEM image shows that the surface is uniformly coated (Fig. 5.7(a)) although the presence of second-phase particles still produces a contrast, based presumably on the combined effects of surface topography and the different electron scattering from Cu rich 125 —I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 200 400 600 800 1000 1200 1400 1600 1800 Kinetic Energy (eV) Figure 5.7 M i r r o r - p o l i s h e d 2 0 2 4 - A l sample after exposure to Z P O coat ing solu t ion at 5 0 ° C for 3 m i n : (a) w i d e - v i e w S E M micrograph; (b) coated A l - C u - M g part icle (P3) ; (c) A u g e r spectra measured from different locat ions: (1-2) par t ic le ' s surface, and ( 3 - 5 ) posi t ions o n matr ix close to P 3 . 126 particles relative to the matr ix [192]. X P S detects an A l 2p s ignal at 74.4 e V , un l ike the situation for coat ing o n the rougher sample (Sect ion 5.3.1). A c c o r d i n g l y it is conc luded that the Z P O coat ing i s thinner o n the mi r ror -po l i shed sample, and such an observat ion h ighl ights that the amount o f coat ing, for otherwise constant condi t ions , is s ignif icant ly inf luenced b y the degree o f surface roughness. Table 5.2 Elemental compositions (atomic%) measured by SAM point analysis from the coated surface shown in Fig. 5.7(c): Al-Cu-Mg particle P3 at locations 1,2; nearby matrix at locations 3-5.* 1 O Z n A l P 51.6 26.3 0.0 22.1 2 51.4 34.3 0.0 14.3 3 44.9 22.4 24.3 8.4 4 49.3 23.1 14.2 13.4 5 32.2 30.6 18.2 19.0 * C contr ibutions have been deleted for these comparisons . These observations are reinforced by further analysis at the microstructural l eve l ( F i g . 5.7(b,c)). A u g e r point measurements f rom the A l - C u - M g part icle (P3) , and its surrounding matr ix , show that the coat ing covers both mat r ix and part icle, a l though its thickness varies w i t h locat ion. The A l K L L s ignal f rom the substrate is not detected from the part icle, but it is from the nearby matr ix . It i s not iced that the Z n / A l ratio measured from the matr ix , at distances 0 .5 ,1 and 2.5 p m from the particle (Table 5.2), is lowest at the loca t ion closest to the particle. That is consistent w i t h a conc lus ion that the coat ing thickness decreases i n order: A l - C u - M g part icle > mat r ix further away from the particle > matr ix adjacent to the part icle. The crysta l l ine Z P O particles formed o n top o f P3 have a d imens ion -0 .1 p m . Dep th profi les ( F i g . 5.8) were taken from the center o f an A l - C u - M g particle. Z n , P and O are seen at the surface, w h i l e A l and C u are the p r imary bu lk components . Nevertheless, the 127 amount o f C u is indicated to be non-uni form w i t h depth. F o r example , a higher C u s ignal is found after sputtering for 700 s compared w i t h the C u signal after sputtering for 2000 s w h e n the bu lk compos i t ion is reached. A l s o , it is noted that there is an increased presence o f C u at the coating-particle interface, w h i l e the A l concentrat ion is s t i l l l o w there. T h i s observat ion is consistent w i t h C u enrichment at the Z P O - p a r t i c l e interface. A l t h o u g h the mir ror -po l i shed sample has been treated under condi t ions that are expected to fo rm a crystal l ine Z P O coat ing across the w h o l e surface, it is noted that some A l - C u - M g particles are s t i l l covered by an amorphous coat ing, w h i c h is s imi l a r to those described i n Sect ion 5.3.2. F igure 5.9(a) shows an A l - C u - M g part icle (P4) that exhibi ts crysta l l ine coverage at its edge, but not over most o f its central area. A u g e r measurements at chosen locat ions show the presence o f Z n , P and O , and hence a z i nc phosphate coat ing ( F i g . 5.9(b)), and the elemental amounts are g iven i n Table 5.3. A Z n / P ratio o f about 1.6 is indicated for the crystal l ine phase, w h i l e the lower value o f 0.7 is obtained f rom the amorphous phase. Table 5.3 Elemental compositions (atomic%) measured by SAM point analysis from different regions of Al-Cu-Mg particle P4 showing ZPO coatings with crystalline (C) and amorphous (A) characters. c C O Z n P Zn/P 32.8 37.3 18.5 11.4 1.6 A 0.0 54.8 18.2 27.0 0.7 5.3.4 Sputtering and Re-oxidation F o r coat ing purposes, the pre-treatment o f a metal surface is important since i t can have a major influence o n the coat ing characteristics. The s ignif icance o f surface roughness has already been ment ioned, but another factor o f l i k e l y importance is the amount o f ox ide present at the 128 0 500 1000 1500 2000 2500 Time (s) Figure 5.8 S A M depth profi les measured f rom Z P O - c o a t e d A l - C u - M g particle o n mirror -pol i shed 2024-A1 sample after so lu t ion appl ied at 5 0 ° C for 3 m i n . 129 C/5 "S vx C P L M M 200 400 600 800 1000 1200 1400 1600 1800 K i n e t i c E n e r g y ( e V ) Figure 5.9 A l - C u - M g part icle (P4) o n mir ror -pol i shed 2024-A1 sample after exposure to Z P O coat ing so lu t ion at 5 0 ° C for 3 m i n : (a) S E M micrograph ; (b) A u g e r spectra taken f rom (1) crysta l l ine region, and (2) flat amorphous region. 130 surface. Mechanical polishing can remove much of the native oxide, but new oxide forms rapidly on a subsequent exposure to air. The low-temperature oxidation process at different alloy micro-regions is not well known. As far as the zinc phosphating process is concerned, the amount and nature of oxide at different local sites is important since it needs to be removed, during the etching stage, in order to initiate the formation of new coating. In general it is expected that a thinner and more porous oxide layer would be removed faster, and hence more easily provide a fresh surface for the deposition of new coating. Figure 5.10(a) shows SAM measurements taken from a Al-Cu-Mg particle, and from the adjacent matrix, after the ZPO coating is removed by Ar + sputtering (spectra 1 and 3), and after the sample has been exposed for 48 h to ambient oxygen in the ultrahigh vacuum (spectra 2 and 4). After the initial sputtering, only the Al KLL signal is detected from the matrix, while, as expected, the Cu LMM, Mg KLL and A l KLL signals are observed from the particle. After the gentle re-oxidation, the O KLL appears, and it is clear that its intensity is higher from the matrix than from the particle. After the 48 h exposure, the O/Al ratio equals 1.63 for the matrix and 0.89 for the particle, so emphasizing that the matrix is oxidized to a greater extent. Further study of the A l KLL spectrum (Fig. 5.10(b)) shows that, as the oxide builds up on the matrix, the fine-structure intensity (kinetic energy range 1280-1380 eV) decreases, and the main peak at 1391.5 eV shifts to lower kinetic energy by 4.1 eV to become consistent with the A l state. In contrast, no comparable change occurs for the main component of the Al KLL signal from the particle, and this gives further support to the conclusion that the initial stage of the oxidation process is much slower on the particle than on the matrix. It is assumed that after mechanical polishing a thinner oxide layer is likely to be present at the surface of the Al-Cu-Mg particles, compared to the matrix, and therefore that the particle's oxide can be removed more easily during the initial etching in the acidic phosphating solution. 131 Cfl 5 Cfl (a) A l K L L C u L M M -I u -I 1 1 1 u 400 500 600 700 800 900 1000 1100 1200 1300 1400 Kinetic Energy (eV) Cfl +-> s Cfl c 1220 1260 1300 1340 1380 1420 1460 Kinetic Energy (eV) Figure 5.10 A u g e r spectra measured from: (a) A l - C u - M g part icle before and after ox ida t ion (spectra 1 and 2 respect ively); matr ix before and after ox ida t ion (spectra 3 and 4), and (b) A l K L L spectra for mat r ix before and after ox ida t ion (spectra l and 2) and A l - C u - M g part icle before and after ox ida t ion (spectra 3 and 4). 132 5.3.5 Coating at pH 3 The influence of a lower pH in the coating bath was examined on 2024-A1 samples. Figure 5.11 shows SEM micrographs of a mirror-polished alloy that has been coated in ZPO solution at pH 3.0 for 3 min at 50°C. Under these conditions, severe etching of the sample surface occurs during the coating deposition, and as a consequence the grain structure becomes visible in the SEM micrograph (Fig. 5.11(a)). The areas containing Al-Cu-Mg and Al-Cu-Fe-Mn particles (Fig. 5.11(b,c)) reveal that the surface of the matrix is rough, and contains a small number of ZPO crystals, while the surfaces of both second-phase particles are smooth and appear uncoated. Figure 5.12(a)) shows Auger measurements taken from the matrix (Figure 5.12(a), spectrum l) and central areas of both particles (Figure 5.12(a), spectra 2 and 3). Zn and P signals are detected from all measured locations indicating the presence of a ZPO coating, but Al KLL and Cu L M M are measured from the matrix (Table 5.4). Table 5.4 Elemental compositions (atomic%) measured by SAM point analysis from different regions of the coated surface shown in Fig. 5.11 obtained after using the coating solution at pH 3.0. Matrix C 0 Ni Zn Al P Cu 54.6 19.1 6.0 7.1 5.7 3.2 4.3 Al-Cu-Mg 44.5 22.9 10.9 12.4 3.9 5.4 * Al-Cu-Fe-Mn 50.9 23.4 9.4 13.4 2.9 * * * element not detected It is concluded that the coating thickness is lower on the matrix compared to the second-phase particles, and that the latter are more likely covered by an amorphous ZPO, as already observed for coating at pH 4.0. For given coating conditions, the increased etching helps the dissolution of Cu from the substrate, which then reprecipitates back on the surface and produces the Cu-rich ZPO coating. It also appears that the Cu reprecipitation occurs on the matrix rather 133 Figure 5.11 S E M micrographs o f Z P O coat ing obtained o n a l l oy sample us ing coat ing bath w i t h p H 3.0: (a) selected area at l ower magni f ica t ion ; (b) r eg ion w i t h A l - C u - M g part icle; and (c) region w i t h A l - C u - F e - M n part icle. 134 200 400 600 800 1000 1200 1400 1600 1800 2000 Kinetic Energy (eV) 600 700 800 900 1000 1100 Kinetic Energy (eV) Figure 5.12 A u g e r spectra measured f rom different locat ions for 2024-A1 sample coated at p H 3: (a) w i d e scan for mat r ix (spectrum l ) , A l - C u - M g part icle (spectrum 2), A l - C u -F e - M n particle (spectrum 3); and (b) same as (a) but shorter energy range. 135 than on the particles, and that is consistent w i t h their e lec t rochemical natures. F igure 5.12(b) shows the spectral region cover ing the Z n L M M and C u L M M A u g e r signals. The N i L M M signal is detected at 845.3 e V . N i 2 + is c o m m o n l y added to the coat ing bath as an accelerator, i n order to speed up the coat ing format ion and enhance cor ros ion protection. M i x e d Z n / N i phosphates are thought to be formed o n ga lvanized steel substrates dur ing this coat ing process [193], but its role i n the phosphating o f a l u m i n u m is not comple te ly understood. The N i / Z n ratio is found to be dependent o n the loca l microstructure; speci f ica l ly , measured values vary f rom 0.74 for the matr ix , to 0.81 and 0.67 for the A l - C u - M g and A l - C u - F e - M n particles respectively. 5.3.6 Z P O C o a t i n g A f t e r N a O H / N a F E t c h i n g A paral le l study has been done i n an attempt to ga in in i t i a l informat ion for the influence o f a l loy grain structure o n the phosphating process. The gra in structure becomes v i s ib l e after e tching o f a mir ror -po l i shed 2 0 2 4 - A l a l l oy i n N a O H / N a F solut ion (F ig . 5.13(a)). The grains w i t h different orientation are presented as areas w i t h var ious gray contrasts, w h i l e the second-phase particles appear as bright whi te spots. F igure 5.13(b) shows S E M micrograph o f the same microarea after immers ion i n the Z P O coat ing solu t ion ( p H 4.0, 2 5 ° C ) for 5 m i n . It is not iced that the grains w i t h l ighter contrast i n F i g . 5.13(a) have become even l ighter after the coat ing deposi t ion, and they show evidence o f c rack ing . E D X measurements have been taken from various microregions shown i n F i g . 5.14, and it is conc luded that the Z P O coat ing is present o n top o f a l l analyzed grains. Larger Z n and P signals, and therefore higher Z n / A l and P / A l ratios, are measured from the lighter A , B , C and H regions compared to the darker D , E , F and G regions (Table 5.5), indica t ing that more coat ing is deposited o n the former grains. 136 Figure 5.13 S E M micrographs o f (a) mir ror -pol i shed and N a O H / N a F etched 2 0 2 4 - A l a l loy sample and (b) after Z P O coating. 137 Figure 5.14 (a) C r a c k e d Z P O coat ing o n grains w i t h different orientation, and selected locations for E D X measurements; (b) grain boundary before coat ing; and (c) same region as i n (b) but after coating. 138 Table 5.5 Elemental weight ratios determined by EDX from different regions of the coated surface shown in Fig. 5.14. Z n / A l A B C D E F G H I 0.187 0.213 0.190 0.140 0.146 0.129 0.149 0.168 0.132 P / A l 0.154 0.177 0.155 0.118 0.126 0.110 0.123 0.144 0.110 C u / A l 0.189 0.204 0.207 0.182 0.197 0.175 0.186 0.182 0.172 In addi t ion, the amount o f C u detected across the sample surface (i.e. ~ 1 0 w t%) is twice that f rom the jus t -pol i shed a l loy , but the C u / A l ratios appear to be higher f rom the l ighter and cracked grains compared to those f rom the darker regions. Tab le 5.6 gives a tomic elemental compos i t i on obtained f rom X P S measurements taken from: mi r ror -po l i shed a l l oy sample designated as M P , the mir ror -po l i shed and Z P O - c o a t e d ( M P Z P O ) sample, the mir ror -po l i shed and etched ( M P E ) sample, and the mir ror -pol i shed , etched and Z P O - c o a t e d ( M P E Z P O ) sample. O n l y trace amounts o f C u are detected from the M P , M P Z P O and M P E samples, but its concentrat ion is increased for the M P E Z P O sample, w h i c h is consistent w i t h the E D X observations. The coat ing obtained at 2 5 ° C o n the mi r ror -po l i shed sample is th in , as indicated by the large A l s ignal present. H o w e v e r , A l and M g are measured from the M P E ( F i g . 5.15(a,b)) and M P E Z P O samples ( F i g . 5.15(c,d)). The A l 2p nar row scan i n F i g . 5.15(a) shows o n l y the A l 3 + component at 74.5 e V b i n d i n g energy, w h i l e that measured from M P E Z P O ( F i g . 5.15(c)) is lower i n intensity and consists o f both A l 3 + and A l ° components. The surface reg ion o f the M P E Z P O sample shows a lower M g content (Table 5.6 and F i g . 5.15(d)). A c c o r d i n g to the results presented, i t is conc luded that the etching i n N a O H / N a F solut ion produces a th ick m i x e d A l - M g ox ide coat ing o n the a l l oy surface and that it interferes w i t h the Z P O coat ing process. It is possible that its thickness varies w i t h gra in orientation; after immers ing i n ac id ic Z P O solut ion, c rack ing occurs and that is more prominent o n grains hav ing a l ighter contrast. D u r i n g the coat ing process, the m i x e d ox ide dissolves , but some is s t i l l present 139 Binding Energy (eV) Figure 5.15 A l 2p and M g 2p spectra measured from mir ror -po l i shed and etched sample, M P E (a) and (b) respect ively; and mir ror -pol i shed, etched and coated sample, M P E Z P O (c) and (d) respectively. 140 at the interface between the alloy and ZPO layer after the coating procedure is completed. It is noticed that a larger Cu signal is detected from the MPEZPO sample compared to that from the MPE sample. This is believed to result from the larger Cu enrichment that follows from both the chemical treatment in NaOH/NaF solution and the ZPO coating process. Table 5.6 Elemental compositions (atomic%) measured by XPS for mirror-polished alloy (MP), for mirror-polished and coated alloy (MPZPO), for mirror-polished and etched alloy (MPE) and for mirror-polished, etched and coated alloy (MPEZPO) samples. Zn MP MPZPO MPE MPEZPO * 2.5 * 5.2 Cu 0.3 0.3 0.4 1.6 F * 0.8 2.7 3.7 0 44.9 37.1 51.3 49.4 C 18.3 43.5 19.2 20.4 P * 1.7 * 3.8 Al 36.5 14.1 11.7 6.1 Mg * * 14.7 9.8 * element not detected It appears that the etching-induced chemical modification of the 2024-A1 alloy surface significantly interferes with the coating mechanism, and that its influence cannot be separated from the influence of grain orientation for the conditions applied. More work is needed to establish the relation between the ZPO coating process and the underlying grain orientation. 5.4 Discussion This work shows mat the formation of a zinc phosphate conversion coating on a mechanically-polished 2024-A1 alloy is strongly affected by alloy microstructure, especially the presence and distribution of second-phase particles. When the 2024-A1 alloy is polished, 141 intermetallic compounds, involving the elemental combinations Al-Cu-Mg and Al-Cu-Fe-Mn, become exposed to create an electrochemically heterogeneous surface, and different processes occur at these micro-regions on contact with the ZPO coating solution. A crystalline zinc phosphate coating initiates preferentially at the Al-Cu-Mg second-phase precipitates, rather than at those of Al-Cu-Fe-Mn type, and the initial nucleation of crystalline deposit starts at the edge rather than the center of a second-phase particle. In the initial stages of ZPO formation, a thin and partially amorphous coating can cover the Al-Cu-Fe-Mn particles, as well as some of the A l -Cu-Mg particles, although it is thicker in the latter case. As the coating process proceeds, crystalline ZPO covers the whole surface, but its thickness varies at the different microregions. Earlier study of pitting corrosion on the 2024-A1 alloy characterized the two main types of second-phase particles according to their behaviors in a corrosive environment. The Al-Cu-Fe-Mn particles are cathodic, and promote localized dissolution of the surrounding matrix, while the Al-Cu-Mg particles are anodic and they oxidize and dissolve [21,194]. Corrosion potentials associated with these particles have been determined by electrochemical testing of the bulk intermetallic compounds [195,196]. It was confirmed that the Al-Cu-Mg particles have a negative corrosion potential with respect to the matrix, while the Al-Cu-Fe-Mn constituents are more noble than the matrix. However, further study indicated that, after exposure to a corrosive environment, all particles eventually become cathodic in nature and therefore promote matrix dissolution [194]. It has been observed that the Al-Cu-Mg particles experience dealloying due to the preferential dissolution of Mg and A l , so leaving the surface rich in Cu and thus altering its electrochemical nature compared with that of the matrix [194,197]. It is believed that the dissolution of the Cu, and its re-precipitation [158,197,198], also contributes to the changed nature of the Al-Cu-Mg particles, as well as to an increased corrosion rate for the matrix. Recent studies by scanning Kelvin probe force microscopy [192,199,200] show that in air all the 142 particles have noble Volta potentials compared to the matrix but, upon exposure to a chloride environment, the Al-Cu-Mg particles become active and dissolve after an induction time. The zinc phosphate coating process starts with etching of the surface oxide, and this is followed by a subsequent precipitation of the coating. It is expected that the natures of the oxide layers formed above the various microstructural regions differ compared to that above the matrix [1,201,202], although a systematic study has not yet been undertaken. The presence of alloying elements in the oxide layer can modify its properties, for example making it more semiconducting rather than insulating [1]. In addition, it is known that Al-Mg alloys develop Mg-rich oxide films with high porosity and poor corrosion protection ability [146,203]. Due to the high Mg content in Al-Cu-Mg particles, it is likely that Mg oxide or a mixed Al-Mg oxide can exist at, and provide some protection to, the particle's surface. However, the nature of the oxide on the particle is inevitably different from that covering the neighboring matrix, and the protective influence is only effective until the local surface is broken in some way (e.g. by scratching or dissolution). When a 2024-A1 alloy sample is exposed to the acidic ZPO solution, it is expected that oxide on top of the Al-Cu-Mg particles dissolves faster than that covering the matrix. A consequence of the preferential dissolution of Mg and A l from the particle is an enrichment of Cu at the surface, and this has a dual effect on the subsequent ZPO precipitation to form the coating. First, the particle's surface becomes rougher as the dealloying creates morphologically favored sites for nucleation of the coating crystals, Second, Cu promotes the cathodic reaction for hydrogen evolution [204], with a consequent increase in the local pH. This drives the precipitation of coating compared to other nearby regions. The fact that the coating initiates at the edge of the particle, and eventually covers the whole surface, is likely to be associated with the faster oxide dissolution and dealloying occurring at the border region, due to galvanic coupling between particle and matrix. This conclusion is reinforced by the SEM image of a coated Al-Cu-Mg particle taken at tilted angle 143 (Fig. 5.6(a)) that shows the existence of a trench around the particle, an indication of severe matrix dissolution. Also the Auger point analysis made after the coating is completed showed that its thickness is least in the matrix region closest to the particle. In addition, SAM depth profiles from a coated particle confirm that Cu enriches at the interface, between the ZPO coating and Al-Cu-Mg particle, where the Cu concentration is approximately twice that of the bulk intermetallic compound. Enrichment by Cu, and other alloying elements, has been shown to occur in the alloy matrix after anodizing and other chemical pre-treatments [205-207], and it also affects the morphology and distribution of a non-commercial ZPO coating on 2024-Al alloy [99]. Figure 5.16 summarizes schematically the mechanism which we believe operates for the initiation of a crystalline ZPO coating on the Al-Cu-Mg particles. Not all Al-Cu-Mg particles become coated to the same extent during a constant exposure to the coating solution. Some small portions remain covered by an amorphous coating. This could result from incomplete oxide dissolution in the initial coating stage, and the provision of sites with local pH insufficient for precipitation of crystalline Z ^ ^ O ^ * 4 H 2 O . It is not yet clear whether the amorphous coating can serve as a precursor for the formation of crystalline ZPO or whether, once formed, it blocks further particle activity. One previous study on the mechanism of ZPO formation on aluminum proposed that the first step involves formation of an aluminum phosphate layer on top of the aluminum oxide/hydroxide that precedes the crystalline ZPO precipitation [189]; another view is that so-called "Zn sponges" are deposited first on the aluminum surface and that they subsequently act as nucleation sites for the crystalline ZPO coating [188]. Since some particles have a partially amorphous coating, even after the main coating process is completed, it is concluded that the coating initiation does depend on the particular microarea surface composition. In principle, the depth of near-surface particles is also likely to be a factor in influencing local coating behavior, but the size, distribution and activity of 144 I n c o a t i n g s o l u t i o n : M g 2 + , A P + c u r i c h 1. a m o r p h o u s c r y s t a l l i n e Z P O Z P O 3. \ i A l - C u - M g 4. c o m p l e t e p a r t i c l e Z P O c o v e r a g e A l - C u - M g Figure 5.16 Schematic indica t ion o f mechanism for the in i t ia t ion o f Z P O coat ing o n A l - C u -M g second-phase part icle: (1) before coat ing the mat r ix and particle are covered w i t h different oxides . (2) Af te r i m m e r s i o n i n so lu t ion , e tching gives C u enrichment especia l ly d r iven by galvanic coup l ing between part icle and matr ix . (3) E x p o s e d C u - r i c h surface regions act as cathodic sites for coat ing precipi ta t ion. (4) O n comple t ion the particle surface is covered w i t h Z P O w h i l e a C u - r i c h interface exists between coat ing and part icle. 145 nearby particles may additionally contribute to the chemistry applying at a particular microarea of interest. According to our observations, the formation of crystalline ZPO coating on the Al-Cu-Fe-Mn particles that are cathodic in nature is delayed during the initial stage of precipitation until the coating is established on the more active Al-Cu-Mg particles. This could be associated with differences in their respective oxide layers, and specifically that those on the Al-Cu-Fe-Mn particles may provide better resistance to the initial etching compared to the oxide on top of the Al-Cu-Mg particles. But as the process proceeds, eventually the entire surface becomes coated (Fig. 5.7(a)). It is clear that a reduction in the pH of the coating bath to 3.0 has a significant influence on the ZPO coating obtained on the 2024-Al alloy. The conditions have not been optimized for the precipitation of tertiary zinc phosphate, but the coating obtained at pH 3.0 is thinner compared to that obtained from the bath at pH 4.0 (Section 5.3.3). The former has a poorer crystalline coverage and is more amorphous in nature. The chemical composition for coating at pH 3.0 indicates the presence of N i 2 + at all measured locations, while Cu is detected only from the matrix. Further study is needed in order to determine the role of Ni in the coating chemistry of aluminum alloys, as well as the mechanism of Cu redistribution during the phosphating process. In this study, no alloy pre-treatment other than mechanical polishing has been done prior to immersion in the ZPO coating solution. Therefore, all observations relate to coating initiation that has not been promoted by the presence of nucleation reagents (e.g. as provided by a Ti pre-treatment step [84-86]). Future work will focus on analysis of the mechanism of coating formation on aluminum alloy surfaces pre-treated in different ways. But even more basic research is needed for the natures of the different micro-regions of the 2024-A1 alloy surface 146 prior to application of the coating process. For example, better knowledge is needed for the natures of the oxides involved. 5.5 Concluding Remarks The research in this Chapter attempts to strengthen an underdeveloped area in materials science, namely that associated with how chemically different microstructural regions of an alloy surface react when exposed to a common environment. The work emphasizes the initiation of zinc phosphate conversion coatings deposited on to a 2024-A1 alloy sample, and the following observations have been made on coated samples through study with various surface analysis techniques: 1. The 2024-Al alloy microstructure strongly affects the initial formation of a crystalline ZPO coating. Such a coating is observed to initiate on the Al-Cu-Mg second-phase particles, rather than on the alloy matrix or the Al-Cu-Fe-Mn-containing particles. After the process is completed, the ZPO coating covers the whole sample, as viewed by SEM. 2. At the initial stage of ZPO coating, the Al-Cu-Fe-Mn second-phase particles are covered with a thinner and less crystalline (i.e. more amorphous) type of coating. The differences in coverage and thickness of the ZPO coating on this type of particle, compared to the coating on the Al-Cu-Mg particles, may result from differences in the natures and etching rates of the oxide layers initially present. 3. The crystalline ZPO coating nucleates at the boundary region between Al-Cu-Mg particles and alloy matrix; that is where the etching rate is likely to be greatest, as driven by the associated galvanic coupling. The selective dissolution of Mg and A l results in Cu-rich sites, which become progressively more cathodic, thereby ftirther driving the precipitation of the ZPO crystalline coating. The latter occurs across the whole of the particle's surface; correspondingly the coating-particle interface is left in a Cu-rich form. 147 4. Some Al-Cu-Mg particles are coated by an amorphous ZPO coating. At this point it is not known what factors lead to the formation of this type of coating (e.g. initial incomplete oxide removal by etching), nor is it known how the initial amorphous deposit affects the nature of the subsequent coating. 5. The A l K L L Auger spectra measured from different microstructural regions on the alloy surface show different shapes, and this emphasizes that the electronic nature of A l varies when in the matrix and when bonded to the other alloying elements in the intermetallic compounds (second-phase particles). 6. The coating formed at pH 3 is thinner than that formed at pH 4; the former coating also has a poorer crystalline coverage and higher amorphous content. At pH 3, there is a build up of Cu on the matrix regions. The presence of N i (from an additive in the coating solution) is detected at all measured sites, but its concentration varies with the microstructural location. 7. The grain orientation appears to affect the coating process, although it is hard to separate that effect from the chemical modifications occurring during the etching process. 8. Initial observations related to the comparison of the oxidation processes on the alloy matrix and Al-Cu-Mg particle under U H V conditions, after coating and sputtering, indicate that the matrix builds up an oxide layer faster than the particle. 148 Chapter 6 Concluding Remarks 6.1 New Results The objective o f the research described i n this thesis is to help establish scientif ic pr inc ip les to guide the development o f n e w coat ing procedures for the cor ros ion protect ion o f a l u m i n u m a l loys . The part icular system investigated has been the 2024-T3 a l loy , used for aerospace applicat ions, w i t h the ul t imate goal to learn about the differ ing chemistr ies occur r ing at the different microstructural regions o f the a l l oy surface w h e n exposed to a reactive environment, such as a coat ing solut ion. T h i s requires emphas iz ing the in i t ia t ion o f the coat ing process, since i t is l i k e l y that the nature o f a f ina l protective layer is inf luenced b y its in i t i a l fo rm. T h i s approach has been fo l l owed to learn h o w the surface microstructure influences the format ion o f z i nc phosphate ( Z P O ) convers ion coatings and the adsorption o f organosilanes o n the 2024-A1 a l loy . Resul ts i n this thesis show that the compos i t ion and dis t r ibut ion o f second-phase particles strongly affect the in i t ia t ion o f both Z P O coat ing and the adsorption o f bis-1 ,2-(triethoxysilyl)ethane ( B T S E ) , w h i l e the influence o f the microstructure is m u c h less prominent for the deposi t ion o f y-aminopropyl t r ie thoxysi lane ( y - A P S ) silane. In addi t ion, the nature o f the ox ide layer at l oca l micro-regions , and the details o f galvanic c o u p l i n g between the particles and matr ix , appear to be important for both coat ing processes. In the ac id ic Z P O coat ing solut ion, r emova l o f the protective ox ide film at A l - C u - M g intermetal l ics occurs q u i c k l y . A s a result, a preferential d i sso lu t ion o f M g and A l occurs f rom the surfaces o f these second-phase particles, and their natures change f rom being anodic to cathodic. The latter dr ives the precipi ta t ion o f the crystal l ine Z P O coat ing, and there is a C u enrichment at the part icle-coating interface. In contrast, an amorphous Z P O coat ing is formed at 149 the A l - C u - F e - M n particles. T h i s is be l ieved to result from an incomplete d isso lu t ion o f the protective ox ide at the latter intermetall ic compounds , and these loca l oxides have different properties from those o n the A l - C u - M g particles. It fo l lows that an increased density o f A l - C u -M g particles at a 2 0 2 4 - A l a l loy surface should be benef ic ia l for the in i t ia t ion o f the Z P O coat ing. A s the coat ing process proceeds, the Z P O coverage increases across the a l l oy surface, but the compet i t ion between coat ing deposi t ion and substrate etching, as governed b y the galvanic coup l ing , lowers the coat ing thickness at the par t ic le-matr ix interface regions. A s imi la r process is responsible for the non-uni form deposi t ion o f B T S E o n to the mechanica l ly -po l i shed a l loy . W h e n exposed to ac id ic B T S E solut ion, ma t r ix areas surrounding the second-phase particles suffer an increased ox ide etching, w h i c h ul t imately affects the silane deposi t ion. A s a result, B T S E adsorbs more o n matr ix areas further away from the particles, as w e l l as o n the particles themselves, compared to the amount deposited at the par t ic le-matr ix interface regions w i t h i n a few particle diameters. In consequence, the higher the density o f second-phase particles at the a l l oy surface, the larger the surface area w i t h reduced coverage. In addi t ion, the properties o f the ox ide f i lms at the loca l intermetal l ic regions appear to govern the adsorption o f B T S E o n them. U p o n a i r -oxidat ion, the thickness o f the ox ide f i l m increases and more B T S E adsorbs, but the coat ing becomes very non-uni form. The ox ide properties and B T S E coverage appear to be governed b y factors other than the dis t r ibut ion o f second-phase particles; poss ib ly the l oca l gra in orientat ion ul t imately determines the B T S E adsorption. The influence o f a l l oy microstructure o n the adsorption o f y - A P S appears s m a l l or negl ig ib le . D u e to an ab i l i ty to fo rm hydrogen bonds through the amino groups, y - A P S forms a f i l m o f un i fo rm thickness across a large sample area, but these films are k n o w n to p rov ide a poor cor ros ion protection. A p p l i c a t i o n o f silanes i n a post-treatment rinse o f phosphated a l l oy samples m a y represent a p romis ing alternative to the convent ional use o f ch romic ac id . The coatings after a B T S E rinse show comparable cor ros ion protect ion abil i t ies to those r insed i n dilute ch romic ac id . 150 Protect ion at the part icle-matr ix interface regions is especia l ly increased b y the B T S E post-treatment, but the failure mechan i sm for this coat ing appears different f rom that post-treated by dilute ch romic ac id . T h e research i n this thesis has been described i n three papers; one has been publ i shed [208], and two have been submitted for pub l ica t ion [209,210]. 6.2 Future Directions T h i s w o r k has contributed some n e w knowledge for the va ry ing chemistr ies at different microstructural regions o f a 2024-A1 a l loy , and for some aspects o f si lane and Z P O coat ing processes, but there is s t i l l a need for further research. F o r example , better understandings are needed for the influences o f the var ious surface treatment processes, such as ox ida t ion i n different cont ro l led environments, as w e l l as the effects o f anod iz ing , o r o f ac id and a lkal ine etching, o n the nature and compos i t i on o f the oxides at the different microstructural regions. Further, more knowledge is needed for h o w differences i n the oxides are reflected i n the adsorption o f B T S E and other silanes, especia l ly regarding the format ion o f direct S i - O - M chemica l bonds. In addi t ion, studies are needed for the degradation mechanisms o f coatings at the different microstructural regions w h e n exposed to corrosive environments . Future directions for the research o n z inc phosphate coatings o n 2024-A1 a l loy w i l l inc lude studies o n h o w surface modif ica t ions by the T i c o l l o i d pre-treatment [84-88] affect the different microstructural regions, and h o w it relates to coat ing in i t ia t ion. M u c h effort is s t i l l needed to b u i l d up a fu l l understanding o f the phosphating process, i nc lud ing o n different A l a l loys and the effects o f n e w post-treatment rinses. These ideas extend other types o f protective coatings. A l t h o u g h there is strong scientif ic cur ios i ty i n the research described here, one mot iva t ion is certainly to develop n e w coatings that can replace the C r ( V I ) compounds that have t radi t ional ly been used for protective purposes. F o r example , Ce-based convers ion coatings 151 provide another possible route, and a recent publication [211] in that context described how coating formation on 2024-A1 alloy is influenced by both surface microstructure and the details of the pre-treatment. 152 References 1. D . G . A l t e n p o h l , A l u m i n u m : Technology , A p p l i c a t i o n s and Envi ronment A Prof i le o f a M o d e r n M e t a l , A l u m i n u m Assoc ia t ion , Inc. Wash ing ton , D . C . 1998. 2. J .R . D a v i s (Ed. ) , C o r r o s i o n o f A l u m i n u m and A l u m i n u m A l l o y s , A S M International, Mater ia l s Park, O H , 1999. 3. S. W e r n i c k , R . Pinner and P . G . Sheasby, The Surface Treatment and F i n i s h i n g o f A l u m i n i u m and its A l l o y s , v o l 1-2, F i n i s h i n g Publ ica t ions L T D . Teddington , M i d d l e s e x (1987). 4. J .R . D a v i s (Ed. ) , A S M Special ty Handbook A l u m i n u m and A l u m i n u m A l l o y s , A S M International, Mater ia l s Park, O H , 1993. 5. K . R . V a n H o r n (Ed. ) , A l u m i n u m V o l . 1. Properties, P h y s i c a l M e t a l l u r g y and Phase Diagrams , A m e r i c a n Society for Meta ls , Meta l s Park, O H , 1967. 6. S . L . C h a w l a and R . K . Gupta , Mater ia l s Select ion for C o r r o s i o n C o n t r o l , A S M International, Mate r ia l s Park, O H , 1993. 7. W . D . Cal l i s te r , Jr., Mater ia l s Science and Eng inee r ing an Introduction, John W i l e y & sons, Inc. N Y , 2000. 8. M e t a l s Handbook 8 t h edi t ion v o l 8, Meta l lography , Structures and Phase Diagrams, A m e r i c a n Socie ty for Meta l s , Meta l s Park, O h i o , 1973. 9. L . F . M a n d o l f o , A l u m i n u m A l l o y s : Structure and Properties, But terwoths, L o n d o n , 1976. 10. A . F . W e l l s , Structural Inorganic Chemis t ry , O x f o r d U n i v e r s i t y Press, L o n d o n , 1962. 11. G . E . T h o m p s o n and G . C . W o o d , i n : Cor ros ion : A q u e o u s Processes and Passive F i l m s , J . C . S c u l l y (Ed. ) , A c a d e m i c Press, L o n d o n , pp 205-329 (1982). 12. G . E . Thompson , K . S h i m i z u and G . C . W o o d , Nature, 286 (1980) 4 7 1 . 13. G . E . Thompson , P . Doher ty and G . C . W o o d , J. Electrochem. Soc. 129 (1982) 1515. 14. G . M . B r o w n , K . S h i m i z u , K . Kobayash i , G . E . T h o m p s o n and G . C . W o o d , Corrosion Science, 34 (1993)2099 . 15. M . Pourba ix , A t l a s o f Elec t rochemica l E q u i l i b r i a i n Aqueous Solu t ions N A C E , Houston , 1974. 16. M . G . Fontana, C o r r o s i o n Engineer ing , M c G r a w - H i l l B o o k C o m p a n y , 1986. 17. D . A . Jones, Pr inc ip les and Prevent ion o f Cor ros ion , E n g l e w o o d C l i f f s N J : Prentice H a l l , 1996. 153 18. K . Wefers , Aluminum, 57 (1981) 722. 19. J .R . S c u l l y , D . E . Peebles, A . D . R o m i g , Jr., D . R . Frear and C R . H i l l s , Metall. Trans. A, 2 3 A (1992) 2642. 20 . C . B l a n c , B . L a v e l l e and G . M a n k o w s k i , Mater. Sci. Forum, 217-222 (1996) 1559. 2 1 . G . S . C h e n , M . G a o and R . P . W e i , Corrosion, 52 (1996) 8. 22 . C - M . L i a o , J . M . O l i v e , M . G a o and R . P . W e i , Corrosion, 54 (1998) 4 5 1 . 23 . R . T . F o l e y , Corrosion, 42 (1986) 277 . 24. T . Bies tek and J . Weber , C o n v e r s i o n Coat ings , Po r t i cu l l i s Press, R e d h i l l , 1976. 25 . C Ede leanu and U . R . Evans , Trans. Faraday Soc. 47 (1951) 1121. 26. M . W . K e n d i g , A . J . Davenpor t and H . S . Isaacs, Corros. Sci. 34 (1993) 4 1 . 27. J . K . H a w k i n s , H . S . Isaacs, S . M . H e a l d , J . Tanquada, G . E . T h o m p s o n and G . C . W o o d , Corros. Sci. 27 (1987 ) 391 . 28. H . A . K a t z m a n , G . M . M a l o u f , R . Baue r and G . Stupian, Appl. Surf. Sci. 2 (1979) 416 . 29. F . W . L y t l e , R . B . Greegor , G . L . B i b b i n s , K . Y . B l o h o w i a k , R . E . S m i t h and G . D . Tuss , Corros. Sci. 37 (1995) 349. 30. J . Zhao , G . S . F ranke l and R . L . M c C r e e r y , J. Electrochem. Soc. 145 (1998) 2258. 31 . J . Zhao , L . X i a , A . Sehgal , D . L u , R . L . M c C r e e r y and G . S . F ranke l , Surf. Coat. Tech. 140 ( 2 0 0 1 ) 5 1 . 32. D . A . D i x o n , N . P . Sadler and T . P . Dasgupta, J. Chem. Soc. Dalton Trans. 23 (1993). 33 . R . L . T w i t e and G . P . B i e rwagen , Prog. Org. Coat. 33 (1998) 9 1 . 34. S . M . C o h e n , Corrosion 51 (1995) 71 . 35. B . R . W . H i n t o n , Met. Finish. 89 (1991) 15. 36. J . C o m y n , i n : A . J . K i n l o c h (Ed. ) , Structural Adhes ive s : Developments i n resins and Pr imers , E l sev i e r A p p l i e d Science Publ ishers L T D , L o n d o n , 1986 pp 269 . 37. E . P . P lueddemann, S i lane C o u p l i n g Agents , P l e n u m Press, N e w Y o r k , 1991. 38. K . L . M i t t a l (Ed. ) , Si lanes and Other C o u p l i n g Agen t s , V S P , Utrecht , 1992. 39. W . J . v a n Oo i j and T F . C h i l d , Chem. Tech. 28 (1998) 26. 40 . T . F . C h i l d and W . J . v a n O o i j , Trans IMF, 77 (1999) 64. 4 1 . W . J . v a n O o i j , D . Q . Z h u , G . Prasad, S. Jayaseelan, Y . F u and N . Teredesai , Surf. Eng. 16 (2000) 386. 42 . W . J . v a n O o i j , J . Song and V . Subramanian, ATB Metallurgie: Acta Technica Belgica, 37 (1997) 137. 154 43. J.D. Venables, D.K. McNamara, J.M. Chen, T.S. Sun and R.L. Hopping, Appl. Surf. Sci. 3(1979) 88. 44. A J . Kinloch, J. Mater. Sci. 15 (1980) 2141. 45. K.W. Allen, J. Adhes. 21 (1987) 261. 46. K.W. Allen, Int. J. Adhes. Adhes. 13 (1993) 67. 47. F.M. Fowkes, J. Adhes. 4 (1972) 155. 48. S.R. Cain, J. Adhes. Sci. Technol. 4 (1990) 333. 49. F.M. Fowkes, Physiochem. Aspects Polym. Surf. Proc. Int. Symp. 2 (1983) 583. 50. F.M. Fowkes, J. Adhes. Sci. Technol. 1 (1987) 7. 51. Z. Pu, W.J. van Ooij and J.E. Mark, J. Adhesion Sci. Technol. 11 (1997) 29. 52. F.J. Boerio, C A . Gosselin, J.W. Williams, R.G. Dillingham and J.M. Burkstrand, in: H. Ishida, G. Kumar (Eds.), Molecular Characterization of Composite Interfaces, Plenum, New York, 1985, pp 171. 53. Y. L. Leung, Y. P. Yang, P. C. Wong, K. A. R. Mitchell and T. Foster, J. Mat. Sci. Lett. 12 (1993) 884. 54. S.R. Morrison, The Chemical Physics of Surfaces, Plenum Press, New York, 1990, pp 285. 55. J.C. Bolger and A.S. Michaels, in: Interface Conversion for Polymer Coatings, P. Weiss and G.D. Cheever (Eds.), American Elsevier Publishing Company, Inc., New York, 1968. 56. E. McCafferty and J.P. Wightman, J. Colloid Interf. Sci. 194 (1997) 344. 57. J. Fang, B. J. Flinn, Y. L. Leung, P. C. Wong and K. A. R. Mitchell, J. Mat. Sci. Lett. 16 (1997) 1675. 58. U. Bexell and M . Olsson, Surf. Interface Anal. 31 (2001) 212. 59. M . Gettings and A. J. Kinloch, J. Mat. Sci. 12 (1977) 2511. 60. M Gettings and A. J. Kinloch, Surf. Interface Anal. 1 (1979) 189. 61. M . L. Abel, R. P. Digby, I. W. Fletcher and J. F. Watts, Surf. Interface Anal. 29 (2000) 115. 62. U. Bexell and M . Olsson, Surf. Interface Anal. 31 (2001) 223. 63. F. J. Boerio and J. W. Williams, Appl. Surf. Sci. 9 (1981) 19. 64. F. J. Boerio, C. A. Gosseilin, R. G. Dillingham and H.W. Liu, J. Adhes. 13 (1993) 159. 65. J. Ondrus and F. J. Boerio, J. Coll. Interf. Sci. 124 (1998) 349. 66. J. D. Miller and I. Ishida, J. Phys. Chem. 86 (1987) 1593. 155 67. Y . L . Leung, M . Y . Zhou, P. C. Wong, K . A . R. Mitchell and T. Foster, Appl. Surf. Sci. 59 (1992) 23. 68. T.W. Coslett, BP 8667 (1906). 69. T.W. Coslett, Fr 376536 (1907). 70. T.W. Coslett, USP 870937 (1907). 71. T.W. Coslett, DRP 209805 (1909). 72. De-Bussy,BP43(1864). 73. A . B . Brown and W.P. Brown, DRP 6998 (1879). 74. G.Ross, BP 3119 (1869). 75. W. Rausch, The Phosphating of Metals, A S M International, Ohio, 1990. 76. D.B. Freeman, Phosphating and Metal Pre-Treatment, Industrial Press Inc, 1986. 77. G. Lorin, Phosphating of Metals: Constitution, Physical Chemistry and Technical Applications of Phosphating Solutions, Hampton Hi l l : Finishing Publications, 1974. 78. S. Spring, Preparation of Metals for Painting, Reinhold Publishing Corporation, New York, 1965. 79. T.S.N. Sankara Narayanan, Corr. Rev. 12 (1994) 201. 80. J. Donofrio, Met. Finish. 100/16A (2002) 89. 81. H.S. Bender, G.D. Cheever and J.J. Wojtkowiak, Prog. Org. Coat. 8 (1980) 241. 82. D. James and D.B. Freeman, Transactions of the Institution of Metal Finishing, 49 (1971) 79. 83. T. Dong Van, N . Bui, K . V u Quang and F. Dabosi, British Corrosion Journal, 29 (1994) 305. 84. P.E. Tegehall, Colloids Surf. 42 (1989) 155. 85. P.E. Tegehall, Colloids Surf. 49 (1990) 373. 86. I. Van Roy, H.Terryn and G. Goeminne, Trans IMF, 76 (1998) 19. 87. I. Van Roy, H.Terryn and G. Goeminne, Colloids and Surfaces A, 136 (1998) 89. 88. I. Van Roy, H.Terryn and G. Goeminne, ATB Metallurgie, 37 (1997) 161. 89. H . Ishii, O. Furuyama and S. Tanaka, Met. Finish. 91 (1993) 7. 90. W. Machu, in: Interface Conversion for Polymer Coatings, P. Weiss and G.D. Cheever, (Eds.), American Elsevier Publishing Company, Inc. New York (1968) pp 128. 91. W.L. Masterton, E.J. Slowinski and C L . Stanitski, Chemical Principles With Quantitative Analysis, Saunders College Publishing, Philadelphia, 1986. 92. W.J. Wulfson, Korroziya i Borba s Nei, 5 (1939) 1. 156 93. V . Cupr and J.B. Pelikan, Metalloberflache, 19 (1965) 187-192,230-231. 94. A . Kozlowski, Galvano, 41 (1972) 345. 95. A . Kozlowski, Galvano, 42 (1973) 153. 96. 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. 97. L . Shi, Zinc Phosphating on 6061-T6 Aluminum Alloy, M.Sc Thesis, University of British Columbia (2000). 98. W.H. Kok, X . Sun, L. Shi, K . C . Wong, K .A .R . Mitchell and T. Foster, J. Mat. Sci. 36 (2001)3941. 99. X . Sun, W.H. Kok, K . C . Wong, R. L i , K .A.R . Mitchell and T. Foster, ATB Metallurgie 40-41 (2000/2001) 503. 100. Y . L . Leung, XPS Studies of Adhesion at Organosilane/Aluminum Interfaces, M.Sc Thesis, The University of British Columbia, 1992. 101. M . Kono, X . Sun, R. L i , K . C . Wong, K.A.R. Mitchell and T. Foster, Surf. Rev. & Letters 8 (2001) 43. 102. L L . Mueller and J.R. Galvele, Corr. Sci. 17 (1977) 179. 103. J.R. Galvele and S.M. DeMicheli, Corr. Sci. 10 (1970) 795. 104. A . Garner and D. Tromans, Corrosion, 35 (1979) 55. 105. P. Campestrini, E .P .M. van Westing, H.W. van Rooijen and J.H.W. de Wit, Corr. Sci. 42 (2000)1853. 106. P. Leblanc and G.S. Frankel, J. Electrochem. Soc. 149 (2002) B 239. 107. T.J. Warner, M.P. Schmidt, F. Sommer and D. Bellot, Z. Metallkd. 86 (1995) 7. 108. M . A . Alodan and W.H. Smyrl, J. Electrochem. Soc. 145 (1998) 1571. 109. V . Guillaumin and G. Mankowski, Corr. Sci. 41 (1999) 421. 110. T. Suter and R.C. Alkire, J. Electrochem. Soc. 148 (2001) B 36. 111. D. Briggs and M.P. Seah (Eds.), Practical Surface Analysis, vol 1, John Wiley & Sons Inc. 1990. 112. J.D. Andrade (Ed.), Surface and Interfacial Aspects of Biomedical Polymers, vol 1, Surface Chemistry and Physics, Plenum Press, 1985. 113. D. Brune, R. Hellborg, H.J. Whitlow and O. Hunderi (Eds.), Surface Characterization, A Users's Sourcebook, Wiley-VCH, 1997. 114. J.C. Riviere and S. Myhra (Eds.), Handbook of Surface and Interface Analysis, Marcel Dekker, Inc. New York, 1998. 157 115. J.C. Vickerman (Ed.), Surface Analysis-The Principal Techniques, John Wiley & Sons Inc. 1997. 116. H . Hertz, Ann. Physic 3 1 (1887) 983. 117. A . Einstein, Ann. Physic 1 7 (1905) 132. 118. H . Robinson and W.F. Rawlinson, Phil Mag. 2 8 (1914) 277. 119. K . Sieghbahn, C.N. Nordling, A . Fahlman, R.Nordberg, K . Hamrin, J. Hedman, G. Johansson, T. Berrnark, S.E. Karlsson, I. Lingren and B. Lindberg, ESCA: Atomic, Molecular and Solid State Studies by Means of Electron Spectroscopy, Almqvist and Wiksells, 1967. 120. P. Auger, Comp. Rend. 1 7 7 (1923) 169. 121. P. Auger, Comp. Rend. 1 8 0 (1923) 65. 122. P. Auger, Ann. Phys. 6 (1926) 183. 123. A . Rosseland, Z. Phys. 1 4 (1923) 172. 124. A.T. Hubbard (Ed.), Surface Imaging and Visualization, CrC Press, Inc. 1995. 125. M.P. Seah and W.A. Dench, Surf. Interf. Anal. 1 (1979) 2. 126. G.A. Somorjai, Chemistry in Two Dimensions: Surfaces, Cornell University, New York, 1981. 127. C D . Wagner, W . M . Riggs, L .E . Davis, J.F. Moulder and G.E. Muilenberg (Eds.), Handbook of X-Ray Photoelectron Spectroscopy, Perkin-Elmer Corporation, 1979. 128. V G Scientific Auger Handbook, V G Scientific Limited, West Sussex, England, 1989. 129. L .E . Davis, N . C MacDonald, P.W. Palmberg, G.E. Riach and R.E. Weber, Handbook of Auger Electron Spectroscopy, Perkin-Elmer Corporation, Eden Prairie, 1976. 130. Y . L . Leung, Surface Studies of Planar Model H D N Catalysts, PhD. Thesis, The University of British Columbia, 1998. 131. LR. Barkshire, M . Prutton and D.K. Skinner, Surf. Interface Anal. 1 7 (1991) 213. 132. L . Reimer, Scanning Electron Microscopy Physics of Image Formation and Microanalysis, Springer-Verlag Berlin Heidelberg, 1985. 133. D.A. Shirley, Phys. Rev. B 5 (1972) 4709. 134. Thermo V G Scientific, Microlab 350 Operating Manual. 135. Web site: http://vv^vw.feibeamtech.com/pages/schottkv.html 136. J.J. Bozola and L.D. Russell, Electron Microscopy: Principles and Techniques for Biologists, Jones and Bartlett Publishers, 1999. 158 137. D.B. Williams, Practical Analytical Electron Microscopy in Materials Science, Philips Electronic Instruments Inc. Electron Optics Publishing Group, 1984. 138. J.C. Russ, Fundamentals of Energy Dispersive X-ray Analysis, Butterworths & Co Ltd., 1984. 139. Energy Dispersive X-ray Microanalysis, Thermo N O R A N , 2001. 140. D.A. Jones, Principles and Prevention of Corrosion, Prentice-Hall, Inc. 1996. 141. E.E. Stansbury, R.A. Buchanan, Fundamentals of Electrochemical Corrosion, Materials Park, OH, 2000. 142. A .C . Fisher, Electrode Dynamics, Oxford University Press, 1996. 143. V . Subramanian, W.J. van Ooij, Corrosion, 54 (1998) 204. 144. K . L . Mittal (Ed.), Adhesion Aspects of Polymeric Coatings, Plenum, New York, 1983. 145. J.S. Solomon, N.T. McDevitt, Thin Solid Films, 84 (1981) 155. 146. L . Kozma, I. Olefjord, Mater. Sci. Technol. 3 (1987) 860. 147. H . Habazaki, K. Shimizu, P. Skeldon, G.E. Thompson, G.C. Wood and X . Zhou, Trans IMF, 75 (1997) 18. 148. X . Zhou, G.E. Thompson, P. Skeldon, G.C. Wood, K . Shimizu and H . Habazaki, Corr. Sci. 41 (1999) 1089. 149. C. Towler, R.A. Collins and G. Dearnaley, J. Vac. Sci. Technol. 12 (1975) 520. 150. K . K . Soni, D.B. Williams, J .M. Chabala, R. Levi-Setti and D.E. Newbury, Oxid. Metals, 37 (1992) 23. 151. F. Gesmundo, F. Viani, Y . Niu and D.L. Douglass, Oxid. Metals, 39 (1993) 197. 152. T.S. Sun, J .M. Chen, J.D. Venables and R. Hopping, Appl. Surf. Sci. 1 (1978) 202. 153. P.W. Atkins, Physical Chemistry, Freeman, New York, 1990. 154. J.S. Quinton and P. Dastoor, Surf. Interface Anal. 28 (1999) 12. 155. G.A. Parks, Chem. Rev. 65 (1964) 127. 156. E.P. Plueddemann, G.L. Stark, Modern Plastics, March (1974) 74. 157. M . A . Petrunin, A.P. Nazarov and Yu.N. Mikhailovskii, Protection of Metals, 26 (1990) 749. 158. R.G. Buchheit, M . A . Martinez and L.P. Montes, J. Electrochem. Soc. Ul (2000) 119. 159. R.G. Buchheit, R.P. Grant, P.F. Hlava, B. Mckenzie and G.L. Zender, J. Electrochem. Soc. 144(1997) 2621. 160. J. Quinton, L . Thomsen and P. Dastoor, Surf. Interface Anal. 25 (1997) 931. 161. J. Quinton, P. Dastoor and W. Allison, Surf. Sci. 402-404 (1998) 66. 159 162. J.S. Quinton, P.C. Dastoor, Surf. Interface Anal. 30 (2000) 21. 163. G. Gorecki, Met. Finish. Organic Finishing Guidebook 98 (2000) 97. 164. R. Seidel, K-D. Brands and K - H . Gottwald, US Patent 5,391,240. 165. W.J. Claffey and A.J . Reid, US Patent 4,650,526. 166. W.J. Claffey and A.J . Reid, US Patent 4,656,097. 167. J. Affinito, US Patent 5,662,746. 168. G. Gorecki, US Patent 5,531,820. 169. N . Tang, W.J. van Ooij and G. Gorecki, Prog. Org. Coat. 30 (1997) 255. 170. S. Maeda and M . Yamamoto, Prog. Org. Coat. 33 (1998) 83. 171. S. Maeda, M . Yamamoto, T. Asai, H . Asano and H . Okada, in: Organic Coating Science and Technology, G.D. Parfit and A . D Patis (Eds.), vol 7, Marcel Decker, New York, 1984, pp 223. 172. A . E . Hughes, R J . Taylor and B.R.W. Hinton, Surf. Interf. Anal. 25 (1997) 223. 173. J . A . Treverton and N.C. Davies, Met. Technol. 4 (1977) 480. 174. T. Hara, M . Terasaka and S. Doi, Nippon Kokan Tech. Rep. 43 (1985) 40. 175. L . X ia and R.L. McCreery, J. Electrochem. Soc. 145 (1998) 3083. 176. H . Ishii and K . Hiroshi, U.S. Pat. 5,795,407 (1998). 177. K . K . Meagher and B.B. Cuyler, U.S. Pat. 6,530,999 (2003). 178. M . L . Sienkowski and G.J. Cormier, U.S. Pat. 6,368,426 (2002). 179. M . Kawakami, N . Kobayashi and K. Oyama, U.S. Pat. 6,179,934 (2001). 180. C E . Rossio, U.S. Pat. 5,868,874 (1999). 181. T. Sugama, U.S. Pat. 5,604,040 (1997). 182. R. Seidel, H-D. Speckmann, K - D . Brands, G. Veldman and R. Mady, U.S. Pat. 5,401,381 (1995). 183. H . Ishii, J. Surf. Finish. Soc. (Japan) 48 (1997) 961. 184. X . Sun, D. Susac, R. L i , K . C . Wong, T. Foster and K.A.R . Mitchell, Surf. Coat. Tech. 155 (2002) 46. 185. W.F. Heung, P . C Wong, K.A.R. Mitchell and T. Foster, J. Mat. Sci. Lett. 14 (1995) 1461. 186. W.F. Heung, Y.P. Yang, M . Y . Zhou, P.C. Wong, K .A.R . Mitchell and T. Foster, J. Mat. Sci. 29(1994) 3653. 187. W.F. Heung, Y.P. Yang, P . C Wong, K .A.R . Mitchell, and T. Foster, J. Mat. Sci. 29 (1994) 1368. 160 188. B. Cheng, S. Ramamurthy and N.S. Mclntyre, J. Mat. Eng. Perf. 6 (1997) 405. 189. M . Handke, A . Stoch, S. Sulima, P.L. Bonora, G. Busca and V . Lorenzelli, Mat. Chem. 7 (1982) 7. 190. G.D. Cheever, J. Paint. Tech. 39 (1967) 1. 191. P.H. Siddiqui, M . Khalid, A . A l i and A . ul Haq, Proc. 5th International Symposium on Advanced Materials (1997) 726. 192. V . Guillaumin, P. Schmutz and G.S. Frankel, J. Electrochem. Soc. 148 (2001) B163. 193. N . Satoh, Surf. Coat. Tech. 30 (1987) 171. 194. R.P. Wei, C - M . Liao and M . Gao, Metall. Mater. Trans. A, 29 (1998) 1153. 195. R.G. Buchheit, J. Electrochem. Soc. 142 (1995) 3994. 196. R.G. Buchheit, L.P. Montes, M . A . Martinez, J. Michael and P.F. Hlava, J. Electrochem. Soc. 146(1999) 4424. 197. R.G. Buchheit, R.P. Grant, P.F. Hlava, B . McKenzie and G.L. Zender, J. Electrochem. Soc. 144(1997) 2621. 198. N . Dimitrov, J.A. Mann and K . Sieradzki, J. Electrochem. Soc. 146 (1999) 98. 199. P. Schmutz and G.S. Frankel, J. Electrochem. Soc. 145 (1998) 2285. 200. P. Schmutz and G.S. Frankel, J. Electrochem. Soc. 145. (1998) 2295. 201. M.C . Reboul, T.J. Warner, H . Mayet and B. Baroux, Mat. Sci. Forum, 217-222 (1996) 1553. 202. C E . Moffitt, D . M . Wieliczka and H.K. Yasuda, Surf. Coat. Tech. 137 (2001) 188. 203. A.J . Kinloch, H.E. Bishop and N.R. Smart, J. Adhesion, 14 (1982) 105. 204. A.J . Bard, Encyclopedia of Electrochemistry of the Elements, Marcel Dekker, New York, 1984. 205. H . Habazaki, M . A . Paez, K . Shimizu, P. Skeldon, G.E. Thompson, G.C. Wood and X . Zhou, Corr. Sci. 38 (1996) 1033. 206. M . A . Paez, T .M. Foong, C T . N i , G.E. Thompson, K . Shimizu, H . Habazaki, P. Skeldon and G.C. Wood, Corr. Sci. 38 (1996) 59. 207. H.H. Strehblow, C M . Mellisar-Smith and W . M . Augustyniak, J. Electrochem. Soc. 125 (1978)915. 208. D. Susac, X . Sun and K.A.R . Mitchell, Appl. Surf. Sci. 207 (2003) 40. 209. D. Susac, C W . Leung, X . Sun, K . C . Wong and K.A.R . Mitchell, submitted to Surf. Coat. Tech. 210. D. Susac, X . Sun, R. L i , P . C Wong and K.A.R. Mitchell, submitted to Appl. Surf. Sci.. 161 211. P. Campestrini, H . Terryn, A . Hovestad and J.H.W. de Wit, Surf. Coat. Tech. (article in press). 212. C S . Barrett and T.B. Massalski, The Structure of Metals, Crystallographic Methods, Principles and Data, Pergamon Press, Toronto ON, 1980. 162 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061188/manifest

Comment

Related Items