UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An investigation of the penetration of thin aluminum oxide films by liquid bismuth Allday, William John 1961

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

Item Metadata

Download

Media
831-UBC_1961_A7 A5 I6.pdf [ 5.59MB ]
Metadata
JSON: 831-1.0105831.json
JSON-LD: 831-1.0105831-ld.json
RDF/XML (Pretty): 831-1.0105831-rdf.xml
RDF/JSON: 831-1.0105831-rdf.json
Turtle: 831-1.0105831-turtle.txt
N-Triples: 831-1.0105831-rdf-ntriples.txt
Original Record: 831-1.0105831-source.json
Full Text
831-1.0105831-fulltext.txt
Citation
831-1.0105831.ris

Full Text

AN INVESTIGATION OF THE PENETRATION OF THIN ALUMINUM OXIDE FILMS BY LIQUID BISMUTH. by WILLIAM JOHN ALLDAY A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n the Department of MINING AND METALLURGY We accept t h i s thesis'" as conforming to the standard required from candidates f o r the degree of MASTER OF APPLIED SCIENCE. Members of the Department of Mining and Metallurgy. THE UNIVERSITY OF BRITISH COLUMBIA A p r i l , 196l . In presenting t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree that permission f o r extensive copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted'by the Head of my Department or by h i s r e p r e s e n t a t i v e s . I t i s understood tha t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n permission. Department of Mining and Metallurgy The U n i v e r s i t y of B r i t i s h Columbia, Vancouver $, Canada. Date April k, 1961 - i -ABSTRACT The penetration of t h i n aluminum oxide f i l m s by l i q u i d bismuth was studied between 35° and 500°C. A c o r r e l a t i o n was sought between the type and thickness of the oxide f i l m , and the time and nature of the attack. Mech-a n i c a l and electropolished, and anodized f i l m s of d i f f e r e n t thicknesses were exposed to l i q u i d bismuth e i t h e r by immersing a p l a t e i n the l i q u i d and measuring the contact angle or by melting a drop of bismuth on a s p e c i f i c area of a p l a t e . E l e c t r o p o l i s h e d surfaces had no resistance to attack by the bismuth. P i t t i n g and edge attack occurred with a l l other surface preparations. The anodized f i l m s often separated from the aluminum under thermal stresses and allowed the bismuth to spread under the f i l m , sometimes removing i t e n t i r e l y . Attempts were- made to c o n t r o l the number and nature of defects i n the oxide f i l m , but the d i f f i c u l t y of t h i s i s shown by the s c a t t e r i n the r e s u l t s . Wo.attack on aluminum oxide i t s e l f i s l i k e l y and no d i f f u s i o n of bismuth through.the oxide was found at the temperatures used. The conclusion was reached that the bismuth penetrated the oxide f i l m only at points where there was a high concentration i n the f i l m of some impurity oxide (such as n i c k e l or copper oxides) that reacts with bismuth. - i i -ACKNOWLEDGEMENT The author i s indebted to Professor W. M. Armstrong, Dr. J . B. Raynor and Mr. M. J . Finlayson f o r t h e i r supervision and encouragement, and to Mr. R. Butters f o r his t e c h n i c a l advice and assistance. The many valuable discussions with Dr. A. C. D. Chaklader and Mr. K. G. Davis are also grate-f u l l y acknowledged. The work was financed by Research Grants DRB 7510-32 and DRB 7510-36 provided by the Defence Research Board of Canada. - i i i -TABLE OF CONTENTS Page I. INTRODUCTION . 1 A. Uses of L i q u i d Metals 1 B. Corrosion by L i q u i d Metals . . . . . . . . . . . . . 3 C. Previous Investigations on L i q u i d Bismuth 5 D. Previous Work on Thin Oxide Films on Aluminum 8 E. S p e c i f i c Aims of the Present I n v e s t i g a t i o n 13 I I . EXPERIMENTAL Part 1. Ik A. Properties of L i q u i d Bismuth Ik B. Laboratory Handling of L i q u i d Bismuth l 6 C. D e s c r i p t i o n of Apparatus 20 D. Experimental Procedure . 28 E. Theory 29 F. Analysis of Results 31 I I I . EXPERIMENTAL • Part 2 kO A. Description of Apparatus kO B. Experimental Procedure kl C. Analysis of Results k2 D. Discussion of Results V7 IV. EXPERIMENTAL Part 3 . 52 A. D e s c r i p t i o n of Apparatus 52 B. Experimental Procedure 55 C. Analysis of Results 56 - i v -Page V o CONCLUSION . o o . o . . . . . . '. o . . . . . . . „ . . . „ . 6 l VI. RECOMMENDATIONS FOR FURTHER WORK . . . . . . . . . . . . . . . 6h VII. APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . 65 VIII. BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . 71 - V -FIGURES No. Page 1 = Surface tension of l i q u i d bismuth, under d i f f e r e n t atmospheres . 6 2 . V a r i a t i o n of equilibrium, contact angle with temperature f o r titanium and zirconium . . . . . . . . . . . . . . . . . . 7 3 = Rate of oxide f i l m growth on aluminum . . . . . . . . . . . . . 1 0 4 . Growth of oxide on aluminum. . . . . . . . . . . . . . . . . . . 1 0 5 o F i l m growth i n ammonium t a r t r a t e e l e c t r o l y t e . „ . „ . . . . . 1 2 6. Water vapour adsorption by molecular sieves . . „ . . . « . . . 1 8 7 o Magnetic balance assembly . . . . . . . . . . . . . . . . . . . . . 2 1 8 . Magnetic balance c i r c u i t diagram . . . . . . . . . . . . . . . 2 3 9 . C a l i b r a t i o n of magnetic balance . . . . . . . . . . . . . . . . 24 1 0 . Drawing o f the glass apparatus . . . . . . . . . . . . . . . . . 25 1 1 . Plate attachment . . . . . . . . . . . . . . . . . . . . . . . 2 . 8 1 2 . Forces on the p l a t e i n l i q u i d bismuth 3 0 1 3 « T y p i c a l p l o t of contact angle v time . 3 4 1 4 . Wetting behaviour of Alcan 2S . . . . . . . . . 3 6 15. Wetting behaviour of mechanically polished 9 9 « 9 8 $ Aluminum . . 3 7 1 6 . Mechanically p o l i s h e d 9 9 - 9 8 $ aluminum after.immersion i n bismuth 3 8 1 7 . Mechanically polished 9 9 » 9 8 $ aluminum a f t e r immersion i n bismuth 4 3 1 8 . V e r t i c a l s e c t i o n of p l a t e 1 9 showing bismuth under the oxide 1 9 . Section of plate 3 3 showing p i t t i n g and intergranular attack . 4 4 2 0 o Plates 2 4 , 2 7 and 2 3 a f t e r immersion i n l i q u i d bismuth at 4 0 0 ° C . 46 / 2 1 . Plates 25, 2 1 and 1 2 a f t e r immersion i n l i q u i d bismuth . . . . . 46 2 2 . Plate 3 2 a f t e r immersion i n l i q u i d bismuth f o r 2 0 minutes at ^ 0 0 ^ 0 o o o o o o o o o o o o o © o o » e c » o o o a o * 5 ^ - v i No. Page 2 3 o Section of plate 3 2 , showing a defect In the oxide layer . . 5 0 2 4 . P a r t i a l section of tube furnace . . . . . . . . . . . . . . 5 3 2 5 . Section of bismuth drop, showing general shape . . . . . . 5 6 2 6 0 Defect i n oxide f i l m on plate D2 after penetration by l i q u i d bismuth „ . . . . . . . . 0 0 . 0 . 0 0 . . . „ 5 9 - v i i -TABLES No. Page I. Analysis of Bismuth . . . . . . . . . . . . . . . . . . . . lh I I . P h y s i c a l Properties of Li.quid. Bismuth . . . . . . . . . . . 15 I I I . S o l u b i l i t y of Bi203 In L i q u i d Bismuth . . . . . . . . . . . . 15 IV. Analyses of Aluminum Used . . . . . . . . . . . . . . . . . . 2.7 V. Sample C a l c u l a t i o n of Contact Angle . . . . . . . . . . . . . 33 M INVESTIGATION OF THE PENETRATION OF THIN ALUMINUM OXIDE FILMS BY LIQUID BISMUTH I. INTRODUCTION With "the coming of nuclear power, i n t e r e s t has grown i n the use of l i q u i d metals as heat-transfer media. The metals considered u s e f u l f o r t h i s purpose are those with melting points below 660°C, and include alum-inum, bismuth, cadmium, gallium, lead, l i t h i u m , magnesium, mercury, pot-assium, sodium, t i n and zinc, and the sodium-potassium and lead-bismuth eu t e c t i c a l l o y s . The use of l i q u i d metals may be r e s t r i c t e d by'the extent to which they corrode the s o l i d metals i n contact with them. To date most of the i n -ve s t i g a t i o n s i n t h i s f i e l d have been concentrated on sodium, potassium and t h e i r a l l o y s . Bismuth was chosen f o r the present i n v e s t i g a t i o n because of i t s i n c r e a s i n g importance. I t was decided to study i t s corrosive a c t i o n on aluminum as t h i s metal i s widely used i n e x i s t i n g nuclear reactors as a con-t a i n i n g m a t e r i a l . A. Uses of L i q u i d Metals P r i o r to the advent of nuclear power, sodium was the only l i q u i d metal used with any success, with the exception of the General E l e c t r i c Company's use, i n 1922, of mercury vapour, i n place of steam, to increase the e l e c t r i c power generating e f f i c i e n c y of turbines-1-.?^„ Sodium has been used since 1928 to cool valves i n a i r c r a f t engines^, cores f o r magnesium die cast-ing, and f o r many other a p p l i c a t i o n s i n v o l v i n g heat-transfer at high temp-eratures. In 19^3, the Dow Chemical Company operated a system containing l i q u i d sodium and recovered as much as one and a h a l f m i l l i o n Btu/hr (440kw) . - 2 -L i q u i d metals have various applications i n nuclear r e a c t o r s . As heat-transfer media, whether used as s t a t i c or dynamic coolants, l i q u i d metals have the advantage of good thermal conductivity, s t a b i l i t y under conditions of r a d i a t i o n and high temperature, and low vapour pressure; used as a f u e l c a r r i e r , whether as a solvent or dispersant of the f u e l , l i q u i d metals are s u i t a b l e f o r continuous processing; they present no problems of dimensional s t a b i l i t y and they possess good heat-transfer p r o p e r t i e s . They are also s u i t a b l e f o r use i n breeder blankets and as r a d i a t i o n s h i e l d s . Mercury was used as the coolant i n the Los Alamos f a s t reactor. NaK, a sodium-potassium a l l o y , has been used f o r thousands of hours i n the EBR (Experimental Breeder Reactor). Sodium i s the coolant f o r the SIR, the EBR-II, the SRE (sodium graphite) and i n Great B r i t a i n f o r the Dounreay experimental f a s t breeder'reactor of the U. K. A. E. A. Bismuth has been proposed f o r the LMFR ( L i q u i d Metal F u e l Reactor) power breeder. Russia i s i n t e r e s t e d i n the p o s s i b i l i t i e s of bismuth-lead a l l o y s . Nuclear reactors i n which l i q u i d sodium i s used s u f f e r from some disadvantages'^: -a) The e f f i c i e n c y of the s o l i d uranium f u e l rods i n the reactor w i l l u l t i m a t e l y decrease as f i s s i o n proceeds and f i s s i o n products accum-ulate . b) A higher operational temperature would be advantageous f o r higher heat-transfer e f f i c i e n c y and t h i s can only be obtained i f a coolant of higher b o i l i n g point i s used. The vapour pressure of sodium at T00°C i s 100 ramJ. I f the uranium f u e l and the f i s s i o n products were soluble i n a l i q u i d of high b o i l i n g point, the above disadvantages would be eliminated. Much work has been done on systems where a continuous cycle of f u e l , d i s s o l v e d i n a l i q u i d , - 3 -passes through the core of the reactor and then through some form of counter-current extraction process to remove the f i s s i o n products. A d d i t i o n a l quan-t i t i e s of f i s s i l e m a t e rial can he added at i n t e r v a l s and concentrations are e a s i l y c o n t r o l l e d , thus g i v i n g optimum e f f i c i e n c y . Of the possible solvents, only l i q u i d bismuth f u l f i l s these conditions and i s , i n f a c t , almost the id^.al m a t e r i a l . L i q u i d metal f u e l l e d reactors have been described i n d e t a i l by Frnst 8 Q and others ' . B. Corrosion by L i q u i d Metals L i q u i d bismuth i s one of the most v e r s a t i l e of a l l solvents. The only m e t a l l i c elements that can be regarded as i n s o l u b l e belov 600°C are beryllium, molybdenum, tantalum and tungsten. This r a i s e s obvious problems so f a r as containing vessels f o r l i q u i d bismuth are concerned. I t i s probably true to say that any l i q u i d metal w i l l completely wet any s o l i d metal provided that there i s nothing to prevent true metal-to-metal contact at the l i q u i d m e t a l / s o l i d metal i n t e r f a c e . A l l previous exper-ience i s i n agreement with t h i s assumption but i t i s d i f f i c u l t to v e r i f y completely"^. In a l l p r a c t i c a l cases, the presence of an oxide f i l m on the s o l i d metal w i l l prevent immediate wetting. Corrosion i s prevented i f : -a) the s o l i d metal does not dissol v e i n the l i q u i d metal. b) an impervious b a r r i e r can be introduced which prevents contact be-tween the l i q u i d metal and the s o l i d metal. This can either be the oxide of the s o l i d metal or a l a y e r of a compound such as zirconium n i t r i d e . For example, i n the l i q u i d bismuth-steel system, 250 p.p.m. Zirconium d i s s o l v e d i n the bismuth reacts with the nitrogen i n the s t e e l forming a hard, impervious, p r o t e c t i v e f i l m of zirconium n i t r i d e on the s t e e l . This prevents the i r o n i n the s t e e l from d i s s o l v i n g i n the bismuth^. - k -Corrosion appears to occur only when wetting takes place. The degree of wetting i s determined by the following f a c t o r s : -a) The s o l u b i l i t y of the s o l i d metal i n the l i q u i d metal, e.g. copper i s wetted by bismuth, and i s soluble to the extent of 8 atomic %> at 500°C 1 1 > 1 2 . b) Attack on metals of very low s o l u b i l i t y i n l i q u i d bismuth (such as niobium) may be increased by a l l o y i n g with a metal that i s soluble -1 o in'bismuth (copper., f o r example) . c) I n t e r m e t a l l i c compound formation at the i n t e r f a c e , e.g. zinc i s wetted by sodium at a l l temperatures above the melting point of sodium be-1k cause of the formation of the compound NaZn^c, d) The a b i l i t y of the l i q u i d metal atoms to d i f f u s e through the oxide f i l m on the s o l i d metal. This i s the mechanism postulated by Iberson^^ f o r the wetting of calcium by sodium. The s i z e of the sodium atom i s such that i t could pass through the calcium oxide l a t t i c e . e) Add i t i o n of an element to the l i q u i d metal to promote wetting, e.g. i r o n i s r e a d i l y wetted by sodium i f 0.1 wt.% calcium i s added - t h i s i s because the oxides of i r o n are reduced more e a s i l y by calcium than by sodium-^. ; I t w i l l be seen that the value of l i q u i d bismuth i n the f i e l d of nuclear reactors i s not so much as a coolant, but as a solvent f o r many elements. From t h i s , i t i s apparent that the surface properties of bismuth, the i n t e r -actions between l i q u i d bismuth and s o l i d metals at the s o l i d / l i q u i d i n t e r f a c e , and the influence of oxide fi l m s at t h i s i n t e r f a c e , are a l l of importance i n the study of corrosion by l i q u i d bismuth. C. Previous Investigations on L i q u i d Bismuth Many measurements of the surface tension of l i q u i d bismuth have been made from 1868 to the present time, using a v a r i e t y of methods in c l u d i n g drop weight, maximum bubble pressure, and du Wquy r i n g , under vacuum and atmospheres of carbon dioxide, hydrogen, and nitrogen, and obtaining values of the surface tension of 465 to 269 dynes/cm. Raynor"^ c a r r i e d out a survey of past exper-iments, and determined a new value using the drop volume method with c o r r e c t i o n factors by Harkins and Brown^. Raynor obtained a value f o r the surface tension of l i q u i d bismuth of 378.0 dynes/cm. at 300°C and a temperature c o e f f i c i e n t of -0,069 dynes/cm/°C. No v a r i a t i o n of the value of the surface tension or the temperature c o e f f i c i e n t was found when atmospheres of argon, nitrogen, hydro-gen or carbon dioxide were used i n place of a vacuum (see Figure l ) 1 ^ . A comprehensive study of the wetting behaviour of a number of trans-i t i o n element metals by l i q u i d bismuth was undertaken by Raynor 1^ using the Wilhelmy v e r t i c a l p l a t e technique. The wetting of the following metals was studied;-a) Titanium and zirconium. Wetting occurred at 650 and 550°C respect-i v e l y (see Figure 2) and was governed by the a b i l i t y of the oxide f i l m to d i s s o l v e i n the s o l i d metal thus exposing metal atoms f o r d i r e c t contact with bismuth. b) Vanadium, niobium and tantalum. Won-wetting conditions predominated due to the resistance of the metal oxide to breakdown by l i q u i d b i s -muth. Wetting of niobium and tantalum was reported at 1000°C-'-^ and of vanadium at 6^0°C^} but no suggestions were made as to the mech-anism. c) Iron. Wetting occurred at 500°C by a process i n v o l v i n g the trans-formation of the surface oxides from Fe,0), to FeO. FIGURE 1. Surface tension of l i q u i d bismuth under d i f f e r e n t atmosph d) Cobalt. Solution of cobalt i n bismuth, was prevented "by the oxide f i l m which acted as a b a r r i e r to wetting. e) N i c k e l and copper were wetted r e a d i l y due to the reduction of t h e i r surface oxides by bismuth. A s i m i l a r study to Raynor's has been c a r r i e d out by Iberson^-5 on the surface and i n t e r f a c i a l properties of liquid sodium against s o l i d trans-i t i o n metals. Much research has been undertaken on l i q u i d bismuth i n conjunction with i t s a n t i c i p a t e d use i n l i q u i d metal f u e l r eactors. Among the problems studied were the formation of i n t e r m e t a l l i c .compounds i n l i q u i d bismuth , 20 21 p a r t i c u l a r l y with the f u e l elements uranium and thorium ' , and the osion of s t e e l s i n l i q u i d bismuth, both systems i n v o l v i n g mass transfer^-* - > } c o r r -22 osion of s t e e l s i n l i q u i d bismuth, both i n isothermal and non-isothermal D. Previous Work on Thin Oxide Films on Aluminum The oxide f i l m on aluminum has been found to be very p r o t e c t i v e against most forms of corrosive attack f o r a number of reasons summarized by Gulbransen and Wysong^^:_ a) the oxide/metal volume r a t i o i s greater than one. b) the oxide i s very stable thermodynamically with respect to decomp-o s i t i o n , reduction and s o l i d phase reactions with other metals (the f r e e energy of formation of AI2O3 at 500°C i s -227 Kcals per mole of 26 O 2 , only l i t h i u m , magnesium, and calcium have lower f r e e energies of formation) c) the r e l a t i v e l y high melting point (2015°C) and b o i l i n g point- (2980°C) of the oxide. - 9 -d) r e l a t i v e l y few oxide structures may form. Al^O-^ i s the only stable oxide at ordinary temperatures. I t has been observed frequently ^7-31 that the very f i r s t oxide f i l m formed on aluminum exposed to a i r at room temperature i s amorphous, apparently co n s i s t i n g of an aggregate of groups of two molecules of A l g O ^ 2 ^ jprom the width of the d i f f u s e bands obtained from electron d i f f r a c t i o n by t h i n alum-inum oxide f i l m s , Harrington and Welson^5 c a l c u l a t e d that the grain s i z e of the o f i l m as formed, i f c r y s t a l l i n e , was of the order of 12A, l e s s than twice the o edge of the u n i t c e l l of y-Al203 (7.90 A). There i s l i t t l e d i f f e r e n c e between an amorphous material and a c r y s t a l l i n e m a terial of t h i s g r a i n s i z e . In dry a i r at room temperature the oxidation rate of aluminum i s i n i t i a l l y very high but r a p i d l y l e v e l s o f f (see Figure 3 ) . Hass 2^ states t;hat f i l m growth stops a f t e r a month at a thickness of X. However, V e r n o n ^ found that when the primary f i l m on aluminum had ceased to thicken, sudden breaks oc-curred i n the weight-increment/time curve (see Figure k). He postulated that they are caused by the occurrence of cracks or f i s s u r e s i n the primary f i l m , exposing f r e s h aluminum surface f o r oxidation. Measurements of the f i n a l thickness of the oxide f i l m formed on a l -o o 27-30,37 urainum at room temperature vary between 20 A and 100 A . The d i f f e r -ences are probably due to v a r i a t i o n s i n surface roughness and hence area. A t h e o r e t i c a l value f o r the oxide f i l m thickness, based on the quantum mechanical "tunnel e f f e c t " f o r conduction of electrons through the f i l m , that has been de-r i v e d by M o t t 3 ^ gives a maximum thickness of kO A at room temperature. Although oxide films formed on aluminum at room temperature are 25 29 30 38 39 amorphous, those formed on molten aluminum are c r y s t a l l i n e -^AJ-202 ' ' ' ' - 10 -FIGURE k. Growth of oxide on aluminum at ordinary temperatures 37, - 11 -While some inve s t i g a t o r s have found f i l m s formed at 620°C to he amorphous, others } , using e l e c t r o n d i f f r a c t i o n , have found c r y s t a l l i n e y-A^C^ i n previously amorphous f i l m s a f t e r heating f o r one hour at 500°C or s i x hours at ! UOO°C. The formation of amorphous or c r y s t a l l i n e f i l m s i s determined by the r e l a t i v e speeds of oxidation and c r y s t a l l i z a t i o n . Many in v e s t i g a t o r s have studied a n o d i c a l l y produced oxide f i l m s on aluminum. These fi l m s f a l l i n t o two classes, those formed i n solutions i n which the oxide i s soluble, and those formed i n solutions i n which i t i s i n s o l -uble. In the f i r s t case the f i l m has a t h i n , constant thickness, b a r r i e r l a y e r adjacent to the metal and a porous l a y e r which increases i n thickness with time, the voltage and current density remaining constant. This process, using s u l -phuric a c i d , phosphoric a c i d or chromic acid, i s widely used i n industry to form t h i c k , p r o t e c t i v e oxide coatings on aluminum. When anodic f i l m s are formed i n an e l e c t r o l y t e which does not dissolv e the oxide, the r e s u l t i n g f i l m i s found to he amorphous and f r e e from pores. At constant voltage the current r a p i d l y f a l l s to a very small value, the f i n a l f i l m thickness depending on the forming v o l t a g e ^ . From t h i s Hunter and Fowle^l have developed a method f o r measuring the thickness of b a r r i e r f i l m s or f o r pro-ducing these f i l m s to a required thickness. An e l e c t r o l y t e of 3$ t a r t a r i c a c i d , adjusted to a pH of 5.5 with ammonium hydroxide at 24°C, i s used. In t h i s s o l u t i o n the current density very r a p i d l y f a l l s t o about 10/oA/cnr2 (see Figure 5) and a f i l m , having a thickness of ik A/applied v o l t , i s formed. A l t e r n a t i v e l y " the f i n a l (leakage) current may be determined by anodizing a clean specimen to a thickness approximately equal to that on a f i l m to be measured, and then inc r e a s i n g the voltage a p p l i e d to the unknown specimen i n the e l e c t r o l y t e u n t i l the p r e v i o u s l y determined leakage current i s reached. The f i l m thickness i n 1100 - 12 -- 13 -o Angstrom units i s then fourteen times the applied voltage. For pure aluminum the leakage currect v a r i e s only s l i g h t l y over a wide voltage range, and at t h i s stage i t was assumed that the f i l m was completely formed. Further i n -v e s t i g a t i o n by Hunter and Towner has shown that the f i n a l current i s more than j u s t a leakage, and that the f i l m a c t u a l l y continues to grow by the forma- . t i o n of porous oxide (see Figure 5)p t h i s would i n d i c a t e that the oxide i s not completely i n s o l u b l e i n the e l e c t r o l y t e . During the f i r s t twelve minutes of •coating time, the b a r r i e r thickness increases i n inverse proportion to the cur-rent flow, a f t e r which both values remain s u b s t a n t i a l l y constant, the growth of the o v e r a l l f i l m thickness continuing by porous oxide formation. E. S p e c i f i c Aims of the Present I n v e s t i g a t i o n The o r i g i n a l aim of t h i s i n v e s t i g a t i o n was to study the wetting be-haviour of aluminum i n l i q u i d bismuth by the Wilhelmy v e r t i c a l p l a t e technique used by Raynor"'"^ i n h i s s i m i l a r i n v e s t i g a t i o n s with the t r a n s i t i o n element metals. I t was hoped that the wetting behaviour would i n d i c a t e the mechanism by which bismuth penetrated the aluminum oxide f i l m , i f , i n f a c t ^ i t was able t o . Aluminum was chosen f o r t h i s study, since i t was known that i t was soluble i n l i q u i d bismuth to the extent of 0.01$ at 300°C and l.k<f> at 600°C^, but that i t was protected by an oxide f i l m which was very stable thermodynam-i c a l l y with respect to decomposition, reduction and s o l i d phase reactions with other metals, as mentioned previo u s l y . Thus any breakdown of the oxide f i l m by the bismuth i t s e l f was u n l i k e l y . With aluminum i t would also be p o s s i b l e to a s c e r t a i n the e f f e c t of oxide f i l m thickness, since by anodizing, the f i l m thickness could be increased i n an e a s i l y c o n t r o l l e d manner. - ih -I I . EXPERIMENTAL - Part 1 A.. Properties of L i q u i d Bismuth The bismuth used i n the present i n v e s t i g a t i o n was supplied i n the form of four pound bars by The Consolidated Mining and Smelting Company of Canada Limited. The bismuth was of "intermediate grade" and had the follow-ing a n a l y s i s : -TABLE I  Analysis of Bismuth . Lot Ko. Analy s i s , parts per m i l l i o n Ca Cu A l Fe Mg Fb S i Ag Wi Cd [ HPM 1+03 -HPM.U52 3 0.1 - 0.1 0.1 - 2 0.5 0.1 0.1 - 0.2 0.1 - 0.1 0.1 - 1.0 The bismuth also contained a small quantity of oxygen. However, metal of t h i s p u r i t y i s very s a t i s f a c t o r y f o r any experiments i n v o l v i n g sur-face tension measurements. S o l i d bismuth has a s i l v e r y appearance. I t i s extremely b r i t t l e and thus i s e a s i l y broken i n t o small pieces with a hammer. I t i s covered with a t h i n oxide f i l m which i s pro t e c t i v e i n a i r . Some of the p h y s i c a l properties of l i q u i d bismuth are l i s t e d i n Table I I , with those of sodium and mercury added f o r comparison. Since the density and v i s c o s i t y of mercury are s i m i l a r to those of l i q u i d bismuth, i t may be used f o r t r i a l runs at room temperature. - 15 -TABLE I I Phy s i c a l Properties of L i q u i d Bismuth Ph y s i c a l Property Bismuth Sodium Mercury Melting Point, °C 271 98 -39 B o i l i n g Point, °C 1477 883 357 Contraction on f r e e z i n g -3^32 2.5 3.6 S p e c i f i c Gravity at melting point, gm/cc 10.6k 0.93 13.65 V i s c o s i t y at melting point, centipoises 1.662 0.684 1.85 Atomic Weight 209 23 201 Pure l i q u i d bismuth i s l i k e mercury i n appearance. However, i t r e -acts r e a d i l y with traces of oxygen and water* vapour to form the black mono-oxide, BIO. .The presence of a d d i t i o n a l oxygen Or water vapour leads to the formation of the more stable, yellow sesquioxide, 6120^. Excess a i r or oxygen leads d i r e c t l y to the yellow sesquioxide." At temperatures greater than 600°C, an orange a l l o t r o p i c form of the sesquioxide e x i s t s . The sesquioxide i s s o l -uble i n l i q u i d bismuth, the s o l u b i l i t y i n c r e a s i n g with temperature (see Table I I I ) . I f l i q u i d bismuth i s heated i n the absence of a i r or water vapour, i t w i l l d i s -solve any oxide f i l m and present a clean, m e t a l l i c surface, provided the surface area/volume rat i o , i s small enough. TABLE I I I S o l u b i l i t y ..of .'Bi203 i n "Liquid :'Bismuth Temperature, °C 272 300 400 500 600 S o l u b i l i t y , p.p.m. of Oxygen 0.17 0.34 3.4 12.0 29.O - 16 -Nitrogen, hydrogen, argon and helium do not react with, and do not •kk dissolve i n l i q u i d bismuth . For t h i s reason they are s u i t a b l e f o r use as a pr o t e c t i v e atmosphere. B. Laboratory Handling of L i q u i d Bismuth Some of the problems encountered when handling l i q u i d bismuth i n the laboratory are mentioned below. 1) Container vessels f o r l i q u i d bismuth The most convenient material f o r holding l i q u i d bismuth i s pyrex glass, which i s not attacked below 500°C= Above t h i s temperature, however, bismuth reacts with the softened glass to form a yellow opalescent substance, which cannot be removed. Pyrex glass was used f o r a l l containment vessels f o r l i q u i d bismuth and f o r a l l a n c i l l a r y glassware. St a i n l e s s steels or chrome-vanadium, low a l l o y s t e els are the only p r a c t i c a l metals f o r holding l i q u i d bismuth. Tantalum, molybdenum, chromium and b e r y l l i u m are reported to be su i t a b l e f o r temperatures up to 750°C or higher^, "but are i m p r a c t i c a l f o r most uses. Platinum and nic k e l - c o n t a i n i n g a l l o y s d i s s o l v e i n l i q u i d bismuth.. 2) Protective atmospheres f o r l i q u i d bismuth Unlike sodium, bismuth w i l l react with dry oxygen, so, not only must a l l moisture be removed from the protecting gas, but also as much oxy-gen as p o s s i b l e . Of the previously mentioned pr o t e c t i v e gases, the follow-ing were t r i e d : -a) helium. This, however, had too high an oxygen content. A t h i c k oxide f i l m formed on the surface of the l i q u i d bismuth, and could not be disso l v e d by r a i s i n g the temperature. - 17 -b) nitrogen. This contained l e s s oxygen than the helium, but would not keep the l i q u i d metal surface b r i g h t as long as--c) nitrogen-hydrogen mixtures or p l a i n hydrogen, which could be de-oxidized with a c a t a l y t i c hydrogen p u r i f i e r . Throughout the experiments a 3 to 1 nitrogen-hydrogen mixture was used. The nitrogen and hydrogen were fed from separate c y l i n d e r s , through separate flow meters, before being mixed and passed through an Engelhard "Deoxo" hydrogen p u r i f i e r to convert any oxygen to water. A t o t a l flow rate of approximately four cubic feet/hour was used. Gas used as a p r o t e c t i v e atmosphere f o r l i q u i d bismuth must be he d r i e d extremely w e l l . Previous experiments ' have show that molecular sieves are superior to sulphuric acid, anhydrous calcium sulphate or c h l o r i d e , s i l i c a g e l , a c t i v a t e d alumina, magnesium perchlorate and phosphorus pent-oxide f o r the removal of water vapour from gasses (see Figure 6 ). "Linde" (tlrfion Carbide) molecular sieves, type 5A, contained i n U-tube, 120 cms long, 2.5 cms i n diameter, with glass taps at each end, were used to remove moisture from a l l gases used i n the experiments. Raynor"^ found i t necessary to use a l i q u i d bismuth bubbler to r e -move oxygen from the argon that he used as a protective atmosphere f o r l i q u i d bismuth. However, using a nitrogen-hydrogen mixture, p u r i f i e d as described, i t was possible to keep the surface of l i q u i d bismuth clean f o r as long as desired (up to 5 hours). Despite the precautions required to keep a clean surface on l i q u i d bismuth, there appears to be no f i r e r i s k since the oxide f i l m , once formed, - 18 -J* -p a O ft > Q -20 --6o -80 L. I Activated Alumina 1 Molecular Sieve S i l i c a Gel 5 10 15 20 Wt$ water adsorbed. 25 Dynamic gas drying. - 19 -i s p r o t e c t i v e , and the l i q u i d metal can he s a f e l y poured i n a i r . For t h i s reason i t i s much easier to work with than l i q u i d sodium, potassium or sodium-potassium a l l o y . 3. F i l t r a t i o n of l i q u i d bismuth The oxides of metals, present as impurities i n the bismuth, are insoluble i n the l i q u i d metal, though the metals themselves may be s l i g h t l y s o l u b l e . By f i l t e r i n g bismuth at _the melting point most of the impurities are removed and the oxygen content i s reduced to 0.17 p. p. m. F i l t r a t i o n was c a r r i e d out using a Pyrex glass f i l t e r s t i c k , kO cms. long and 2.5 cms i n diameter, with a sint e r e d glass disc (EC grade) fused i n t o the tube at i t s lower end, and having a gas by-pass tube carrying a tap "A" (see Figure 10). Lumps of bismuth were placed i n the f i l t e r s t i c k and melted with a Bunsen burner while gas passed down the by-pass tube. The f i l t e r d i s c sup-ported the melted bismuth, u n t i l the by-pass tap "A" was closed and the gas pressure f o r c e d the bismuth through the f i l t e r , l eaving behind the s o l i d im-p u r i t i e s . k) Assembly of apparatus As i t was important to prevent contamination of the bismuth by grease, only those ground-glass j o i n t s through which no l i q u i d bismuth passed were greased with s i l i c o n e grease, applied sparingly. Before melting the bismuth, the apparatus was purged with gas to remove a l l a i r . -Also the system was maintained at a s l i g h t p o s i t i v e pressure to prevent ingressrof a i r during the experiment. - 20 -Due to the expansion of bismuth on freezing,. the l i q u i d metal was poured i n t o a s t a i n l e s s s t e e l beaker at the end of an experiment to prevent breakage of glass apparatus. A l l glass apparatus was cleaned with concentrated n i t r i c a c i d , followed by r i n s i n g with water and alcohol, and drying before reuse. C. Des c r i p t i o n of Apparatus A magnetic balance was used to measure the forces acting on a t h i n aluminum p l a t e as i t was p u l l e d through a l i q u i d bismuth surface. From these measurements i t was possible to c a l c u l a t e the receding contact angle. l ) The magnetic balance Figure 1 gives a diagrammatic representation of the magnetic balance assembly. An "Alnico" permanent magnet (A), 3/l6 inch i n diameter and 1 l / 2 inches long, had rods of l / l 6 inch diameter a u s t e n i t i c s t a i n l e s s s t e e l glued to each end, p a r a l l e l to i t s a x i s . This was suspended v e r t i c a l l y i n -side a c o i l (B) of approximately 2000 turns of 2h s'.w.g. lacquered copper wire, wound oh a Perspex:.former. The magnet was located l a t e r a l l y by two f l a t , s p i r a l springs (c) of t h i n copper-beryllium a l l o y , which provided very l i t t l e r e s t r a i n t i n a v e r t i c a l d i r e c t i o n . The metal plates were suspended from the bottom of the magnet ass-embly by a Pyrex glass f i b r e (D). This was prepared by p u l l i n g a piece of glass tube to the required diameter and f u s i n g i n t o each end a short length of tungsten wire (to f a c i l i t a t e attachment t o balance and plate clamp). The combined weight of the magnet assembly, suspension f i b r e and p l a t e was supported by a stop (E). D FIGURE 7» Magnetic balance assembly. Scales F u l l s i z e . - 22 -To measure the forces acting on the p l a t e , the t o t a l downward force was balanced by a n upward force on the magnet produced by a current flowing i n the c o i l (B). The magnet assembly f l o a t e d off. the stop (E) and i t s upper end touched the f i x e d contact (F). The contact was mounted i n an i n s u l a t e d block (G) and was adjustable to give a t o t a l v e r t i c a l movement of the magnet assembly of .005 inches. The contact (F) was connected through a neon bulb to the l i n e terminal of the 110 v o l t mains, and the magnet assembly Was grounded, so that when the upward force j u s t exceeded the downward fo r c e , the neon l i g h t came on. The force was measured by moving the coarse and the f i n e adjustments (see Figure 8) u n t i l the neon l i g h t j u s t came on. The current flowing could be determined by measuring the voltage drop across a standard 0.1 ohm r e s i s t o r with a T i n s l e y Portable Potentiometer. The balance w a s c a l i b r a t e d by hanging weights on i t and noting the voltage drop at the b a l -ancing current. The c a l i b r a t i o n curve i s shown i n Figure 9° The accuracy of the balance was + 20 mg. The complete balance was f i x e d to a brass nut (H) which moved v e r -t i c a l l y on the screw rod (K) of ten threads per inch, so that one turn of the knurled d i s c (L) moved-the balance v e r t i c a l l y 0.1 inches. The d i s c had a c l i c k stop operating once per r e v o l u t i o n . The complete assembly was clamped to a v e r t i c a l aluminum "T" beam. By loosening t h i s clamp, the assembly could v be moved v e r t i c a l l y through a range of two f e e t . 2) The glass apparatus The whole of the apparatus was made of Pyrex glass (see Figure 10). I t consisted of the main container, 2 l/2 inches i n diameter and 8 l/2 inches t a l l , with two upswept side-arms and a two-piece v e r t i c a l extension^ down - 23 -12 Volts Coarse -vww-18 Ohms Adjustment AAAM/WvWvVV1 22 Ohms Ammeter ,0-1 Amps. 18 Ohms Fine Adjustment 0-500 Ohms. Magnetic Balance. 8 Ohms Standard Resistor A / W 0.1 Ohms. Tin s l e y Portable Potentiometer FIGURE 8. Magnetic balance c i r c u i t diagram. Gas I n l e t A (—1 F i l t e r S t i c k t<\ Suspension Fi b r e - 26 -which the suspension f i b r e passed. One side-arm c a r r i e d the f i l t e r s t i c k , previously described, capable of holding 850 grams of bismuth. The other provided access f o r a glass sheathed chromel-alumel thermocouple, which dipped i n t o the l i q u i d bismuth. The l i q u i d bismuth was contained i n the Pyrex glass v e s s e l , and was protected from oxidation by a mixture of nitrogen and hydrogen, which entered through the f i l t e r s t i c k and passed out through the v e r t i c a l exten-sion. This extension was reduced i n diameter towards the top to the smallest diameter that would allow f r e e passage of the suspension f i b r e . 3. Heating and temperature c o n t r o l . The bottom section of the glass v e s s e l containing the l i q u i d bismuth was heated by a 200 watt "Electrothermal" heating mantle (Type MJ 1504 EX 6) . This mantle was connected to the power supply through a "Wheelco" Model 402 i n d i c a t i n g c o n t r o l l e r , which also accommodated the thermocouple. The Model 402 c o n t r o l l e r i s an off-on or two-position temperature c o n t r o l l e r with an adjustable time-proportioning device to minimize c o n t r o l v a r i a t i o n at the c o n t r o l set point. The c o n t r o l action occurs when the meter i n d i c a t o r i s at the red c o n t r o l set p o i n t e r . The instrument has a range of 0 to 800°C and during the experiments the temperature v a r i e d no more than 2°C ei t h e r side of the set temperature. he Preparation of plates (  "Superpurity" aluminum (99-98$0 and Alcan 2S a l l o y were used i n the experiments. The analyses of these material are give i n Table IV» - 27 -TABLE IV  Analyses of Aluminum M a t e r i a l Analysis i n percent Cu Fe S i .Mg Mn T i Zn Cr Wi V 'Superpurity' .005 Trace .008 - Trace - - .008 -Alcan 2S .02 .52 .Ik .010 .05 <,010 <.010 <.010 .01 Balance - Aluminum The a l l o y specimens were cut from l6 gauge (1.5 mm) Alcan 2S " U t i l i t y " sheet, which had been r o l l e d down to O.h mm. The "superpurity" aluminum was av a i l a b l e i n ingot form. A slab, approximately 6 cms by 2.5 cms by 1 cm, was cut from the centre of an ingot section and was r o l l e d down to approximately 0.0k cms t h i c k , with one i n t e r -mediate anneal i n a i r at 500°C f o r one hour. The plates were cut to s i z e from the 0.0k cms sheet with a g u i l l o t i n e , and the edges rounded by p o l i s h i n g with emery. The dimensions of the aluminum plates used were approximately 3-2 cms by 1.2 cms by 0.0k cms. Two d i f f e r e n t surface preparations were used f o r the wetting experiments:-a) Mechanical p o l i s h i n g under kerosene with 0, 2/0 and 3/0 emery paper, followed by degreasing i n chlorethane. b) as (a), but followed by e l e c t r o p o l i s h i n g f o r 2 minutes at 2k v o l t s i n a s o l u t i o n of 15$ p e r c h l o r i c a c i d i n a c e t i c a c i d . The plates were then washed i n water and al c o h o l . - 28 -A l l specimens were kept i n a dessicator containing molecular sieves u n t i l ready f o r use. Before each experiment, the dimensions and weight of each plate'were noted. The plate was then attached to the suspension wire by means of a clamp (see Figure 11) so that i t hung v e r t i c a l l y . 0 •Suspension Wire -Weight Plate Clamp -Plate FIGURE 1 1 . Plate attachment. The high density of l i q u i d bismuth presented a problem. During t r i a l runs the aluminum p l a t e tended to f l o a t and d r i f t e d to the side of the v e s s e l , despite the comparitive r i g i d i t y of the Pyrex f i b r e suspension. How-ever, i t was found that, i f a brass weight was added above the p l a t e clamp, the aluminum p l a t e would hang v e r t i c a l l y . D. Experimental Procedure The apparatus was flushed out with N2-H2 mixture f o r 1 1 /2 to 2 hours and the l i q u i d bismuth v e s s e l was preheated to about 300°C. The - 29 -bismuth, i n the f i l t e r s t i c k was melted with a Bunsen burner and f i l t e r e d by cl o s i n g the tap i n the side arm. The l i q u i d bismuth f i l l e d the v e s s e l to a depth of about 5 c m s • I t was heated to 500°C to dissolve any surface oxide before adjusting the temperature to the desired l e v e l . With the nut (H) (see Figure 7 ) one turn below the top of i t s t r a v e l , the whole magnetic balance assembly was lowered on the "T" beam u n t i l the bottom of the plate j u s t touched the surface of the l i q u i d bismuth. I t was then clamped to the beam. The pl a t e (plus the magnetic balance) was then lowered one inch by turning the disc (L) through ten turns, and the timing was st a r t e d . A scale ( M ) alongside the r a i s i n g and lowering screw i n d i c a t e d the p o s i t i o n of the plate,, A f t e r one minute the pl a t e was withdrawn 0.1 inches at a time (one t u r n of the dis c ( L ) as i n d i c a t e d by the c l i c k - s t o p ) , the force on the p l a t e was balanced at each stop and the potentiometer reading noted. When the pl a t e was c l e a r of the l i q u i d , an "out-reading" was taken to i n d i c a t e any l o s s or gain i n the weight of the p l a t e . The plate was immediately r e -immersed and the above procedure repeated every ten minutes. One set of readings, from the s t a r t of r a i s i n g to reimmersion i n the bismuth, took from three to four minutes depending on how much adjustment of the balance and potentiometer was needed. The f a c t that the bottom of the pl a t e was im-mersed f o r longer than the top was accounted f o r i n c a l c u l a t i n g the r e s u l t s . E. Theory The forces acting on the pl a t e are represented i n Figure 12 and are: -a) The upward force, F grams, due to the tension i n the suspension wire. This i s also the force recorded by the magnetic balance. - 30 -Surface tension force. FIGURE 12. Forces on the plate i n l i q u i d bismuth. b) the v e r t i c a l component of the surface t e n s i o n ^ a c t i n g at a contact angle of 9 to the p l a t e , g i v i n g a value of ycos9 dynes/cm. c) the net weight Wgms = t o t a l weight-bouyancy force The t o t a l weight includes the weight of the p l a t e , clamp, and sus-pension f i b r e . . The bouyancy forc e = t(w) (x) (p). where t = thickness of pLate i n cms. w = width of pla t e i n cms. x = length of p l a t e immersed i n bismuth i n cms. jo = density of l i q u i d bismuth. (The value used was that of hi i Bircumshaw and was 10.03 gms/cc at the melting p o i n t , the temperature c o e f f i c i e n t being -0.00.3 gms/cc/°C). - 31 -Resolving v e r t i c a l l y , we have; 2(t + w) ycos 6 = F - W g From the above equation and Raynor's"1"^ value of the surface tension of l i q u i d , bismuth, the contact angle 9 can be c a l c u l a t e d . Accuracy of r e s u l t s The accuracy to which the contact angle could be determined depended on a number of factors;- the approximate error f o r each i s given below:-a) thickness and width of plate +0.2$ b) net weight of pl a t e +1.0$ c) surface tension. The value used had an accuracy of + 1.0$. Any v a r i a t i o n of temperature would produce a change of only +0.01$ per °C. d) the magnetic balance reading had an accuracy of + 20 mg throughout the range, g i v i n g much lower percentage accuracy at low readings, i . e . when the contact angle was greater than 90°. The accuracy v a r i e d between 2$ and 0.5$. The net error may be up to 7$; but would be lower than t h i s f o r con-t a c t angles near 0°, as the percentage error i n the magnetic balance reading i s a minimum at t h i s p o i n t . F. Analysis of Results Before proceeding i t i s necessary to define the d i f f e r e n t degrees of wetting that are used i n the discussion of r e s u l t s . D e f i n i t i o n of wetting a) Wetting and p a r t i a l wetting w i l l imply contact angles l e s s than 9°°. - 32 -b) Complete wetting w i l l imply a contact angle of zero and i s recog-nized by the presence of l i q u i d metal adhering to the s o l i d metal. c) Non-wetting w i l l imply a contact angle greater than 90° and i s recognized by the absence of l i q u i d metal adhering to the s o l i d metal. o Contact angles were measured at various temperatures between 350 C and 500°C f o r mechanically p o l i s h e d Alcan 2S and mechanically and e l e c t r o -p o l i s h e d "superpurity" aluminum. The influence of immersion time on contact angle was examined at each temperature. The r e s u l t s were c a l c u l a t e d using the equation given i n Section E. Table V shows a t y p i c a l c a l c u l a t i o n of the contact angle. The p o s i t i o n column gives the reading of the v e r t i c a l p o s i t i o n scale (M i n Figure 7) and i n d i c a t e s the depth of immersion of the aluminum p l a t e i n the l i q u i d bismuth i n units of 0.1 inches. The f i r s t three columns are those recorded during the experi-ment. The contact angle versus time r e l a t i o n s h i p was p l o t t e d e i t h e r as a group of curves each representing a d i f f e r e n t p o s i t i o n on the p l a t e , or as a curve representing the average contact angle over the whole p l a t e . Figure 13 shows a p l o t of the f i r s t type. Each curve shows the change of contact angle with time at a p a r t i c u l a r l e v e l on the p l a t e (only four are shown f o r c l a r i t y ) , and each point represents the average contact angle at t h i s l e v e l . I t w i l l be observed that the curves follow the same pattern. - 33 -TABLE V Sample C a l c u l a t i o n of Contact Angle Plate 5 Mechanically polished Alcan 2S 500°C Time mins. Position- Potentiometer Reading M V F gms gms S gms Cos 0 e Immersion Time-mins. 1.00 9 39-65 1.162 1.978 -.816 -.969 165.5 1.00 8 40.15 1.279 2.088 -.809 -.96O 164. 1.37 7 41.03 1.486 2.201 -.715 -,8k$ 148. 1.75 6 4l.4o 1.572 2.311 -.739 -.877 151.5 2.12 5 41.85 1.678 2.423 -.7-45 -.884 152. 2.50 4 42.19 1.758 2-533 -•775 -.920 157. 2.87 3 42.97 1.941 2.645 -.704 -.836 146.5 3.25 2 •^ 3-79 2.133 2-757 -.624 -.741 138. 3.62 4.00 l M+-53 2.307 2.868 -.561 -.666 131.5 4.00 y= 362.0 dynes/cm; p = 9°73 gms/cc. w => 1.101 cms. t = 0.0409 cms. 'Position' i s the reading of the pointer on the balance and represents the length of p l a t e immersed i n un i t s of 0.1 inches. S = F - W = ycos 0 .2(w + t) g F = Magnetic balance reading W = Wet weight .= T o t a l weight - Bouyancy force Immersion Time, minutes. - 35 -Curves of the second type are shown i n Figures 14 and 15". The contact angle p l o t t e d i n t h i s case i s the average over a l l posi t i o n s f o r one withdrawl of the p l a t e . This curve w i l l run through the centre of the group of corresponding curves of the f i r s t type. (Compare the p l o t s f o r p l a t e l 4 i n Figures 13 and 15) Figure 14 shows the v a r i a t i o n of contact angle with time at temp-eratures of 400, 450 and 500°C f o r plates of mechanically pol i s h e d Alean 2S. I t w i l l he seen that the I n i t i a l contact angle i s the same i n each ease hut that i t decreases f a s t e r with immersion time at the higher temperatures. The wetting behaviour of mechanically polished "superpurity" aluminum at temperatures of 350> 400, 450 and 500°C i s shown i n Figure 15. The i n i t i a l increase i n the contact angle i n the case of the two lower temp-erature curves was probably due to a s l i g h t increase i n the f i l m thickness caused by a pick-up by the aluminum of oxygen di s s o l v e d i n the l i q u i d bismuth. A s i m i l a r e f f e c t was observed by R aynor^ on niobium, tantalum and vanadium. The I n i t i a l i n c r e a s e i n the curve at 450°C i s probably due to the same cause. The maxima of these curves tended to shorter times and the slope increased uniformly with increasing temperature. The curves i n Figures l 4 and 15 appear to show a steady decrease i n contact angle with immersion time. However, the contact angle shown Is a c t u a l l y the average angle over the whole p l a t e , or, i n the case of Figure 13, over each l e v e l on the p l a t e . V i s u a l observation of the plates (see Figure 16) showed that at any l e y e l on the p l a t e , conditions would vary from complete wet-t i n g , to non-wetting. - 37 -o FIGURE 1$. Wetting behaviour of mechanically polished 99.98$ aluminum. 0 500°C P l a t e 15. A 450°C " 8. o Loo°C " 6. • 350°C " 14. J 1 1 1 i _ ko 60 80 100 Immersion Time, minutes. 9 8 Figure l6 (b) Mechanically polished 99-98$ aluminum a f t e r exposure to l i q u i d bismuth at various temperatures and times l i s t e d below. a) Plate lk 350°C 123 minutes " 18 kOO°Q 76 6 kOO°C YLk b) Plate 9 H-50°C 23 8 450°C 3k - 39 -At f i r s t stained areas woiild appear, followed by areas with, bismuth adhering and .areas showing attack, often edge attack; a l l these--areas would be completely wet by the l i q u i d bismuth. The steady decrease i n contact angle shown i n Figures. Ik and 15, therefore, represents not a general decrease, hut a spread-ing of areas on the plate that are completely wetted. At .the lower temperatures, .actual attack was u s u a l l y l i m i t e d to the edge of the plate (Figure 16, a and b ) . With increased temperature more severe attack occurred with considerable s o l -u t i o n of the aluminum (Figure l6 (b)) although some areas pf the p l a t e were not wetted. A f t e r s o l u t i o n of one tenth or more of the aluminum p l a t e i n the bismuth, the contact angle measurements had to be stopped, since the l i q u i d bismuth surface became covered with an aluminum oxide l a y e r , even though I t contained no more than 10 p. p. m. of di s s o l v e d aluminum. This concentration was also s u f f i c i e n t to prevent the formation of interference colours i n the oxide f i l m when the l i q u i d bismuth was poured i n a i r at the end of the experiment. Three plates that had "been e l e c t r o p o l i s h e d as described e a r l i e r , were teste d and a l l were found to wet completely and uniformly, with .uniform s o l u t i o n and thinning of the p l a t e . Since contact angle measurements could give no more than the rate of increase of wetted areas, i t was decided to abandon t h i s experimental technique and develop a procedure whereby a c o r r e l a t i o n could be achieved between the time at which the aluminum p l a t e was f i r s t attacked by the l i q u i d bismuth and the surface c o n d i t i o n of the aluminum oxide f i l m . - i+o -I I I . EXPERIMENTAL - Part 2 A. D e s c r i p t i o n of Apparatus The same apparatus was used as i n part 1, with the exception of 'the magnetic balance, which was replaced by a l / l 6 inch diameter s t a i n l e s s ( s t e e l rod clamped d i r e c t l y to the nut H (Figure 7) at the top and with the p l a t e clamp s i l v e r soldered to the bottom end. l ) Preparation of plates The aluminum plates were of the same s i z e and m a t e r i a l as those used before. The "superpurity" aluminum plates were given the following surface treatments:-a) Mechanical p o l i s h i n g under kerosene on 0, 2/0 and 3/0 emery paper followed by degreasing i n chlorethane. b) Cleaning by immersion f o r f i v e minutes i n a s o l u t i o n of 35 m l / l i t r e phosphoric a c i d (85$) and 20 gms/litre chromium oxide at 100°C and f o r f i v e minutes i n a 50 gm/litre sodium .carbonate s o l u t i o n at 80°C. They were then anodized i n a 3$ ammonium t a r t r a t e s o l u t i o n 41 as described by Hunter and Fowle . The following voltages were used:-o 1Q v o l t s g i v i n g an approximate f i l m thickness of 140 A 0 15 "volts" g i v i n g an approximate f i l m thickness of* 210 A o 20 v o l t s g i v i n g an approximate f i l m thickness of 280 A . o 30 v o l t s g i v i n g an approximate f i l m thickness of 420 A c) Cleaning as i n (b) and anodizing i n 15$ sulphuric a c i d at 15 v o l t s and 20°C to give a porous oxide f i l m approximately 14 microns 51 t h i c k . One such f i l m was sealed by immersion i n b o i l i n g water f o r 30 minutes. ~ kl -B. Experimental Procedure The apparatus was flushed out with N2-H2 mixture and the bismuth f i l t e r e d as described i n part 1. The aluminum p l a t e , clamped to the bottom end of the s t a i n l e s s s t e e l rod, was lowered u n t i l i t j u s t touched the l i q u i d bismuth surface, and the r a i s i n g and lowering mechanism was clamped to the "T" beam. The plate was then immersed i n the l i q u i d bismuth to a depth of 1 inch by means of the screw mechanism (K i n Figure 7.) • At the end of the f i r s t minute the pl a t e was quic k l y withdrawn from the l i q u i d bismuth f o r examination and immediately reimmersed. This was repeated at the end of the f i r s t f i v e minutes and every f i v e minutes f o r the f i r s t hour, and then every ten minutes u n t i l the p l a t e was attacked. As the pl a t e was withdrawn, i t s surface was c a r e f u l l y examined f o r s t a i n i n g , adhering bismuth, p i t t i n g or any other forms of attack which were noted together- with the time at which they occurred. A l l the plates tested by immersion i n l i q u i d bismuth are l i s t e d i n Appendix I, together with d e t a i l s of the t e s t i n g conditions and the type of attack that occurred. In the column headed "Time" i s given the time o f immersion at which attack by the l i q u i d bismuth was b e l i e v e d to have s t a r t e d . This time was extremely d i f f i c u l t to estimate as many of the plates had b i s -muth adhering to the surface i n small drops. I f the pl a t e was immersed i n concentrated n i t r i c a c i d a f t e r the t e s t , any bismuth on the pl a t e was r a p i d l y d i s s o l v e d and i n many cases i t was found that the aluminum p l a t e was p i t t e d where the bismuth had been located. I t was impossible to t e l l whether the bismuth adhered to the pl a t e because the f i l m had already been penetrated or whether penetration occurred a f t e r the bismuth adhered to the plate and had established close contact with the f i l m . I t i s believed, however, that the former i s the case. - k2 -C. Analysis of Results In general, the aluminum plates were attacked, i n one of two ways, or i n some cases a combination of the two. In one case the plates f i r s t became stained, p a r t i c u l a r l y round the edges; t h i s was followed by bismuth adhering to the surfaces and edges of the plates i n small drops (see Figure l 6 ) . The f i r s t v i s i b l e signs of attack were one or more small notches In the edges of the p l a t e s . In some cases the adherent bismuth could e a s i l y be removed with a f i n g e r n a i l , although, as mentioned previously, i n others, p i t s were found when the bismuth was d i s s o l v e d . The rate at which the alumi-num d i s s o l v e d i n the l i q u i d bismuth was slow and there was only a small number of points of attack, often not more than three or four. In the second type of attack, the bismuth, on penetrating the alumi-num oxide f i l m , would r a p i d l y spread under the f i l m . The compressive stresses i n the f i l m , induced by the oxide/metal volume r a t i o being greater than one, caused the oxide f i l m to wrinkle. On withdrawing the p l a t e from the l i q u i d bismuth, the bismuth trapped under the oxide f i l m would run to the bottom of the p l a t e and form a bulge i n the f i l m (see Figure 17)-When a p l a t e that had been attacked i n the l a t t e r way was sectioned, a l a y e r of bismuth could be seen between the aluminum and the oxide f i l m . Figure 18 shows a v e r t i c a l s e c t i o n of the bottom of p l a t e 19 where the bismuth has c o l l e c t e d i n a bulge under the oxide f i l m . The mechanically polished specimens generally showed a f a i r l y consistent form of attack i n the l i q u i d bismuth. - k3 -15 19 2 0 FIGURE 17. Mechanically polished 99.98$ A l a f t e r immersion i n l i q u i d bismuth at the following temperatures. a) Plate 15 500°C f o r Ik minutes. b) " 19 500°C f o r 15 minutes. c) 20 500°C f o r 10 minutes. FIGURE 18. A v e r t i c a l section of plate 19, mechanically polished 99.98$ A l a f t e r immersion In l i q u i d bismuth f o r 15 minutes at 500°C. showing bismuth under the oxide f i l m at the bottom of the p l a t e . Magnification 11OX. - kh The Alcan 2S specimens showed the f i r s t type of attack i n a l l cases Staining of the p l a t e was followed by bismuth adhering to the surface, and f i n a l l y by edge attack. On d i s s o l v i n g o f f the bismuth, i t was found that most plates were p i t t e d underneath. In the case of two plates that were sectioned, intergranular attack was also v i s i b l e In the aluminum around the p i t s (see Figure 19). Figure 19 Section of p l a t e 33, mechanically p o l i s h e d Alcan 2S, a f t e r immer-sion i n l i q u i d bismuth f o r 15 minutes at ^00°Cf showing p i t t i n g and intergranular attack. Magnification 99QX. The Alcan 2S plates were attacked more quickly at the higher temp-eratures, but there was considerable s c a t t e r i n the values of the* time taken to i n i t i a t e attack, which cannot be e n t i r e l y accounted f o r by the uncertainty In the precise moment at which the attack commenced. - 45 -The "superpurity" aluminum plates were attacked In the same way as the a l l o y specimens at 35° and 400°C, hut with l e s s p i t t i n g on the surface of the p l a t e , and no s i g n of i n t e r g r a n u l a r attack (see Figure l 6 a ) . At ,450°C, the s t a i n i n g was followed "by penetration of the f i l m at a number of places and non-uniform attack on the aluminum'under the f i l m (see Figure 16 b ) . o At 500 C the second type of attack predominated with r a p i d spreading of the bismuth under the oxide f i l m and f a i r l y severe l o c a l attack.(see Figure 17)« Again there was the wide s c a t t e r i n the values of the time taken f o r attack to s t a r t . The r e s u l t s f o r the 99-98$ 'aluminum, anodized i n ammonium t a r t r a t e solution,'showed even l e s s r e p r o d u c a b i l i t y than previous r e s u l t s . Even o o though the minimum f i l m thickness used was 140 A compared with 20-40 A f o r the mechanically p o l i s h e d specimens, no s i g n i f i c a n t increase i n the r e -sistance to attack was produced, except f o r the t h i c k e r f i l m s (280 and 420 A). The second type of attack, i n v o l v i n g complete removal of the oxide f i l m followed by r a p i d s o l u t i o n of the aluminum, occurred more frequently i n these t e s t s and was not confined to the higher temperatures. The oxide f i l m .often f l o a t e d o f f the p l a t e i n bulk and l a y on the surface of the bismuth on e i t h e r side of the p l a t e i n large pieces. However, when the f i l m was not detached, the attack was very l o c a l i z e d , often appearing to spread from only one or two points, with l i t t l e s t a i n i n g or adherent bismuth elsewhere. Figures 20 and 21 show the types of attack that ocpurred on anodized specimens.. Four specimens were anodized i n 15$ sulphuric a c i d to give a f i l m - thickness of approximately l 4 microns-'-'". One p l a t e t e s t e d at 400°C and one at 500°C showed no attack a f t e r 3 and 5 hours r e s p e c t i v e l y . Another p l a t e was - he -FIGURE 21. Plates 25, 21, and 12 (lj+O % anodized f i l m ) af£er immersion i n l i q u i d "bismuth at 50t C f o r 25 minutes, 450 C f o r 9 min-utes and 42 minutes r e s p e c t i v e l y . - kl -held j u s t above the l i q u i d bismuth surface at 500°C to allow i t to heat up before entering the hot l i q u i d . As i t was lowered i n t o the l i q u i d metal, the aluminum oxide f i l m f l o a t e d o f f the p l a t e . The plate was withdrawn before i t was completely immersed, and a dark s t a i n could be seen r a p i d l y t r a v e l l i n g up the p l a t e . On sectioning the p l a t e , i t was found that bismuth had spread over the whole p l a t e under the oxide f i l m , presumably by cgjpillary a c t i o n since bismuth w i l l wet unprotected aluminum. The f o u r t h specimen was sealed by immersing i n b o i l i n g water f o r t h i r t y minutes a f t e r anodizing. When tested, the oxide f i l m was penetrated and bismuth spread under the f i l m , causing extensive cracking i n the t h i c k oxide f i l m . D. Discussion of Results , 1 The r e s u l t s have shown that the attack on aluminum by l i q u i d bismuth occurred i n two ways. When penetration of the oxide f i l m l e d to immediate spreading of the bismuth under the f i l m , i t was thought at f i r s t that t h i s was caused by r a p i d s o l u t i o n of the surface of the underlying aluminum. I f , however, the bismuth was di s s o l v e d o f f the p l a t e , i t was found that attack on the aluminum was confined to a small area around a break i n the oxide f i l m . In places where access by f r e s h bismuth was l i m i t e d , the rate of attack was very slow, due to the small s o l u b i l i t y of aluminum i n bismuth (about 1$ at 500 C) . For bismuth to t r a v e l up to l/2 inch under an unbroken oxide f i l m , at the rate observed, by d i s s o l v i n g the aluminum would, therefore, be im-probable. I f , however, the oxide f i l m had already separated from the alumi-num the bismuth could r a p i d l y spread under the f i l m by c a p i l l a r y a c t i o n . - 48 -The c o e f f i c i e n t of thermal expansion of an anodic f i l m on aluminum' i s approximately one f i f t h of that of the "base metal. When an ano<li z e (l p l a t e i s heated, a considerable shear stress w i l l occur at the metal/oxide i n t e r f a c e . This stress could produce a separation of the metal and oxide along the i n t e r -f ace. Furthermore, the stress w i l l be increased by shrinkage of the oxide due to dehydration. I t i s probable that t h i s condition occurred i n many of the plates tested. As would be expected t h i s phenomenon was more prevalent among the plates with t h i c k oxide f i l m s . These plates would also be more susceptible to cracking of the f i l m under the s t r e s s , which would account f o r the very r a p i d breakdown and removal of the oxide f i l m i n some cases. This was demonstrated by the p l a t e that was allowed to heat up before im-mersion i n the bismuth and whose oxide f i l m f l o a t e d o f f as the p l a t e was immersed f o r the f i r s t time. The performance of anodized plates that were not attacked i n t h i s way c o n s t i t u t e d f u r t h e r evidence i n support of the separation of the oxide from the aluminum under thermal s t r e s s . I f the a c t i o n of bismuth was r e -sponsible f o r the separation of the f i l m , separation would occur whenever the bismuth penetrated the f i l m under conditions that l e d . t o separation on other p l a t e s . However, some plates were attacked under these conditions and were d i s s o l v e d i n a number of places without any spreading of bismuth under the f i l m (see Figures 20 and 21). I t would appear that the removal of the oxide f i l m i n the l i q u i d bismuth or the spreading of the bismuth under the f i l m , was not a d i r e c t consequence of the attack on aluminum by l i q u i d bismuth, but rather of the thermal stresses induced by heating the p l a t e . The separation of the oxide from the aluminum allowed the bismuth to run between the metal and the oxide when i t penetrated the f i l m . - 4 9 -When plates that had only been attacked by p i t t i n g or edge attack were examined, i t was noticed that the attack was concentrated at a number of d i s c r e t e points on the surface or edges of the p l a t e . On some plates there -were only two or three such points. As mentioned previously, i t i s extremely u n l i k e l y that bismuth could attack aluminum oxide chemically and no evidence of any r e a c t i o n was observed i n these experiments. The type of attack noted was consistent with the theory that flaws were present i n the oxide l a y e r through which the bismuth penetrated. 48 Stern and Uhlig , have also postulated that aluminum i s attacked through flaws i n the oxide f i l m . They subjected aluminum p l a t e s , that had been anodized i n ammonium t a r t r a t e and sulphuric acid, to attack by b o i l i n g carbon t e t r a c h l o r i d e . They found that corrosion d i d not s t a r t uniformly over the e n t i r e surface of the aluminum, but rather at a few small points. Once started, the r e a c t i o n spread r a d i a l l y u n t i l the corroding areas met. Edge attack was also present i n some specimens.. These r e s u l t s also i n d i c a t e the presence of defects i n the oxide f i l m . On examining plates with thick-,, anodized oxide l a y e r s . i t was p o s s i b l e to see defects i n the f i l m . Figure 22 shows a hole i n the oxide f i l m , the l i g h t area surrounding the hole i s an area of oxide that has been l i f t e d o f f the underlying bismuth l a y e r . Figure 23 shows, i n section, a hole through the oxide l a y e r ( c i r c l e d ) . There i s bismuth above and below the oxide l a y e r . The aluminum i s at "the bottom 'of •" the photograph. FIGURE 22. Plate 32 (anodized i n 15$ HgSO^ and sealed) a f t e r immersion i n l i q u i d "bismuth f o r 20 minutes at 500°C, showing a hole i n a region of r a i s e d oxide. Magnification 60X. FIGURE 23. Section of plate 32 (anodized i n 15$ H2S0^ and sealed) a f t e r immersion i n l i q u i d bismuth f o r 20 minutes at 500°C, showing a defect i n the oxide f i l m . Magnification 1200X. - 51 -The defects i n the oxide f i l m could he of several types. They could he areas where the f i l m i s very t h i n , such as sharp edges on the p l a t e , scratches or p i t s i n the surface. P a r t i c l e s embedded i n the surface during r o l l i n g or p o l i s h i n g , or non-metallic i n c l u s i o n s i n the metal i t s e l f would l e a d to d i s c o n t i n u i t i e s i n the f i l m . Concentrations of impurity atoms on the surface would produce areas of the oxide of the impurity element i n the f i l m . I f the impurity element was r e a d i l y attacked by l i q u i d bismuth, these areas of impurity oxide would allow penetration by the bismuth. Cracks i n the oxide f i l m , e i t h e r thermally or mechanically induced, would l e a d to r a p i d penetration. A t h i r d s e r i e s of experiments was designed to give more information on the a c t u a l penetration of the f i l m . I t was hoped that by the e l i m i n a t i o n of as many of the sources of defects as p o s s i b l e , the type of defect r e -sponsible could be determined. The S e s s i l e drop method was chosen to elim-inate edge defects, and to allow the bismuth to be positioned on or away from v i s i b l e defects i n the p l a t e . - 52 -IV. EXFERIMMTAL - Part 3 S e s s i l e Drop Experiments A small lump of bismuth was placed on a t h i n aluminum p l a t e and heated i n a tube furnace. The time at which the bismuth penetrated the oxide f i l m was noted. This method had the advantage that only a small area of the pl a t e was attacked and the bismuth could be placed on or away from obvious defects. Also, four plates could be tested at the same time. A. Desc r i p t i o n of Apparatus l ) The tube furnace A conventional tube furnace was used; i t was 15 .inches long and had a 1 inch i n t e r n a l diameter Vycor tube, 36 inches long, running through i t (see f i g u r e 24). One modif i c a t i o n was made by blocking the top h a l f of the furnace up l / 4 inch to provide a 5 inch.long s l o t aeross the furnace, g i v i n g a view through the tube. This s l o t enabled the plates to he observed with-out d i s t u r b i n g the tube. An i l l u m i n a t e d white background was placed behind the s l o t to throw the drops of bismuth on the plates i n t o s i l h o u e t t e . The tube, complete with p l a t e s , could be withdrawn to allow closer examination of the surface of the p l a t e s . A graphite boat, 5 inches long and of a semi-circular cross-section, held four aluminum p l a t e s . A chromel-alumel thermocouple i n a s i l i c a sheath was positioned over the graphite boat i n the centre of the furnace. The thermocouple was .connected to the Wheelco temperature c o n t r o l l e r used i n the previous experiments. The furnace element was connected to the mains through a Variac and a r e l a y . The r e l a y was operated by the temperature Tube Furnace FIGURE 2h, P a r t i a l section of tube furnace. - 54 -c o n t r o l l e r . The temperature over the centre 4 inches of the furnace v a r i e d no more than 10°C from the set temperature. P u r i f i e d hydrogen was passed through the tube from the r i g h t hand end t o the l e f t , where i t was burnt. 2) Gas p u r i f i c a t i o n Hydrogen was used f o r a l l the s e s s i l e drop experiments, although the tube was fl u s h e d with nitrogen at the beginning and end of a run to eliminate the r i s k of an explosion. The hydrogen was p u r i f i e d , E S : before, by passing .through a c a t a l y t i c p u r i f i e r to remove oxygen and molecular sieves to remove moisture. A flow of 1 cubic foot/hour was maintained during experiments. A f a s t e r flow (3 cubic feet/hour) was used f o r f l u s h i n g out. 3) Preparation of plates Great care was taken i n the preparation of the plates to prevent the incorporation of defects. In some cases i t was found that d i s s o l v e d hydrogen caused b l i s t e r i n g of p a r t i a l l y r o l l e d sheet during annealling. To remove hydrogen the aluminum was remelted and carbon rods soaked i n carbon t e t r a -c h l o r i d e were dipped i n t o the molten metal. The aluminum was cast as a f l a t slab i n a copper mold. The surface was machined to remove any t h i c k oxide before r o l l i n g . New r o l l s were used and great care was taken to prevent g r i t being r o l l e d i n t o the surface of the aluminum. The sheet was given an i n t e r -mediate and a f i n a l anneal at 550°C f o r one hour. The f i n a l thickness of the sheet was Q*25 mm. The sheet was cut i n t o plates 2.4 cms square. "Superpurity" aluminum only was used f o r the s e s s i l e drop experiments. A l l plates were chemically pol i s h e d and anodized. Various e l e e t r o p o l i s h i n g - 55 -solutions were tested, but the best surface f i n i s h was given by chemical p o l i s h i n g . However, i t was not possible to obtain a p l a t e completely f r e e of p i t s with any method. Alcoa Bright Dip was used c o n s i s t i n g of 80$ phosphoric acid, 14$ a c e t i c a c i d and 6$ n i t r i c a c i d . The plates were immersed f o r about 10 minutes at 70 - 80°G. The plates were anodized i n the ammonium t a r t r a t e e l e c t r o l y t e . o described i n the i n t r o d u c t i o n . A non-porous f i l m with a thickness of 14 Aper applied v o l t i s obtained a f t e r anodizing f o r 12 minutes. The plates were wast\ed i n water, alcohol and acetone, and then driedj. They were kept i n a dessicator u n t i l required. Defects were d e l i b e r a t e l y Introduced mechanically i n some plates e i t h e r before or a f t e r anodizing to determine t h e i r e f f e c t on the penetra-t i o n of the bismuth. B. Experimental Proqedure Small lumps of bismuth, approximately 2.5 mm i n diameter, were washed i n d i l u t e n i t r i c a c i d to remove any oxide; they were then washed i n water, acetone and d r i e d . Four aluminum plates with the same f i l m thickness were placed on the graphite boat. One lump of bismuth was placed on each p l a t e i n a previously selected p o s i t i o n , e i t h e r on or away from any v i s i b l e defects. The boat was pushed to the centre of the furnace and the thermo-couple was i n s e r t e d i n the tube so that i t was above the centre of the boat. The furnace tube was flushed with nitrogen f o r 20 minutes and hydro-o gen f o r 30 minutes before the furnace was switched on. A temperature of 500 C was used f o r a l l experiments, t h i s temperature being reached about 10 minutes a f t e r switching on the furnace. - 56 -The bismuth lumps melted, and formed nearly s p h e r i c a l drops (see Figure 25), no wetting of the oxide occurring. With only one or two excep-ti o n s , the surfaces of the drops were clean and free from oxide, and remained t h i s way for the duration of the longest t e s t (75 hours). FIGURE 25 Section of bismuth drop on plate D4 (140 $ oxide f i l m ) , showing general shape of drops. MagnJ.fication 20 X. The bismuth drops were examined through the s l o t i n the furnace every f i v e minutes for the f i r s t houn , and thereafter at increasing Inter-v a l s . I f possi b l e , the furnace tube was quickly withdrawn every hour to examine the surface of the p l a t e s . In nearly every case, however, penetration of the f i l m was i n d i c a t e d by the bismuth drop r a p i d l y f a l l i n g , the bismuth spreading over the surface of the plate under the oxide f i l m . C. Analysis of Results D e t a i l s of a l l plates tested, i n c l u d i n g f i l m thickness, the time taken f o r the bismuth to penetrate the f i l m , and comments on the form of attack are given i n Appendix I I . - 57 -Once again there were wide v a r i a t i o n s i n the time taken by the bismuth to penetrate the oxide f i l m . In many cases, penetration was followed by im-mediate spreading of the bismuth under the oxide f i l m , again i n d i c a t i n g .that the f i l m had separated from the aluminum even though the rate of heating was much, lower than i n previous experiments. In a few cases, however, the bismuth took several hours to spread showing that separation had not occurred. In four cases, the bismuth was placed so that i t covered one or more small p i t s on the surface of the p l a t e s . Once, the bismuth penetrated the f i l m before 500°C was reaqhed, i n d i c a t i n g a defect that went completely through the f i l m . The other f i l m s were not penetrated a f t e r 7h 3/k hours, 28 hours and 8^  1/2 hours ( i n the l a t t e r case, a l l plates were penetrated when the temperature c o n t r o l r e l a y jammed and the temperature went to over 550°c). Plates with no v i s i b l e defects were attacked a f t e r times varying from K7 minutes to 9 hours 45 minutes. I t would appear that defects that can be seen with the unaided eye (or a magnifying glass) do not lower the pro-t e c t i v e power of the f i l m . (Fine cracks or pores i n the f i l m i t s e l f cannot be seen even under high magnification as the oxide f i l m i s transparent). Two plates had very sharp p i t s introduced with a s t e e l s c r i b e r , one before and one a f t e r anodizing. Both plates were attacked at about the same time. S i m i l a r l y , two plates were scratched with a s t e e l scriber,one before and one a f t e r anodizing. The pl a t e scratched .before anodizing was attacked a f t e r 9 hours, the one a f t e r anodizing was unattacked a f t e r 12 l / 2 hours. In a l l cases the bismuth was placed on the defect. I t would appear that exposure to the atmosphere healed the break i n the f i l m providing p r o t e c t i o n as good as that of the anodized f i l m . - 58 -Two plates were severely deformed by bending through l80° over a 5/l6 inch diameter rod to crack the f i l m . The plates were then f l a t t e n e d and a drop of bismuth placed on what was the outside of the bend. The fil m s were penetrated a f t e r 37 and 67 minutes. Of the two con t r o l p l a t e s , one f a i l e d a f t e r k7 minutes, the other was unattacked a f t e r lh hours. o Four plates were prepared with a 700 A oxide f i l m . I t was thought that f i l m s of t h i s thickness might crack under the thermal stresses induced by heating. I f cracking occurred, i t would almost c e r t a i n l y be during the heating of the p l a t e . None of the plates, however, showed any penetration by bismuth a f t e r 20 hours. o I t w i l l be noticed that two plates (ihO A fil m s ) were unattacked a f t e r 7h hours and three (280 A fil m s ) a f t e r 28 hours. On sectioning the 7h hour plates through the bismuth drop, no sign of attack or chemical action could be seen. Unfortunately, i t was im p r a c t i c a l to carry out t e s t s f o r longer periods than 75 hours. The only mechanism, however, by which the bismuth could penetrate the oxide f i l m i n times of t h i s order, would be by d i f f u s i o n of bismuth atoms through the aluminum oxide f i l m . To see i f the d i f f u s i o n of bismuth through the f i l m was a probable mechanism, three sapphire windows of d i f f e r e n t orientations were heated at 500°C f o r hQ hours with a l/h inch diameter drop of bismuth on each. The windows were washed qui c k l y with d i l u t e n i t r i c a c i d to remove any bismuth that had condensed on the surface. The drops of bismuth were knocked o f f and the surface of the sapphires pol i s h e d l i g h t l y with Brasso to remove any adhering bismuth or bismuth oxide. The area under the bismuth drop was then spectrographically analysed f o r bismuth. - 59 -None was found i n d i c a t i n g that the bismuth had not d i f f u s e d i n t o the sapphire. I t i s , therefore, extremely u n l i k e l y that the bismuth could d i f f u s e through the aluminum oxide f i l m on the p l a t e s . An a d d i t i o n a l sapphire window was heated with the three previously mentioned. This window had no bismuth placed on i t , and was not cleaned before a n a l y s i s . A trace of bismuth (less than 10 p. p. m.) was found, i n d i c a t i n g that a monolayer of bismuth had condensed on the plare even o though the vapour pressure of bismuth at 500 C Is l e s s than 1 mm. Hg. o FIGURE 26 Defect i n oxide f i l m on plate D2 (140 A film) a f t e r penetration by l i q u i d bismuth. Magnification 150 X. On examining plates that had been attacked by the bismuth, i t was found that the oxide f i l m was wrinkled as i n previous t e s t s . In places the bismuth that had spread under the f i l m could be seen through a hole i n the - 60 -f i l m (see Figure 26). The only explanation f o r these holes i s that the bismuth had attacked a weak part of the f i l m from below. The only type of defect that would be susceptible to such attack i s an area of impurity oxide i n the aluminum oxide f i l m . The surface density of such holes was approximately the same as that of p i t s i n the f u l l y immersed p l a t e s . > - 6 l -V. CONCLUSION The r e s u l t s are discussed at the ends of the appropriate experimental sections. The technique of measuring the v a r i a t i o n of contact angle with 16 temperature and time of immersion In l i q u i d bismuth as used by Raynor to i n -dicate the breakdown of the oxide f i l m s on t r a n s i t i o n metals, was impract-i c a l f o r use with aluminum. A decreasing contact angle d i d not i n d i c a t e a general decrease, but an increase i n the number of i s o l a t e d areas at which wetting had occurred. A l l oxide f i l m s produced on the aluminum were p r o t e c t i v e to a l i m i t e d degree, with the exception of those on e l e c t r o p o l i s h e d specimens. The l a t t e r provided no p r o t e c t i o n against the l i q u i d bismuth. This was to 49 be expected since Gy-Hollo ^ found the f i l m on e l e c t r o p o l i s h e d aluminum to be both t h i n and porous. The aluminum was attacked by the l i q u i d bismuth i n two ways, by p i t t i n g on the surface and edges of the p l a t e s , and by bismuth spreading under the f i l m ( i n some cases causing i t to f l o a t o f f ) . There was no ex-perimental evidence that the aluminum oxide i t s e l f was subjected to general attack. S e s s i l e drop experiments on anodized plates and sapphire windows showed that bismuth does not d i f f u s e through the oxide f i l m under the exper-imental conditions employed. The p i t t i n g type of attack was prevalent at the lower temperatures and with the thinner f i l m s . Only a small number of points of attack was observed on each p l a t e . The nature of the attack i n d i c a t e d that the bismuth was penetrating the oxide f i l m through defects. S e s s i l e drop experiments showed that p i t s i n the aluminum under the oxide f i l m were not n e c e s s a r i l y sources of defects i n the f i l m , and that scratches through the f i l m q u i c k l y healed i n a i r and d i d not constitute a weakness. Mechanical defects i n the p l a t e or f i l m do not diminish the p r o t e c t i v e power of the f i l m , with the exception of cracks produced by large deformations. I t i s most probable that concentrations of impurities e x i s t e d i n the aluminum. These concentrations would l e a d to areas of the oxide f i l m having a composition corresponding to the oxide of the impurity. The main impurities i n the aluminum (see Table IV) are copper, titanium, magnesium and n i c k e l , of these, copper and n i c k e l are attacked by l i q u i d bismuth at temperatures below 500°C, t h e i r oxides being reduced by bismuth. These concentrations cannot be seen i n the plates used i n the experiments even under high magnification, but much i n d i r e c t evidence e x i s t s f o r t h e i r presence. 50 Davis has found a d e n d r i t i c structure i n the ingot from which the plates were made, thus segregation could be expected i n the ingot. I t i s b e l i e v e d that the r o l l i n g and annealing would not be s u f f i c i e n t to destroy the segregation completely. The Alcan 2S plates with a l l o y i n g additions had many more p i t s than the equivalent "superpurity" aluminum p l a t e s . Since bismuth cannot attack or d i f f u s e through a pure aluminum oxide f i l m , any defect i n such a f i l m would have to go completely through the f i l m , and the bismuth could penetrate immediately. The f a c t that some fil m s were penetrated only a f t e r 9 hours can only be explained on the basis of bismuth slowly r e a c t i n g with an area of impurity oxide. - 63 -The second type of attack, where the bismuth spread r a p i d l y under the oxide f i l m , was almost c e r t a i n l y caused by the separation of the aluminum and i t s oxide along the oxide/metal i n t e r f a c e . Shear stresses r e s u l t i n g from the diffe r e n c e i n the c o e f f i c i e n t of thermal expansion of aluminum and i t s oxide and from dehydration of the oxide f i l m , would be generated during heating of the p l a t e , and would be greater In the t h i c k e r f i l m s . Pene-t r a t i o n of the f i l m could occur by the mechanism j u s t described f o r p i t t i n g , by cracking of the f i l m along the edges of the p l a t e where a stress r a i s e r e x i s t s , or craze cracking on the surface of the p l a t e . - 6k -VI. RECOMMENDATIONS FOR FURTHER WORK The results- of the present i n v e s t i g a t i o n are q u a l i t a t i v e and several l i n e s of i n q u i r y might he followed that would r e i n f o r c e the present f i n d i n g s . The r e s u l t s of varying the impurity content of the aluminum might throw a d d i t i o n a l l i g h t on the mechanism of penetration of the f i l m by l i q u i d bismuth. I f the mechanism postulated i n the present i n v e s t i g a t i o n i s co r r e c t , zone r e f i n e d aluminum, or a t h i n aluminum f i l m condensed on a glass s l i d e , should e x h i b i t exceptional resistance to attack. By varying the oxide f i l m thickness and the rate of heating i t should be possible to obtain information on the bond strength between alumi-num and i t s oxide f i l m . - 65 -VII. APPENDIX I A. Alcan 2S. Mechanically polished. Temp. °C Plate No. Time Mins. Type of Attack 4oo 2 50 S t a i n i n g and s l i g h t attack along one edge 450 4 3h 40 10 Staining, bismuth adhering to p l a t e , s l i g h t attack at two points on one edge. Sta i n i n g , bismuth adhering, edge and surface attack at many points. 500 5 33 35 9 5 12 Staining, bismuth adhering, severe attack at sur-face l e v e l , s l i g h t attack at two points on one edge. Staining, bismuth adhering, edge attack at many points, surface p i t s . Bismuth adhering, many surface p i t s under bismuth, s l i g h t edge attack. B. "Superpurity" Aluminum . E l e c t r o p o l i s h e d . Temp* °C Plate No. Time Mins. Type of Attack 350 11 0 Immediate wetting followed by d i s s o l u t i o n of the p l a t e . 4oo 10 0 As 11. 450 7 0 As 11. - 66 -C. "Superpurity" Aluminum. Mechanically p o l i s h e d . Temp. °C Plate Wo. Time Mins. Type of Attack 350 14 43 Staining, bismuth adhering, s l i g h t edge attack. 400 450 6 18 8 9 63 10 8* ^ — Staining, bismuth adhering, edge attack at three points, surface p i t t i n g . Staining, bismuth adhering, s l i g h t edge attack at l i q u i d l e v e l . Staining, followed by edge attack at many places spreading to surface of p l a t e . As 8.. 500 15 19 20 5 10 0 Quick penetration of f i l m , with spreading under f i l m causing wrinkling, considerable d i s s o l u t i o n of plate. Staining, bismuth adhering, leading to l i q u i d l e v e l attack and spreading of bismuth under f i l m . Immediate penetration of f i l m with spreading under f i l m . Attack near top of p l a t e . D. "Superpurity" Aluminum. Anodized i n 15$ H2SQJ4. f o r 30 minutes, ( f i l m thickness approximately 14 microns) Temp. °C Plate No. Time Mins. Type of Attack 400 30 180 S l i g h t s t a i n i n g near l i q u i d l e v e l - no attack. 500 31 300 No attack. 41 32 0 20 Oxide removed on immersion of bottom of p l a t e , b i s -muth spread under f i l m on unimmersed part. Sealed f o r 30 mins. Penetration and spreading of b i s -muth under f i l m with cracking of f i l m and s o l u t i o n of aluminum. - 67 -E. "Superpurity" Aluminum. Anodized i n Ammonium Tartrate . Temp.. Plate Time Fi l m 1 °c 0 Type of Attack Wo. Mins. Thickness A 350 26 40 140 S l i g h t edge attack at "bottom corner , 4oo 23 5 140 ._,,„ 1 Removal of oxide f i l m from most of pl a t e , s l u t i o n and thinning of p l a t e 24 10 140 Considerable edge attack at one poin t . no other attack. 27 0 140 Immediate wetting followed by removal of oxide f i l m from most of p l a t e . 39 55 140 S l i g h t s t a i n i n g , very small area of edge attack. 450 12 30 140 Complete removal of oxide f i l m from most of plate-considerable s o l u t i o n of aluminum. 21 0 140 Immediate wetting with removal of oxide f i l m and s o l u t i o n of aluminum. 22 36 210 Attack under parts of f i l m leading to detachment of bottom of p l a t e . 500 25 20 140 Edge attack at several points near bottom of plate spreading Inwards. 36 18 280 Removal of oxide and s o l u t i o n over small area at bottom of p l a t e . 37 105 420 Bismuth adhering to centre of pl a t e with p i t s underneath. 4o 175 280 Small amount of bismuth adhering, f o l -lowed by complete removal of oxide f i l m - 68 -APPENDIX I I . A l l plates were of "superpurity" aluminum, chemically polished, and o anodized i n ammonium t a r t r a t e s o l u t i o n f o r 12 minutes (10 minutes f o r 70 A f i l m s ) . Temperature of t e s t i n each case was 500°C. Plate No. F i l m o Thickness A Time to Penetrate Comments C 1 70 3 hrs. C 2 8 1/2 hrs. Relay j ammed,T > 550°C. C 3 8 1/2 hrs. As C 2. C 5 8 .1/2 hrs. As C 2. Bismuth placed on p i t i n plata„ E 1 70 14 hrs. + No attack. E 2 37 mins. Mechanically cracked. E 3 47 mins 0 E 4 67 mins. Mechanically cracked. - 69 -Plate No. Film o Thickness A Time to Penetrate Comments A 3 140 9 3/4 hrs. A 4 A 5 5 mins. 3 1/2 hrs. Penetration before 500°C reached. t. B i s -muth drop on f i n e p i t s i n p l a t e . A 6 9 3/4 h r s . D 2 D 3 140 5 hrs. 4 h r s . Bismuth penetrated and spread slowly under f i l m over a period of 24 hours. As D 2. Very l i t t l e attack on aluminum. D h :D 5 74 3/4 hrs. plus 74 3/4 hrs. plus No- attack. Bismuth drop on f i n e p i t s i n p l a t e . No attack. K 1 K 2 K 3 K 4 140 ^5 hrs. a/5 hrs. A/9 hrs • 12 1/2 hrs. plus Mechanically p i t t e d before anodizing, bismuth placed on p i t . Mechanically p i t t e d a f t e r anodizing, bismuth placed on p i t . Plate scratched before anodizing, b i s -muth close to scratch. Plate scratched a f t e r anodizing, b i s -on scratch. No attack. - TO -' Plate F i l m Time to o Comments Wo. Thickness A Penetrate B. 1 280 28 1/3 hrs. No attack. Bismuth drop on small p i t i n plus p l a t e . B 2 28 1/3 hrs. No attack. plus B 3 5 h r s . 3 hours taken f o r bismuth to spread under f i l m . B k 28 1/3 hrs. No attack. plus H 1 700 20 hrs. + No attack on any of the four plates a f t e r H 2 20 h r s . + 20 hours. I f thermal cracking of the H 3 20 hrs. + f i l m had occurred, penetration would H k 20 hrs. + have taken place w i t h i n a few hours. - 71 ~ VIII. BIBLIOGRAPHY 1. Hackett, H. N. Trans. Am. Soc. Mech. Engrs., 64, 6kl ( l 9 4 2 ) . 2. Smith, A. R. and Thompson, E. S. Trans. Am. Soc. Mech. Engrs., 64, 625 (1942). 3- Heron, S. D. U. S. Patent 1670965, May 22, 1928. 4. Bennett, F. C. Dow Chemical Co., Report 14854, Oct. 1948. 5- C h i r k i n , J . Nucl. Energy, 5, 124 (1957). 6. Novikov, J . Nucl. Energy, 4, 385 (1957). 7. Lyon, R. N. ' l i q u i d Metals Handbook". U. S. A. E. C. Rep. Ko. NAVEXOS P-733 (1952) Ch. 2. 8. Boadle. Atomics, 8, 4 l and 83 (1957). 9. F r o s t . Nucl. Eng., 1, 33k and 373 (1956). 10. Bailey, Fox and Watkins. U. K. A. E. R. E. Rep. No. M/TN6 ( l 9 5 l ) . 11. Bondi. Chem. Revs., 52, 4l7 (1953). 12. Hansen. "Constitution of Binary A l l o y s " . (McGraw-Hill 2nd Ed. 1958). 13. F r o s t et a l . Peaceful Uses of Atomic Energy, Paper P/27O (1958). 14. Addison, Addison, Kerridge and Lewis. J . C. S. 1454 (1956). 15. Iberson, E. Ph. D. Thesis, Nottingham (1959). 16. Raynor, J . B. Ph. D. Thesis, Nottingham (1959). 17. Harkins and Brown. J . A. C. S. 4 l , 499 (1919). 18. Parkman and Shepard. U. S. A. E. C. Rep. No. ORO-45 (1951). 19. Barton, P. J . and Greenwood, G. W. J . Inst. Met. 86, 504 (1957-58). 20. T e i t e l , R. J . et a l . Nucleonics, 12 (7) , 14 (1954). 21. Greenwood, G. W. and Sharpe, B. A. E. R. E. Rep. M/R 2250 (1954). 22. Graham, L. W. and Wilson, G. W. J . Iron St. In s t . 193 Pt I I , 103 (1959) 23. Weeks, J . R. et a l . Proc. Inst. Conf. Peaceful Uses Atomic Energy (Geneva 1955) 9, 3^1. - 7 2 -24. Dawe, D.. .W., Parry, G. I . and Wilson, G. W. J . Iron'St. Inst. 1 9 0 , 2 7 1 ( 1 9 5 8 ) . 2 5 . Gulbransen, E. A. and Wysong, W. S. <J. Phys. Chem. 5 _ 1 , IO87 ( 1 9 V 7 ) . 2 6 o Richardson, F. D. and J e f f e s , J . H. E. J . Iron St. Inst. 1 6 0 , 2 6 l ( 1 9 4 8 ) 2 7 . Hass, G. Z. Anorg. Chem., 254, 9 6 ( 1 9 V ? ) . 2 8 . K e l l e r , F. and Edwards, J . D. In t . Conf. Surf. React, p. 2 0 2 (1948). 2 9 . Preston, G. D. and Bircumshaw, L. L. P h i l . Mag. 2 2 , 6 5 4 ( 1 9 3 6 ) . 3 0 . de Brouchere, L. J . Inst. Met. 7 1 , 1 3 1 (1945). 3 1 . Raether, H.,C. R. Acad. S c i . , P a r i s , 2 2 7 , 1247 (1948). 3 2 . Wilsdorf, H. G. F. Nature, London, l 6 8 , 6 0 0 ( l 9 5 l ) . 3 3 - Cabrera, N. and Mott, N. F. Rep. Progr. Phys. 12, 1 6 3 (1948-49). 34. Mott, N. F. Trans. Farad. Soc. 3 5 , 1 1 7 5 . 3 5 ° Harrington, R. A. and Nelson, H. R. Trans. Amer. Inst. Min. Met. Eng. 1 3 7 , 1 2 8 (1940). 3 6 . Kubaschewski, 0 . and Hopkins, B. E. "Oxidation of Metals and A l l o y s " (Butterworth's 1 9 5 3 ) p. 1 3 1 . 3 7 . Vernon, W. H. J . Trans. Farad. S o c , 2 3 , 1 1 7 . 3 8 . Hunter, M. S. and Fowle, P. J . Electrochem. Soc. 1 0 3 , 482 ( 1 9 5 6 ) . 3 9 - S t e i n h e i l , A. Ann. Fhysik, 1 9 , 4 6 5 (1934). 40. Hass, G. J . O p t i c a l Soc. Amer. 3 9 , 5 3 2 ( 1 9 4 9 ) . 41. Hunter, M. S. and Fowle, P. J . Electrochem. Soc. 1 0 1 , kdl ( 1 9 5 4 ) . 42. Hunter, M. S. and Towner, P. F. J . Electrochem. Soc 1 0 8 ( 2 ) , 1 3 9 ( 1 9 6 1 ) . 43. Kelman, et a l . U. S. A. E. C. Rep. No. ANL-4417 ( 1 9 5 0 ) . 44. Kubaschewski, 0 . Z. Elektrochem. kh, 1 5 2 ( 1 9 3 8 ) . 4 5 . Addison, C. C , Iberson, E. and Raynor, J . B. Chem. and Ind. 9 6 ( 1 9 5 8 ) . 46. "'Linde' molecular sieves f o r s e l e c t i v e adsorption" (B.-D. H. Ltd., Poole, England). 47. Bircumshaw, L. L. P h i l . Mag. 12, 5 9 6 ( 1 9 3 1 ) • - 73 -48. StejrrijM. and U h l i g , H. H. J . Electrochem. Soc. 99, 381 and 389 (1952). 49. Gy-Hollo' Mme M. Memoires S c i e n t i f i c de l a Rev. Met. 57, 23 (i960). 50. Davis, K. G. U n i v e r s i t y of B r i t i s h Columbia. Private communication. 51. Spooner, R. C. J . Electrochem. Soc. 102, 156 (1955). 

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.831.1-0105831/manifest

Comment

Related Items