UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Unbalanced magnetrons Clarke, Glenn A. 1990

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

Item Metadata

Download

Media
831-UBC_1990_A7 C52.pdf [ 4.27MB ]
Metadata
JSON: 831-1.0074522.json
JSON-LD: 831-1.0074522-ld.json
RDF/XML (Pretty): 831-1.0074522-rdf.xml
RDF/JSON: 831-1.0074522-rdf.json
Turtle: 831-1.0074522-turtle.txt
N-Triples: 831-1.0074522-rdf-ntriples.txt
Original Record: 831-1.0074522-source.json
Full Text
831-1.0074522-fulltext.txt
Citation
831-1.0074522.ris

Full Text

UNBALANCED MAGNETRONS by GLENN A. CLARKE B.Sc, Queens University, 1988 THESIS SUBMITTED IN PARTIAL FULFILLMENT THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES ENGINEERING PHYSICS We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1990 ° Glenn A. Clarke, 1990 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT In recent years the 'unbalanced* magnetron sputter source has been shown to be an important advancement for applications requiring ion-assisted, t h i n f i l m deposition. Diamondlike coatings and o p t i c a l multilayers are examples of coatings which can be s i g n i f i c a n t l y improved by ion bombardment of the growing f i l m . The purpose of t h i s thesis i s to develop a better understanding of the main engineering factors which determine the ultimate performance of the unbalanced magnetron. Two pieces of experimental apparatus were designed for t h i s study. A scanning magnetometer system was constructed to measure the magnetic f i e l d pattern of the magnetrons, which allowed for a detailed analysis i n terms of magnetic f i e l d l i n e s . A multiple, plasma probe assembly was developed to measure the discharge c h a r a c t e r i s t i c s for various unbalanced magnetron configurations. The r e s u l t s showed that the ion flux to the substrate was highly dependant on the development of a secondary discharge away from the target surface. The secondary discharge was generated through confinement of i o n i z i n g electrons i n a magnetic b o t t l e along the magnetron axis. The ion flux to the substrate was observed to be approximately independent of the pressure and target material but highly dependant on the i i discharge current and target thickness. The ion/deposition flux r a t i o increased with decreasing target-to-substrate distance. However, the useful deposition area decreased under these conditions as the ion currents became highly focused about the axis. i i i TABLE OF CONTENTS Abstract i i Table of Contents i v L i s t of Tables v i L i s t of Figures v i i Acknowledgements ix Chapter I. Introduction 1 A. Magnetron Sputtering 1 B. Thin Film Microstructure 7 C. Methods of Zone T Thin Film Production . 8 D. History of Unbalanced Magnetrons 10 E. Objective of Thesis 13 Chapter I I . Theory 15 A. Introduction 15 B. Plasma Theory 15 C. Plasma Probe Theory 17 D. Gas Phase C o l l i s i o n s 20 E. Electron Motion i n a Plasma 21 F. Planar Magnetron Discharge 24 Chapter i l l . Magnetic Measurement System 26 A. Introduction 26 B. Magnetic F i e l d Line P l o t t i n g 28 C. Measurement System 32 1. Design 32 2. Alignment and C a l i b r a t i o n 37 3. Magnetic F i e l d Scan Tests 40 Chapter IV. Plasma Probe 42 A. Introduction 42 B. Measurement System 42 1. Design 42 2. Alignment 45 3. Operation 46 Chapter V. Experimental Details 47 A. Introduction 47 B. Results from Varying the Strength Between the Annular and Centre Magnet . 48 1. Magnetron Configurations 48 2. Data C o l l e c t i o n and Processing .. 49 3. Magnetic F i e l d Line Plots 50 i v 4. Plasma Probe Measurements 55 5. Results with large plasma probe . 56 6. Results with small probes 57 7. Discussion 60 C. Centre Piece Geometry and Target Thickness E f f e c t s 68 1. Introduction 68 2. Experimental r e s u l t s 71 3. Discussion 71 D. Plasma Charact e r i s t i c s of Unbalanced Magnetrons 74 1. Introduction 74 2. Target material 74 3. Discharge current 75 4. Pressure 76 E. Ion/Deposition Flux Ratio 76 1. Introduction 76 2. Deposition flux measurements .... 77 3. Ion/Deposition flux maximization 79 Chapter VI. Conclusion 82 References 85 Appendix 87 v L i s t of Tables Table Page 5.1 Magnetic Configurations 49 5.2 Discharge c h a r a c t e r i s t i c s measured by the large probe at 200 mA 57 5.3 Results for the large probe for magnetrons 5 and 6 ... 70 v i L i s t of Figures Figure Page 1.1 Planar magnetron sputtering system 2 1.2 Cathode assembly 3 1.3 Unbalanced magnetron 10 2.1 I-V c h a r a c t e r i s t i c s of a plasma probe 18 2.2 Potential v a r i a t i o n near a negative electrode 19 3.1 C i r c u l a r magnetron 27 3.2 Cross-section of a rectangular magnetron 27 3.3 Elect r o n i c s of magnetic f i e l d measurement system ... 33 3.4 Mechanical aspects of magnetic f i e l d measurement system 36 4.1 Plasma probe system 4 3 4.2 E l e c t r i c a l components of plasma probe 44 5.1 Magnetron 1 51 5.2 Magnetron 2 51 5.3 Magnetron 3 52 5.4 Magnetron 4 52 5.5 Magnetron 5 53 5.6 F i e l d strength along magnetron axis 54 5.7 Ground currents for magnetrons 1 to 5 57 5.8 Ion currents for magnetrons 1 to 5 58 5.9 Radial dependence of ion currents for magnetron 1 .. 59 5.10 Radial dependence of ion currents for magnetron 5 .. 59 v i i 5.11 Secondary magnetic bot t l e f i e l d l i n e 65 5.12 Magnetron 6 69 5.13 F i e l d strength along axis f o r magnetrons 5 and 6 ... 69 5.14 Ground currents for magnetrons 5 and 6 70 5.15 Ion currents for magnetrons 5 and 6 71 5.16 Ion currents versus discharge current 75 5.17 Deposition rates along magnetron axis 78 5.18 Ion/Deposition flux r a t i o along magnetron axis 80 5.19 Ion currents versus r a d i a l p o s i t i o n (normalized) ... 80 v i i i ACKNOWLEDGEMENTS F i r s t , I would l i k e to give my thanks to my research supervisor, Dr. R.R. Parsons, for a l l h i s help, ideas, and encouragement throughout the course of my th e s i s . I would also l i k e to thank a l l those i n the lab that have lent t h e i r support, including Peter Mulhern for h i s timely advice, and most e s p e c i a l l y Norman Osborne for h i s constant input during t h i s project. I must also acknowledge the mechanical and e l e c t r o n i c technical s t a f f i n the Physics department and Jack Bosma i n p a r t i c u l a r for t h e i r h e l p f u l advice and assistance during the design and construction of the experimental apparatus. F i n a l l y I would l i k e to give my appreciation for the f i n a n c i a l assistance of R.R. Parsons, the University of B r i t i s h Columbia and the B.C. Science Council. i x CHAPTER I. INTRODUCTION A. MAGNETRON SPUTTERING Magnetron sputtering has developed over the past twenty years into a predominant method for the deposition of t h i n films. Though each deposition technique has i t s advantages, magnetron sputtering allows for a p a r t i c u l a r l y wide range of applications. The sputter process i s e s s e n t i a l l y one of momentum transfer whereby the material of i n t e r e s t for deposition, or target, i s bombarded by energetic p a r t i c l e s , usually ions. This causes the ejection of target atoms, a portion of which then condense onto a substrate, usually positioned between 4 to 12 cm from the target surface. Unlike other methods, which use chemical and/or thermal techniques f o r d i s s o c i a t i o n , almost any material can sputtered. This makes the process extremely v e r s a t i l e . Another major advantage of magnetron sputtering i s the ease to which i t lends i t s e l f to scale-up. Research that i s done on small, r e l a t i v e l y inexpensive systems can be applied to the development of large, i n d u s t r i a l ones. Other a t t r a c t i v e factors include f i l m surface smoothness, the a b i l i t y to control f i l m uniformity and thickness, and the a b i l i t y to deposit over a large area. A schematic of a planar magnetron system i s shown i n F i g . 1.1 and a more detailed drawing of the cathode assembly, i n Fig . 1.2. There are a number of other geometries i n use, such 1 Introduction / 2 INERT R E A C T I V E GAS GAS CATHDDE ASSEMBLY VACUUM CHAMBER SUBSTRATE HOLDER SUBSTRATE HIGH VACUUM PUMP Figure 1.1 Schematic of a planar magnetron sputtering system. The cathode assembly i s shown i n greater d e t a i l i n Fi g . 1.2 as a c y l i n d r i c a l configuration 1, but the planar ( f l a t target) system i s the only type considered i n t h i s t h e s i s . During the deposition process the vacuum chamber i s i n i t i a l l y pumped down to a pressure of approximately 1 x 10"6 Torr. An i n e r t gas, such as argon, i s then introduced to r a i s e the pressure to 1 - 10 mTorr. Often a reactive gas such as nitrogen or oxygen i s also employed, to produce a d i e l e c t r i c f i l m , e.g. a Introduction / 3 Figure 1.2 Schematic of the cathode assembly of a planar magnetron system. Also included i s the e l e c t r i c a l biasing. n i t r i d e or oxide of the target material. A f t e r the gas i s at the desired pressure, a negative pot e n t i a l i s applied to the target while the chamber and part of the target assembly are usually grounded, which creates a plasma discharge. The applied voltage can be either dc or r f . The l a t t e r mode i s required for i n s u l a t i n g targets. Positive ions i n the plasma bombard the target surface, r e s u l t i n g i n the ejection of both neutral atoms and secondary electrons, as well as other species such as negative ions. A portion of the ejected atoms condenses onto the substrate of inter e s t , producing a f i l m . An important factor i n the sputtering process i s Introduction / 4 sustaining the plasma discharge. In order to achieve t h i s , the loss of ion-electron pairs needs to be overcome. This loss r e s u l t s from ion-electron recombination on the chamber walls, ion n e u t r a l i z a t i o n on the target, and electron loss through the grounded surfaces. The loss of ion-electron pairs can be overcome through the emission of secondary electrons during the process of ion bombardment of the target. The secondary electrons are i n i t i a l l y accelerated away from the target surface gaining k i n e t i c energy through the e l e c t r i c f i e l d . They then can undergo c o l l i s i o n s with the gas atoms, which often involve e x c i t a t i o n and/or i o n i z a t i o n . The newly created ions, i n turn, bombard the target surface producing more secondary electrons, making the process s e l f - s u s t a i n i n g . The sputtering method was i n i t i a l l y developed without the use of magnetic f i e l d s 2 . This method i s s t i l l used i n practice; however, i t has a number of problems which l i m i t s i t s range of applications. The f i r s t of these r e s u l t s from a tradeoff between e f f i c i e n t use of the secondary electrons for i o n i z a t i o n and the deposition rate. At pressures low enough to prevent s i g n i f i c a n t scattering of the deposition flux the mean free path of the electrons can be comparable to the target to anode distance. As a r e s u l t , there may not be enough i o n i z a t i o n to sustain the discharge. This problem can be overcome at higher pressures but at the cost of a reduced deposition flux. A second problem i s that the growing f i l m i s Introduction / 5 subjected to bombardment by highly energetic electrons with an energy comparable to the cathode voltage; e.g. 500 V. Such bombardment often causes f i l m damage. These problems can be overcome through the a p p l i c a t i o n of a magnetic f i e l d p a r a l l e l to the target surface. Under these conditions the electrons are s t i l l i n i t i a l l y accelerated away from the surface, but are forced back towards the target by the Lorentz force. The electrons undergo c y c l o i d a l motion about a centre which d r i f t s i n a d i r e c t i o n perpendicular to both the e l e c t r i c and the magnetic f i e l d ( E x B motion 3). For the planar magnetron assembly shown i n F i g 1.2, the source of the f i e l d consists of an annular and centre magnet mounted on a high permeability pole piece. As the outer magnet completely surrounds the inner magnet, the d r i f t has a closed path. This traps the electrons close to the target surface and increases t h e i r e f f e c t i v e path length. This leads to s u f f i c i e n t i o n i z a t i o n to sustain the discharge at much lower pressures than for non-magnetic methods. Ideally a l l of the magnet f i e l d l i n e s o r i g i n a t i n g on the annular magnet should return to the centre i n order to maximize the electron trap (Fig 1.2) . This i s not possible i n practice for a planar magnetron due to f r i n g i n g e f f e c t s , though i t can be approximated by equalizing the strength of the centre and annular magnets. The above configuration i s sometimes referred to as a balanced magnetron. However, even under i d e a l conditions, the trap i s s t i l l imperfect as Introduction / 6 electrons can escape through energy gained through electron-electron c o l l i s i o n s , interactions with the i n s t a b i l i t i e s i n the plasma 4, and along the f i e l d l i n e s near the centre of the target. This leads to a plasma i n the v i c i n i t y of the substrate. The problem can be p a r t i a l l y overcome by the placement of the ground s h i e l d so as to i n t e r s e c t the f i e l d l i n e s such as that i n Fig 1.2. Under these conditions the ground s h i e l d w i l l provide an e f f e c t i v e return path to ground for the electrons due to confinement of the electrons across the f i e l d l i n e s . This r e s u l t s i n a system with e f f i c i e n t i o n i z a t i o n at low pressures and with a plasma discharge predominately i n the v i c i n i t y of the target. A d i f f e r e n t type of plasma discharge r e s u l t s i f the annular magnet i s made much stronger and/or with a larger area than the centre magnet( Fig 1.3). Under these conditions the substrate can undergo s i g n i f i c a n t electron and ion bombardment5. The electron flux i s usually of a low energy and, therefore, can be repelled by applying a small negative bias ( -10 to -35 V) to the substrate. The ion bombardment can a c t u a l l y be advantageous i f not c r u c i a l for c e r t a i n applications. This i s best understood by f i r s t examining the microstructure of t h i n films that r e s u l t s under a number of d i f f e r e n t deposition conditions. Introduction / 7 B. THIN FILM MICROSTRUCTURE Films that are deposited under conditions where the growth temperature i s less than about 30% of the melting point of the deposited material have what i s referred to as a zone 1 microstructure 6. At these temperatures the f i l m i n i t i a l l y grows around preferred nucleation s i t e s located at areas of substrate inhomogeneties and roughness. Atoms reaching these s i t e s w i l l i n i t i a l l y be loosely bounded to the f i l m l a t t i c e and are referred to as adatoms. The mobility of the adatoms i s low under zone 1 conditions and they w i l l not be able to t r a v e l between the nucleation s i t e s . As the nucleation s i t e s grow i n siz e they w i l l prevent further depositing atoms from reaching the substrate through shadowing e f f e c t s . Hence the microstructure i s characterised by long columns separated by s i g n i f i c a n t voids. The r e s u l t i n g f i l m has a large surface roughness and properties that are quite unlike that of the bulk material and i s often unsuitable for many applications. Different microstructures r e s u l t at higher deposition temperatures. For temperatures between 30 and 50% of the melting point the mobility of the adatoms increases to the point where they can undergo s i g n i f i c a n t d i f f u s i o n on the grain boundaries. The f i l m microstructure then consists of columnar grains separated by i n t e r c r y s t a l l i n e boundaries and i s referred to as zone 2. For substrate temperatures above 50%, the r e s u l t i s a zone 3 microstructure where d i f f u s i o n within the grains leads to a f i l m characterized by equiaxed Introduction / 8 grains. Though zones 2 and 3 have desirable properties for a number of t h i n f i l m applications, the deposition conditions required often makes t h i s method of f i l m growth impractical, e s p e c i a l l y i f the substrate has a melting point lower than that of the f i l m or can be damaged at high temperatures. Fortunately high q u a l i t y films can be grown at low temperatures (e.g. 80°C) i f the substrate i s subjected to p a r t i c l e bombardment during the deposition process. The bombarding p a r t i c l e s , either ions or energetic neutrals, w i l l both impart energy into the growing f i l m , increasing the adatom mobility, and forward sputter the f i l m material. The above processes r e s u l t i n a microstructure consisting of densely packed fibrous grains. These films, referred to as zone T, have smooth surfaces, high densities and properties close to that of the bulk values. Examples of applications for zone T films include wear r e s i s t a n t metallurgical coatings 7 , and diamondlike t h i n f i l m s 8 . C. METHODS OF ZONE T THIN FILM PRODUCTION There are a number of methods currently i n use for the deposition of zone T t h i n films. One of these i s an alte r n a t i v e sputtering technique which uses a separate ion source 9 to provide target bombardment instead of a plasma discharge. There are two main variati o n s of the above method which also provide bombardment of the substrate. The f i r s t i s Introduction / 9 to simply supply another ion source, t h i s time directed at the substrate. The second source should supply p a r t i c l e s at a lower energy than the f i r s t i n order to prevent resputtering of the deposited f i l m . Another method positions the substrate so that the ions directed towards the target w i l l also bombard the substrate, though at a more oblique angle. The disadvantage of growing zone T films i n t h i s manner i s that ion beam sputtering does not lend i t s e l f as e a s i l y to scale up as that of magnetron sputtering. Ion bombardment can also be achieved i n conventional magnetron sputtering as incomplete trapping of the secondary electrons r e s u l t s i n a plasma i n the proximity of the substrate. The plasma can be further enhanced by a l t e r i n g the ground s h i e l d so that i t does not intersect the f i e l d l i n e s . The energy of the bombarding ions can be controlled through biasing of the substrate. The main problem with t h i s method i s that the ion flux, t y p i c a l l y between 5 to 10% of the deposition f l u x 1 0 , i s often too low to produce the desired f i l m properties. The magnetron system of F i g . 1.3 i s an a t t r a c t i v e a l t e r n a t i v e to the systems described above. I t o f f e r s both ease of design and scale up and a high ion/deposition flux r a t i o ( t y p i c a l l y between 2 to 10 for metal tar g e t s ) . This system, often referred to as an unbalanced magnetron, has the p o t e n t i a l to make magnetron sputtering even more wide ranging than i t presently i s . A summary of the state of the art for Introduction / 10 Figure 1.3 Schematic of a cathode assembly with an unbalanced magnetron. Here the ground s h i e l d has been cut back so that the f i r s t surface that i n t e r s e c t s the f i e l d l i n e s i s the substrate. the unbalanced magnetron i s discussed i n the next section. D. HISTORY OF UNBALANCED MAGNETRONS There had been very l i t t l e research done on the co r r e l a t i o n between the magnetic f i e l d configuration and the ion bombardment of the substrate i n magnetron sputtering p r i o r to 1986. I t had been confirmed, as mentioned previously, that the ion flux was t y p i c a l l y 5%-10% of the deposition f l u x for most dc magnetrons. There had also been some investigations on the e f f e c t s of ion bombardment of the substrate through Introduction / 11 the use of external magnetic f i e l d s with r f magnetron sources 1 1 , 1 2. For dc magnetrons the only major work involved the a p p l i c a t i o n of a magnet behind the substrate 1 3. An increased ion flux was observed i f the magnet was placed so as to aid the f i e l d of the annular magnet. The f i r s t papers 5 , 1 4 to deal s p e c i f i c a l l y with unbalanced magnetrons were published i n 1986. The papers made reference to two types of unbalanced magnetrons, one where the magnetic flux from the centre magnet was much greater than that for the annular magnet and the second for the opposite case. The f i r s t supplied very low ion fluxes and w i l l not considered further i n t h i s text. The second provided a considerable ion flux at the substrate. The magnetron discharge had been analyzed by the use of a large probe 5, 10 cm i n diameter, located along the magnetron axis. In addition i t had a number of smaller probes placed along a radius of the larger one to provide s p a t i a l information. For a magnetron and ground s h i e l d configuration such as that i n F i g . 1.3 the large probe c o l l e c t e d a l l of the discharge current when grounded, and ion currents as high as 10 mA/cm i n the centre when biased at -100V. This resulted i n a ion/deposition flux r a t i o between 2 and 10 for most m e t a l l i c targets. Further investigations revealed that the ion flux was highly dependant on the discharge current and f a i r l y independent of the gas pressure and target composition. There were two basic mechanisms c i t e d to explain these Introduction / 12 e f f e c t s . One was that the f i r s t surface to i n t e r s e c t the f i e l d l i n e s was the probe, rather than the ground s h i e l d . When the probe was grounded i t would serve as the anode of the system and c o l l e c t the discharge current. The ions d r i f t with the electrons to preserve the charge n e u t r a l i t y of the plasma. At -100V bias the electron flux i s e f f e c t i v e l y repelled, leaving only ion bombardment. I t was also noted that out along the magnetron axis there was a c o n s t r i c t i o n of the f i e l d l i n e s which created a magnetic mirror. This was thought to provide an electron trap, leading to further i o n i z a t i o n away from the target surface. The mechanisms involved i n ion production could not be discussed any further than t h i s due to a lack of information on the quantitative aspects of the magnetic f i e l d . Though the above-mentioned would demonstrate the p o t e n t i a l of the unbalanced magnetron, there was a drawback i n that the ion d i s t r i b u t i o n was highly concentrated about the magnetron axis. With the probe located at a distance 6 cm from the target face the ion flux decreased by almost a factor of f i v e at 2 cm along the radius of the probe. The d i s t r i b u t i o n became more spread out further along the axis, but along with an o v e r a l l reduction i n the ion flux. There were a series of publications 1 5" 1 7 which dealt with the modification of f i l m properties with the use of unbalanced magnetrons. Data were given to show the t r a n s i t i o n from a zone 1 to a zone T microstructure with increased ion Introduction / 13 bombardment. The most recent paper 1 8 to date, before t h i s thesis, also dealt with the nature of the plasma discharge for unbalanced magnetrons. The r e s u l t s showed an increasing ion f l u x to the substrate as the magnetic configuration became increasingly unbalanced towards the annular magnet. I t also showed that the ion flux became more concentrated about the magnetron axis under these conditions. I t did not, however, o f f e r any further analysis of the mechanisms involved or give any quantitative d e t a i l s of the magnetic f i e l d configuration. E. OBJECTIVE OF THESIS In previous investigations of unbalanced magnetrons, the description of the magnetic f i e l d has been q u a l i t a t i v e as indicated i n F i g . 1.3 and discussed i n part D. Therefore, only a q u a l i t a t i v e description of the mechanisms involved i n ion production has been possible. I t would be advantageous to be able to understand these mechanisms i n a more quantitative manner. The arguments given above could be validated and i t would be possible to i d e n t i f y any other factors that might be involved. This would be a great aid i n the design of more e f f i c i e n t unbalanced magnetrons and i n scale up. Though a complete quantitative description would be i d e a l , i n p r actice t h i s i s quite d i f f i c u l t to achieve. Plasma discharges are d i f f i c u l t to analyze and conventional magnetron Introduction / 14 sputtering i t s e l f i s not completely understood. However, despite the lack of a thorough understanding, magnetron sputtering has grown into a wide and diverse f i e l d . Often a small increase i n understanding of the o v e r a l l properties has lead to a wider range of applications. The same should be possible for unbalanced magnetrons. This thesis, then, w i l l take an engineering approach to the analysis of unbalanced magnetrons. The emphasis w i l l be on aiding unbalanced magnetron design and scale up through a semi-quantitative analysis of the phenomena. Also, i t i s intended to aid the t h i n f i l m researcher through a thorough investigation of the properties of unbalanced magnetrons. This w i l l include how to maximize both the ion/deposition r a t i o and the e f f e c t i v e area of substrate bombardment. In order to accomplish these goals, two pieces of experimental apparatus were designed; one to measure the magnetic f i e l d and one to measure the r e s u l t i n g discharge c h a r a c t e r i s t i c s . As the discharge i s very dependant on the nature of the magnetic f i e l d , there was a natural coupling between the r e s u l t s of these two experiments. Therefore, when a parameter was varied, such as the r a t i o of strengths of the magnetron's annular and centre magnet both experiments were conducted before another parameter was changed. This method resulted i n a more immediate understanding of the mechanisms involved and helped determine the d i r e c t i o n of the next set of t e s t s . CHAPTER II. THEORY A. INTRODUCTION This chapter discusses the theory of magnetron discharges. An understanding of basic physics of magnetrons i s necessary for both the design of the measurement systems outlined i n Chapter 1, and for an analysis of the observed phenomena i n l a t e r chapters. In keeping with the objective discussed i n Chapter 1, the theory discussed i n t h i s chapter i s outlined only rather than being developed with a great deal of mathematical d e t a i l . The topics are broken down into f i v e sections. The f i r s t gives some general plasma theory necessary for the understanding of the remainder of the topics. The second deals with the theory of plasma probes i n reference to the probe designed i n t h i s thesis. Next there i s a discussion on gas phase c o l l i s i o n s , with emphasis on those which produce ions. The fourth topic i s electron motion i n a plasma i n the presence of e l e c t r i c and magnetic f i e l d s . F i n a l l y the nature of electron trapping i n magnetic f i e l d s used i n planar magnetrons i s discussed. B. PLASMA THEORY A plasma i s a gas with e s s e n t i a l l y three components, neutral atoms, ions and electrons. The i o n i z a t i o n can be achieved either by heating or, as i s the case with magnetron 15 Theory / 16 discharges, through the application of an e l e c t r i c f i e l d . In an ide a l plasma, the ions and electrons w i l l have equal densities and a Maxwellian energy d i s t r i b u t i o n . The average energy of the electrons i s usually much higher than that of the ions i n a magnetron plasma as the f i e l d can transfer energy to l i g h t e r p a r t i c l e s more e f f i c i e n t l y . Typical average energies 1 9 for electrons are 2-8 eV while those f o r ions, around 0.03 eV. As the ions and the electrons have a v e l o c i t y d i s t r i b u t i o n there i s an impingement flux of each species on any surface i n the plasma. This flux i s given by J = -^ p, (2.1) where n i s the p a r t i c l e density and c i s the p a r t i c l e mean speed. For a plasma with predominately f i r s t i o n i z a t i o n products, the ion and electron densities are approximately equal. Therefore, the impingement flux of the electrons i s greater than that of the ions due to the higher energy of the electrons. Plasmas have properties s p e c i f i c a l l y due to the presence of free charged p a r t i c l e s . One i s best i l l u s t r a t e d by considering the e f f e c t of placing a e l e c t r i c a l l y i s o l a t e d surface i n the plasma. I n i t i a l l y , the greater electron flux causes a negative charge b u i l d up on the surface. This bias w i l l then reduce the electron flux u n t i l i t i s equal to the ion f l u x . The net current to the surface i s then zero and the Theory / 17 surface i s biased negative with respect to the plasma. A p o s i t i v e l y charged region e x i s t s i n front of a surface which has a negative bias with respect to the plasma 3. This region l i m i t s the extent to which a dc e l e c t r i c f i e l d w i l l penetrate into the plasma. The length of the f i e l d i s dependant upon both the electron energy and density. A grounded surface w i l l also be biased negative with respect to the plasma, though to a lesser extent than one that i s i s o l a t e d . This bias i s known as the plasma p o t e n t i a l . Another property of plasmas i s that of ambipolar d i f f u s i o n . I f there i s a net d r i f t of the electrons from the plasma an e l e c t r i c f i e l d w i l l r e s u l t . This f i e l d w i l l both retard the motion of electrons and accelerate that of the ions. The r e s u l t i s that the ions and the electrons w i l l d r i f t together. The e f f e c t s of ambipolar d i f f u s i o n on the impingement flux i s discussed i n the next section. C . PLASMA PROBE THEORY F i g . 2.1 shows the id e a l current density-voltage c h a r a c t e r i s t i c s for a plasma probe. When a large enough negative bias i s applied to the probe, only p o s i t i v e i o n i c current i s c o l l e c t e d . This current w i l l be proportional to the ion flux that impinges upon the boundary of the p o s i t i v e space charge region, and i s c o l l e c t e d up to the plasma p o t e n t i a l . As the bias becomes less negative, the probe c o l l e c t s a small electron flux consisting of those electrons Theory / 18 with an energy high enough to overcome the negative \ _ i - v p o t e n t i a l . The electron and ion fluxes to the probe are equal at Vf, the probe f l o a t i n g p o t e n t i a l . When the probe i s biased at the plasma pot e n t i a l , V , i t w i l l c o l l e c t a l l of Figure 2.1 Ideal current density-p voltage c h a r a c t e r i s t i c of a the electron impingement plasma probe. flux and s t a r t to repel most of the ions. In practice, the behaviour of plasmas with respect to probes i s more complicated than discussed above. The additional factors that have to be taken into consideration are discussed below for three points on the j-V curve of i n t e r e s t to t h i s t h e s i s . These are the ion current at -100 V bias, the f l o a t i n g p o t e n t i a l , and the ground current. At -100 V the detector has a s u f f i c i e n t negative bias to repel nearly a l l of the electrons i n the plasma but not enough to cause s i g n i f i c a n t secondary electron emission from the bombarding ions 5. This w i l l give a measurement of the ion flux that w i l l reach a substrate. However, unlike the case described above, i n practice t h i s i s several orders of magnitude greater than the random ion f l u x . The discrepancy i s due to the ambipolar nature of the plasma, which was neglected i n the arguments above. Before i t was assumed that Theory / 19 Plasna -kT./2e Positive Space Charge Region the e l e c t r i c f i e l d d id not penetrate further than the point where the ion and electron densities became equal. In fact there i s a quasi-neutral t r a n s i t i o n region 2 0 known as the Bohm Figure 2.2 Potential v a r i a t i o n sheath which i s shown i n near a negative electrode. The value of zero corresponds to the Fig 2.2. This has the plasma p o t e n t i a l . e f f e c t of increasing the energy of the ions at the boundary of the p o s i t i v e space charge. I f the ions i n the f i e l d free region are assumed to have k i n e t i c energies much less than that of the electrons then, for a magnetic f i e l d free plasma, the ion current density 1 9 to the probe i s (2.2) where n e i s the electron density i n g/cm3, T e the electron temperature i n degrees Kelvin, and m- the mass of the ion i n grams. The ion flux, then, i s dependant upon both the density of the plasma and the electron temperature. At Vp, the f l o a t i n g p o t e n t i a l , the electron and ion flux that reach the detector are equal. A derivation of the f l o a t i n g p o t e n t i a l using 2.2 and assuming a Maxwellian d i s t r i b u t i o n for the electrons gives 1 9 Theory / 20 kT. 2e £ln| (-J0J—) \2.3/n e / - (2.3) The f l o a t i n g p o t e n t i a l i s dependant on the electron temperature. A grounded detector c o l l e c t s a l l of the random electron flux with a energy greater than the plasma p o t e n t i a l . This current can be s i g n i f i c a n t enough to disturb the plasma and i n practice probes operating i n t h i s region are made as small as possible. Assuming that the probe i s small enough not to disturb the plasma, the current density to ground i s D. GAS PHASE COLLISIONS There are a number of c o l l i s i o n s i n a discharge process. The most important of these are those producing ions, which allows the plasma to be both generated and sustained. This can occur when an electron has a k i n e t i c energy equal to or greater than the i o n i z a t i o n energy of an atom. The l i k e l i h o o d of t h i s depends on the cross-section p r o b a b i l i t y of i o n i z a t i o n . For argon the i o n i z a t i o n threshold i s at 15.8 eV and the maximum cross-section occurs around 100 eV. A gas phase c o l l i s i o n can also cause e x c i t a t i o n . This i s of importance as the cross-sections for e x c i t a t i o n are s i m i l a r to those of i o n i z a t i o n . Regions of i o n i z a t i o n glow from the (2.4) Theory / 21 photons emitted from the relaxation of excited atoms and ions. I t i s u n l i k e l y that the photons are caused by the recombination of ions and electrons as the requirement that energy and momentum need be conserved make t h i s highly improbable for a two body c o l l i s i o n . In practice most recombination processes occur on the chamber walls at pressures employed i n magnetron sputtering. E. ELECTRON MOTION IN A PLASMA This section examines the e f f e c t s of e l e c t r i c and magnetic f i e l d s on the motions of electrons. For s i m p l i c i t y , a single electron i s considered and the e f f e c t s of other charged p a r t i c l e s and c o l l i s i o n s are ignored. The c o l l e c t i v e behaviour i s considered i n the next section. The electron motion i s then determined by **-°(g+PxS). (2.5) at m In a uniform B f i e l d the electrons w i l l be unaffected i n the d i r e c t i o n of the f i e l d and w i l l o r b i t the f i e l d l i n e s with a radius of 1 zg- 3.37 (W±)1/2/B cm, (2.6) where B i s i n Gauss, and Wx i s the energy i n eV's of the electron motion perpendicular to the f i e l d . I f an e l e c t r i c f i e l d perpendicular to B i s applied, then a d r i f t v e l o c i t y 1 r e s u l t s i n the d i r e c t i o n perpendicular to both such that Theory / 22 vd-l08EjB cm/sec, (2.7) where E ± i s the f i e l d i n volts/cm. I f B i s not s p a t i a l l y constant, as i s the case for f i e l d s i n planar magnetron systems, then a number of other d r i f t currents r e s u l t . These are derived 3 by considering an electron to have a v e l o c i t y v - V ^ x f . , (2.8) in a magnetic f i e l d B- Bc+ (f g - V)B c, (2.9) —» where Bc i s the f i e l d at the centre of the o r b i t of the electron, and ? g the pos i t i o n of the electron r e l a t i v e to the centre of i t s o r b i t . Then, substituting these equations into 2.5 and assuming that the magnitude of the magnetic f i e l d i s much greater than that of {fg-V)Bcl a number of f i r s t order d r i f t currents r e s u l t . The d r i f t currents are easiest to v i s u a l i z e with the use of magnetic f i e l d l i n e s . Magnetic f i e l d l i n e s are tangential to the f i e l d and the flux between the l i n e s i s a constant. One of the f i r s t order e f f e c t s i s that an electron w i l l follow the curvature of the f i e l d , rather than cross f i e l d l i n e s . I f VB i s p a r a l l e l to the magnetic f i e l d , then a force acts to e i t h e r retard, i f the f i e l d l i n e s are converging, or accelerate, i f the f i e l d l i n e s are diverging, the electron motion. In the absence of an e l e c t r i c f i e l d the electrons Theory / 23 conserve t h e i r magnetic moment, nm, such that VLm-myo/B, (2.10) where v Q i s the v e l o c i t y perpendicular to the f i e l d . I f the d r i f t v e l o c i t y i s ignored, and only the o r b i t a l v e l o c i t y and the v e l o c i t y p a r a l l e l to the f i e l d , v p i s considered, then conservation of energy leads to E- \me(v* + v p 2) . (2.11) If the magnitude of the magnetic f i e l d increases then so must the perpendicular v e l o c i t y . However, i f the energy remains constant, the v e l o c i t y p a r a l l e l to f i e l d has to decrease. I f the increase i n B i s large enough, the p a r a l l e l v e l o c i t y w i l l be reduced to zero and the electron i s r e f l e c t e d back. I f we consider the f i e l d at two points i n space with magnitude B1 and B2 respectively (with B2 less than and B3) , then the maximum energy of an electron that can be r e f l e c t e d from the mirror i s E - m°V°?B* . (2.12) I t would appear that, provided the o r b i t a l v e l o c i t y i s large enough, electrons of any energy can be trapped by the magnetic mirror. However, the r e s t r i c t i o n i s that the value of v Q must be small enough for the i n i t i a l assumptions, which allowed for only f i r s t order e f f e c t s i n a f i e l d not s p a t i a l l y constant, to hold. Theory / 24 F. PLANAR MAGNETRON DISCHARGES The theory of planar magnetron discharges has been outlined i n Chapter 1. In t h i s section the theory w i l l be developed i n more d e t a i l . As stated previously, i n a planar magnetron discharge, the secondary electrons are i n i t i a l l y accelerated away from the target surface, but are forced back towards the target through the Lorentz force. Assuming a low emission energy, which i n practice i s a few electron v o l t s 2 1 , the furthest distance an emitted electron w i l l t r a v e l from the target surface i s 1 9 where VT i s the target voltage. I f the electron t r a v e l s further than the extent of the target sheath, Vd i s equal to the plasma p o t e n t i a l . Ideally the electrons should return to the target with the same k i n e t i c energy as that upon emission. However, for t y p i c a l magnetron sputtering conditions with pressures of approximately 1 Pa and f i e l d strengths of approximately 300 gauss, only about h a l f the electrons return to the surface due to energy losses through c o l l i s i o n s and interactions with plasma i n s t a b i l i t i e s 2 2 . The electrons then undergo c y c l o i d a l motion around a centre which d r i f t s about the axis of the magnetron. The electrons are large l y contained i n the region where the f i e l d l i n e s are p a r a l l e l to the target surface due to magnetic mirror e f f e c t s (Fig 1.2). d = — 2m„ (vd-vT) Theory / 25 Of i n t e r e s t i s the mechanisms through which electrons escape the magnetron trap. The c l a s s i c a l theory for conduction across a magnetic f i e l d predicts a negative space charge region i n front of the target i n magnetron discharges due to the electrons having a lower mobility across f i e l d l i n e s than the ions. In practice i t i s believed 2 3 that due to electron-electron interactions the actual d i f f u s i o n c o e f f i c i e n t i s much higher than the c l a s s i c a l predictions allowing for an electron mobility greater than that of the ions. Di f f u s i o n of t h i s nature has been observed by Rossnagel et. a l . 2 4 when determining the r a t i o between the d r i f t and discharge current. CHAPTER III . MAGNETIC MEASUREMENT SYSTEM A. INTRODUCTION A system was designed and b u i l t to measure the magnetic f i e l d of a c i r c u l a r magnetron, an example of which i s shown i n Fig. 3.1. As the magnetic f i e l d i s symmetric i n the angular d i r e c t i o n the system had only to measure i n an r-z plane. The measurement system was automated as i t was f e l t that i n the long term t h i s would be the most e f f i c i e n t method of data c o l l e c t i o n . The state of the art of microcomputers, interface cards, and electronics i s such that i t i s r e l a t i v e l y easy to design a system capable of handling large amounts of data and taking accurate and repeatable measurements. Though the data c o l l e c t i o n was r e l a t i v e l y straightforward, the presentation of the data was not and had to be considered before the system was designed i n any d e t a i l . The problem lay i n how to graphically represent a two dimensional vector f i e l d . From the discussion of electron motion i n a f i e l d i n Chapter 2, i t would appear advantageous to do so by p l o t t i n g the f i e l d l i n e s . Magnetic f i e l d l i n e p l o t s would lend themselves well i n discussions on the electron motion i n the plasma discharge. Magnetic f i e l d l i n e s represent the actual f i e l d i n two ways. F i r s t the f i e l d l i n e s are tangential to the f i e l d at each point i n space. Secondly, the spacing between the f i e l d l i n e s i s inversely proportional to the f i e l d strength or, 26 Magnetic Measurement System / 27 Figure 3.2 Cross-section of an i n f i n i t e l y long rectangular magnetron symmetric about the y-z plane. Magnetic Measurement System / 28 stated otherwise, magnetic flux i s a constant between f i e l d l i n e s . Therefore, plots of t h i s nature w i l l show both d i r e c t i o n and the r e l a t i v e f i e l d strength f o r each point i n space. As magnetic f i e l d l i n e s cannot be measured d i r e c t l y , the data had to be processed i n order to produce the f i e l d l i n e p l o t s . The method of processing i s discussed i n the next section. B. MAGNETIC FIELD LINE PLOTTING The purpose of t h i s discussion i s to show how magnetic f i e l d l i n e s can be derived from Maxwell's equations and how a car e f u l consideration of the coordinate system can simplify the task. Consider Maxwell 1s equations governing a system i n the absence of an e l e c t r i c f i e l d Vxi?= — J, (3.1) c V-B-O. (3-2) In our system the current density, J, i s zero and the medium i s one of constant permeability. Eq. 3.1 can then be written as VxB = 0 . (3.3) Therefore, the magnetic f i e l d must be the gradient of a scaler p o t e n t i a l , which, i n analogy to e l e c t r o s t a t i c f i e l d s , i s Magnetic Measurement System / 29 referred to as the magnetic scaler p o t e n t i a l . This leads to S--V<|>m. (3.4) As the f i e l d of inter e s t possesses symmetry i n the angular d i r e c t i o n , the l a s t equation reduces to two dimensions. Though the problem w i l l eventually have to be solved for c y l i n d r i c a l coordinates, i t i s easiest to f i r s t consider a s i m i l a r problem i n the Cartesian system, that of a f i e l d from a i n f i n i t e l y long rectangular magnetron such as that i n Fig 3.2. Here B - -  m , B - -  m . (3.5) Bx ' y dy I f the equipotential surfaces of the magnetic scaler p o t e n t i a l are considered, from Eq. 3.4 i t can be seen that the l i n e s perpendicular to these w i l l always be i n the d i r e c t i o n of the magnetic f i e l d . From the Cauchy-Riemann r e l a t i o n s , for any well behaved scaler f i e l d , * m(x,y) , such as that above, there exists a conjugate f i e l d , T m(x,y), such that dx dy ' dy dx whose contours are perpendicular to those of * m(x,y). 7 m(x,y) i s referred to as the imaginary scaler p o t e n t i a l and i s related to the magnetic f i e l d as dH B = -  m , B = m . (3.7) By ' y Bx I f the change i n Ym i s a constant, as i t w i l l be between the contours of Tm, then from Eq. 3.7 the flux (B-fidAi i s a Magnetic Measurement System / 30 constant. Therefore both requirements for magnetic f i e l d l i n e s are s a t i s f i e d by the contours of the imaginary magnetic scaler p o t e n t i a l . Hence contour plots of T|]1(x,y) w i l l give the magnetic f i e l d l i n e s . The soluti o n to 3.7 i s Y m(x,y) -Vm(x.O) -fyBx(x.y')dy'. (3.8) Jo or T.U.y ) -Y a(0,y) + f*By(x'.y) dx*. (3.9) If we l e t x = 0 i n 3.8 and substitute into the 3.9, the re s u l t i s VJx.y) =T m(0,0) -| o yB x(0,y /)ciy / + fXBy(x',y)dxf. (3.10) ¥ m(0,0) i s an ar b i t r a r y constant and can be set to zero without a f f e c t i n g the r e l a t i v e values of the contours. The f i r s t i n t e g r a l can also be set to zero through a consideration of the symmetry about the y-axis (Fig 3.2) ; i . e . the value of Bx on the y-axis i s zero. This s i m p l i f i e s equation 3.10 to VJx,y) -foXBy(x',y)dx'. (3.11) In examining equation 3.11 we can see that the in t e g r a l represents the flux through an area i n the x-z plane located at y of length 0 to x, normalized i n z. To i l l u s t r a t e how the contours of Y m(x,y) represent the f i e l d l i n e s F i g 3.2 should be considered again. As both the magnetic f i e l d and 7 are functions of x and y they can be Magnetic Measurement System / 31 represented on the same set of axes, as w i l l be done i n the next argument. Near the o r i g i n By i s negative. Inspection of Eq. 3.11 shows that, near the o r i g i n , as x increases Ym(x,0) decreases, with the distance between i t s contour l i n e s along the x-axis inversely proportional to the magnetic f i e l d strength. Further along the axis the d i r e c t i o n of By changes and the value of Tm(x,0) becomes more p o s i t i v e . The contours w i l l then take the same values of those further down the axis. Between the f i r s t f i e l d l i n e displayed on the x-axis at point 'a' and the o r i g i n there i s a fixed amount of flux. This f i e l d l i n e also intersects l i n e x' at 'b'. At 'b' both the value of 7 and the flux between the f i e l d l i n e and the m axis are the same. I f flux crossed the y-axis then the f i r s t i n t e g r a l i n 3.10 would have to be employed as a correction factor for the value of 7 m . However, due to the system's symmetry, the magnetic f i e l d i s always p a r a l l e l to the y-axis, hence the y-axis i t s e l f i s a contour l i n e . In a c y l i n d r i c a l system such as F i g . 3.1 the system i s angularly symmetric and there i s no Br component along the z-axis. In t h i s way i t i s analogous to Fig 3.2. However the d i f f e r e n t geometry r e s u l t s i n an area component of rdrdB instead of dxdz. Therefore, the i n t e g r a l that represents the flux through an area i n the r - 8 plane located at z of length 0 to r, normalized i n 6 i s Magnetic f i e l d p lots can be produced for a c i r c u l a r Magnetic Measurement System / 32 VJr.z) - jJr'Bz(x',z)dr'. (3.12) magnetron by f i r s t measuring Bz i n an r-z plane. The value of T m can then be determined at each point through numerical integration. The contours of 7m can then be plotted through a v a r i e t y of commercial graphics packages. C. MEASUREMENT SYSTEM 1. D e s i g n A system was developed that could measure the magnetic f i e l d i n a two dimensional g r i d . For a c i r c u l a r magnetron t h i s would map out the f i e l d i n an r-z plane, taking advantage of the angular symmetry. Though only Bz would have to be measured to produce the f i e l d plots, the system was designed to be able to measure either component i n case measurements of Br were desired. A block diagram of the el e c t r o n i c control system i s shown i n Fig 3.3 and a schematic of the mechanical aspect i n F i g 3.4. The l e v e l of accuracy required had to be considered before the system was designed i n d e t a i l . In keeping with the objectives discussed i n Chapter 1, i t was decided that any calcul a t i o n s would only required measurements accurate to a few percent. For the production of f i e l d l i n e p l o t s , the data was integrated. This reduced the e f f e c t of any random errors i n the system though care had to taken with respect to any Magnetic Measurement System / 33 Computer Corrtrol Program Lab Pac Interface Card PORTA PdRTB • PORTC Axial Stepping Motor Controller r Radial Stepping Motor Controller Axial Radial Stepping Stepping Motor Motor Multiplexer Prog. Gain Amp, Pre-Amp 8-Blt ABC H a l l P r o b e Constant Volt-age Source C12 Volts) Figure 3.3 Block Diagram of the ele c t r o n i c s of the magnetic f i e l d measurement system. systematic errors. The f i r s t source of systematic error considered was the detector i t s e l f . I t had to give repeatable r e s u l t s but needed only be accurate to a few percent. Therefore, a Hal l probe, which was accurate to about +/- 1%25, was deemed s u f f i c i e n t . A magnetometer was designed s p e c i f i c a l l y for t h i s system. There was a Lab Pac interface card already available i n the lab; however, the commercial meters immediately on hand were found to be incompatible with the interface card for the magnetometer system. Though the meters gave a voltage output Magnetic Measurement System / 34 proportional to the measured f i e l d , t h e i r maximum was less than that which could be read by the card's 8-bit A-D convertors. Also, i t was impossible to set the meters gain e l e c t r o n i c a l l y , which li m i t e d the dynamic range. Rather than order a new meter or interface card i t was decided that from the point of view of time and expense that i t would be best to take advantage of the inexpensive H a l l probe chips that have recently become available. These chips turned out to be simple to use and an amplifier was custom b u i l t for them allowing for a good deal of f l e x i b i l i t y when i n t e r f a c i n g the system with other components. The H a l l probe chosen was a Sprague UGN-3501K. The probes output was both temperature (l%/5 deg) and input voltage (1%/lvolt) dependant. The input voltage dependence was e a s i l y remedied with a constant voltage source accurate to 0.03 v o l t s . The temperature s e n s i t i v i t y was dealt with by operating i n a controlled environment where the temperature did not vary by more than 2 degrees Celsius. The output of the chip was also non-linear for applied f i e l d s of more than 1000 Gauss. The non-linearity was corrected for by comparing the output of the Sprague probe to that of a commercial meter's i n the presence of a magnetic f i e l d . The f i e l d was supplied by a large Helmholtz c o i l , which produces a uniform f i e l d i n the v i c i n i t y of i t s axis. The outputs of the two detectors were compared for a range of values between 0 and 3000 Gauss. The subsequent values measured by the Sprague Magnetic Measurement System / 35 probe could then be corrected for through cubic spline f i t t i n g . Though t h i s increased the d i g i t a l error, at the maximum f i e l d measured (2400 Gauss) the non-linearity was only 25%. Also the vast majority of data points were below 1000 Gauss. The output of the H a l l probe was passed through a preamplifier, which had an adjustable gain and o f f s e t control. The signal l e v e l was increased with the use of an instrumentation amplifier with a programable gain of 1, 3, 10 and 30. The output of the amplifier was sent to an 8 b i t analog to d i g i t a l converter interfaced with the computer. The measurement system had a dynamic range from 1 to 3000 Gauss and an accuracy between 1 to 3% , depending on the input signal l e v e l and the amplifier gain. The analog-to-digital convertor and three 8 b i t control ports were supplied by a Lab Pac interface card which could be programmed with a high l e v e l language. Ports A and B were set as output ports to program the stepping motor c o n t r o l l e r s while Port C had four input and four output l i n e s for c o n t r o l . Two of the output l i n e s were used to control the amplifier gain while the remainder, combined with the input l i n e s , provided the "handshaking" for the stepper motor c o n t r o l l e r s , which were b u i l t by the UBC Physics e l e c t r o n i c s shop. The mechanical system allowed for motion i n a plane i n two perpendicular d i r e c t i o n s . The stepping motors were attached to a d r i v i n g shaft of 20 threads per inch and had a Magnetic Measurement System / 36 Lower Guide B a r , Magnet ron Ho lde r S tepp ing M o t o r s D e t e c t o r , Ho lde r T e f l o n Guide 7 ^ Upper Guide B a r 1 Driving S h a f t s Figure 3 .4 Schematic of the mechanical aspects of the measurement system. resolution of 200 steps per revolution. Theoretically, t h i s made the po s i t i o n accuracy better than +/- .01 mm. There was, however, a s l i g h t wobble i n the system which r e s t r i c t e d the accuracy to approximately +/- .05 mm. The d r i v i n g shaft moved a t e f l o n guide which rode along two p a r a l l e l bars. Only one end of the underside of the t e f l o n guide had a grove i n which a bar would f i t , while the other side was mill e d f l a t . Therefore only the bar that f i t into the groove had to be accurately aligned. The software of the system was written as a l i b r a r y of routines run by a main c o n t r o l l e r program. Each routine did a single function such as advance the stepping motor a sp e c i f i e d number of steps or performed part of the "handshaking". These could then be linked together by the main program to run a scan of the desired choice. Magnetic Measurement System / 37 2. Alignment and C a l i b r a t i o n The error i n the measurement of the f i e l d was, at best, about 1 %. This was used as a guide when considering how p r e c i s e l y the system needed to be c a l i b r a t e d and aligned. I t was also kept i n mind that the systematic errors had to be kept small enough not to noticeably a f f e c t the contour plots of Ym • m The f i r s t problem to be dealt with was to ensure that the two t e f l o n guides ran perpendicular to one another. The task was s i m p l i f i e d with only one of the two supporting rods acting as a guide bar. Both t e f l o n guides were milled as rectangular blocks to within the smallest possible tolerance ( +/- .002 " ). The guiding groves were then cut p a r a l l e l to t h e i r adjacent edge. The top guiding bar was mounted into the top assembly so that i t would l i e p a r a l l e l to the edge, which, i n turn was l i n e d up with edge of the lower t e f l o n guide that was perpendicular to i t s groove. I t was found that the two guides were perpendicular to within a 0.5 of a degree. The next step was to a l i g n the H a l l probe detector so that i t s axis was p a r a l l e l to the upper guide bar. This could be accomplished by supplying a magnetic f i e l d p a r a l l e l to the lower guide bar then adjusting the angle of the detector u n t i l i t s output was zero. A Helmholtz c o i l was used to generate a f i e l d of 50 Gauss. One end of a s t r a i g h t rectangular bar was placed along the side of the two c o i l s and Magnetic Measurement System / 38 the other end l i n e d p a r a l l e l with the lower guide bar. The angle of the detector was then adjusted u n t i l i t s output was +/- .5 Gauss. This aligned the detector to within +/- .5 degrees. As the earth's magnetic f i e l d was a factor at t h i s l e v e l , the current to the c o i l was turned o f f and the amplifier o f f s e t adjusted, then the current turned back on and the process repeated u n t i l the detector gave the same reading i n both cases. The detector was then c a l i b r a t e d . The axis of the detector was set p a r a l l e l to the Helmholtz f i e l d ; the preamplifier, to unity gain; and the instrumentation amplifier, to a maximum gain of t h i r t y . The f i e l d was increased i n 2.5 Gauss steps up to 60 Gauss. A l e a s t squares f i t was performed to determine the output voltage to f i e l d r a t i o of the system. The preamplifier voltage was then adjusted so that the output l e v e l of the amplifier for 1 Gauss at maximum gain was equal to the minimum voltage difference that the AD convertors could detect, 39.4 mV. The f i e l d would subsequently be measured to within an integer value, which allowed for e f f i c i e n t data storage. The magnetron holder was then adjusted so that i t s axis was aligned p a r a l l e l to the axis of the detector. This task was s i m p l i f i e d by i n i t i a l l y s e t t i n g the face of the magnetron holder p a r a l l e l to the upper guide bar. To accomplish t h i s , one end of a square was placed against the upper guide bar while the other end was positioned approximately 1 cm from the Magnetic Measurement System / 39 magnetron holder. Then, by the use of a p a i r of c a l l i p e r s and placing the square to either side of the magnetron holder the face was aligned. The magnetron holder was then rotated 90 degrees aligning i t s axis p a r a l l e l to that of the detector. The height of the magnetron was set so that i t s axis was at the same l e v e l as that of the detector. A c y l i n d r i c a l piece with a small indent i n i t s centre was machined on a lathe, then placed i n the centre of a magnetron set i n the holder. A l i n e on the detector holder was etched at the same l e v e l as the centre of the detector. A small square was placed up against the end of the s t e e l piece and a machined block was used to set i t s height to that of the etch i n the detector holder. The height of the magnetron was then adjusted so that the indent would be at the same height as the top of the str a i g h t edge. Once the magnetron holder was at the correct angle and height, the i n i t i a l s t a r t i n g p o s i t i o n of the detector could be determined. The s t e e l centre piece was l e f t i n place from the height alignment. A square was placed against that side of the piece which was away from the main assembly. The detector was positioned u n t i l the edge of i t s holder came i n contact with the side of the square. By measuring the distance of the detector centre from the holder • s edge and the centre piece diameter the detector could be positioned r a d i a l l y . This was checked through a scan across the magnetron face which should be symmetric about the o r i g i n . To record the Magnetic Measurement System / 40 proper po s i t i o n , the distance between the lower t e f l o n guide and the assembly edge along the guide bar was measured. This was used to rezero the detector. For positioning along the axis a mark 3 mm from the edge of a square was made. The end of the square was placed against the face of the centre piece and the assembly positioned so that the mark bisected the depth of the detector. 3. F i e l d Scan Tests The system was i n i t i a l l y designed without the use of shaft encoders i n order to simply the electronics and the assembly. Scanning tests were then employed to see i f the system would run with the desired l e v e l of accuracy without d i r e c t p o s i t i o n measurement. Backlash was a problem that had to be corrected. To avoid t h i s the data were always c o l l e c t e d i n the same d i r e c t i o n of motion, with the detector moving away from the magnetron. In a scan the assembly would c o l l e c t measurements i n the a x i a l d i r e c t i o n , then move the required step i n the r a d i a l d i r e c t i o n , return to the magnetron face, then begin another set of measurements. This l e t the stepper motor with the l e a s t load do the most of the work. In order to eliminate the backlash before beginning the next set of measurements i n the a x i a l d i r e c t i o n , the detector would move an extra step towards the magnetron face then back again. I t was found that Magnetic Measurement System / 41 by moving the detector i n t h i s manner that 10 scans along the axis at the same r a d i a l p o s i t i o n produced the same r e s u l t s . The p o s i t i o n accuracy was also tested by doing a scan of 65 by 65 mm and measuring the i n i t i a l and f i n a l p ositions. The f i n a l p o s i t i o n was accurate to within +/- 0.1 mm. There was a small amount of d r i f t i n the voltage output. I f the system was l e f t on for a day before beginning a scan the d r i f t was less than 1 Gauss per day. A scan to produce a f i e l d l i n e p l o t took about 10 hours. CHAPTER IV. PLASMA PROBE A. INTRODUCTION A plasma probe was designed and b u i l t to operate at the three points on the j-V curve discussed i n Chapter 2. The re s u l t s would be coupled with the magnetic f i e l d information to help understand the properties of unbalanced magnetrons. The most complete study of t h i s nature to date was done by Window et. a l . 5 . In order to be able compare with t h e i r r e s u l t s a s i m i l a r probe was b u i l t . The probe was not a conventional one i n that i t was quite large, and would a f f e c t the plasma discharge. However, the probe was designed to take the place of the substrate and i t s holder and hence became part of the discharge system. The purpose of the probe was predominately to measure the e f f e c t s of the plasma on the substrate, such as through ion bombardment. However, i t was s t i l l possible to analyze the plasma to a c e r t a i n extent at the more negative values as has been done i n other studies 2 6. B. MEASUREMENT SYSTEM 1. Design As stated i n the Introduction, the probe was s i m i l a r to that i n Ref. 5. A 10 cm diameter copper disk was used as a probe f o r the average behaviour of the discharge while f i v e smaller probes of 0.8 cm diameter provided s p a t i a l 42 Plasma Probe / 43 CATHODE DRIVING ROD ASSEMBLY GUIDING RDD H n a DETECTOR '— BACKING PLATE y LEVELING SCREWS - FEEDTHROUGH CLAMPING BOLT RADIAL ADJUSTMENT CHAMBER DDDR STEPPING MDTDR Figure 4 .1 Plasma Probe System information. The f i r s t of the smaller probes was positioned i n the centre of the large probe with the others placed at 1 cm i n t e r v a l s along a radius. Though the probe was very s i m i l a r to that of Ref 5, the supporting mechanism (Fig 4.1) d i f f e r e d as i t had to be adapted to d i f f e r e n t sputtering chamber. In Ref. 5 the probe was attached to s l i d i n g seal, positioned d i r e c t l y under the magnetron. Instead, i n our system, advantage was taken of a rotary feedthrough which normally turned a substrate table by means of a stepping motor outside of the chamber. The probe assembly could be moved along the magnetron's axis by means of a threaded d r i v i n g rod attached to the Plasma Probe / 44 ® ± - -100 V Figure 4.2 Schematic of e l e c t r i c a l components of the plasma probe feedthrough, while a guide rod would provided the necessary counter torque to allow for motion i n the a x i a l d i r e c t i o n only. A c i r c u l a r indent i n the chamber door around the feedthrough allowed for a simple clamping system for the guiding rod. An assembly to hold the detector was designed. For support, the probes were mounted on a 5 mm deep nylon in s u l a t i n g backing plate which, i n turn, was fastened into a copper holder. The detector holder was mounted onto a s l i d e i n the main assembly through three l e v e l l i n g screws which would allow the probe to be aligned p a r a l l e l to the magnetron face. The s l i d e i t s e l f could move along the assembly frame, which allowed the probe to be positioned along the chamber radius. The angular p o s i t i o n of the probe could be adjusted by Plasma Probe / 45 the placement of the base plates. Due to the i n s u l a t i n g backing plate, the probes were e l e c t r i c a l l y i s o l a t e d from the chamber. A schematic of the e l e c t r i c a l setup i s F i g 4.2. I t should be noted that during a measurement a l l of the probes were at the same po t e n t i a l . 2. Alignment The f i r s t step i n the alignment process was to ensure that the large probe was p a r a l l e l to the target face. This presented some d i f f i c u l t y i n that the probe assembly was mounted onto the chamber door. However, as the door i n the closed p o s i t i o n was p a r a l l e l to the target face, the probe was aligned to the chamber door instead. During the alignment process, the door was held as v e r t i c a l as possible to avoid any problems r e s u l t i n g from the s h i f t i n g of the centre of gravity. Next, the centre of the probe was aligned along the magnetron axis. This was done quite simply by machining a c y l i n d r i c a l piece that f i t snugly into the opening of the ground s h i e l d . A hole was d r i l l e d i n the centre of the piece i n which a f e l t marker would j u s t f i t . The probe p o s i t i o n could be adjusted r a d i a l l y by the s l i d e on which the detector holder was mounted and angularly by ro t a t i n g the entire assembly. Then by cl o s i n g the magnetron door and observing the r e s u l t i n g mark on the detector face i t could be Plasma Probe / 46 centred to within +/- .5 mm accuracy. When a set of tests were taken for comparison purposes, the assembly was l e f t aligned i n the chamber for the entire set of t e s t s . 3. Operation I n i t i a l l y the system ran with an acceptable wobble i n both the a x i a l and r a d i a l d irections of approximately +/- .25 mm. However, a f t e r a few sputtering runs, some of the thread of the main assembly wore away and an unacceptable wobble developed. This was corrected by the use of a rubber band between the main assembly and the base plate. A f t e r t h i s the system ran as smoothly as before. CHAPTER V. EXPERIMENTAL RESULTS A. INTRODUCTION The systems described i n Chapters 3 and 4 were used to measure the magnetic f i e l d and the r e s u l t i n g discharge c h a r a c t e r i s t i c s for a series of magnetrons. The r e s u l t s are presented i n t h i s Chapter along with an analysis of the observed phenomena. The tests were broken down into four sections. The f i r s t set of experiments involved examining the e f f e c t s of a l t e r i n g the r e l a t i v e strengths of the centre and annular magnets. For the second set of experiments, the e f f e c t s of target thickness and the magnetic f i e l d configuration were examined. The magnetic f i e l d was further altered through changing the geometry of the centre piece. In the t h i r d set of experiments, an optimum magnetron configuration was used while various sputtering parameters were varied. The magnetron had an annular magnet and a mild s t e e l centre. The charged p a r t i c l e fluxes were measured as a function of sputtering gas pressure, discharge power, and target composition. The f i n a l set of tests examined the ion/deposition flux r a t i o . The deposition rates and ion current were determined as a function of a x i a l p o s i t i o n . 47 Experimental Results / 48 B. RESULTS FROM VARYING THE RELATIVE STRENGTHS BETWEEN THE ANNULAR AND CENTRE MAGNET 1. Magnetron C o n f i g u r a t i o n s Five d i f f e r e n t magnetic configurations were examined i n t h i s set of t e s t s . The basic geometry i s shown i n F i g 3.1. In a l l cases the annular magnets were neodymium based a l l o y with an outer diameter of 4.95 cm and were 0.64 cm deep. The configurations had a c y l i n d r i c a l centre piece of either neodymium a l l o y magnet or mild s t e e l . The dimensions of the centre piece were the same for each magnetron with a radius and depth of 0.64 cm. The difference between the magnetic configurations i s shown i n Table 5.1. The f i e l d strength of the centre magnet was measured on the axis at a distance 3 mm from the surface. The centre magnet of the f i r s t magnetron was f u l l y magnetized while those for magnetrons 2 and 3 were demagnetized through heating, to 50% and 25% of t h e i r maximum value. The f i e l d strengths of the annular magnet were measured for t h e i r greatest value at a distance 3 mm from the surface. The component measured was Bz. The magnets were both mounted on a mild s t e e l pole piece during the measurement process. The flux from the surface of the magnets were also calculated for purpose of comparison. The method of c a l c u l a t i o n i s discussed i n the Appendix. The r a t i o of the flux crossing the surface Experimental Results / 49 T a b l e 5.1 Magnetic C o n f i g u r a t i o n s Centre Piece Annular Magnet Magnetic Config. Material F i e l d Strength Inner radius F i e l d Strength (Gauss) (cm) (Gauss) 1 Neo. Al l o y 2300 2.05 950 2 Neo. A l l o y 1800 2.05 930 3 Neo. Al l o y 1350 2.05 910 4 Mild Steel 700 2.05 900 5 Mild Steel 1450 1.65 1500 (when isolated) between the annular magnet and centre magnet for configuration 1 i s , to within a few percent, 2. The r a t i o of the flux between the annular magnets of magnets 5 and 1 i s 1.5. 2. Data C o l l e c t i o n and P r o c e s s i n g A l l of the magnetic f i e l d measurements were taken with the same procedure. A g r i d of data points spaced 1 mm apart i n the a x i a l d i r e c t i o n and 0.5 mm r a d i a l l y was c o l l e c t e d for the value of Bz. The g r i d extended 6.4 cm along the axis and 3.2 cm i n the r a d i a l d i r e c t i o n and started on the axis 3 mm from the face of the magnets. Before each scan, the o f f s e t of the amplifier was adjusted to cancel the e f f e c t s of the earth's magnetic f i e l d and any zero o f f s e t . The magnetic configuration was then placed i n Experimental Results / 50 the holder and the detector positioned manually to ensure r e p e a t a b i l i t y . The s t a r t i n g p o s i t i o n was approached i n the same d i r e c t i o n of motion as that of the data c o l l e c t i o n to avoid any problems r e s u l t i n g from backlash. After the data was collected, i t was processed i n order to produce magnetic f i e l d l i n e s p l o t s . The non-linearity of the detector was corrected for through cubic spl i n e i t e r a t i o n . The data was then integrated along the radius at each a x i a l p o s i t i o n through the use of Simpson's Rule which requires a middle value for each added " s t r i p " . This produced the values for the imaginary magnetic scaler p o t e n t i a l for a 1 mm2 g r i d of 64 by 32 mm. The pote n t i a l f i e l d was r e f l e c t e d about the magnetron axis, then added to the o r i g i n a l f o r the purpose of presentation. 3. F i e l d Line Plots The magnetic f i e l d l i n e s of configurations 1 to 5 are presented i n Figs 5.1 through 5.5. As discussed i n Chapter 3, these are ac t u a l l y the contours of Tm. A cross-section of the magnets i s also presented shown both to scale and i n i t s p o s i t i o n r e l a t i v e to the f i e l d . For the purpose of l a t e r arguments, the positions of the target used i n the sputtering run are also shown. These have been represented by dotted l i n e s above the magnetron. In each case the target i s 3 mm th i c k expect for magnetron 5, where a 8 mm target i s also displayed. For magnetron 1 the placement of the ground s h i e l d Experimental Results / Figure 5.1 Magnetron 1. This configuration i s used i n conventional magnetron sputtering. Shown also i s the ground s h i e l d configuration. Figure 5.2 Magnetron 2. The strength of the centre magnet has been reduced to 50% of that of magnetron 1. Experimental Results / Figure 5.3 Magnetron 3. The strength of the centre magnet has been reduced to 25% of that of magnetron 1. Figure 5 .4 Magnetron 4. Here the centre magnet has been replaced by a mild s t e e l piece. Experimental Results / 53 Figure 5 . 5 Magnetron 5. The annular magnet i s replaced by one of a greater strength. used during a sputtering run i s presented. The flux between the f i e l d l i n e s i s constant f o r a l l f i v e p l o t s , 5 x 10"6 Wb. This was achieved p a r t i a l l y by assigning equally spaced values for the contours. Extra care had to be taken, however, as the imaginary scaler p o t e n t i a l was r e f l e c t e d about the axis. The value of T on the axis was m zero. To s a t i s f y the requirement that flux was constant between the f i e l d l i n e s , the p o s i t i v e and negative values were equally spaced about zero. The flux between the f i r s t p o s i t i v e contour, which was the f i r s t to return the rear of the magnetron, and the axis was then h a l f of that between the contour l i n e s . The flux between t h i s f i e l d l i n e and i t s r e f l e c t i o n was then equal to the flux between the other f i e l d Experimental Results / 54 400 <a C l A -400 -4-> 6 C <" -800 M W 3 -1200 QJ • rH .a - i 6 o o QJ PI «>-2000 s -2400 0 5 10 15 20 25 30 35 40 45 50 55 60 65 Position (cm) Figure 5.6 Magnetic f i e l d strength along the axis for magnetrons 1 to 5. The inset shows an increased scale for the f i e l d . l i n e s . The a x i a l component of the magnetic f i e l d along the axis i s shown i n F i g . 5.6. As the r a d i a l component here i s zero t h i s i s the strength of the f i e l d . When the magnetic f i e l d i s v i s u a l i z e d i n three dimensional space, the f i e l d l i n e s are rotated about the axis, rather than extended outwards from the page as i n a cartesian coordinate system. This does create some d i f f i c u l t y i n interpretation, as equally spaced f i e l d l i n e s away from the axis do not represent a constant magnetic f i e l d due to the r a d i a l factor i n Eq. 3.12. This i s taken into considerations i n l a t e r arguments. Experimental Results / 55 4. Plasma Probe Measurements The charged p a r t i c l e flux r e s u l t i n g from the use of magnetrons 1 through 5 i n a sputter run was measured with the apparatus described i n Chapter 4. Each of the experiments were run i n the same manner. F i r s t the detector was placed i n i t s s t a r t i n g p o s i t i o n r e l a t i v e to the magnetron face. As the assembly was fastened to the chamber door, the probe was i n i t i a l l y positioned to the door. This, i n turn, would po s i t i o n i t with respect to the magnetron face once the door was closed. During the positioning process, the door was raised u n t i l i t was open ju s t enough to allow for measurement inside the chamber, preventing any error a r i s i n g from a s h i f t i n g centre of gravity. The probe was then properly positioned i n the same d i r e c t i o n of motion i t would take during the sputtering run to avoid problems with backlash. The chamber was closed and pumped down to pressure of approximately 5 X 10"5 Torr. Argon gas was then introduced at a pressure of 1 Pa. The discharge current was held constant at 2 00 mA during the run. A copper target, 3 mm thick, was employed. Measurements were taken at 1 cm i n t e r v a l s along the magnetron axis from 3 to 11 cm. The data c o l l e c t e d at each a x i a l p o s i t i o n included both average and s p a t i a l information. The ground current was measured for both the large and the small central probe. I t should be noted that when measurements involving the large Experimental Results / 56 probe were taken, the smaller probes were also included i n the c i r c u i t . The probes were then i s o l a t e d e l e c t r i c a l l y to determine the f l o a t i n g p o t e n t i a l . F i n a l l y a -100V bias was applied and the ion currents were measured for both the ent i r e system and each of the smaller probes. In Chapter 1 i t was discussed how the placement of the ground s h i e l d was also an important factor i n the nature of the discharge. For magnetron 1 two ground shields were used. The f i r s t , referred to as the r e s t r i c t i v e s h i e l d , intersected the f i e l d l i n e s between the annular and centre magnet. The second, referred to as the open shie l d , was cut back away from these f i e l d l i n e s . The second configuration was converted to the f i r s t through the addition of a l i p i n i t s opening. The opening of the second configuration and the l i p are shown i n F i g . 5.1. When the p a r t i c l e fluxes were examined for magnetrons 2 to 5 the open s h i e l d was always used. 5. Results for the Large Plasma Probe Results from the measurements taken at an a x i a l distance of 6 and 10 cm with the large plasma probe are shown i n Table 5.2. A set of r e s u l t s are displayed for Magnetron 1 for each ground s h i e l d configuration. The discharge voltage increased as the centre magnet's strength was reduced, while the annular magnet was constant. The use of magnetron 5 decreases the power requirements to below that of magnetron 1. The probe c o l l e c t s about h a l f of Experimental Results / 57 Table 5.2 Discharge c h a r a c t e r i s t i c s measured by the l a r g e probe f o r magnetrons 1 t o 5. Probe Probe Probe Magnet. Ax i a l Disch. s e l f - ground - 100 V Config. Dist. Voltage bias current current (cm) (V) (V) (mA) (mA) 1 6 334 -17.0 94 -2.9 (res.) 10 332 -15.6 76 -2.6 1 6 332 -20.1 190 -9.0 (open) 10 334 -15.3 136 -5.3 2 6 342 -20.0 193 -9.8 10 345 -14.8 144 -4.9 3 6 361 -20.0 193 -10.3 10 361 -14.0 151 -5.0 4 6 397 -20.5 200 -13 . 6 10 397 -10.7 151 -6.3 5 6 307 -29.8 200 -39.1 10 310 -17.7 190 -18.4 the discharge current at 6 cm for magnetron 1 with the r e s t r i c t i v e s h i e l d but nearly a l l with the open configuration. A l l or nearly a l l of the discharge current (200 mA) i s co l l e c t e d at t h i s p o s i t i o n for the remainder of the magnetrons. 6. R e s u l t s f o r the Small Probes. The r e s u l t s for the smaller probes are shown i n Figs. 5.7 to 5.10. F i g . 5.7 displays the ground currents for the f i v e magnetron configurations. In F i g . 5.8 the ion currents are displayed along the axis for each of the magnetron assemblies. For Figs. 5.9 and 5.10 the r a d i a l dependence of the ion currents are shown for magnetrons 1 and 5 respectively. Experimental Results / 58 o 5 • 4 Position (cm) Figure 5 .8 Ion currents to the small central detector along the axis for magnetrons 1 to 5. Experimental Results / 59 0.00 2 3 Position (cm) o 4 cm • 7 cm A 10 cm Figure 5.9 Radial dependence of the ion currents for magnetron 1. The data i s shown for an a x i a l distance of 4, 7, and 10 cm. Figure 5.10 Radial Dependence of the ion currents for magnetron 5 at a x i a l distances of 4, 7, and 10 cm. Experimental Results / 60 7. Discussion Figs. 5.1 through 5.6 demonstrate how the magnetic configuration a f f e c t s the magnet f i e l d . The differences i n the f i e l d between the magnetrons can be observed primarily i n two regions. The f i r s t region i s the magnetic tunnel which consists of those f i e l d l i n e s going between the annular magnet and the centre magnet. The second region i s a magnetic funnel caused by the c o n s t r i c t i o n of f i e l d l i n e s about the magnetron axis. The magnetic tunnel sustains the primary discharge. As the magnetron becomes increasingly unbalanced from 1 and 4, both the r e l a t i v e amount of f i e l d l i n e s i n the tunnel and the extent of the tunnel from the target surface decreases. When the inner radius of the annular magnet i s increased from magnetron 4 to 5, the o v e r a l l flux i s increased and a greater number of f i e l d l i n e s i n the tunnel i s observed. However the tunnel extends approximately the same distance from the target surface as before. The tunnel acts to contain the discharge i n the region where the f i e l d l i n e s are p a r a l l e l to the target surface through magnetic b o t t l e e f f e c t s . The confinement i s due to the f i e l d strength being greater along the tunnel f i e l d l i n e s i n the proximity of the annular magnet and magnetron centre than i n the region where the f i e l d l i n e s are p a r a l l e l to the target. A set of investigations were done to examine the Experimental Results / 61 effectiveness of the tunnel for both primary trapping and confining the electrons i n the region where the f i e l d l i n e s are p a r a l l e l to the target surface. The f i r s t i n v e s t i g a t i o n looked at the r a t i o of the distance to which an emitted electron would i n i t i a l l y t r a v e l from the target between the d i f f e r e n t magnetron configurations, i n order to give an in d i c a t i o n of the r e l a t i v e effectiveness of the primary traps. I f we assume that the electron t r a v e l s beyond the p o s i t i v e space charge i n front of the target and that the absolute value of the plasma potential i s much less than that of the discharge voltage then, from equation 2.12, the r a t i o of the distances for two magnetrons i s approximately (5.1) The discharge voltage increases by approximately 25 % between magnetrons 1 and 4 while the f i e l d strength i n the v i c i n i t y of the target drops by about h a l f . Therefore, from magnetrons 1 to 4, we would expect the emitted electrons to extend further from the target. For magnetron 5 we would expect the electrons to be contained l a r g e l y i n the same region as magnetron 1. The next investigation examined the r e l a t i v e effectiveness of the tunnel for each magnetron to confine electrons to the region where the f i e l d l i n e s were p a r a l l e l to the target surface. I t was d i f f i c u l t to compare the magnetic mirror i n the v i c i n i t y of the magnetron centre, as, though the f i e l d i n Experimental Results / 62 the region of the discharge was decreasing for the more unbalanced cases, so was the f i e l d at the magnetron centre. For magnetron 5, however, the f i e l d i n the discharge region i s approximately that of magnetron 1 while from Table 5.1 the f i e l d i s les s at the magnetron centre. Therefore the tunnel of magnetron 5 i s less e f f e c t i v e i n containing electrons from the centre region. For magnetrons 1 to 4, the f i e l d strength i n the v i c i n i t y of the annular magnet i s approximately constant. As the f i e l d strength i n the region of the discharge decreases for the more unbalanced cases, the annular magnetic mirror i s less e f f e c t i v e for the more balanced magnetrons for these configurations. For magnetron 5, the f i e l d i n the main confinement region i s approximately the same as that for magnetron 1, however the f i e l d i n the annular region i s much greater. Therefore, the annular magnetic mirror of magnetron 1 i s les s e f f e c t i v e than that for magnetron 5. With the open ground s h i e l d configuration, the discharge electrons can escape both from the regions i n the v i c i n i t y of the annular magnet and from the magnetron centre. Therefore for the more unbalanced magnetrons the electrons have a greater p r o b a b i l i t y of escape along the magnetron axis. The magnetic funnel becomes more constricted about the axis as the magnetron becomes more unbalanced. This, combined with the greater p r o b a b i l i t y of the discharge electrons escaping along the magneton axis, should r e s u l t i n the Experimental Results / 63 discharge current becoming more focused under these conditions. From Table 5.2 i t can be seen the large detector c o l l e c t s a l l or p r a c t i c a l l y a l l of the discharge current at 6 cm along the magnetron axis. In Fig 5.7 i t can be seen that at t h i s point the discharge current i s becoming more focused about the axis as predicted. Table 5.2 shows that the o v e r a l l ion current to the large detector increases as the magnetron becomes more unbalanced. From Eq. 2.2, i t i s shown that the ion currents increase i f either the density of the plasma increases or the electron temperature increases. From Eq. 2.3, and assuming the plasma poten t i a l i s constant, only magnetron 5 shows a s i g n i f i c a n t increase i n electron temperature. I t i s possible that the electron temperature i s r i s i n g s l i g h t l y from magnetrons 1 to 4 as the plasma pote n t i a l i s not known and Eq. 2.4 i s only an approximation. However, because of the square root dependence of the ion current on electron temperature, t h i s i s not the dominant factor i n any of the cases. Therefore i t would appear that the increase i n ion current i s l a r g e l y due to an increase i n the plasma density i n the region of the probe. A possible explanation for the small increases i n ion current for magnetrons 1 to 4 i s that the primary discharge i s contained further from the target i n the more unbalanced cases. Though the probe i s at the same p o s i t i o n on the magnetron axis, i t i s closer to the actual discharge and the o v e r a l l plasma density i s greater. Experimental Results / 64 An increased plasma density i n the v i c i n i t y of the probe could also be due to ion generation i n the region away from the magnetic tunnel. One possible mechanism for t h i s increase i s a more energetic electron flux d i f f u s i n g from the tunnel. The higher energy electrons have a greater p r o b a b i l i t y of inducing i o n i z a t i o n on t h e i r path to the detector. From Table 5.2 i t can be seen that from magnetrons 1 to 4 the discharge voltage i s increasing, i n d i c a t i n g that the primary trapping i s less e f f i c i e n t . The f l o a t i n g potential for magnetrons 1 to 4 at 6 cm does not, however, indicate a much more energetic electron flux i n t h i s region. For magnetron 5, the f l o a t i n g p o t e n t i a l indicates a more energetic electron flux i n the v i c i n i t y of the detector, however t h i s cannot be explained through less e f f i c i e n t trapping as the tunnel i s more e f f i c i e n t than that for magnetron 4. An increase plasma density i n the v i c i n i t y of the probe could also be due to ion generation i n the region away from the magnetic tunnel. Window et a l . 5 have suggested that a secondary discharge could be contained by a magnetic b o t t l e i n the funnel region. The magnetic f i e l d measurements performed i n t h i s thesis are consistent with a magnetic b o t t l e i n t h i s region. F i g . 5.11 shows the f i e l d strength along one of the magnetic f i e l d l i n e s i n the funnel region near the axis. There i s a l o c a l minimum, B2 and a l o c a l maximum, B3. The magnitude of B2 decreases for f i e l d l i n e s c loser to the axis. The s p a t i a l extent of the secondary magnetic b o t t l e was Experimental Results / 65 300 (j-> to ZD X Ld Cd I— CO o o t i < 200 100 was that B2 was at least 2 0% less than B3 DISTANCE ALONG FIELD LINE ( c m ) Figure 5.11 Magnetic f i e l d strength along a secondary magnetic b o t t l e f i e l d l i n e . investigated for configurations 1, 4, and 5. The f i e l d strength along the f i e l d l i n e s was determined through the use of a computer program written for t h i s task. The c r i t e r i o n for a f i e l d l i n e to be considered part of the secondary b o t t l e The r e s u l t s showed that as the magnetron became more unbalanced, the secondary magnetic b o t t l e moved closer to the target. The a x i a l p o s i t i o n of B2 for each magnetron was, to within a millimetre, the same as where the f i e l d went to zero i n F i g 5.6. The absolute value of was, to within a few Gauss, the same as the l o c a l maximum i n Fig 5.6. The shape of the secondary b o t t l e was roughly a cone with the wider end cl o s e s t to the target. An i n t e r e s t i n g r e s u l t was that the s p a t i a l dimensions Experimental Results / 66 of the secondary magnetic b o t t l e was p r a c t i c a l l y the same i n each case. For magnetron 1 the dimensions were a large and small radius of approximately 8 and 6 mm respectively and for magnetron 5 approximately 7 and 5 mm. The length of the secondary b o t t l e along the magnetron axis was approximately 1.2 cm. Unlike the tunnel, the secondary magnetic b o t t l e must be supplied with energetic electrons from another source; i . e . , from the primary discharge. The magnitude of plasma generation i n the funnel region i s dependant, therefore, on the a b i l i t y of the f i e l d to channel electrons from the tunnel to the bo t t l e , the a b i l i t y of the b o t t l e to trap electrons above the i o n i z a t i o n threshold of the sputtering gas, and the a v a i l a b i l i t y of electrons with the required energy. Typical electron energies for the discharge from e f f i c i e n t magnetrons such as 1 are i n the range from 2 to 5 eV 1 9. I t should be expected that as the magnetron becomes less e f f i c i e n t the average electron temperature increases. Therefore there are electrons with energies s u f f i c i e n t for i o n i z a t i o n for a l l of the magnetons used i n t h i s section. The s p a t i a l extent of the magnetic mirror i s approximately the same for a l l the magnetrons used i n t h i s section. The funnelling of the discharge then favours larger ion generation i n the secondary magnetic bo t t l e as the magnetron becomes more unbalanced. As mentioned previously, the maximum energy that a trapped Experimental Results / 67 electron can possess i n the secondary magnetic b o t t l e i s an important factor i n ion generation. Eq. 2.6 shows that the o r b i t a l radius i s dependant on both the electron energy and the magnitude of the magnetic f i e l d . The o r b i t a l radius should not be larger than the dimensions of the secondary magnetic b o t t l e . To determine approximately what the maximum electron energy that could be trapped the c r i t e r i o n was that the electron o r b i t a l radius could not be larger than the minimum radius of the secondary b o t t l e . Using the value of B3 for the magnetic f i e l d value, the maximum energies that could be trapped for magnetrons 1, 2, 3, 4, and 5 were found to be 7, 11, 17, 25, and 175 eV respectively. Notwithstanding the approximation used, i t i s evident there i s a large increase i n trapping e f f i c i e n c y i n going from magnetron 4 to 5. This increased trapping i s expected to r e s u l t i n an increased ion density. Three other factors lend support to an e f f e c t i v e i o n i z i n g mechanism away from the primary discharge f o r magnetron 5. The f i r s t i s the greater electron temperature indicated by the lower probe s e l f bias i n Table 5.2. This indicates the presence of a plasma discharge closer to the probe. Secondly, magnetron 5 has a lower discharge voltage than magnetron 1 despite no i n d i c a t i o n of having a more e f f e c t i v e primary tunnel. This can be explained by a proportion of the ions generated i n the secondary magnetic b o t t l e d r i f t i n g back Experimental Results / 68 towards the target and helping to support the discharge. The t h i r d factor i s the decrease i n ion current f o r magnetron 5 only i n F i g . 5.8. The a x i a l p o s i t i o n of B3 i s at approximately 2.5 cm. This indicates that the probe i s i n t e r f e r i n g with the secondary discharge. C. CENTRE PIECE GEOMETRY AND TARGET THICKNESS EFFECTS 1. Introduction From the discussions i n part B, i t became apparent that the ion current that bombards the substrate was l a r g e l y dependant on the development of the secondary magnetic b o t t l e away from the target. Attempts were made to change the f i e l d configuration through a l t e r i n g the geometry of the centre pole piece. Another mechanism used to a l t e r the magnetic f i e l d configuration was to change the thickness of the target. This would a l t e r the magnetic tunnel but not the funnel. The r e s u l t s for two target thicknesses and two magnetron configurations are shown i n t h i s section. The two target thicknesses were 3 and 8 mm. The f i r s t magnetron used was magnetron 5 of the l a s t section. The second, magnetron 6, had the same annular magnet and the base of the centre piece was the same as that for magnetron 5, however, at 0.32 cm along the axis, magnetron 6's centre piece tapers inward u n t i l i t comes to a point. The distance of the point from the base i s 0.64 cm. Experimental Results / 69 Figure 5.12 Magnetron 6 -2400 -120 10 15 20 25 30 35 40 45 50 55 60 65 Position (cm) Figure 5.13 Magnetic f i e l d strength along the axis for magnetrons 5 and 6. Experimental Results / 70 Table 5.3 R e s u l t s f o r the l a r g e probe f o r magnetic c o n f i g u r a t i o n s 5 and 6 w i t h a 3 and 8 mm t a r g e t . Probe Probe Probe Magnet. Ax i a l Disch. s e l f - ground - 100 V Config. Distance Voltage bias current current (cm) (V) (V) (mA) (mA) 5 6 -307 -29.8 200 -39.1 3 mm 10 -310 -17.7 190 -18.4 6 6 -320 -26.7 198 -34.3 3 mm 10 -320 -12.3 185 -12.4 5 6 -501 -23.6 198 -28.8 8 mm 10 -506 -11.7 187 -14.1 6 6 -635 -25.0 200 -48.7 8 mm 10 -629 -11.1 189 -24.6 J I I I I I I I I i 1 I L 3 4 5 6 7 8 9 10 11 12 Position (cm) F i g u r e 5.14 Ground currents to the central detector along the magnetron axis for magnetrons 5, 6 for target thicknesses of 3 and 8 mm. Experimental Results / 71 7 6, 8 m m 5, 8 m m 6, 3 m m 5, 3 m m 0 3 4 5 6 7 8 Position (cm) 9 10 11 12 F i g u r e 5.15 Ion currents along the axis for magnetrons 5, 6 for target thicknesses of 3 and 8 mm. 2. Experimental R e s u l t s The f i e l d r e s u l t i n g from magnetron 6 i s displayed i n F i g . 5.12. Both target thickness are displayed i n the same manner as that for magnetron 5. The magnetic f i e l d along the axis i s shown i n Fig . 5.13 for magnetrons 5 and 6. The r e s u l t s for the large probe are presented i n Table 5.3 and for the small central probe i n Figs. 5.14 and 5.15 3. D i s c u s s i o n There i s l i t t l e difference i n the magnetic funnel between 5 and 6. The main difference appears that for magnetron 6 the minimum of the funnel i s two millimetres c l o s e r to the Experimental Results / 72 magnetron. The tunnel of magnetron 6 contains fewer f i e l d l i n e s than that of magnetron 5 and extends out from the face by a less distance. The f i e l d strength along the l a s t f i e l d l i n e of the tunnel i s the same for both magnetrons to within a few percent. The secondary magnetic b o t t l e has the same s p a t i a l dimensions for both magnetrons with that of magnetron 6 being 2 mm closer to the magnetron. The maximum energy of the electrons that can be confined i n the b o t t l e i s s l i g h t l y higher than that for magnetron 5 as indicated i n F i g . 5.13. The most int e r e s t i n g r e s u l t i s that for a 3 mm thick target, magnetron 5 i s the most e f f e c t i v e i n producing ions while f o r a 8 mm thick target i t i s magnetron 6. In each case the larger ion generation i s from the magnetron with the greater discharge current along the axis (Fig. 5.14). For the 3 mm target, the discharge voltages i n both cases are low, in d i c a t i n g e f f e c t i v e primary trapping. Magnetron 5 generates more ion current than that of magnetron 6. As mentioned previously t h i s appears to be due to a greater discharge current along the magnetron axis. As the funnel appears to be of the same width for both magnetrons, the tunnel i s most l i k e l y responsible for the difference i n discharge currents long the axis. From F i g . 5.13 i t can be seen that the strength of the magnetic f i e l d i n the v i c i n i t y of the centre i s approximately the same for both magnetrons. However, due to the reduction i n high permeable Experimental Results / 73 material i n the centre piece of magnetron 6 along the edges, the f i e l d strength i s less i n the region between the centre and annular magnet close to the target surface. Therefore, the discharge for magnetron 6 w i l l extend further from the target. This w i l l be a region of a weaker magnetic f i e l d than that f o r magnetron 5 and the magnetic tunnel w i l l be more e f f e c t i v e i n keeping the discharge from the centre. The magnetic tunnel becomes much les s e f f i c i e n t f or magnetron 5 when the 8 mm target i s employed and even more so for magnetron 6. These factors are indicated by the discharge voltages shown i n Table 5.3. Also, for both magnetrons, the discharge current along the axis i s much larger than i n the case with the 3 mm target. The i n e f f i c i e n t trapping and the large discharge currents along the axis indicate that the discharge i s being supported more by the secondary discharge than when the 3 mm target i s employed. This could explain why the ion current i s greater for magnetron 6. As magnetron 6 i s much less e f f i c i e n t , i t r e l i e s more on the secondary discharge than 5. This i s evident from the greater a x i a l discharge current. Another factor i s that with the larger discharge voltage and less e f f i c i e n t trapping, the average energy of the electrons which escape the primary trap i s greater. The t o t a l ion current i s less for magnetron 5 than when a 3 mm target i s used. A possible explanation i s that the ov e r a l l plasma density i s l e s s . With the higher discharge Experimental Results / 74 voltage, the ions impacting upon the target w i l l produce more secondary electrons. Therefore the o v e r a l l plasma density required to sustain the discharge i s l e s s . D. PLASMA CHARACTERISTICS OF UNBALANCED MAGNETRONS 1 . I n t r o d u c t i o n The purpose of t h i s section was to examine the discharge c h a r a c t e r i s t i c s for magnetron 5 under a number of sputtering conditions. The p a r t i c l e fluxes were measured i n the same method as that of section B and C while the target material, sputtering gas pressure and discharge current were varied. 2. T a r g e t M a t e r i a l Two target compositions were examined, copper and g r a p h i t i c carbon. The pressure was held constant at 1 Pa. and the discharge current was 100 mA. The discharge voltage increased from 500 v o l t s for the copper target to 650 v o l t s for the carbon. This i s due to the decreased value of the secondary electron emission c o e f f i c i e n t for carbon. The difference between the remainder of the measurements were within experimental error (approximately 5%). Therefore, to within these bounds, the plasma c h a r a c t e r i s t i c s appear independent for these two target materials. Similar r e s u l t s have also been noted between copper and s i l i c o n 1 4 . Experimental Results / 75 0 5 L i i i i i i i " 0.5 1.0 1.5 2.0 2.5 Log of D i s c h a r g e C u r r e n t Figure 5.16 Logarithmic p l o t of ion currents to the central probe versus discharge current. Magnetron 5 i s used with an 8 mm carbon target. The probe i s positioned at a distance of 6 cm along the magnetron axis. 3. Discharge Current The discharge current was varied over a range of 7 to 150 mA. The pressure was held constant at 1 Pa and the probe was positioned 6 cm along the magnetron axis. A logarithmic p l o t of the ion current from the small central detector biased at -100 V versus discharge current i s shown i n F i g . 5.16. The r e s u l t s were plotted l o g a r i t h i m i c a l l y to present the data at lower discharge currents more c l e a r l y . The r e s u l t s were also normalized which, on a logarithmic plo t , has the e f f e c t of s h i f t i n g the y-intercept only. A l i n e a r f i t was included with the data. Experimental Results / 76 The f i t appears to be reasonable and has a slope of approximately unity. This indicates that the ion currents along the magnetron axis are d i r e c t l y proportional to the discharge current over t h i s range. These r e s u l t s compare well with that of Ref 14. 4. Pressure The pressure was varied over the range of .25 to 2 Pa. The discharge current was 50 mA. and the target was carbon. The probe was positioned 6 cm along the magnetron axis. The power requirements of the system went from 27 to 30 watts, as the pressure increased. This i s due to the decreased i o n i z a t i o n cross-section of the plasma and i s also observed when conventional magnetrons are employed. The f l o a t i n g p o t e n t i a l stayed constant at about - 26 v o l t s i n the range from 1 to 2 Pa, but increased sharply as the pressure was lowered to .5 and .25 Pa to - 32 and - 42 v o l t s respectively. However, both the ground and ion currents stayed constant to within error. This indicates that for carbon, the ion flux i s independent of pressure i n t h i s range. Similar r e s u l t s have also been noted for copper and s i l i c o n 1 4 . E. ION/DEPOSITION FLUX RATIO 1 . Introduction The purpose of the next set of tests was to examine the Experimental Results / 77 ion/deposition flux r a t i o for magnetron 5. The goal was to determine how to both maximize the ion/deposition flux r a t i o and the e f f e c t i v e deposition area. 2 . Deposition Flux Measurements The deposition flux during a sputtering run was measured. Copper was chosen as a target material as i t had been used i n the ion flux measurements and had a high sputtering rate. The deposition rates were determined by weighing glass substrates before and a f t e r deposition. Special substrate holders were designed for t h i s experiment, with two purposes i n mind. The f i r s t was to mechanically clamp the substrates i n place, avoiding such methods as taping which would leave a residue. Secondly, they would expose the same area, eliminating the error due to d i f f e r e n t substrate si z e s . The holder was 1/8 " thick and had a 1 " hole milled i n deep enough to accommodate the substrate. A l i p covered the top of the substrate so that an area of 7/8 " diameter was exposed. The l i p was cut back at an angle of 45 degrees to prevent any shadowing e f f e c t s . The deposition t e s t went as follows. The glass substrates were weighed immediately a f t e r cleaning on a scale accurate to 10"5 grams. The contact areas of the substrate table and holder were cleaned with isopropyl alcohol. The substrate was then fastened to the substrate table and any dust on the substrate blown o f f . The deposition time was chosen so that Experimental Results / 78 2.5E1 6 r (N E y 2.0E16 -o CO \ | 1 .5E1 6 -o 2 1.0E1 6 -o _^ o % 5.0E1 5 -o CL Q> Q O.OEO • 1 ' ' ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 5 6 7 8 9 10 11 12 13 Position (cm) Figure 5.17 Deposition rates along the axis for magnetron 5 for a discharge current of 200 mA and a copper target. the films would weigh at le a s t 10"3 grams. This was t y p i c a l l y about ten minutes at a discharge current of 200 mA. After deposition, the f i l m and substrate were blown free of dust and weighed immediately. The deposition flux was examined as a function of a x i a l distance. The error of measurement was estimated to be +/- 1.5 x 10"4 grams. The deposition rates were measured along the magnetron axis between 5 and 12 cm i n increments of 1 cm. The pressure was held constant at 1 Pa and the discharge current at 200 mA. At t h i s pressure there i s not a s i g n i f i c a n t scattering factor. Therefore we would expect the deposition rate along the axis to go approximately as 1/z2 where z i s the distance along the axis. To t e s t the above model a l e a s t squares f i t Experimental Results / 79 was performed on the logarithms of the a x i a l p o s i t i o n and deposition rate. The r e s u l t s were Value of n i n l / z n i s 2.0 +/- »05 F i g 5.17 shows the r e s u l t s along with the best f i t l i n e . The data has been presented i n units of atoms/s-cm2. This was determined by d i v i d i n g the mass of the f i l m by the substrate area, deposition time and the atomic mass of copper. 3. Ion/Deposition flux maximization The r e s u l t s from the small r a d i a l probes, F i g . 5.10, showed that the ion current i s l a r g e l y focused about the magnetron axis. The ion/deposition flux r a t i o was determined along the magnetron axis from 5 to 11 cm. The ion flux was determined by d i v i d i n g the currents to the small central probe by the area of the probe and the charge on the ions. The determination of the deposition flux was discussed above. The r a t i o along the magnetron axis i s shown i n F i g . 5.18. The r e s u l t s show that the ion/deposition r a t i o increases for substrate positions closer to the target. Also of importance i s the useful deposition area for high ion/deposition flux r a t i o . F i g . 5.19 shows the normalized r a d i a l d i s t r i b u t i o n of the ion currents f o r magnetron 5 at various positions along the magnetron axis. The r e s u l t s show that the d i s t r i b u t i o n becomes more focused at positions closer to the target. The reduction i n the ion/deposition r a t i o away from the axis i s p a r t i a l l y compensated by the reduced E x p e r i m e n t a l R e s u l t s / 80 < cc X O Q. Ld Q 3.0 I 1 1 1 1 1 1 1 1 1 1 1 r O 2.5 2.0 O 1.5 1 .0 ^ 0.5 0.0 _i i i i i_ _i i i_ 6 7 8 9 AXIAL POSITION (cm) 10 11 12 Figure 5.18 I o n / D e p o s i t i o n f l u x r a t i o a l o n g t h e a x i s f o r m a g n e t r o n 5 . 1.2 UJ 1 .0 0.0 O 4 cm • 7 cm A 1 0 cm 0 1 2 3 4 5 RADIAL POSITION (cm) Figure 5.19 I o n c u r r e n t s v e r s u s r a d i a l p o s i t i o n f o r m a g n e t r o n 5 . T h e i o n c u r r e n t s a r e n o r m a l i z e d f o r c o m p a r i s o n p u r p o s e s . Experimental Results / 81 deposition flux. At pressures low enough to prevent scattering, the deposition flux follows approximately a cos 26 r e l a t i o n s h i p 2 7 , where cos8 i s equal to r a t i o of the r a d i a l distance of the point of deposition from the centre of the substrate (which i s on the magnetron axis) and the a x i a l distance of the substrate from the target. Using t h i s approximation, the ion/deposition r a t i o was found to decrease at a r a d i a l distance from the centre of the substrate of 2 cm to 23, 41 and 70% of that at the centre , for a substrate at a distance from the target of 4, 7, and 10 cm, respectively. CHAPTER VI. CONCLUSION The objectives of the thesis have been achieved. The mechanisms involved i n ion production for unbalanced magnetrons have been investigated i n a more quantitative manner than previously studied. This has lead to an understanding of the r o l e of the secondary discharge i n the funnel i n the production of high ion fluxes at the substrate. The investigations have also shown how to maximize the ion/deposition r a t i o through placement along the magnetron axis and the considerations necessary to maintain an e f f e c t i v e area of substrate bombardment. As stated above, the r e s u l t s presented and discussed i n t h i s thesis show that large currents to the substrate are highly dependant on the development of a secondary discharge away from the magnetic tunnel. This discharge i s dependant on the a b i l i t y of the secondary bo t t l e to trap electrons above the i o n i z a t i o n potential of the sputtering gas. When designing unbalanced magnetrons, attention must be given to both the s p a t i a l dimensions of the secondary b o t t l e and the magnitude of the l o c a l f i e l d maximum (B3) i n order to achieve large ion/deposition flux r a t i o s . Care must also be taken with respect to the target thickness as t h i s e f f e c t s the tunnel, which, i n turn, e f f e c t s the secondary discharge. The optimum magnetron chosen i n t h i s thesis was configuration 5 with a 3 mm target. Configuration 82 Conclusion / 83 6 with a 8 mm target produced higher o v e r a l l ion currents at the substrate, but had two disadvantages over the previous design. The f i r s t was the greater power requirements. The second was the factor that small changes i n the tunnel lead to large changes i n ion currents indicated by the differences i n the discharges between magnetrons 5 and 6. This causes problems during the sputtering process as the target thickness decreases over time. A system could be designed to move the magnetics during the deposition process, but at a loss of s i m p l i c i t y . The ion/deposition flux r a t i o was shown to be dependant on the a x i a l p o s i t i o n of the substrate with the r a t i o increasing towards the target. However, at the shorter substrate to target distances the useful deposition area i s less and a tradeoff e x i s t s between the ion/deposition r a t i o and larger deposition areas. Moving the substrate further from the target also has the disadvantage of reducing the deposition rate. A possible simple method that would allow larger deposition areas closer to the target i s to a l t e r the magnetic f i e l d beyond the secondary discharge. I f the f i e l d l i n e s were less focused beyond t h i s point, the ion currents would be more evenly d i s t r i b u t e d r a d i a l l y . The desired a l t e r a t i o n of the f i e l d l i n e s could be achieved through the correct positioning of a disk magnet behind the substrate to reduce the magnitude of B2 and increase the magnitude of Bp. Care would have to be Conclusion / 84 taken both to avoid creating another magnetic b o t t l e , as t h i s would increase the l o c a l i z a t i o n of the ion currents, and to avoid decreasing the trapping e f f i c i e n c y of the secondary magnet b o t t l e to below that which could trap i o n i z i n g electrons. REFERENCES 1. J.A. Thornton and A.S. Penfold, i n "Thin Film Processes" (J.L. Vossen and W. Kern) Chap. 2 , Academic Press, New York, 1978. 2. J.L. Vossen and J . J . Cuomo, i n "Thin Film Processes" (J.L. Vossen and W. Kern) Chap. 1 , Academic Press, New York, 1978. 3. J.G.Linhart," Plasma Physics",North-Holland Publishing Co., New York, 1961. 4. W.Knauer, J . Appl. Phys. 33 , 2093 (1962). 5. B. Window and N. Sawides, J . Vac. S c i . Technol. A4, 196 (1986). 6. J.A.Thornton, Ann. Rev. Mater. S c i . 7, 239 (1977). 7. J.E. Sundgren and H.T.G. Hentzell, J.Vac.Sci.Technol. A4, 307 (1986). 8. N. Sawides, J . Appl. Phys. 5 9 , 4133 (1986). 9. H.R. Kaufman, J . J . Cuomo and J.M.E.Harper, J . Vac. S c i . Technol. 21, 405 (1982). 10. S. S c h i l l e r , V. Heisig and K. Goedicke, Thin S o l i d Films 40, 327 (1977). 11. T. Minami, H. Nanto, and S.Takata, Appl. Phys. Lett. 41, 985 (1982). 12. S. Onishi, M. Eschwei, and W.C.Wang, Appl. Phys. Lett. 38, 419 (1981). 13. D.B. Fraser and H.D. Cook, J . Vac. S c i . Technol. 14, 147 (1977). 14. N. Sawides and B. Window, J. Vac. S c i . Technol. A4, 504 (1986). 15. B. Window and K.H. Muller, Thin S o l i d Films 171, 183 (1989). 16. A.G. Spencer, K. Oka, R.P. Howsen and R.W.Lewin, Vacuum 38, 857 (1988). 17. R.P. Howson, A.G. Spencer, K. Oka, and R.W. Lewin, J . Vac. S c i . Technol. A7, 1230 (1989). 85 References / 86 18. B.Window and G.L. Harding, J . Vac. S c i . Technol. A 8 , 1277 (1990). 19. B. Chapman, " Glow Discharge Processes", John Wiley and Sons, New York (1980). 20. D. Bohm and E.P. Gross, Phys. Rev. 75, 1851 (1949). 21. E.S. McDaniel," C o l l i s i o n Phenomina i n Ionized Gases", Ch. 13. Wiley, New York, 1964. 22. W.Knauer, J . Appl. Phys. 3 3 , 2093 (1962). 23. D. Bohm, E.H.S. Burhop, and H.S.W. Massey, i n "The Charact e r i s t i c s of E l e c t r i c Discharges i n Magnetic F i e l d s " (A. Guthrie and R.K. Wakerling, eds.), pg. 13. McGraw H i l l , New York 1949. 24. S.M. Rossenagel and H.R. Kaufman, J . Vac. S c i . Technol. A 5, 88 (1987). 25. P. Horowitz, W. H i l l , " The Art of E l e c t r o n i c s " , pg. 1007. Cambridge University Press, New York 1989. 26. L. Holland and G. Samuel, Vacuum, 3 0 , 267 (1980). 27. R. Glang, i n " Handbook of Thin Film Technology" (L.I. Maissel and R. Glang), McGraw-Hill, New York 1970. 28. J.D. Jackson, " C l a s s i c a l Electrodynamics" Chap. 5, John Wiley and Sons , New York 1975. APPENDIX Flux from a magnet face Though the calculations for a magnetic f i e l d at a given point i n space i s quite d i f f i c u l t , a much simpler task i s to determine the net flux leaving the face of a magnet. This can be shown by considering a c y l i n d r i c a l magnet with the same coordinate system as Fi g . 3.1 with a radius, R, and a depth, b. For a magnet with a magnetization M(x') the solution to the vector potential i s given by 2 8 assuming that M i s constant throughout the volume then As the magnet i s magnetized along i t s z axis we have that VxM(x') = 0 therefore M z x z = 0 and M zxr~M 0 87 /88 Therefore AU)-rrr^-nrdz'dQB Jo J - i l x - x 1 and i t can be seen by symmetry that Ae i s a function of r and z only. Now consider the flux through one of the faces of the magnet <j) - j B-nds - 271 j RBZ (r 7, 0) r'di' Now, i n c y l i n d r i c a l coordinates we have that z i dr 58 And as Af = 0 we have (t>=27r./ —, ^ , r'dr' ° 27rA> (a, 0) •'o r ' 6\r' The value of Ae i s a n a l y t i c i n z and can be solved numerically for 8 . The flux from the surface of an annular magnet i s also quite simple to solve for. From above we can see that ^=2%AQ(R, 0) -2UAQ(ROI 0) 

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.831.1-0074522/manifest

Comment

Related Items