Untitled

Download as pdf or txt
Download as pdf or txt
You are on page 1of 491

This item was submitted to Loughborough University as a PhD thesis by the

author and is made available in the Institutional Repository


(https://dspace.lboro.ac.uk/) under the following Creative Commons Licence
conditions.

For the full text of this licence, please go to:


http://creativecommons.org/licenses/by-nc-nd/2.5/
/3 LL :z:. b /'YD,' _ 2::,£';/166//8-2>
,

LOUGHBOROUGH
UNIVERSITY OF TECHNOLOGY
LIBRARY
AUTHOR/FILING TITLE

~ _________ ____ §_'-~IJ._Q_T_/_-- JLA ___________ _

ACCESSION/COPY NO,

--\70L~NO~----~-- -c-!lsI-!-~~I-~-~------- --------


THE EFFECT OF PARTICLE SIZE ON

ELECTROSTATIC POWDER COATINGS

by

RICHARD A. SIBBETT

A Doctoral Thesis

Submitted Ln partial fulfilment of the requirements

for the award of

Doctor of Philosophy

of the Loughborough University of Technology

September 1982

Supervisor ·P.J. Lloyd


Chemical Engineering Department

© by Richard A. Sibbett, 1982.

" ..
<

LeUgl1bOrough Univeroltr r
-.
.f T e<:MGloIY U~r ... y 4

Clan
~ ~'l-

Ace.
Ale. 1\ ~Sl ~ 19-z..
I

•• '0:"
To MUM and DAD
---11--

for their unceasing support

. _. . -.- ". '.~ \-- .......-


,,,..-,"':

"',

'.
ACKNOWLEDGEMENTS

The author wishes to express grateful thanks for the assistance,

encouragement and guidance from his supervisor, John Lloyd. -

Thanks also to ,the Sponsors of the Coordinated Powder Research

Programme (and Chairman Mr. Peter Thompson, NRDC) for their financial

support and also to all the individual representatives of the

sponsoring companies who gave helpful suggestions, including:

Ault and.Wiborg Paints Ltd

Berger Paints Ltd

Bink Bullows Ltd

Carrs Paints Ltd

Blundell Permaglaze Ltd

Devilbiss Company Ltd

Donald Macpherson Powder Ltd

Drynamels Ltd

GEMA, SA

ICI plc, Paints Division

International Paint Company Ltd

The Nordson Corporation

Volstatic Ltd

and special thanks to Aerostyle Ltd and TAM-SAMES et Cie for the loan

of spray equipment.

Sincere gratitude is also acknowledged to the following:

Mr Neale for the help with development of the goniophotometer and

spray rigs.

Mr Leo Moore and his team for help in the construction of rigs.

Mr Geoff Boyden for his advice and help with cine photography and

photographing of experimental equipment .

. Mrs Jackie Harrison for the reproduction of C1ne films.


Mr Graham Gerrard, Brian Negus (Nag routine) and the staff of the

Loughborough University Computer Centre.

The Department of Materials Engineering, Loughborough University

for the loan and use of the Talysurf 10 and Salt Spray booth.

Dr Peter Cardew (ICI), Ur Ian Parker (IeI), and Ur Roward English


(L.U.T.) for helpful suggestions and discussions.

Mr Bill Mackrodt and his team for use of the VAX computer at

ICI plc, Runcorn.

Mrs Margaret Redican for her patient help with the plotting of

computer results.

ICI plc for the use of computer facilities and other equipment.

Late Professor A.W. Bright, Dr. S. Singh and Dr. J.F. Rughes

(Elec. Eng. Dept., University of Southampton) for their useful

discussions.

Professor D.C. Freshwater for the use of the Chemical Engineering

research facilities and various members of his department for their

helpful discussions.

Mrs Chris Steven for the excellent typing of the thesis.

Finally, thanks to my wife, Ann, for not only her help with

the mounting of all the numerous photographs, computer plots, and

preparation of printing boards, but her patience, support and

encouragement throughout the period of writing up.


Synopsis

In the electrostatic powder coating industry an increase in

deposition efficiency and a controlled decrease in film thickness

are constantly strived for. If adjustments to the size distribution

of powders are made in order to improve the process in any way it is

important that the excellent physical properties of the coating are

not sacrificed.

The aim of this study is to gain an understanding of the effects

of particle size on the packing and final film properties of electro-

statically sprayed coatings so that-recommendations can be made to

increase the effectiveness of the process.

Experimental studies involved the production of stoved sample

coatings of various thicknesses obtained by spraying narrow size

ranges of a commercial powder using a reproducible coating method.

A testing programme was developed to assess the effect of particle

size on the physical properties of the films produced. A three way

comparison of physical property, size of sprayed powder and coating

thickness has been compiled and results are discussed. It was

observed that particle size had no significant physical effect on film

properties.

A method for determining a spreading factor, representative of

the flow characteristics of a given powder, was developed. Observations

from these stoving experiments suggested that the excellent flow

properties of the powder used in earlier experiments accounted for

the small variations in those results. However'" photomicrographs of

typical coatings showed that large particles were dominant in the


upper layers of packings. A loss of fines was identified from size

analysis of various powders on coated plates compared to the original

feed materials.

Computer plots simulating the packing of particles on a

substrate showed similar trends. Integration of the trajectory

equation for individual par.ticles, represented by means of a force

balance, was carried out with consideration of all interparticulate

electrostatic forces of the packed layer together with field and

aerodynamic forces. Simulations of packings of monosized and size

distributed powders sprayed at various transport air velocities show

that, as a packing increases in thickness, various sizes behave

differently. The results from these theoretical experiments give

indications for the explanation of previous results.


CONTENTS

INTRODUCTION

CHAPTER 1 - ELECTROSTATIC POWDER COATING - THE BASICS

1.1 Introduction 5
1.2 Electrostatic Powder Coating Process 8
1. 2.1 Powder Supply 8
1. 2. 2 Spray Gun 9
1. 2. 3 Spray Booth 10
1. 2. 4 Powder Recovery System 11
1. 2. 5 Stoving Ovens 12
1. 2. 6 Powder 13
1. 2. 7 Colour Change 17
1. 2. 8 Substrate Preparation 18
1.3 Previous Research 19
1.4 Levelling in Powder Coating 24
1.5 Practical Problems 25
1.6 Summary 27

CHAPTER 2 - A REVIEW OF TEST METHODS

2.1 Introduction 28
2.2 Gloss 29
2.3 Film Thickness 32
2.3.1 Destructive Methods 33
2.3.2 Non Destructive Methods 34
2.4 Roughness 35
2.5 Environmental and Accelerated Weathering 37
2.6 Adhesion 38
2.7 Hardness 40
2.8 Abrasion and Flexibility 41
2.9 Chip or Impact Resistance 42
2.10 Porosity 43
2.11 Summary 45

CHAPTER 3 -

3.1 Introduction 47
3.2 Spray Equipment 48
3.2.1 Static Gun 48
3.2.2 Problems Encountered 49
3.2.3 Commercial Spray Equipment 50
3.2.4 Experimental Method 51
3.2.5 Pretreatment of Sample Plates 54
3.3 Testing Methods and Equipment 55
ii

3.4 Thickness Measurement 56


3.4.1 Experimental Method 56
3.4.2 Experimental Results 57
3.5 Gloss Test 57
3.5.1 Specular Gloss 57
3.5.2 Goniophotometric Curve 58
3.5.3 Instrumentation 59
3.5.4 Experimental Method 60
3.5.5 Results and Discussion 62
3.6 Surface Roughness 65
3.6.1 Instrumentation 66
3.6.2 Experimental Method. 66
3.6.3 Results and Discussion 67
3.7 Adhesion 68
3.7.1 Experimental Method 69
3.7.2 Results and Discussion 69
3.8 Corrosion 70
3.8.1 Experimental Method 71
3.8.2 Results and Discussion 71
3.9 Porosity 72
3.9.1 Results and Discussion 72
3.10 Conclusions 73
3 • 11 Summary 75

CHAPTER 4 - PARTICLE STOVING AND PACKING OBSERVATIONS

4.1 Introduction 76
4.2 Stoving Observations 76
4.2.1 Experimental Observations 77
4.2.2 Comparison of Different Types of Powder 79
4.2.3 Determination of Spreading Factor 79
4.2.4 Experimental Results 80
4.2.5 Discussion of Results 81
4.3 Particle Packing Considerations 82
4.3.1 Effect of Formulation on Powder Characteristics 82
4.3.2 Experimental Observations 84
4.3.3 Quantification of Observed Effect 85
4.3.4 Experimental Results 88
4.3.5 Discussion of Results 88
4.4 Summary 89

CHAPTER 5 - TRAJECTORY EQUATIONS

5.1 Introduction 91
5.2 The Force Balance 92
5.3 Charge on a Particle 94
5.4 Electric Field 97
iii

5.5 The Forces 99


5.5.1 Drag Force 99
5.5.2 Flow Model 102
5.5.3 Field Force 105
5.5.4 Interparticulate Forces 106
5.5.5 Image Force 106
5.5.6 Repulsive Forces 107
5.5.7 Attractive Forces 107

5.6 Other Representations of Forces 109


5.7 Comparison with the 3 Dimension Case 113
5.8 Summary 114

CHAPTER 6 - COMPUTER SIMULATION

6.1 Introduction 115


6.2 NAG Routines 116
6.3 Program Description 119
6.3.1 Problems Encountered 120
6.3.2 Main Subroutine 122
6.3.3 Function Subroutine 123
6.4 Testing of Program 125
6.4.1 Drag Coefficient 125
6.4.2 Starting Velocities 126
6.4.3 Error Testing - Setting of Tolerance Bound 128
6.4.4 Particle 'Splaying' 129
6.4.5 Further Effects 129
6.5 Experimental Results and Discussion 130
6.5.1 No Field Force 131
6.5.2 Monosized Powders 132
6.5.3 Size Distributed Powders 139
6.5.4 Theoretical Charged Layer· 141
6.5.5 Effect of Gravity 144
6.6 Conclusions 145
6.7 Summary 147

CHAPTER 7 - FINAL DISCUSSION AND SUMMARY

7.1 Aims of the Investigation 148


7.2 Summary of Results of Investigations 151
7.2.1 Testing of Properties 151
7.2.2 Packing and Stoving 152
7.2.3 Trajectory Equation 152
7.2.4 Computer Simulation 153
7.3 Limitations of this Work 155
7.3.1 Experimental 155
7.3.2 Theoretical 155
7.4 Future Work 156
7.5 Final Concluding Remarks 159
iv

References

Nomenclature

APPENDIX A - GLOSS TEST RESULTS

APPENDIX B - ROUGHNESS TEST RESULTS

APPENDIX C - MICROSCAL SIZE DISTRIBUTION ANALYSIS RESULTS

APPENDIX D - COMPUTER PROGRAM LISTING AND PRINTOUT

APPENDIX E - COMPUTER SIMULATION PLOTS


INTRODUCTION
Consideration of conserving energy, resources, money and

controlling pollution has never been more important (A.D. Moore,

1973). The electrostatic powder coating industry is one which has

developed because of its many advantages 1n saving materials, energy,

being non-polluting and generally producing a very good protective

coating (R. Lever, 1978; J.D. Toff, 1979). In the last 10-15 years

its growth has accounted for approximately 10% of the world protective

finishing market (Anon, 1979).

However, in spite of the long list of advantages to its credit,

as shown in Table 11, electrostatic powder coating has some major

disadvantages which have curtailed its growth in the large commercial

industries (W. Crisp, 1973; E. Miller, 1974; D. Payne, 1973;

K.C. O'Neill, 1977). One of the largest users of protective coatings

is the automobile industry and as yet electrostatic powder coating

has not had a large impact on this market (S.L. Mason, 1974;

E.W. Drew, 1977; M. Cowley, 1974).

The car industry first used a spray gun 1n 1912 for applying

primer coats to car bodies, but although this method is used for

applying both primer and finishing coats the application is in

liquid form. This has many drawbacks.

Thick films have to be built up in layers and the process is

very wasteful. Sludge systems to remove and dispose of the waste

are therefore required. Even after many advances in controlling the

system it is still apt to air and water pollution. The final coat

is susceptible to corrosion. In these respects powder coating is

highly advantageous.
Table Il

Advantages of Electrostatic Powder Spraying Process

System Advantages

No solvents used hence no pollution or fire risks.

Overspray can be reused by incorporating recycling system.

Highly skilled labour is not necessary.

Cleaning of booth is more easily carried out.

Builds up thick films in one coat.

Lower capital outlay and energy costs.

Less preparation required for surface.

Film Quality Advantages

Very good adhesion.

High abrasion, scratch and impact resistance.

Extremely high chemical resistance.

Very good electrical Insulation.

Orange 'peel and wrap around effects.


- 2 -

However, the electrostatic powder coating system has its own

disadvantages. Problems are encountered in the production of thin

films and also when colour changes are required. Unlike liquid paint

systems, powder systems must be totally cleaned out when a colour

change is required, and in practice this can be quite frequent. This

is necessary to avoid any contamination of the next colour by airborne

particles. In the paint industry colour change is effected by a

simple change of paint supply, since oversprayed paint will not become

airbourne again.

In powder coating the colour change is both labour intensive

and time consuming. Although the usage of powder can be up to 99%

efficient by use of a recycling system, the actual deposition

efficiency is only about 40-60%. An increase in this efficiency to

95%, by the manipulation of powder characteristics and operating

conditions would remove the necessity for colour change.

The second problem is the production of thin films. Although

powder coating can very efficiently give final films of 50-150um thick

in one coat, a more economic thickness of 25-50 u m is required. The

associated high quality finishes must also be maintained. This would

lead to a considerable saving on raw material costs.

Investigations into the mechanisms involved in powder coating

show that a spraying system has a self limiting coating thickness

(J.A. Bassett et a1, 1975; J.A. Cross et a1, 1980). This is due to

the repulsion of charged particles away from the charged packed layer

of particles, and due to the onset of back ionisation. However, this

limiting effect only occurs at thicknesses above 100um and after long

coating times.
- 3 -

It is with these problems in mind that the work described here

~s directed. The effects of particle size and size distribution on

the properties of electrostatically sprayed powder coatings have been

investigated. It is important that any change in the feed powder

that is made to give rise to an improved coating process in terms of

economy does not lead to a decrease in the standard of finish. The

effects of particle size on the physical properties of various

thicknesses of commercially applied coatings have been investigated

using a range of test methods. A review of the generally accepted

test procedures for testing organic coatings is given following a

brief outline of the electrostatic powder coating process (Chapters

1 and 2).

Size fractions (lO~m wide) of a commercial powder were sprayed

under similar operating conditions and at various coating thicknesses.

These were then stoved and tested for differences in physical

properties. Results are expressed in terms of sprayed particle size,

film thickness and physical property as a three way comparison. The

powder" used in the experiments was a standard epoxy resin type as

often used in the coating industry and supplied by powder manufacturers.

No significant differences were found in the results for changes in

particle size sprayed or film thickness.

Exploratory experiments were carried out to investigate the

reasons behind these results. Fundamental stoving experiments and

microscopic observations of coatings (unfused) showed interesting

results. Studies of the flow properties of powders suggested that

a powder that has good flow characteristics would not be subject to

effects due to particle size. Packing experiments have shown that


- 4 -

when a powder with a wide Sl.ze distribution (e"g. commercial) is

sprayed a loss of fines takes place and large particles preferentially

deposit on the top of thick coatings.

Theoretical investigations based on these results were directed

towards the -understanding of how powders with different size

distributions pack on a substrate. In particular a comparison of

mono sized and size distributed powders was made. A computer program

has been developed using the force balance given in Chapter 5 and is

described in Chapter 6. The program simulates the landing of particles

on a- substrate taking into account field and aerodynamic forces

acting on them, and also all interparticulate electrostatic forces

due to ':the._ charged powder layer. Results are presented in the form

of computer plots.

It has not been an intention of this study to observe changes

in deposition efficiency with different sized particles that have

been sprayed. This work has studied any effects that particle Sl.ze

has had on the properties of the stoved films and also any changes

in the way in which particles are deposited on a substrate or powder

layers. In doing so a better understanding of the importance of

particle size on the electrostatic powder process has been obtained.


CHAPTER 1

ELECTROSTATIC POWDER COATING - THE BASICS

1.1 Introduction

1.2 Electrostatic Powder Coating Process

1. 2.1 Powder Supply

1. 2. 2 Spray Gun

1. 2. 3 Spray Booth

1. 2.4 Powder Recovery System

1. 2. S Stoving Ovens

1. 2.6 Powder

1. 2. 7 Colour Change

1. 2. 8 Substrate Preparation

1.3 Previous Research

1.4 Levelling in Powder Coating

1.S Practical Problems

1.6 Summary
- 5 -

1.1 Introduction.

It was only in 1953 that the first commercial powder coating

was made by an American factory dipping heated metal parts into·

powder (S.T. Harris, 1976). Since then, and especially in the past

15 years, powder coating, and in particular electrostatic powder

spraying, has grown to take over about 10% of the world market in

the finishing industry (P.G. Delange, 1978).

The reasons for the increase in powder coating· in var~ous

coating industries can easily be seen.. When raw material prices are

escalating it is not surprising that new techniques, such as powder

coating, which improve the overall efficiency are considered as

replacements. The industry has further been enhanced by the need to

reduce air pollution and sustained by the possibility of better or

equal finishes at lower costs. However, the increase in popularity

also raises the interest in more control over finishes, namely

producing 1-2 mruLsmooth coatings.

The obvious advantages of using dry powder coating have already

been shown in Table 11. In the early stages of the industry

manufacturers were conscious of the great deal of work and research

that was necessary in order to break into various markets. Apart from

the obvious disadvantage of colour change, which gives difficulty due

to the easy contamination of materials and equipment, many other

problems had to be dealt with.

Development of recovery facilities, equipment design and powder

specifications was required. Since the beginning of the seventies

the range of powders available to coating companies has become

extensive (J.D. Tierney, 1975). An understanding of formulation

on final film properties has been obtained. Powders now come in


- 6 -

many colours and hence can be· used in a vast amount of applications.

However, the effects of the particle characteristics and

electrical properties (excluding formulation) on the packing and

properties of the dry and fused films is far from clear. It is with

respect to the understanding of the process and the effects due to

particle characteristics that this work is aimed. Firstly it is

necessary to know something about the process itself and the problems

. encountered.

There are two main types of electrostatic powder coating

equipment:-

(1) Electrostatic fluidised bed

High voltage wires are introduced into a fluidised bed to give

electrostatic charge to particles. The article to be coated is

positioned in a high voltage space and is earthed. The article does

not have to be preheated as in non-electrostatic fluidised bed coating

and it is not actually dipped into the bed. The particles are

attracted to the object as it approaches and a uniform layer of

coating is obtained. However, with larger particles this gives

varying coating thicknesses between the top and bottom. Particle

size and voltage have significant effects on the coating deposition

and thickness. Electrostatic effects will be preferential to a certain

size range (between 20 and 80 ~m) and so fresh powder is required

otherwise coating thickness will decrease.

(2) Electrostatic powder spraying.

In electrostatic powder spraying charged powder in an air stream

is directed at the earthed object so as to bring the powder adjacent

to the surface. In order to accomplish a successful coating the

method comprises of a combination of elements.


- 7 -

(i) Delivery of powder

(ii) Charging of powder

(iii) Transfer of powder from charging region to obj ect

(iv) Adhesion of powder to object

(v) Stoving of coating

(vi) Recovery of powder

This is shown diagrammatically in Figure 1.1.

The fluidised bed method is limited by being a fixed position

applicator. Spraying of powder overcomes this. Spraying plants vary

immensely depending on the type of articles that are to be coated

(S. Kut, 197.Ia., 1972, 1974; M. Drury, 1974). They can either be

large scale; fully automated robot systems spraying large articles

which are transported by conveyors, or small scale; articles are

coated by a hand gun and carried to an oven. However the basis of

the equipment in each case is very similar. One such arrangement

is shown in Figure 1.2.

The powder is transported by compressed air from a hopper or

reservoir to a sp·ray gun. The powder then passes one or more high

potential electrodes in the gun whence the particles are charged.

The mechanism of charging particles will be discussed in detail later.

The particles are propelled forwards by means of electrostatic forces

and the transport air towards the object, which is earthed and

enclosed in a booth. The charge on the particle attracts it to the

substrate where it is held by various adhesion forces. The powder

layer is then stoved in an oven to form a continuous coating.

Oversprayed powder is recovered by a recycling system.


Finished Product

H.T. Stoving
Supply Oven

Spray Booth
r --- - -- - - - - - - - - - I- - -- "I
I I
I
I
Gun Target I
I
I

-f--- - - r - - - - - - - - - - - -

Powder Recycling of
Feed Hopper Powder

Air Supply

Figure 1.1. Diagram of Typical EPC Process


Air Fan

Bag Filter I~
t

Cyclone
Recovery

Air
kec~cle

Figure 1.2. Electrostatic Powder Coating Equipment


- 8 -

Each section of the process will now be described.

1.2 Electrostatic Powder Coating Process

1.2.1 Powder Supply

It is very important that the amount of powder supplied

to the powder gun and hence onto the object is carefully controlled.

Starting and stopping of the powder supply system must be possible

without surging or blocking.

The most popular system incorporates a venturi feeder. possible

types are shown in Figure 1.3 (a),(b) and (c). Air is supplied to the

injection part of the venturi and this causes powder to be drawn

from the hopper which is then suspended in the air. To aid powder

being drawn out of the hopper the bed can be agitated by an external

vibrator (1.3(a)). A further modification is to incorporate a motor

driven blade and a screw feeder within the powder (3(b)). The delivery

·of the powder is then more easily controlled.

Another way of aiding the exit of powder is by fluidising the

bed structure, 1.3(c). Air is passed through the powder via a gauze

at the bottom of the bed and the venturi pump is then immersed in

the hopper. Each air supply can be controlled by a solenoid valve

which is connected to the mai.n generating system or to the trigger

of the spray gun.

Where multiple guns are ~n operation a venturi pump valve

combination is used for each gun. This gives interdependence of

each gun. For multiple arrays .of guns a single supply cylinder can

be used.

The output of the feeder system ~s obviously a function of the

number of guns in use and connected to it. In general the output

from a single spray gun is about 2-5 gms/second of powder, the lower

(
_Powder Storage
~-- Hopper
Vibrator

Compressed
Air Powder to Gun

(a)

~ Hopper
Powder

Agitator

Screw
Feed
Airi
Air ---~~=====~ c::.: ;,:... _. ·. :.:;i;·-.,-.---~- Powder

r"==....,-- Drive
Unit

(b)

Powder
Level

Compressed
Air

Fluidised
Air -"----r--

(c)

Figure 1.3. Typical Powder Feed Systems


- 9 -

the output the harder it ~s to obtain a consistent flow.

1.2.2 Spray Gun

The nature of the gun is very important since it is

responsible for charging and spraying the particles.

The shape of the spray pattern is controlled by means of a

deflector (stationary or rotating) at the nozzle of the gun

(Plates 1.1 & 1.2). In addition to the powder supply air, there is

a separate air supply to the gun which controls the pattern of the

spray. This air passes through the gun and is emitted at the nozzle

or just before the deflector. Manipulation of this air supply

enables control of the amount to which the powder is 'splayed out'.

The higher the air input then the higher the turbulence and the larger

the spray pattern.

A pointed electrode is positioned near the nozzle or at the

front of the deflector. A resistor of 10 8 n is incorporated in the

body of the gun, 'or comprises the voltage supply cable to the gun,

to act as a safety device. The corona electrode may consist of

pointed wires, whiskers or edges of a disc. The electrical field

around the sharp point must reach the breakdown strength of the air

to produce a stable corona. In general the potential applied is

40-80 kV.

Resistive guns are those in which resistors have been added

(as opposed to stiff guns) so· that discharges and sparking to grounded

objects is prevented. As the gun is brought near a grounded substrate

the voltage will decrease and current increase so that energy output

is constant. A stiff gun would give a constant voltage but the current

would increase rapidly so as to cause sparking at a close distance.

The normal output of a gun is about 10 Kg/hr of powder, with an exit

air velocity of between 2-7 m/so


Plate 1.1. Spray Gun Nozzle

Plate 1 . 2. Spray Gun in Opera t ion (hand- held)


- 10 -

1.2.3 Spray Booth

The spray booth in powder coating is designed so as to

aid the deposition of powder onto an article and to collect the

overspray for recycling. This is unlike the wet paint system which

traps overspray for disposal. Ideally the air in a spray booth would

be perfectly still so that the electrostatic forces and powder

velocity from the gun would control the coating process. In practice

a stream of air across the workpiece is required for several reasons:-

(i) Keeping powder concentration low in the booth (below

safety explosive limits of 10 gms/ cm 3)

(ii) Minimises effects of extraneous draughts from other

areas of plant

(iii) Prevents backdraught of overspray towards operator

(iv) Acts as first stage of the recovery system

Confines.any possible fire

The booths can be made of either steel or plastic. Steel is

least expensive and provides protection in case of fire but becomes

coated with the charged particles, hence problems are increased

during colour changing. Plastic booths are more expensive.

A spray booth is designed so that there is little tendency for

powder to accumulate. Therefore sloping floors are used to give a

funnel shape to the opening of the air withdrawal. The size of the

booth is determined by the parts to be coated, the speed of the

conveyors (if used) and the number of guns. The air movement in the

booth is dependent on the size of the booth and should be about


\
0.3-0.4 m/so The air flow should be sufficient to stop powder

forming high concentrations but not enough to blow the powder cloud

away from the workpiece.


- 11 -

The inside of the booth should be free from ledges, where

powder can collect, and generally be designed so that it is easy to

clean. Access to the booth should be quick and simple, but it should

be ensured that automated booths cannot be entered whilst in operation.

1.2.4 Powder Recovery System

For both economic and environmental reason~ the over

sprayed powder must be recovered with equipment that will separate

powder from the air exhausted from the booth. By removing overspray

an electrostatic powder system removes complaints of air pollution

and eliminates the need for dispersal equipment. It can also make

the system as a whole up to 99% efficient on powder usage.

Any recovery system consists of several elements that separate

the powder from air, collect it, and return it for use in the system.

In general, two types of equipment are used, namely the cyclone

. separator and bag filters. These are either used separately or in

a combination. Combinations of cyclones and filter cloth systems help

in cases where rapid colour changes are needed.

Bag filters have a very large filtering area for a small

volume and can easily be cleaned by using reverse air flow. The

cyclone is a very simple device in which a spiral air stream is

generated to give the particles a centrifugal force. The large

particles are thrown out to the walls and drop to the bottom of the

cyclone where they are collected. Very small particles escape

collection. If colour changes are required frequently then a cyclone

followed by a bag filter can be used. The powder recovered from the

cyclone (about 85%) is recycled and the powder from the bag filter is

not reused. Hence colour changeover is effected by cleaning only

the spray booth and cyclone. However, the loading on the bag filter

must be carefully monitored as high loading will cause a· drop in air


- 12 -

velocity through the system, which in turn will lead to a loss 1n

efficiency of the cyclone separator.

Other systems include a counter current cyclone and a powder

centrifuge. The recycled powder must be sifted at about 120 um to

remove any agglomp.rates, fibrous contaminants and dirt etc. The

collected powder is either directly fed back to the feed hopper or

accumulated for later addition in bulk. A typical system has been

shown in Figure 1.2.

1.2.5 Stoving Ovens

The type of stoving that is required to produce a good

powder coating film is dependent on the type of powder being used.

The curing time of any powder is determined by the temperature of the

oven but is a function of the formulation of the powder and its

cure and gel characteristics. It is essential that the stoving

cycle for a particular powder is closely followed for each and every

particle. Problems can be encountered with large heavy objects

where the base metal takes a long time to heat up. With heat applied

from the powder surface to the substrate it is possible to fuse the

upper layers into a continuous film and leave the lower layers unfused.

This results in very poor film qualities and in severe cases leads to

flaking off of large sheets of the coating i.e. no adhesion.

When considering the type" of oven to be used it is necessary

to consider the size and type of articles to be stoved; whether a

conveyor line or batch process; amount of air supply required; and

the temperature requirements of the powders to be used.

There are two main types of oven - Chamber or box ovens and

continuous ovens, which are open at both ends. The heat for stoving

can be supplied in a number of ways.


- 13 -

Convection ovens are one of the commonest types. Gas or oil

fired burners are used to heat air which is then transferred to the

enclosure of the oven. The length of the oven depends on the

production rate required, curing time etc. The higher the air flow

rate through the oven the shorter the stoving time, but the air

velocity is generally kept to 1.5 m/s to avoid powder being blown off.

Articles can be stoved by infrared radiation which affords very

fast heating with low air velocity. The hot source is created by

electric elements or gas burners.

Induction heat is when the part to be stoved is exposed to an

alternating electromagnetic field causing eddy currents to be induced

and,therefore, the production of heat within the article. This is

very useful for thin plate and large quantities of small parts, but

it has the disadvantage. of being extremely expensive when stoving

large items, due to the necessity to heat all of the object.

However, this method does eliminate the possibility of bad adhesion

due to incomplete fusing near the substrate since stoving occurs from

·inside to out.

1. 2.6 Powder

The powder formulation and characteristics are of great

importance in coating any particular article. The powder will be

chosen for a particular purpose according to its formulation and the

properties it exhibits in the fusing cycle. There are many types of

powder used in the industry but they can be classified into two main

categories.
- 14 -

(i) Thermoplastic powders

When fused, these powders form a smooth continuous film.

They have very high molecular weights due to long straight chain

hydrocarbons. They have very few chemical groups attached to the

chains and hence have a fairly high resistance to chemical attack.

(ii) Thermosetting powders

These are usually low molecular weight powders which

melt and flow during fusion and undergo chemical conversion at the

same time. Hence, once a thermoset has been cured it cannot be

remelted to a plastic material. These powders are more easily used

in industry giving high quality thin films from the small particle

size range powder, which is difficult to produce for high molecular

weight thermoplastics.

Thermosetting powders crosslink across the polymer backbone

and hence the formulation, curing, flowing and viscosity charac-

teristics can be carefully adjusted to produce a powder for a specific

purpose. Thermosets have several advantages over thermoplastic

powders. They need no primer; have excellent chemical and solvent

resistance; and are capable of higher pigmentation levels. However,

a powder is generally chosen for its individual properties with

regard to a particular system (J. Smarsh, 1972; C. Korf, 1976).

The choice of the resin and hardner is made with respect to the

final coating properties. Apart from the colour contribution made

by the pigment, consideration must be given to the properties of:-

(i) thermal stability

(ii) chemical stability

(iii) hiding power

(iv) particle size

~) electrical properties
- 15 -

The formulator must, therefore, take into consideration the

following essentials:-

(i) Method of production: Pigment loadings and addition of

hardener is dependent on the type of equipment used, be

it batch or continuous.

(ii) Resin/hardener choice: Toxicity and particle size must

be considered with regards to the homogeneity of compounding

operations

(iii) Correct level of gloss and colour

(iv) Choice of grinding process to give correct size

distribution

(v) Storage stability and flowability

(vi) Capability of recycled powder to be used again without

reprocessing.

A typical manufacturing sequence of a powder is given in f.~gure 1. 4.

The final product that is achieved should be a homogeneous compounding

of all the ingredients with a uniform particle size distribution and

each particle having the same composition.

The first stage of the manufacture can be done in one of two

ways. Both have the same objective to mix the ingredients to ensure

homogeneity. Premixing, using Ltumble', horizontal, conical or high

speed mixers and blenders, is followed by hot melt compounding of

the powder by means of Z-blade mixers, heated rolls or continuous

extrusion methods. The Z-blade mixer has the advantage of not needing

any pr emixing'. The mixture is first melted in the blender and when

molten the other ingredients are added slowly, whilst using the blade.

This mixer gives good dispersion and is relatively cheap but it is

difficult to clean and cannot be used for fast curing systems. The

flow agents and hardener are finally added and the mixture is then
Granulate Resins

Weighing Station
Resins Pigments
Additives Hardness

~
/ Premixing

Z blade mixer

~
Extruder

Cooling Band
/

Granulation

Grinding

/ ~
Sieving Classifying !--

~ 7
Packing

EPC Process

Figure 1.4. Powder Manufacture Sequence


- 16 -

cooled rapidly on a roller cooling band.

The semi continuous method of using rollers is not widely used

for powder coatings but can be used for thermoplastic powders.

Extrusion is the most common method of producing powders. The

mixture is passed through a heated screw extruder in which the resin

melts and leaves as a homogeneous mixture. The extruder has the

advantages of being very easy to clean and that it is easy to operate

and control the temperature. The extrudate is quickly cooled and

grinded roughly before passing into the pulverising and classification

section.

The ease of particle size reduction mainly depends on the

molecular weight of the base resin. Hence thermoplastics are extremely

tough and resilient and these types of powders are reduced in size

at low temperatures by freeze grinding or similar methods. The

grinding of the ·particles is otherwise effected by ball, rod and

tube, or hammer mills. Once the material is processed through this

equipment it has a wide particle size range and so a classifier is

used to recycle the large particles back to the grinder. The

classification of the particles is mainly carried out by cyclones or

sieves.

There are many types of powders available. Table 1.2 gives

an outline of the usage of the various commercial powders in use.

Each one is greatly influenced by its formulation and method of

manufacture (5. Kut, 1971c) •

Epoxy powders still account for the largest proportion of

market. However development of powders for the automobile market is

in progress and it is thought that the successful powder will more

likely be a polyester or acrylic.


Table 1. 2

Properties and Usages of Various Types of Commercial Powders

Powder Properties Uses

Epoxy A. Excellent electrical insulation Functional items


outstanding adhesion, impact, Rotors, Stators.
abrasion and scratch resistance Pipe coatings
High temp resistance Wire goods
No primer required Light fi ttings
Thin films

D. Chalking and UV weathering

Polyesters A. Very versatile


High light and heat stability Transformers, control
High exterior resistance boxes
Good gloss & levelling Guard rails, highway
posts
Wire shelving
Hardware
D. Can hydrolyse - lower chemical
resistance
Primer required

Acrylic A. Light and heat stability .Automotive and exterior


Excellent weathering applications
resistance. High gloss. Window frames
Good adhesion Gas heaters
swi tch gear.
D. Primer required

Nylon A. Average environmental prop. Aircraft batteries


V. good mechanical &. wear props. furniture, safety
Good electrical prop. devices, handles,
door hinges, springs,
D. Weak to strong acids valve stems and seats

Cellulose A. Thick films Food racks - trays


Solvent & weather resistance Screl;driver handles
Outdoor signs
D. Needs primer

Vinyl A. Good water immersion and Dishwashers, wire goods


corrosion resist. Outdoor Chain-link fencing
resistance. Good chemical Posts and rails
resistance

D. Primer required
- 17 -

1.2.7 Colour Change

In powder spraying it is absolutely vital to remove the

last powder sprayed from the whole plant before another powder is

used. The time it takes to do this is dependent on the equipment

in use. The design of spray booths is important in that difficult

cleaning places must be minimised. The type of recovery unit is also

important with regards to cleaning. As previously mentioned, bag

filters are not suitable for frequent colour changes.

Therefore the engineering approach to the design of the coating

system has included the combination of one or more of the following

components

(i) Spray equipment

(ii) Spray booth

(iii) Cyclone

(iv) Fabric filter

The arrangement and multiplication of each of these components

is dependent on a series of factors governing the type of use of the

system:-

(i) Number of colours to be used

(ii) Usage time for each colour

(iii) Required change-over time

(iv) Type of system in use

(v) Allowable colour contamination/colours in use.

The ideal approach would be to disregard the oversprayed powder

but this is not economically viable. The problem is to find a

relatively simple and economic system using a high percentage of the

powder, which is easy to operate and not space consuming. Hence there

is a need to understand the basic fundamentals of the system, and to


- 18 -

investigate the effects of powder and operational characteristics

on the efficiency of the process so as to minimise this problem.

1. 2. 8 Substrate Preparation

However much time is spent in determining the best

formulation of a powder it should always be remembered that the

substrate surface must be in proper condition to make a coating

effective. In powder coating, just as in any other coating process,

operations. must be carried out to remove imperfections from the

substrate; create a surface susceptible to bonding; and prevent

corrosion, through a break in the coating, by a chemical protective

coating (5. Kut, t9.71b; H. Stein, 1972; R.A. Ashdown, 1974; W.G. Willows,

1968).

Methods for cleaning the substrate inc1ude:-

(i) Solvent wipe

(ii) Emulsion cleaning

(iii) Alkali cleaning

(iv) Acid cleaning

(v) Abrasive blasting

The choice of method of pretreatment depends on whether the coating

is for internal or external use; the type of coating; and the type

of substrate to be coated. The most popular methods of pretreatment

are:

(i) Spraying iron or zinc phosphate

(ii) Zinc phosphate Dip

(iii) Chromating process

(iv) Anodising
- 19 -

It has been shown that each type of pretreatment can give

different results to adhesion and salt spray tests. These results,

however, are also a function of substrate material. For steel,

chromate coating was found to be very resistant to corrosion and for

both aluminium and steel iron phosphate gave balanced results.

Results have been observed to vary for the flow of the powder

layer during melting on different pretreated. substrates. The surface

condition was found to be responsible for this, grit blasting giving

lower areas of flow CB. Whiting, 1977).

Hence to ensure a coating that has good corrosion resistance

it is necessary to combine powder coating with pretreatment, the

method of which is dependent on the demands on mechanical properties

and use of the coating.

1.3 Previous Research

In the previous sections of this chapter the electrostatic

powder process and the powders used therein have been briefly

described. In order to clarify the reasons for this work there are

certain phenomena and practical problems that should be mentioned.

These are important when considering the effectiveness of the process.

The limitations of the industry at present have been mentioned

in the introduction. Problems of colour change and film control

must be eliminated or simplified so that markets are increased.

This meanS that production of 25-50~m films, with minimal overspray,

must be achieved and yet the excellent film properties that the EPC

process presently possesses must still be maintained. The effect of

particle size and film thickness on the properties of the final film

are therefore investigated and described in detail later. Following

these experiments, research into the way in which different sized


- 20 -

particles pack is reported to try and understand the fundamental

effects already observed. Recommendations can then be made to solving

the above problems when the mechanisms involved are more fully

understood.

There have been several workers that have already made fundamental

investigations into the effect of operational variables on the coating

(A. G61ovoy, 1973abc; G.F. Hardy, 1974; G.D. Cheever, 1975;A. Golovoy, 1975).

These effects have been reported with reference to the film thickness

and to the deposition efficiency. This is the ratio of powder

deposited on a substrate to the amount of powder sprayed from the gun.

Some of the results that have been obtained are shown in Figures

1.5-1.12.

Figure 1.5 shows a basic effect in th~t the deposition efficiency

of the system, as the packing builds up, decreases. This effect is

also shown in Figures 1. 6, 1. 7, 1. 8 and 1. 10 in a similar way.

Deposition efficiency decreases as charging voltage applied to the

electrodes of 'the spray gun is decreased (Figure 1. 8) and is lower

for thicker films.

The charge on ,the particle is a very important factor in these

cases. The theoretical considerations of charging particles and the

forces acting on the particles will be fully discussed in a later

chapter. However it is worth stating here that the size of the

particle. its charge, and the forces acting on the particle are all

inter-related. A change in anyone of these factors can give rise

to one or more of many different effects.

In brief, the charge on a particle is dependent on the particle

size, its electrical properties. the gun potential and the strength of
v 60 kV
... 80
=

»
u

.""U
QJ

." 60
""'""'
'"
.""
0

."'"' 40
'"0
Q
""
QJ

20

o ~ ____ ~ ______ ~ ______ ~ ____ ~~ ______


1 2 3 4
Film Thickness - mil
Figure 1.5. Deposition Efficiency vs Film Thickness at 60 kV

0.8 mil

80
....
»
u
2 mi
.""u
QJ
60
."
""'""'
'"
.""
0 3 mil
40
."'"'
'0"
""
Q
QJ

20 o
Charging Voltage kV
Figure 1.6. Deposition Efficiency vs Charging Voltage for Various
Film Thicknesses
400

300
e;:l.

'"'"
.,.
'"
.....c<J"
200

....
00
...."
'"0
",. 100
u

20 40 60 80
Coating time, secs

Figure 1.7. Coating Thickness vs Coating Time

72 kV

54 kV
.........'"e 6

'"'"
.,. <lJ

.....c"u 4
....
e
.........
'" 2

5 10 15 20
Spraying Time, seconds
Figure 1.8. Film Thickness vs Spraying Time for Various Voltages
- 21 -

the field near the gun. A particle is normally charged by passing

through a corona region that is·produced by application of an H.T.

voltage to electrodes near the end· of the gun. The particles are

designed to have a resistivity of the order 10 14 n cm so that they

hold their charge for the time they are being sprayed and stoved.

In practice the particles will remain on the workpiece for several

hours.

The higher the gun potential the larger the charge on the

particle and hence the greater the initial attraction of the charged

particle to the substrate. However as coating continues the

deposition decreases and the thickness of the film increases only

very slowly. In Figure 1.6 it can be seen that a 2 miLthickc:film

has a much lower deposition efficiency. Basset et al (J.D. Basset et aI,

1975; Y.V. Ting,. 1978) found that the coating was thickness limiting

between 0.25 and 0.45 mm after coating times in excess of 20 secs,

depending on the operating voltage (Figure 1.7). Golovoy however

found that film thickness decreased towards a limit of approx 9 mi1s

in the best case (Figure 1.8)~ The ability of the system to be

limiting can be very useful but occurs at fairly high thicknesses.

The mechanisms involved can be of two types:-

a) Repulsion of particles due to the packed layer of charges,

of the same polarity to the oncoming particle. The particle

is repelled and will then be attracted to less packed areas or

leave as overspray. The 'wrap-around' effect is achieved when

these repelled particles are attracted to the 'blind' side of

the target, purely by electrostatic forces.


- 22 -

b) Back Ionisation. This is due to electrical discharges 1n the

powder layer. Ions of opposite sign leave the surface and

discharge powder being sprayed, such that it is not deposited

(A.W. Bright, 1977).

Observations of back ionisation, shown by an increase in current

to the earthed substrate, take place after a long coating time (above

20 secs), (Y.C. Ting, 1978).

The experiments conducted in this work involve spraying times

much less than 20 secs. In the case of a positive applied voltage

back ionsation is greater. This is due to the fact that electrons

are being emitted from the layer and are able to cause further

ionisation. This is by secondary electron emission from the powder

particles before forming negative ions. The discharge of ions from

the layer is due to dielectric breakdown caused by the increase in

potential within the layer as it builds up.

The 'critical thickness' means the thickness of the unstoved

coating when the potential difference across the powder layer is equal

to the dielectric strength of the powder. Hence when the critical

thickness is exceeded the deposition efficiency is drastically reduced

and the uppermost particles are held on very loosely. Hardy has

shown this with experiments involving the vibrating of sprayed panels

bearing thick powder layers. In the first few seconds of vibration

the majority of weakly charged powder was removed. (G.F. Hardy, 1974).

Similar experiments concerning the adhesion of the powder to

the substrate have been carried out by Ong and Cross (P.H. Ong et a1,

1975; J. Cross, 1975).


- 23 -

Golovoy has shown that the average particle size of the powder

sprayed affects the limiting thickness achieved. Small particles,

with associated high charge to mass ratios, produce thinner films

(Figure 1.9).

Operational variables, such as powder output, velocity, spraying

distance and substrate size, also effect deposition efficiency

(K. Venlet, 1973). The greater the spraying distance the thinner

the films produced (Figure 1.10). This is related to the velocity

gradient of the air (ratio of air velocity at gun exit and spraying

distance). Larger gradients give thinner films (Figure 1.11).

The substrate width compared to the spray diameter has also been

investigated by Golovoy and results are shown in Figure 1.12. The

larger the substrate width compared to the spray diameter, the better

is the deposition efficiency.

The critical thickness of a coating rises with increasing

particle diameter and density, increasing dielectric constant and

decreasing particle charge. However the inverse is true for the

adhesive forces but adhesion is usually adequate under most conditions.

The rate of charge decay of a deposited layer is influenced by the

powder's chemical composition, temperature and relative humidity.

Increase in humidity and temperature cause an increase in rate of

charge decay, but providing the resistivity of the powder is kept


10
above 10 n cm the residual charge should be ample.
121Jm

........e
tIl

6
tIl
tIl

...'c:"
.....CJ
..c: 4
E-<
e
........P« 4.51Jm

15 2
Spraying time (s)
Figure 1. 9. Film Thickness vs S2ra;l:ing Time for Various Sized Powders
,

2 ins

5 ins

........e
tIl

6
tIl
tIl

...'c:"
......c:
CJ

E-<
4
e
....
.....

10 ins

2
15 ··:i:ns

~
5 10 15 20
Spraying time (s)
Figure 1.10. Film Thickness vs S2 ra;l:ing Time for Various iliun-target
Distances
-1
3 sec

8
6 sec
-1

.,
....i3
.-< 6

.,., sec
-1
<lJ
.:
.:.:tJ
.....J: 4
....
i3
....
.-<
~
2

5 10 15 20 25
Spraying time (s)
Figure 1.11. Film Thickness vs Spraying Time at Various Air Velocity
Gradients

80
.... V = 60 kV
>.
tJ
.:
....
<lJ
60
........."
.....
'".:
.......0 40
.....,
0
0..
<lJ
<=>

20

o 5 1

Figure 1.12. P1otof Deposition Efficiency vs Ratio of Substrate Width


and Spray Diameter
- 24 -

r.4 Levelling in Powder Coating

After a powder coating is stoved it tends to exhibit a waviness,

or 'orange-peel', which is due to incomplete levelling during the

bake. The extent of this is dependent on the irregu1atiries in the

powder layer i.e. the way in which the particles have packed, and

therefore the thickness of the coating (G.T. Spitz, 1973; S.E. Orchard,

1962; S.M. Wolpert, 1973; S. Gabriel, 1975).

As a powder coating is stoved the roughness of the surface

will decrease and will eventually stay at a given amplitude. The

wavelength of the waviness does similarly, but increases with time.

The levelling in powder coatings has been discussed by Spitz,

Nix et a1 and Wo1pert (V.G. Nix et aI, 1973). From their experiments

it has been concluded that the ripple· in coatings comes from an

uneven deposition of powder, and that those differences due to

particle size have less effect than those associated with the flow

out of particle clusters (Figure 1.13) (S. Wolpert et aI, 1972;

C.H.J. Klaren, 1976).

In the process of stoving the particles first melt and then

flow out to cover the substrate surface. The cross-linking reactions

then begin and the polymer starts to gel, preventing any further flow

(M.J. Hannon et aI, 1976; A. Quach, 1973). The flow of the particles

is shown in Figure 1.14 with respect to temperature. The time to

gelation is a function of temperature and the type of epoxy powder

being used (shown in Figure 1.15).

Gabriel found that the behaviour of any system during stoving is

determined mainly by the chemical nature of the resin/hardener

combination. For any particular powder the way it cures can be


A Deflection typical of cratering
B Deflection typical of dust particles included in film

1 cm = 2~m

1 cm = 500~m

Figure 1.13. A Typical Powder Coating Profile


GEL

FLOW

_c_ _ _ MELT ,..

Temperature •

Figure 1.14. Plot Showing Stages in the Melt and Spread of an Epoxy
Resin with Increasing Temperature
1000

Slow Curing

Medium Fast Curing

Fast
Curing

10L-_______________________________________________

Temperature •

Figure 1.15. Time to Gelation of Various Epoxy Powders vs Temperature


- 25 -

expressed with a DSC (differential scanning calorimeter) thermogram.

Using this method the amount of energy absorbed or evolved by the

sample is recorded as a function of the sample temperature. The

differential of the amount of power required to keep the sample at

the same temperature as a reference, as expressed as energy per

unit time, is plotted against temperature. These thermograms clearly

show the glass transftion point and melting and curing times of the

powder, since temperature is proportional to time via the heating

rate applied to the reference (Figure 1.16).

From the combination of all these results it can be seen that

the stoving cycle plays a considerable part with reference to the

final properties of the film, as does also the structure of the

particle packing. In the described experiments, the stoving cycle

for all samples is kept constant so as to eliminate any changes in

properties due to this.

1.5 Practical Problems

As previously mentioned, when the film thickness reaches that

of the critical thickness there is an onset of electric breakdown

within the layer. When discharges take place the surface of the layer

becomes rough and cratered. positive ions will stream out towards the

gun corona and neutralise oncoming particles. This causes loosely

bound particles near the surface of the coating. The same effect

occurs when the particle charge is continually low. The driving force

for deposition is small and therefore the packing tends to become

loose and rough.


Cure

Tg
\
-.1-- Melting
\ .-------::.- - -

j
.....
CJ

...<ll
~

'"o
.",

...~~~--------------~----------------~----------------~--------------~---
50 100 150 200 250
Temperature °c

Figure 1.16. Typical DSC Thermogram for an Epoxy Resin Powder


- 26 -

The sprayed particles tend to follow the lines of the electric

field and hence when the voltage on the surface increases due to packing,

oncoming charged particles will deposit onto regions of lower

thickness. Hence the layers tend to grow fairly uniformly on flat

substrates.

However for the case of a can or similar object the shielding

effect of the surrounding conductors gives rise to poor penetration.

This is known as the Faraday Cage effect, which states that any empty

space enclosed within a conductor is free from any field.

Better penetration into a Faraday cage can be obtained by a

number of methods:-

i) reducing turbulence of the transporting air (giving particles

higher momentum)

ii) increasing the particle charge (lowering powder feed rate)

iii) increasing charging voltage (increases field at opening but

also increases particle charge which is more important)

iv) increase particle size (higher momentum).

All these methods effectively mean that the ·particle is entering the

cage by having higher momentum.

At the edges of a substrate the field lines will tend to be

more concentrated and so particles will preferentially deposit here.

Therefore better coverage is obtained with powders rather than liquids

where the surface tension effects pull liquids away from the edges.

Other problems that arise are the space charge effect and

electrode blockage which will effectively quench the corona at the

gun. This causes a drop in charging and can even give zero deposition.

High humidity can cause electrode blockage, powder particles sticking

to the electrode, and also increases the charge decay rate.


- 27 -

1. 6 Sunnnary

A brief description of the electrostatic powder coating process

has been made. The process, its equipment, powders and operational

problems have been discussed. It is evident that there are many

factors that contribute to the way in which both dry and fused coatings

are formed.

At the electrostatic powder coating conference in London 1968

Prof. D.e. Freshwater said:~

"There is a need for research both applied and fundamental into

particle packing and its qualitative effect on the factors outlined.

Until we understand more about relationships involved we shall

not be able to make sensible predictions about the desirable properties

of powders for powder coatings".

It is the intention of this work to study the effects that

particle characteristics and film thickness have on the qualities of

the stoved film and also to investigate the way in which single

particles pack onto a substrate. In doing this it is hoped that a

better understanding of the way in which the EPC process can be

controlled will be gained.


CHAPTER 2

A REVIEW OF TEST METHODS

2.1 Introduction

2.2 Gloss

2.3 Film Thickness

2.3.1 Destructive Methods

2.3.2 Non Destructive Methods

2.4 Roughness

2.5 Environmental and Accelerated Weathering

2.6 Adhesion

2.7 Hardness

2.8 Abrasion and Flexibility

2.9 Chip or Impact Resistance

2.10 Porosity

2.11 Summary
- 28 -

2.1 Introduction

The function of the particle characteristics on the final film

produced in powder coating is very complex. There are many properties

that can be affected by one or a combination of characteristics in

the dry film, formed when spraying, or the conversion of the dry

film to the continuous cured film (V.K. Croutch, 1976). Particle

size can affect packing porosity, thickness, strength and hence give

rise to changes 1n adhesion, roughness, gloss, corrosion resistance

etc. The spraying process and the stoving conditions employed are

also very important parameters when considering the effects of any

one particular property. The types of effects produced will be

briefly discussed later.

With regards to the actual testing of a coating, the tester

himself has to decide what standards are required and what the

meaning of the results of each test represent (A.E. Claxton, 1975).

In practice it is quite possible that experience and knowledge of the

coatings themselves with respect to how they should be applied is

enough to decide upon a particular coating system, or conditions for

a given problem. However, when optimisation and comparison of slight

differences is being observed it is often difficult to assess the

situation without having some quantitative results for a given set of

parameters of the individual coatings. Even then it is necessary to

distinguish the relative merits and importance of these differences

between the results.

For the problem at hand, one of the greatest difficulties is

that a standard commercial powder is being used which is known to give

good finishes under normal spraying conditions. To use any chosen

size distribution of this formulation of powder and to measure the


- 29 -

properties of the final film means that sensitive tests must be

employed to detect any differences in the measured properties. Hence

a review of the standard test methods is in order and their applic-

ability to the given situation is to be considered. (ASTM D345l,

1975) ..

Test methods are constantly being devised year by year and

there are many organisations which review and develop tests and

observe their applications to various sections of industry. Two of

these such organisations are the American Society for Testing af:.

Materials (ASTM) and the British Standards Institution. The main aim

of these organisations is to advise (ASTM Standards, Part 27) on the

possible standard methods for particular purposes which will give

reproducible results within prescribed limits. However, the exact

meaning of these results must also be determined.

The tests on a powder coating can comprise of testing the dry

powder layer and/or the fused film. The following is a brief summary

of the various test methods employed for measuring properties of the

fused film.

2.2 Gloss

It is widely accepted that gloss, besides colour, is one of

the most important features of a coating. There are many methods and

apparatuses for measuring gloss, but one of the most difficult tasks

is to relate from one set of results to another (N.I. Gaynes, 1977).

Gloss in itself is a very complicated property in that changes in

gloss are considered as possible indications of other insufficient

qualities of the film (5. Huey, 1964; H.K. Hammond Ill, 1974). Indeed

gloss is largely a function of smoothness of surface and hence changes

can mean deterioration of other factors in the film. One of the


- 30 -

simplest tests is to compare the image formed when using the film

surface as a mirror. Here it is very easy to compare one film with

another but proper assessment of the film is very difficult

(B. Svoboda et aI, 1962).

Hence the measuring industries have produced many types of

gloss measuring instruments (U. Zorll, 1972; R. Goudie, 1974;

J.A. Fraunhofer(8) , 1972c). First of all it is necessary to state

exactly what type of gloss is to be measured since there are several

types of gloss such as specular gloss, sheen, uniform, contrast gloss

etc.

The basic principle of a gloss meter is to shine a beam or

pencil of light onto the surface of the film at a specified angle

(20, 45, 60 or 85 0 ) and the light reflected at this mirror angle is

registered by a photocell.

The light that is received by the photocell will be made up of

the mirror reflected beam and the diffusely reflected light. Hence

two panels might have the same 60 0 gloss reading but the eye will

notice the haze of the image compared to the distinctness of a

reflected image. Panels with a relatively high degree of haze will

hence have a high gloss rating although the distinctness of image is

low. The 60° gloss meter constantly needs standardising and checking

so that accuracy is maintained.

The correlation between visual impression of gloss and results

obtained with specular gloss meters is not always satisfactory because

of the aforementioned reasons. ASTM D523 and British Standard BS3900

Part D2 (1967) describe how to measure specular gloss. The main

difference between the two methods is that in the UK and Europe the
0
standard angle of incidence is 45 and in the USA and other parts of

the world it is 60 0 •
- 31 -

Sheen is the brightness or gloss of a surface observed at very

low (grazing) angles of incidence and reflection, normally made at


0
70 or greater. There are many devices on the market that are

available for measuring gloss and sheen, each one of them measuring

at a given angle of incidence and so are used for a particular type

of finish. For example, Sheen's 85 0 Specular gloss meter is used for

checking gloss within the eggshell to flat classification (or any


0
panel measuring below 30% on the 60 glossmeter). Hence the higher

the gloss the smaller the angle of incidence that is used.

A more precise gloss measurement, which takes account of the

fine differences between the various types of gloss, must depend on

the determination of the distribution of light scattered from the

coating surface (J.H. Calling et al., 1968). The goniophotometer is

such a device that can show quickly these qualitative differences

between samples (M. Tehan et al., 1974; P.S. Quinney et al., 1971).

In this device the sample is illuminated with a parallel beam

of, light under a fixed angle of incidence and the distribution of

the scattered light is measured for a wide range of angles of

observation using a photocell. The quantity of scattered light

measured can then be plotted against the observed angle (angle of

reflection) so giving a gloss curve. It can generally be said that

narrow curves indicate high gloss and broad curves indicate low

gloss. Th'is curve, also known as the goniophotometric curve, in

general yields sufficient detail about the type of gloss and permits

the derivation of quantities characterising the latter. It has become

common practice to note the maximum and half width of the peak of the

goniophotometric curve (T. Kosbahn, 1964; K. Koffman et al., 1966;

u. Zorll, 1963).

The comparison of reflectance methods and related methods has

been carried out for the study of film surface deterioration due to
- 32 -

weathering (M. Tehan, 1974). Comparisons were made using measurements

of the goniophotometric curve, surface average roughness and attenuated

total reflection. In comparing these methods it was found that

although determining the goniophotometric curve is slow the method is

very precise and the results can be represented by various parameters

in many different ways.

Since the types of finish produced in the following experiments

cannot be known beforehand it is necessary to use such a type of

technique, as the goniophotometer, that enables a broad range of

gloss measurements to be made. This technique, therefore, gives a

broad but precise method of measuring gloss and is therefore well

suited to the experiments to be undertaken.

2 .. 3 Film Thickness

Film thickness is now one of the most regarded properties of

a powder coating film. The coating industry continuallY attempts to

produce thinner films, with a target of about 25~m, but which still

retain all the o~r film properties, and most important of which is

continuity. The shortage of materials and the escalating costs of the

same make it increasingly important to maintain. all the other

properties (corrosion resistance, hiding power etc.) but to decrease

film thickness. However, it can be seen that there must be an

optimum point when the replacement of protective coatings due to

corrosion through thin films outweighS the cost of producing thicker

films •

Hence the instrumentation industry has produced a wide range of

devices for measuring coating thickness whether the films are on

metallic or non-metallic substrates and whether the coatings are


- 33 -

organic or metallic (J. Boxall et al., 1974; J.A. Fraunhofer ,Pt I,

1971) •

The types of thickness measuring methods fall into two

categories :-

2.3.1 Destructive Methods

Microscopy is one of the most simple and absolute

techniques of measuring thickness. A specimen can be mounted in a

perspex or resin block and then ground and polished to expose a

section of the subs.trate and coating. Any materials can be used for

this technique but problems are encountered with soft or very thin

layers.

Another simple method is that of weight change. This is a very

quick method but not so accurate. Usually the coating·is stripped

off the substrate and the surface area of a piece must be accurately

found as well as the density of the coating.

A jet of corrosive solution can be directed at the component

and the time it takes to penetrate to the substrate is recorded ..

The coating thickness is then read off calibration charts but again it

is difficult to predict accurately because different coatings react

at different rates.

The Gardner needle gauge however is simple and relatively

accurate for non-conductive coatings on metal substrates. A needle

penetrates through the coating until a lamp lights when electrical

contact is made with the substrate. A dial gauge is read to obtain

the thickness of the film. The Gardner depth gauge is a similar

apparatus but uses a scratch to penetrate instead of the needle.

The Elcometer thickness gauge measures film thickness by using

a simple dial indicator, such as used in the machinery industry, on a


- 34 -

hole cut in the substrate.

ASTM 01005 uses a similar simple gauge as this but suggests

the removal of the film. If the thickness of the substrate is obtained

before coating then this device can be used non-destructively.

2.3.2 Non-Destructive Methods

Modern quality control in industry calls for quick, easy

to implement, and easily interpretable methods on large, small or

odd-shaped items. In general it is required to have fast and

reproducible non-destructive methods which can also be automated if

desired.

A number of tests rely on the effect on the strength of a

magnetic field or eddy current between a probe and substrate caused

by the separation due to the coating (T.D. Latter, 1976).

One of the simplest forms of the change in the magnetic field

is given in ASTM 01186 where the force that is required to pull off

a bar magnet suspended onto a coating and substrate is determined.

The force is applied by turning a graduated dial which coils a

beryllium copper spiral spring. The thickness is calculated from

calibration curves obtained by using reference thickness standards.

The more sophisticated forms of:thisc' come 'as the Elcome ter Mini tector,

Mikrotest, Magnetic Gauge, Magne Gauge and Permascope ES8 (Anon,

1976; D.R. Birchenough et al., 1972). These work on the principle

that a reduction of flux occurs when the coating i~ introduced into

the magnetic field. It is necessary of course to have the coatings

on a magnetic material substrate.

However, in the eddy current method the thickness of electrically

non-conducting coatings can be measured on electrically conductive


- 35 -

(usually non-magneti"c) substrates. The probe in this case applies a

high frequency field which produces eddy currents in the metallic

substrate and measurement of the reaction of these currents is made

by a test head placed on the surface coating (D. Jarvis, 1974). The

strength of the reaction is dependent on the conductivities of the

coating and base metal and on the thickness of the coating. Hence

test panels must be used for calibration of these instruments.

Typical examples are the Permascope EC8 and Eddytector instruments.

Combinations of these two methods are employed in the Twintector

and Dualscope so that most coatings and substrates can be catered for

(Anon, 1976). Each instrument can operate at a variety of ranges and

are simple to use.

The C-Scope can measure very thin coatings on a conducting

substrate by using the capacitance method. Here the capacitance

formed by a non-conducting coating between a conducting substrate and

probe is a measure of the thickness.

In the precious metals industry the Beta-backscatter method is

widely used and is dependent on the adsorption of radiation. Beta

particles are emitted from an isotope onto the coating and are

scattered back in proportion to the coating thickness and atomic

number of the coating material and substrate (R. Allen, 1973).

2.4 Roughness

The main method for measuring surface roughness is by the

traversing of a stylus across the surface coating. This method has

been used since the early 1940's. Vertical movements of the stylus

relative to a skid which is used to guide the pick up over the work-

piece are converted by a transducer into corresponding changes in an

electric current. These changes are then amplified in an electronic


- 36 -

unit and are used to drive a profile recorder and an averaging meter.

The stylus tip is nominally in the order of about 2~m diameter and so

it will enter very fine grooves, and the large taper angle (90 0 )

enables troughs of up to 100~m to be measured (R.C. Spragg,

D.J. Whitehouse, 1972).

It is generally accepted that the irregularities in the surface

caused by the machine tool used to produce the surface constitute

the roughness whilst the irregularities generated by vibration between

the tool and workpiece are classed as waviness. In the case of powder

coating this can be described as the contribution due to the method of

spraying and the way in which the particles fuse and flow out. The

surfaces are characterised by the mean roughness height which is

expressed as either the Centre-line Average (CLA) in the UK or the

root-mean-square (RMS) in the USA.

A reference length of the surface is taken and a centre line is

drawn through the profile. The sum of all the absolute values of the

height of peaks and depth of troughs from the centre line is

calculated. The sum is then divided by the total number of measurements

within the reference length and the result is expressed in ~m or ~in.

The simplest method of measuring roughness, however, is to use

a modified dial gauge mounted on a flat base which when placed on the

surface coating will rest on all the peaks. The probe of the device

then projects through into the troughs and grooves in the surface,

hence measuring the surface roughness.

Measurement of surface smoothness is generally performed by

interference microscopy. Variations in the surface smoothness will

cause disturbance of interference fringes (width A/2) which are

formed on the surface by means of a monochromatic light source

(wavelength A). The degree of deviation is a measure of roughness.


- 37 -

Very small variations (0.03 - 0.8~m) can be measured by this technique.

However, this test is not used as a routine test in industry.

2.5 Environmental and Accelerated Weathering

Since many coatings are on exposed exterior surfaces accelerated

testing of coatings is required. This is particularly required in the

paint industry where the number of paints that are being developed is

considerable; Therefore there is a need for rapid test methods to

replace the time-consuming atmospheric exposure methods (M.B. Kilcullen,

1975). For many years coatings were rated in terms of "number of years

of Florida exposure" (B. Lindberg, 1975). This type of outdoor

weathering test is performed by using racks which face south and which

hold the coating materials at angles of 30 0 or 45 0 to the horizonta~.

The racks are situated in industrial atmospheres or close to sea

shores etc. SO that they are exposed to different climates and

atmospheres. These are the simplest form of tests but are very long

(in the order of months and years).

The most common indoor accelerated test is the_ salt spray test

(ASTM Bl17-73). Here the test panels are placed in a fog chamber into

which a salt solution (S parts by weight) is sprayed at a temperature


o
of 3S e. In this case the exposures are made in multiples of 24 hrs.

Other forms of this test are a combination of salt and acetic acid

which raises the pH from 6.S - 7.2 to 3.2 - 3.S.

In this test a 'cross' made by a scalpel through the coating to

the substrate can be used to examine the corrosion and blistering

through undercutting of the coating layer. This is also indicative

of the adhesion between the coating and substrate.

ASTM D870 describes the test method of immersion of the panels

by 2/3rds into deionised water at 38 0 e. In each test the sample plate

is observed after testing and the amount of surface degradation and

corrosion is recorded (I.D. Aitken, 1968).


- 38 -

Humidity cabinets are also used in which water is heated to give

the relative humidity in the cabinet above the critical 70% rH level

(C.A. Mitton, 1968). The temperature is also altered depending on

the cycle of the test that is desired. However, this test is a

relatively mild-accelerated test compared to the salt spray test

(J.A. Von Fraunhofer, 1972a).

The accelerated weathering apparatus consists of exposing the

coated panels to an intermittent distilled water-spray and continual

exposure to a high power arc lamp. The plates are sprayed once every

20 mins. The Atlas 'weather-o-meter' model is an example of such

an apparatus in which the power of the arc, with or without filters,

the dew cycle and humidity can all be independently varied.

In conjunction with weathering tests, whether using simple

outdoor environmental or accelerated weathering methods, the optical

and surface properties can be observed as the exposure time of the

tests is changed. Here the gloss, roughness and adhesion parameters

can be monitored with reference to the weathering time.

2.6 Adhesion

One of the most simple adhesion tests is that made by using a

fingernail. This tests the strength of all the adhesive forces such

as cohesion, Van der Waals, molecular forces etc. but is little use

when a quantitative assessment is required (Myers & Long, 1969).

Perhaps the most common adhesion test used today is the cross

cut test (BS 3900:E6). This consists of cutting a succession of

grooves in the coating through to the substrate, 1 mm apart, and then

another set perpendicular to the first ones. The loss of the small

squares of the film during the experiment is compared to a classifi-

cation test chart which will give a classification of O~5. This


- 39 -

test can be extended in the form of the tape test which is carried out

by placing masking tape over the cross hatching and then slowly

removing it. (ASTM D3359 (tape test)). This is also compared to

a classification chart. In each case the load applied to the film

(especially in making the grooves) and the cutting edge itself, must

be carefully controlled to ensure reproducibility.

ASTM method D2l97 is very similar but uSes just parallel

grooves as opposed to cross hatching. In each cut the load applied

must be kept constant. The grooves are made closer and closer

together until the coating between the grooves is torn or lifted

from the substrate.

Cathodic reaction is the basis of the PASS method of determining

adhesion whereby electrolyte is held above a scribed line on the

coating. The amount of coating· ~emoved by the reaction in a certain

time is recorded. Adhesion measured in fundamental units of shear or

tensile strength is the feature of the Hesiometer. It is based on

the fact that coating removal is by four mechanisms:- cutting, chipping,

cracking and peeling, and that each of these mechanisms is indicative

of the ratio of the cohesive to interfacial forces (W.J. McGill, 1976).

The coating is removed by use of a fixed cutting tool which asseSses

the cutting force in relation to film thickness and can accurately

determine the critical thrust force when the removal mechanism switches

from cohesive to interfacial bond failure. The force required to

remove the coating is recorded against distance and results interpreted.

Another adhesion tester, ~n the form of an aluminium 'dolly' which is

cemented to the coating, is used to measure the poundage at which the

coating pulls off from the substrate. This is also used for measuring
- 40 -

the strength of adhesives (M.J. Marshall, D.R. Birchenough, 1972).

The adherence of the film depends on many fundamental forces

and especially on the ability of the coating fluid to thoroughly wet

the surface. Hence the thermodynamics and the kinetics of wetting

come into consideration. Therefore, it becomes extremely difficult

to devise methods to assess adhesion which accurately determine the

influence of these forces and mechanisms without encompassing any other

properties.

2.7 Hardness

Hardness indicates a condition which can be described as the

ability to resist indentation. This is covered by ASTM D1474 which

provides methods for the measuring of the Knupp Hardness No. and the

Pfund Hardness No. (S. Spindel, 1973). The test is basically performed

by applying a load to the surface of the coating by meanS of a

pyramidal shaped diamond having specified facet angles (R. Allen,

1973). The measurements of the resulting permanent impression are

converted to a hardness number. The load is applied slowly (so as not

to be an impact) and held for 18 seconds. The length of the impression

is then measured with a microscope. The pfund No. differs in that a

quartz or sapphire indentor which has a hemispherical tip of specified

diameter is used.

Other types of this apparatus are the Rockwell, Vickers and

Brunell hardness tests.

A simpler approach to hardness testing involves scratching the

surface with pencils tipped with diamonds where the width of the

scratch produced is a measure of the surface hardness (R.R. Blakey,

1977).

Rocking pendulum hardness testers (such as the Sward Hardness


- 41 -

Rocker) give a measure of hardness after a number of oscillations

from a fixed initial amplitude.

Pendulum hardness testers are of similar principle in that the

hardness is measured by the damping effect due to hysteresis losses

of a pendulum fixed to a beam with one or two steel balls placed on

the film. The pendulum rests on the surface to be tested and pivots

on the two steel balls. The time for the pendulum to decay from

some initial amplitude to some final one is a measure of hardness

(J.A. Von Fraunhofer(7) , 1972b}.

A variation of this is the dynamic hardness tester. An indentor

tipped with a tungsten carbide hemisphere strikes the test piece with

a predetermined kinetic energy. The indentor hence loses energy and

the deceleration is detected by a signal to a piezoelectric crystal.

The deceleration force (proportional to output voltage) is related to

a Vickers hardness value.

2.8 Abrasion and Flexibility

Any coating is subject to wear and hence must function as a

protective coating. Mechanical action includes sanding, rubbing,

scraping whilst other forms of wear, such as that of the wind, must

also be taken into account. There are several wear/abrasion tests

available but very few give quantitative and reproducible results.

Many use wheels, nylon brushes or rubber cups that impinge on the

film and are rotated by a reciprocating motor. The Taber Abraser

uses weighted abrasive wheels and the abrasion is measured by either

(a) the number of milligrams of weight lost per 1000 cycles or

(b) the number of cycles to wear down a film to the substrate

(5. 5pindel, 1973; D. Jarvis, 1974).


- 42 -

In ASTM method D-968 the resistance to abrasion is measured-by

the amount of sand required to -wear -through a thickness of film when

the sand falls from a given height onto the panel. A similar device

is one in which grit is blown from a tube to the test surface in an

enclosed cabinet. Wear is measured in terms of microns of surface

thickness removed per l50g of grit.

Flexibility has been defined as the degree to which a dry paint

film is able to conform to movement or deformation of its supporting

surfaces without cracking or peeling. This, of course, involves

adhesion.

Again, as in many tests, thickness of the film plays a very

important_part in the final results. ASTM D-522 refers to a

flexibility test using a mandrel. The panel is bent over a cylindrical

mandrel of a specified size (or a conical mandrel) with the coated

side up. The panel is then examined for cracks. If no cracks are

observed the same procedure is repeated for successively_- smaller

diameters of mandrel (R. Goudie, 1974).

Other methods include the Erickson cupping test where the panel

is deformed using a baIlor round-head plunger and the maximum

deformation before tearing is measured (N.I. Gaynes, 1977).

2.9 Chip or Impact Resistance

There is some interdependence between types of optical properties

but in general it is possible to separate the effects and make

measurements that can be described as colour, opacity or gloss.

Mechanical properties however are almost impossible to isolate, and

measurements of hardness, flexibility and adhesion are also highly

dependent on the thickness and age of the film, and so can indicate

a degree of curing of the film.


- 43 -

Commonly, impact experiments concern the falling of a weight

onto the coating, the type of weight being peculiar to various test

methods. BS 3900:E7 uses a 50 mm diameter steel ball.of mass 900g

that falls from an electromagnet held at a height of 3m from the

coating. The procedure is repeated so that the ball falls on the same·

spot each time and signs of cracking, flaking and detachment are

recorded.

Other methods vary 1n the form of the notching of the specimen

due to a weighted pendulum and the way in which the specimen is

clamped. Such tests are the Charpy (both ends clamped) and the IZOD

(only one end clamped) (N.I. Gaynes, 1977).

The chip resistance test is described in ASTM D-3l70. Stan-

dardised road gravel is projected at the specimens by compressed air

at a pressure of 70 psi. The effect of temperature on the tests is

very marked and so the panels are kept in a refrigerator at lOaF until

required for the test. The lower the temperature the more severe the

test. At the end of the run masking tape is pressed over the panel

and then pulled off slowly to remove any loose chips. The panel is

then referred to a set of pictorial standards.

It is worth noting the similarity in the test methods here to

those of abrasion and adhesion, emphasising the fact that isolation

of mechanical properties is extremely difficult.

2.10 Porosity

Porosity can be a major factor where resistance to corrosion is

the main requirement in that pores may provide a site for corrosion

of the substrate.

Large pores may be readily detected by visual inspection or

under low power magnification, but in general micropores tend to be


- 44 -

present which are more difficult to detect.

Porosity tests are designed to detect discontinuities in the

·coating, such as pores and cracks, after corrosion tests, which try

to reproduce in a short time the type of corrosion that might occur

over several months, have been carried out. Many tests are based.

on the corrosion of the substrate that occurs when the coated metal

is exposed to an aggressive environment. A test atmosphere of 10%

sulphur dioxide and 86% relative humidity can be used in this way

and after 24 hours pores are revealed as brown-black spots of corrosion

product.

A quicker test for steel substrates is that using a 3% sodium

chloride and 0.1% potassium ferricyanide solution. Pores are

revealed by blue spots on a solution soaked filter paper. Distilled

water at 90-9S oC will also produce rust spots in 2-l0h and these can

be evaluated using ASTM method D-6l0, using photographic reference

standards.

A widely used technique for porosity is the electrographic test.

The coating is pressed onto a semi-permeable printing medium (such as

absorbant paper) impregnated with electrolyte and made anodic so that

the metal substrate dissolves. Metal cations are transferred

through the coating discontinuities to the printing medium where they

react with a reagent marking the sites of the pores. Complex shapes

can often be tested by means of gelatine films impregnated with the

detector reagent (N.I. Gaynes, 1977).

Instruments working on the wet sponge principle are simple and

inexpensive as low voltage is supplied (M.J. Marshall et al., 1972).

A voltage is supplied between the substrate and a wet sponge. The


- 45 -

current is used to activate an audible alarm. An alternative method

to this is the Holiday detector which will deal with thicker systems

and applies a voltage between 0-20 kV, again giving an alarm when

a pore is found (V.K. Croutch, 1976).

The electron microscope can also be used to give an indication

of porosity, depending on the type of coating being observed. When

used in the transmission mode (i.e. using X-rays as opposed to

electrons) absorption will take place within· the coating due to the

high atomic number elements in the pigments. Pores all the way

through to substrate and also air pockets will be shown by a lack of

adsorption of the X-rays. A copper target produces the X~rays.

ASTM Method D 1653 describes the measurement of the rate of

evaporation of moisture through the coating using a permeability cup.

The coating is removed from the substrate and the cup exposes 25 cm 2

of film to high humidity on one side and low humidity on the other.

The loss of water is recorded at intervals and used to calculate the

rate of· movement through the film.

Summary

Brief descriptions have been given on the types of test methods

available for.assessing the quality of a film. There are very many

commercially available pieces of apparatus for carrying out· these

methods, all differing slightly, and are too numerous to mention.

However, the basic principles are described and in most cases the

standard method is referenced to either British or American Standards.

A list of the tests and related numbers applicable to powder coatings

is given in Table 2.1.

One of the most important requirements to consider is the


- 46 -

understanding and significance of the test and its results. It is

obvious that the testing and analysis of data obtained for coating

films is very complex. Each individual must determine which tests to

use and how to interpret the results in order t~ optimise the

performance of the particular properties of the product. In this

case changes in powder size distribution giving possible increases

in powder deposition must not be employed if physical properties

of the coating are subsequently forfeited. Therefore, a set of

applicable tests have been chosen for the needs of the experimental

investigations being undertaken.

The next chapter fully describes the tests employed and the

results obtained are presented and discussed. The equipment used for

the production of the test coatings is also described.


Table 2.1. Summary of Related British and American Standard Test Methods

British Standards B.S. 3900

Parts Test

02 Specular Gloss
El Bend Test
E2 Scratch Test
E4 Cupping Test
E5 Pendulum Oamping Test
E6 Cross Cut Test
E7 Resistance to Impact (Falling Ball)
E8 11 11 11 (Pendulum Impact)
E9 Bucholz Indentation
ElO Pull off Test for Adhesion
F2 Resistance to Humidity

American Standard Test Methods


ASTM No's Test
0523 Test for Specular Gloss
0610 Oegree of Rusting Evaluation
Bl17 Salt Spray (Fog) Test
0870 Water Immersion Test
0968 Test for Abrasion Resistance
01005 Measurement of Film Thickness
01186 Ory Film Thickness (Magnetic Type Gauge)
01474 Test for Indentation Hardness
02197 Test for Adhesion
03170 Test for Chip Resistance
03359 Measuring Adhesion by Tape Test
01653 Permeability Cup Method
CHAPTER 3

EXPERH!ENTAL EQUIPMENT AND RESULTS

Summary

3.1 Introduction

3.2 Spray Equipment

3.2.1 Static Gun

3.2.2 Problems Encountered

3.2.3 Commercial Spray Equipment

3.2.4 Experimental Method

3.2.5 Pretreatment of Sample Plates

3.3 Testing Methods and Equipment

3.4 Thickness Measurement

3.4.1 Experimental Method

3.4.2 Experimental Results

3.5 Gloss Test

3.5.1 Specular Gloss

3.5.2 Goniophotometric Curve

3.5.3 Instrumentation

3.5.4 Experimental Method

3.5.5 Results and Discussion

3.6 Surface Roughness

3.6.1 Instrumentation

3.6.2 Experimental Method

3.6.3 Results and Discussion

3.7 Adhesion

3.7.1 Experimental Method

3.7.2 Results and Discussion


Contents contd

3.8 Corrosion

3.8.1 Experimental Method

3.8.2 Results and Discussion

3.9 Porosity

3.9.1 Results and Discussion

3.10 Conclusions
- 47 -

3.1 Introduction

It is proposed that particle size and particle size distribution

might be an important factor in the quality of the final stoved films

of an electrostatically sprayed powder coating. In order to

investigate this proposition it is necessary to observe the effects

of these parameters on various qualities of the final stoved film.

This has been achieved by studying the change in film properties

of a set of coatings produced using different sized fractions of

an epoxy resin powder. By observing the product quality it was hoped

that more information on film formation might identify a better size

distribution for commercial use.

Although the deposition efficiency of powder of different

particle sizes has been investigated in various ways (A. Golovoy, 1973abc)

the effect of particle size sprayed on final film coating properties

has not been studied. Any advantage gained in deposition efficiency,

or in' producing thin films, by manipulation of particle size or

distribution must not be outwe'ighed by a deterioration in film

properties.

Development of a testing program was necessary to quantify any

changes in film properties. In the preceeding chapter an outline of

the types of test methods available for testing organic type coatings

was given. Using this information, a set of test methods were chosen

for assessment of films produced for this work. The test methods

employed were chosen to encompass a range of physical properties.

It was not the intention of these experiments to measure

deposition efficiency of the various classifications of powders used

but rather to assess their effects on the final film properties.


- 48 -

To accomplish this, apart from a test procedure, a reproducible

method of coating sample plates was required. Using this method the

effect of the spraying system was eliminated since coatings were

produced and stoved under identical conditions. The equipment and

method employed for coating will now be described.

3.2 Spray Equipment

In dealing with powders it is difficult to maintain reproducibility.

Slight variations in the way in which a system is operated can produce

a variety of different effects. It was therefore necessary to ensure

that the experimental methods employed in these experiments, whether

in the production of samples or actual testing, were reproducible.

Two approaches for the prOduction of sample plates were employed.

3.2.1 Static Gun

Initially a static gun was used for the coating of

sample panels. This consisted of a ~" diameter perspex tube which

had six needle electrodes inserted through its walls at the end of

the tube. The electrodes were connected to a Bradenburg High Tension

supply' unit. Powder to the tube was fed by means of a vibratory

feeder which poured powder into a funnel. The powder was then drawn

into the tube by compressed air passing through a venturi. This

compressed air acted as the transport medium for the powder from the

gun to the target (see Figure 3.1). A 2" square sample plate was

positioned into the back of a supported 10" square aluminium plate

and held by an electromagnet. The edges of the small plate were

covered by an overlapping brass shim which covered the outside i"


of each side of the smaller plate. The brass shim was used so as to

minimise edge effects (build up of powder near the edges) which give

variations in thickness of film. The target plates were earthed and


Figure 3.1. Static Spray Gun and Booth Design

Earth

6 II.T. needles

Vibrator
r.===============+=======~
10" sq Aluminium
Feeder
Sheet

,...-11....-,,.------1..,
.:
it
!\"
•• ' -_ _ _ _ _ _ _ _ _ _ _ _ ~~ _ __ L_ _ _ _ _¥
Brass
Shim
-- /
r-~~~""1'/
Electromagnet

----------~.~ Vacuum

Earth ___
Needle
.....>--_____ 30 cnl"'S'--____--<.~

II.T. Supply
D.C. Supply
Wooden Booth
Hand-held
Control
- 49 -

held approximately 30 cms away from the end of the spray gun. The

whole assembly was contained in a wooden booth (2' x 3' x 2') of

which the door operated a safety microswitch for the HT power supply.

To avoid time between switching on the HT voltage and supply

of powder to the gun, and hence the presence of excess ions, the

vibratory feeder and voltage were switched by two buttons on a hand

held controller. A weighed amount of sample powder was distributed

over the vibratory feeder and then the compressed air was supplied

to the gun. The feeder and HT voltage were operated by the hand held

controller and the sample coated.

The sample plate was carefully removed using the electromagnet

which was held by a retort stand. The current supplied to the electro-

magnet was switched off and the plate taken by hand to the stoving

oven. The coating was cured in a fan assisted electric oven at a

typical temperature of l80 0 C for 15 mins.

A set of coatings were produced by this method using size

fractions of a commercially used powder. These size fractions were

<lOum, l0-20 um, 20-30 um, 30-40 um, 40-50 um and +50um. The commercial

powder itself was also used. Each set of samples were produced in

the same manner and stoved under identical conditions.

3.2.2 Problems Encountered

The apparatus described was that achieved after problems

of powder feed, sample handling, HT control etc had been overcome.

It was found that this arrangement gave very reproducible coatings

for a given powder when operated under similar conditions. However,

problems persisted with this static gun when trying to control film

thickness.
- so -

When spraying different sized powders at identical air flow"ates

it was observed that the area covered by the spray gun varied to a

very large extent i.e. 'splaying-out'. Therefore, different amounts

of powder for each size fractionated powder were required to produce

films of the same thickness. By spraying a powder with a wide size

distribution it was evident that size distribution effects took place

across the 10" sample plate. Indeed, in an experiment where three

different coloured powders were used to represent three size ranges

«20, 20-40 &. ~Oum), rings of different colours were produced. The

largest sizes being in the centre and the fines at the outside.

These effects gave concern regarding the comparison of coated

samples of the different sized fractions. It also meant that the

production of a set of. coatings of similar thickness was both very

labourious and difficult. A new method was required in which a large

number of samples could be produced using a non-static gun for a

range of film thicknesses.

3.2.3 Commercial Spray Equipment

To combat the aforementioned problems a commercial type

spray booth was kindly donated by Aerostyle Ltd. This consisted of a

metal booth with both side and front opening hinged doors and a

powder extract section at the floor of the booth. A bag filter unit

was connected at the base of the booth to remove overspray. A spray

gun powder supply unit and HT unit was supplied by Tam-Sames et Cie.

The powder supply unit consisted of a fluidised bed hopper with an

air jet ejection system used to supply an air-powder mixture to the

spray gun. A third air control was available to control a supply

of air to the gun nozzle which varied the spray cloud pattern.
- 51 -

Diagrams of the powder feed hopper and spray gun nozzle are shown

in Figures 3.2 and 3.3. Plate 3.1 shows the spray booth and spraying

equipment. Plate 3.2 gives a close up view of the spray gun, powder

supply unit and high tension voltage supply unit used.

3.2.4 EA~erimental Method

In order to produce a set of coatings for testing, the

following method was employed. Pi. 10" x 8" steel plate was covered

with twenty 2" x 2" thin steel plates (1/16" thick) and then placed

on a U shaped electromagnet. After switching on the DC supply to

the electromagnet the assembly was suspended from a metal bar inside

the spray booth, with the plates facing towards the opened door on the

long side (see plate 3.1). The HT unit was set to supply 60 kV when

the spray gun trigger was depressed. The air flow rate to the

fluidised bed was increased to fluidise the 2/3 rds full hopper of

epoxy resin powder (commercial type or size fractioned). The air to

the jet for powder supply and to the gun nozzle were preset to

suitable positions. The spray gun trigger being depressed, powder

was sprayed towards the target for a suitable length of time to

produce a relatively uniform coating on the plates. The spray system

was immobilised and the coated plates, still held by the electro-

magnet, carefully removed and placed in a horizontal position. The

electromagnet was then switched off and the base plate together with

the coated sample plates were placed in a stoving oven. The coatings
o
were stoved for 15 minutes at l80 C, as per the powder manufacturers

recommendations.
Figure 3.2. Spray Gun Powder Feed Honper

Powder +

U" "'""".1 Secondary Alr


to Spray Gun

Air for Venturi



Air Cont::-ol
va7s
\ I

/
,
Fluidised
Bed
\ ~ ~ ~

Distributor
Plate
\ G~
.
....,

Fluidising
• Alr

Figure 3.3. Spray Gun Head

H.T. Electrode
Secondary
Air

H. T. Cabl e

to
-Trigger
--~~~--~-------------
_________________ Powder + Air Mixture
~

Deflector
Plate 3.1 . Electrostatic Spray Booth and Equipment
Plate 3.2 . Spray gun, H. T. Supply Unit and Powder Feed Unit
- 53 -

This process was repeated for a given sample of powder until a

sufficient number of coatings were obtained for further experiments

and tests. The only change that was made was in the type of powder

used.

A standard commercial epoxy resin powder as supplied by IeI

Paints Division, Slough was used as the basis for the experiments.

The powder manufacturers confirmed that this powder was representative

of the type of epoxy powders currently used in the coating industry.

The pigmentation waS white and the powder had a size distribution as

shown in Figure 3.4. Electron micrographs of this powder are shown

in Plates 3.3·and 3.4. It can be seen that the particles are very

irregular in shape. The powde·r was classified into 6 size ranges

using an Alpine Zig-Zag air classifier. The ranges were <lO~m, 10-20~m,

2o-30~m, 30-40~m, 40-50~m and >50~m. The unclassified commercial

powder together with the 6 classified fractions were used to produce

sample coatings.

Since it was not possible to control the thickness of the final

coatings a procedure was adopted in which the thickness sprayed was

controlled by eye. The plates were subsequently measured after

stoving by the method described later. With experience a required

thickness range could be obtained but final classification was made

only after measurement.

On completion of the production of coatings a collection of

over 2500 sample plates were obtained for the seven different powder

grades and these were then selected into groups of film thickness

ranges. Poor quality coatings (i. e.· those that had been disturbed

in some manner) were not included and eventually approximately 10 of

each thickness range were kept for further testing.


Figure 3.4. Size Distribution of'White Epoxy Powder

Coulter Counter Analysis

100

..
.....,
N

..... 80
6
....
...
.c00
.
.... 60

.
;3

....
I>
...
....'" 40
u
~

20

oJ------r-----.----_.r-----r-----r------r----~-----.r-----r-----'1
o 10 20 30 40 50 60 70 80 90 100
Particle Size (\lID)
Plate 3.3. SEM Photograph of Epoxy Resin Powder (x 1000)

Plate 3 . 4. SEM Photograph of Epoxy Resin Powder (x 3000)


- 54 -

This spray gun allowed more control of the production of

coatings. It was not static and therefore problems of size segregation

across the plate were not encountered. This system allowed for faster

production of samples and by coating and stoving twenty at one time

the risk of disturbing the powder layers by movement was minimised.

Edge effects were also reduced since each plate was placed flush

against at least two others.

Although movement of the gun and time of spraying was arbitrary,

each plate was individually measured for thickness and hence

classified into groups.

3.2.5 Pretreatment of Sample Plates

Strips of greased mild steel plate, 2 feet long by 2"

wide and 1/16" thick were employed. These strips were guillotined

into the 2" square sample plate sizes. It was checked that this

method of cutting did not cause any appreciable distortion of the

edges of the plates to ensure that they would lie flat on -the large

sample holding plate. A quantity of plates were cut prior to coating.

The plates were degreased using carbon tetrachloride and dried using

tissues. Each plate was dedusted, to remove any debris from the

tissues, using a compressed air supply. The plates were then stored

in a desiccator ready for coating.

Although general pretreatment methods have been described

earlier, it was decided that in this case pretreatment was not

required. This decision was based on a number of reasons. Firstly,

the mild steel plating was particularly chosen for its quality and

that it had been greased to protect it from any possibility of

corrosion. In these coating trials the samples were not left for any
- 55 -

significant time in either an ungreased state or in a corrosive

atmosphere. Secondly, various methods of pretreatment including

phosphating and zinc coating were tried and compared using the salt

spray test (described later) but very little difference was found

using these materials.

3.3 Testing Methods and Equipment

As previously described, the range of tests available to test

a particular coating on a substrate is very wide and can be applied

to all types of materials and substrates. It is important to find

a series of tests that give quantitative results for the properties

of a coating.

After carefulronsideration of the properties that should be

investigated, to indicate the effects of particle size on the final film

properties of the coating, the following tests have been employed:-

1) Film Thickness

2) Gloss Test including (a) Specular Gloss

(b) Gloss Factor

(c) Scatter

3) Roughness including (a) Average Roughness

(b) Roughness Wavelength

4) Adhesion'

5) Corrosion Resistance

6) Porosity

The problems encountered in determining the significance of

the results from these types of tests are enhanced by the fact that

commercially used powders are being tested, and as such are known to

give generally satisfactory results.


- 56 -

Therefore it was decided to consider the adhesion and corrosion

tests in preference to all the other mechanical tests that could be

employed (i.e. impact, scratch, chipping etc). These tests were used

as 'pass/fail' tests and supplied valuable information as to whether

the stoving cycle had been effectively carried out.

The test method and equipment used in each case will now be

described and results will be presented and discussed with each test

in turn.

3.4 Thickness Measurement

The measurement of thickness is an important factor in this

work as it enables any effect on a physical property to be observed

for both changes in type of powder sprayed and film thickness. This

will give guidelines for the optimal film thickness with regards

economy and retaining advantages of physical characteristics.

The 'Elcometer Minitector' was employed for the measurement of

film thickness (see Plate 3.5).

A probe is placed on the coated plates and the thickness of the

organic coating measured directly from the scale on the meter. This

device can measure coatings on either non-ferric or ferric substrates

by the use of eddy currents. A varying magnetic field induces eddy

currents in the conducting substrate material. The change in the

apparent impedance of the probe coil that induces the eddy currents

is monitored. Changes in the current flow in the coil, produced by

various spacings of it from the metal substrate, are an indication of

the thickness of the coating.

3.4.1 Experimental Method

The correct setting on the gauge is selected for the

thickness of film expected and the type of metal base being used.

The probe is placed on a piece of the uncoated substrate and the


Plate 3 . 5 . Elcometer Thickness Measuring Instrument
- 57 -

device zeroed. The device was calibrated by placing a set of film

standards between the probe and the metal plate. The standards

covered the thickness range of interest and the calibration knob

was used to obtain the correct setting.

After every plate had been measured the various size fractions

were sorted onto plates in different thickness ranges, primarily

in 5~m ranges. The best ten sample plates in each film thickness

range were used for further experiments. This assessment was based

on visual observation for evenness of film etc. For the seven

different types of powder samples sprayed a collection of approximately

1000 plates out of the originally measured 2500 were kept.

3.4.,2 Experimental Results

The five measurements for each film were 'taken

approximately I" in from each corner and also in the centre of the

coating. The average thickness for the coating was calculated and

recorded against the designated sample number. Table 3.1 shows a

summary of all the thickness ranges that have been used, the

number of plates that'were retained in each range, and the average

thickness of the plates in each group. This is done for all seven

powders used in the experiments. This provides a comprehensive

set of samples for the testing of other physical properties.

3.5 Gloss Test

3.5.1 Specu1ar Gloss

Initially the gloss of the coatings was to be assessed

by measurement of the specular gloss only. A Brice Phoenix Light-

scattering photometer would be used to shine a parallel beam' of

light at a fixed angle of incidence onto the substrate. The reflected

light, at an angle of reflection equal to the angle of incidence, is


Table 3.1. Distribution of Sample Plates in Film Thickness Groups

Commercial <10\lm 10-20l'm 20-30\lm 30-401'm 40-5011m +501'm

No. No. No.


Range Range Range . Range No.
. Range R No. Range No. No.
in Ave. 1n Ave. 1n Ave. in Ave. ange .
1n Ave. in Ave.
Range
in Ave.
No. \lm I'm I'm I'm \lm I'm I'm
Range Range Range Range Range Range Range

1 20-24 10 22 10-19 10 16 10-19 10 18 20-29 10 27 20-29 3 29 30-34 9 33 30-39 5 38


2 25-29 10 27 20-24 10 22 20-24 10 22 30-34 10 32 30-34 10 33 35-39 10 37 40-44 10 43
3 30-34 10 33 25-29 10 28 25-29 10 27 35-39 10 37 34-39 10 37 40-44 10 42 45-49 10 47
4 35-39 10 37 30-34 10 32 30-34 10 32 40-44 10 42 40-44 10 42 45-49 10 47 SO-54 8 52
5 40-44 10 42 34-39 10 37 35-39 10 37 45-49 10 48 45-49 10 48 SO-54 10 52 SS-51 10 57
6 45-49 10 47 40-49 9 44 40-44 10 42 SO-54 10 52 SO-54 10 52 SS-59 10 57 60-64 10 62
7 SO-54 10 52 50-59 4 56 45-49 6 46 55-59 10 57 55-59 10 57 60-69 10 63 65-69 10 67
8 55-59 7 56 60-64 5 62 50-59 6 56 60-64 9 62 60-64 6 62 70-79 3 75 70-74 9 71
9 60-64 10 62 65-69 7 67 60-69 10 62 65-69 10 62 65-69 10 67 80-89 7 84 75-79 7 76
10 65-69 10 67 70-79 7 74 70-74 10 71 70-74 7 72 90-94 5 92 80-84 6 81
11 70-79 7 74 80-89 8 84 75-79 8 77 75-79 5 77 95-99 9 97 85-89 7 87
12 80-89 6 85 90+ 9 95 80-90 1(:; 84 80-84 7 82 lOO-nO 5 106 90-94 10 92
13 90+ 6 100 90+ 5 100 85-90 9 87 111+ 10 135 95-99 5 97
14 90-99 8 93 loo-no 9 104
15 100+ 4 102 +110 6 122
- 58 -

then measured by means of a photocell and measured as millivolts.

This method proved to be very inaccurate because alignmenc

of collimator, photocell and sample plate on the staging proved to

be of great importance and difficult. A slight deviation in the

comparison of angles of incidence and reflection gave very large

errors in the specular gloss measurement.

3.5.2 Goniophotometric Curve

If the angle of reflection (position of photocell) is

changed and the reflected light measured for various positions, a

graph can be plotted of intensity of reflected light (mV) against

angle of reflection, for a given angle of incidence. This is

called a goniophotometric curve. From this curve (shown in Figure 3.5)

a value of the specular gloss can be calculated as the height of the

peak at angle of reflection equal to angle of incidence. In this

case a small offset in the setting up of the sample plate can be seen

as a movement of the peak from this angle and can hence be compensated

for. However, this curve also describes the way in which the incident

light is scattered at various angles and from the measurements of

specular gloss and scatter a quantity called the 'gloss-factor' can

be obtained (as defined in the figure).

The plotting of these curves for each sample (measuring

reflectance at various angles) is very laborious and so the photocell

mounting stage was motorised and by use of other electrical equipment

the goniophotometric curve was plotted directly onto a chart recorder.


• IW •
h

I
I
I
............ I
'-.l :--_-\-____
I',
I "
I "
I
I
I
I
I
I
I
. 00 o

Angle of. Reflection C>

h
Gloss factor, GF = -
w

where
h = peak height
w = width of peak at half peak height

Figure 3.5. Goniophotometric Curve Showing Gloss Factor Calculation


- 59 -

3.5.3 Instrumentation

The Brice Phoenix light scattering photometer is generally

used for determining particle size distributions, optical properties

of liquids etc by measuring turbidity and scatter of light through

a light scattering cell. For the goniophotometric studies the

photometer was modified in several ways.

The optical features of the system are shown in Figure 3.6.

(P.P.I., 1963). The light beam from the lamp (L) passes through a

monochromatic filter (Fl) and into the shutter collimating tube (Cl).

The lamp is a high pressure mercury lamp, 85 watts. The intensity

of the light beam is controlled by the shutter (S) and a set of

neutral filters (F2). After passing through an achromatic lens (Ll)

and a pIano-cylindrical lens (L2) light is emitted from the primary

collimating"tube (C2) out of the diaphragm (D2). The light impinges

on the sample plate which is held on a modified mounting table (G).

The light is then transmitted, reflected or scattered, and is

measured by the photocell (PT). The specimen table is calibrated

(Figure 3.7) so that the mounted plate can be rotated through 360 0 •

It is secured by alIen screwS at the position required. The holder

ensures that the front surface of the coating is in the same relative

position to the beam and photocell each time, regardless of the coating

thickness.

The photocell can be rotated by 270 0 by the electric motor

which is connected to the graduated disc assembly (D) by means of

a rubber wheel. This motor turns the photocell at a constant

angular speed so that the time of traverse is proportional to the


Figure 3.6. Optical Features of the Brice Phoenix Light Scattering Photometer

Fl
r , '"
. - -.......
"""
\

---Tt-="'
II . r=:V';'L2/
L "
G
\

@
•.,
If' ! / @-- ~ I
,
I ;.. "- 2
- ~ - _I-_~ CD:1. .l • AN
"
--. 2 [G~T ;'
""
L Cl ". , /

::;.:9'
D--~

Figure 3.7. Modified Mounting Table

~ Sample Plate

Sample Plate ----... - .. ...........


Fixing Screw ,

Angle Adjustment
Screw
-
- 60 -

angle of reflection. The photocell can be traversed in either

direction and has a removable analyser and variable slit aperture

(AN and 04 respectively.). All parts inside the scattering compartment

were painted matt black, including the mounting table, so that no

stray light reflection was considered.

The output from the photocell is fed to a chart recorder so

that the goniophotometric curve, for a given angle of incidence, can

.be directly plotted. The format of this curve has been previously

shown. It can be seen that the curve occurs after a threshold level.

In order to increase the accuracy of this method a low power supply

unit (PSU), used to provide a back e.m.f., and voltmeter are

incorporated; the reasons for this are discussed in the experimental

method. The complete circuit diagram is given in Figure 3.8.

Plate 3.6 shows all the equipment used.

By comparing samples of different finishes i.e. high, medium

and low (matt) gloss, the changes in the shape of the goniophotometric

curv~s can be clearly seen (Figure 3.9). For samples of high gloss

the curve is tall and narrow and for low gloss it is short and wide.

From these curves the gloss factor, which takes into account both

these properties of the curve, can be obtained. The gloss factor

has been defined in Figure 3.5.

3.5.4 Experimental Method

The mercury lamp was given at least half an hour to warm

up so that the emitted light intensity remained constant. The photo-

cell is traversed so that it is directly opposite the beam of light

(exact position is when the output from the photocell is a maximum).

The output is registered on a O.V.M. and the intensity of light is


1
Motor Driven
Goniophotome ter

Power
Digital Supply Unit
Voltmeter

Back E.M.F.

Chart
Recorder
Digital
Voltmeter

Recorded Output

Figure 3.8. Gloss Test Electrical Equipment Arrangement


INTENSITY (mV)

900
- High Gloss (mirror)

LOIN Gloss (matt)


800

700

\
Medium Gloss

600

500
35 40 45 50 55
Angle of Reflection (Deg.)

Figure 3.9. GonioFhotometric Curves for High, Medium


and LOIN Gloss Specimens
.."..,.....,. - - --
,. .,.
- UNIVERSAL LIGHT SCAnERING PHOTOMffiR

••

Plate 3.6. Brice Phoenix Light Scattering Photometer and Related Equipment used for Gloss Measurement
61 -

adjusted by means of the coarse, medium or fine adjustment knobs until

the output is 1.000 Volts.

The sample is then placed in the mounting table at the required

angle of incidence to the light beam and the photocell is traversed

from 0-90 0 • The time taken for the photocell to do this was measured

ten times and compared to the paper feed rate on the chart recorder.

The x axis was calibrated in terms of angle of reflection, degrees/cm.

The y axis was calibrated in mV/cm by comparison with the D.V.M.

Figure 3.10(a) shows a recorder output. By using a back

e.m.f. a curve of the type shown in Figure 3.10(b) is obtained and

hence-gives more accurate results. A back e.m.f. of 0.650V was

typically used.

Five sample plates from each thickness range of each powder

type were tested using this technique. The plates were positioned

in the mounting table and placed in the machine. The photocell was

traversed through the appropriate angles of reflection from the plate

and the goniophotometric curve obtained on the chart recorder. Each

curve was analysed and the height of the peak (above the threshold

value) and the peak width at half peak height were recorded in terms

of centimetres from the recorder paper. These figures were then

converted to mV and degrees and the gloss factor was calculated for

each curve by ratioing the two measured parameters as described

earlier. By initially setting the response of the photocell (to

the incident light on the plate) as an output of 1.00V the measured

peak height represents the percentage change in intensity of light

reflected from the plate.


~

:e
~

...
.= 800
....
,.J
00

...uQJ

.... QJ 600
""'
c::
QJ

""'0
.......co
>. 400
c:
...c:
QJ

....
200

Figure 3.l0(a) and (b). Typical chart recorder outputs from goniophotometer
with (a) no back e.m.f. (above) and (b) with back e.m.f. of
650 mV

750

:>
~
=
...
.=
....co
,.J
~

...
U
QJ

....
QJ

""'c:: 700
'"
......
0

.......co
>.

c:
...'c:"
....

650

45 0
Angle of Reflection
- 62 -

3.5.5. Results and Discussion

The five measurements for.each thickness range were

averaged and the standard deviations and probable errors calculated.

These results are presented in Tables Al to A7 of appendix A. Each

table represents results of the three measured and calculated

parameters for each different powder type used. The film thicknesses

quoted are the averages for the film thickness range that the samples

used come from. The results are plotted in terms of film thickness

versus gloss measurement quantity in Figures Al to A2l.

In general it can be seen from the results tables that the

standard deviations of each batch of five measurements are less than

10% and this gives an indication of the reproducibility of both the

test and the coating method.

The graphs indicate tr.ends in terms of the gloss measurement

with increasing film thickness for either the commercial powder or

one of its size fractions. There is ·generally a large spread of

results from any possible trend in most of the graphs and so for

comparison purposes a linear regression fit has been calculated for

each set of data. This is used to approximately represent the trends

in the results but is not meant as a correlation. The data used

for these fits are given in Table AB for each of the powder types

and measured parameters.

The results of this analysis have been plotted for each of the

powder types in Figures 3.11, 3.12 and 3.13 for changes in thickness

with gloss factor, peak height and peak width respectively. From

these graphs it is possible to compare results of the various powders


Figure 3.11. Gloss Factor vs. Film Thickness for Different Powders S rayed

Linear Regression Fits

35

31
20-30~m
1O- 2Ol1m
30-40~m
+50~m
27
H 40-50~m
....u0
<U

Ul
Ul
23 Conunercia 1
....0
t!l

19
<IOpm

15~----'-----'-~--'-----'-----~
o 10 20
____~____r-____r -____r -__- .
30 40 50 60 70 80
Film Thickness (I1 m) 90 100
Figure 3.12. Peak Height vs. Film Thickness for Different Powders Sprayed

Linear Regression Fits

110
10-20~m~~~~__________________________________________________________~~~~~:::

106
+50~m 30-40~m

~ 40-50~m 20-30~m
~
~

...
.c
102
Conunercial
....
00

~
.:.:<U
Q) 98
~

94
<lO~m

90~------'-------r------'-------r------'-------r------'-------r------'------'
o 10 20 30 1,0 50 60 70 80 90 100
Fi Im Thickness (11m)
Figure 3.13. Peak Width vs. Film Thickness for Different Powders Sprayed'

Linear Regression Fits

5.5

<lOl'm
5.0
~
.
00
QJ
A
~

...
.a
-0
4.5
Commercial
.....
:.:
~
~
QJ 4.0 40- 5011m
30-4011tn
+5 0 11m
3.5 1O- 20 11ffi
20-3011m

3.0

2.5~------~----~r-----~-------r------'-----~r-----~-------r------'------'
o 10 20 30 50 60 70 80 90 100
Film Thickness (1Im)
- 63 -

for different thicknesses of film. By doing so the effect of particle

size sprayed as a function of the gloss parameters for any given

thickness can be observed. Figures 3.14 to 3.16 show these effects

for film thicknesses of 30, SO, 70 and 100~m and the data used for

these plots is given in Table 3.2. From these graphs it is very

obvious that the results for the <lO~m range powder are substantially

poorer than the other powders. The reason for this is probably

due to the fact that particles <lO~m in size tend to be very cohesive

in nature and as such form larger lumps when placed together in the

fluidised bed powder hopper. The particles remain clumped together

when the powder is sprayed, so that after stoving an uneven finish is

obtained.

Figures 3.17 to 3.19 show the best fit lines to the data of

Figures 3.14 to 3.16, ignoring the <lO~m range results for the linear

regression calculations. Table A9 contains the data used, for these

fits.

The raw data obtained for each powder type for varying film

thicknesses show no marked trends in the figures given in Appendix A.

In general the height of peak results, which give indications of

the specular gloss of the surface, tend to increase as film thicknesses

increase. As might be expected the width of the peaks at half peak

height decrease. This measurement gives a measure of the amount of

scatter of light from the surface. The wider the scatter the more hazy

an image would look if the coating was used as a mirror. When an

actual mirror is used in this test the goniophotometric curve is very

tall and narrow depicting that a clear, sharp image would be seen.
Table 3.2. Gloss Test Results for Different Sized
Powders at Various Film Thicknesses

Thickness of Coating (\lm)

Test
Powder Type 30 50 70 100
Parameter

Gloss Factor 24.7 27.0 29.4 32.9

Commercial Peak Height 102.4 104.0 105.7 10B.2

Width 4.20 3.90 3.62 3.lB

Gloss Factor 19.6 21.3 22.7 25.5

10\lm Peak Height 93.5 94.1 94.4 95.5

Width 4.90 4.56 4.12 3.72

Gloss Factor 30.2 31.0 31.9 33.2

1D-20\lm Peak Height 10B.7 10B.5 10B.0 107.4

Width 3.61 3.49 3.37 3.29

Gloss Factor 30.1 30.2 30.3 30.4

20-30\lm Peak Height 106.9 106.0 105.3 104.1

Width 3.54 3.54 3.54 3.54

Gloss Factor 2B.4 2B.1 27.8 27.4

30-40\lm Peak Height 107.0 106.5 106.0 105.2

Width 3.78 3.80 3.82 3.86

Gloss Factor 27.1 27.5 27.8 28.4

40-50\lm Peak Height 104.2 104.7 105.2 105.9

Width 3.8B 3.85 3.81 3.75

Gloss Factor 28.4 28.9 29.3 29.9

+50\lm Peak Height 105.2 105.3 105.4 105.9

Width 3.70 3.66 3.63 3.55


Figure 3.14. Gloss Factor vs. Size of Powder Sprayed

With Various Thicknesses of Coatings

40
.-
0-
Key
30jlm Film
50jlm Film

• - 70jlm Film
0 - 100jlm Fi lm
35
0

'0"' 30
Q
~
'IIIu"'
~

III
III 25
0
.....0
<.!l

0
20

15

lO;----------r---------r---------.---------.---------,----------r-------~
o 10 20 30 40 50 60 7
Average Particle Size (jlm)
Figure 3.15. Peak Height vs. Size of Powder Sprayed

at Various Film Thicknesses

115
.-.
0-
Key

-
30llm Film
50llm Fi lm
70llm Film
110
~
0 _. 100llm Film
~
~0 ~ i
~

...
.c00 105
.....
'"
p::

7J 100
~'"
0
95
;
90

854---------~--------r_------_,r_------_,--------_,--------_.--------_.
o 20 30 40 50 60 70
Average Particle Size (Ilm)
Figure 3.16. Width at Half Peak Height vs. Size of Powder

at Various Coating Thicknesses

5.5 .-
0-
Key
30)lm Film
50)lm Film
~

• - 70)lm Film


00
QJ 5.0
A
'-'
0 - 100)lm Fi Im
...
.c00
..... 0
::c 4.5
QJ

.>:
'"
QJ
Po<

....
'H

::c'"
4.0 •
...
...'"
.c
'1j
.....
:;:
3.5
0
•~
3.0

2.5~------~-------'--------r-------r-------'--------r-------'
o 2 30 40 50 60 70
Average Particle Size ()lm)
Figure 3.17. Gloss Factor vs. Particle Size Sprayed

Various Film Thicknesses

35

33

100llm
..."u
0 31 70llm
50 llm
<U
~ 30llm
III
III

.....0
c.!> 29

27

25~-----------'------------r-----------~--~------~-----------T----------
o 10 20 30 40 50
__
6
Particle Size (Ilm)
Figure 3.18. Gloss Test Peak Height vs. Particle Size Sprayed

Various Film Thicknesses

110

30llm
108
SO\.lm
70 llm
~

~ 106
~

.
.c
100\.lm

.....00
Gl
;t:

-a 104
...
Gl

102

100
1 2 3 4 5 60
Particle Size (Ilm)
Figure 3.19. Gloss Test Peak Width vs. Particle Size Sprayed

Various Film Thickness


4.0

3.8

01)
.
QJ
Q
~
3.6 30\lm
.c 50\lm
....
70\lm
'"
.....
;3: 100jJm
.,.
tU 3.4
QJ
p..

3.2

3.0~----------~~r-------------;-------~----~--------------;-------------~
10 20 30 40 50 60
Particle Size (\lm)
- 64 -

On comparing the results for each powder type it can be seen

that there is little difference between the powders. When it is

considered that a poor, matt finish (poor with respect to gloss)

would have a gloss factor of approximately 3-4 and a mirror finish

would be between 30 and 40, these results represent fairly good

finishes with little difference between them.

The <lO~m size range powder, however, has generally poorer

gloss than the others for the reasons previously mentioned. This

powder would not be recommended for use in commercial systems as

problems of powder flow and dispersion would be encountered.

In Figures 3.14 to 3.16 the change in gloss parameters are shown

for changes in particle size of the sprayed powder. Ignoring the

<lO~m range, the results are not significantly different, although

there does appear a slight decrease in gloss with increase in size.

Overall the best powder range, based on this data, for the highest

gloss would be the 10-20~m range.

However, when the significance of these results are taken into

consideration it can be concluded that these size ranges do not have

any adverse effects on the gloss characteristics of the coating

compared to the commercial powder. Also, the gloss parameters are

not unduly effected by changes in thickness of film.

The visual appearance of a coating is extremely important since

the eye can be very precise in comparing qualitatively a set of

coatings. The visual appearance of the films agreed with the gloss

measurements in that they were all of high quality and none were

significantly different.
- 65 -

3.6 Surface Roughness

The Talysurf 10 is one of the most widely used devices for the

measurement of surface roughness parameters. The assessment of

surface roughness by stylus methods has been in use since the early

1940's and they are used in many national standards. This particular

method has the advantages of being non-destructive, sensitive and

gives quickly readable results (R.e. Spragg and D.J. Whitehouse, 1972).

A surface which is nominally smooth and flat will always

exhibit some roughness and it may also exhibit some waviness or a

combination of both. This is shown in Figure 3.20.

Roughness and waviness are measured by traversing a pick-up

which carries a sharply pointed stylus across the surface. Vertical

movements of the stylus with reference to a datum are measured. The

quantitative assessment of the surface profile is very difficult since

no single number can adequately describe the complex patterns and

irregularities of the surface. The Talysurf 10 gives two values that

relate to roughness and waviness in order to try and describe the

surface profile under observation.

The Roughness Average (Ra) has been internationally accepted

for many years now. It is defined as the arithmetical average of the

departures of the profile above and below the reference line (centre

mean line) throughout the sampling length. Definitions of the

centre line and Roughness Average are given in Figures 3.21 and 3.22.

The roughness average quoted is usually an average of a number of

consecutive measurements taken along the surface.

The roughness average gives no information about the openness

or closeness of the profile. The Average Wavelength parameter (Aa),

however, is a more recent measurement that is based on the spacing

of irregularities along the surface.


Waviness Spacing
Roughness
I I ·

Waviness
Roughness
Spacing

Figure 3.20. Surface Characteristics of a Coating


• L
Areas B + D + F + H+ J = Areas A + C + E + G+ I + K

Figure 3.21. Centre Line Definition

• L

1
Ra = = L

Figure 3.22. Definition of Ra


- 66 -

By Fourier analysis a series of sine waves can be produced to

represent the complex profile waveform.- If the power spectrum is

plotted (individual intensities for each component in the series),

the average wavelength is approximately equal to the weighted mean.

This measurement is not an actual measure of the spacing of the peaks

or of a theoretical waveform because it is derived from the power

spectrum. A comparison of waveforms and their average wavelengths

are shown in Figure 3.23, together with their power spectrums.

By quoting both the roughness average and wavelength parameters

it is possible to obtain a much clearer impression of the type of

surface that is being tested. Figure 3.24 shows a typical range of

results as obtained for the surfaces of rolled pieces of sheet for

the motor car industry. It can be seen that quoting roughness

wavelength is very important when roughness average results are

similar. (R.e. Spragg, D.J. Whitehouse, 1972).

3.6.1 Instrumentation

Vertical movements of the stylus that is traversed across

the workpiece are converted by a transducer into corresponding changes

in an electric current.

These are then amplified electronically to give a surface

profile which is obtained as an average reading, or readings, on a

meter. A schematic representation of the apparatus is shown in

Figure 3.25 and the equipment is shown in Plate 3.7.

3.6.2 Experimental Method

The Talysurf 10 was set up in the skidless mode since

the type of irregularities to be encountered were not known. The

traverse shaft thus acts as the datum and the pick-up assembly is

monitored for movement.


Figure 3.23. Comparison of Roughness Wavelengths
P A)

-
O .lmm
Average
--i io- Wave length
P (A)
Aa

-
O .lmm
--t 10-- Average wavelength
peA) Aa

20 m!
fo-- Average .wavelength l..:~~===:::;.__

Figure 3.24. Typical Values of Ra and Aa

No Classif ica tion


Open Irregular

1 2.2 211.

5).1mL
.lmm

Open Smooth

2 6.9 320

Sharp Peak - Sharp Valley

3 0.64 75

2).1mL
Imm
Open Smooth

4 2.1 361
--- ------------------.,
Mag.
Switch

Ra
meter
Amplifier Filter
/

Pen Shift
Recorder

Meter
Cut Off I
Switch I
L _____ 1 _ _ _ _ _ _ _ _ _ _
I

-
_
...J

P:~k-UP I
Traverse Unit

...... ..~- ._., . ' ..


;' . ,
• '.I',
"
:!.j• \
.. ,j
4"'" ';'- •• '

Figure 3.25. Schematic Arrangement of Talysurf 10


Plate 3. 7. Taylor-Hobson Talysurf 10 Roughness Measurement Device
- 67 -

A supplied test piece was used to check the calibration of the

instrument before each new set of readings. The system is carefully

levelle'd by using the levelling knob in conjunction with the chart

recorder as described in the manufacturers instructions. Four test

lengths were taken for each sample plate, each time the sample being

rotated by 90 0 to measure the roughness in different directions.

The appropriate magnification/sensitivity was used to cover variations

in roughness.

3.6.3 Results and Discussion

The averages and standard deviations were calculated for each

thickness range from the measurements carried out. In total

approximately 3800 measurements were recorded. The overall results

for each powder type are given in Appendix B, Tables Bl to B7.

Plots of these results against film thickness are given in

Figures Bl to B7for'roughness average and in Figures B8 to B14 for

roughness wavelength. Similarly to the results for gloss tests,

the data in these plots have been fitted to straight lines using

linear regression. These fits have been compounded together for

each powder type for roughness average and roughness wavelength onto

Figures 3.26 and 3.27 and the data for the linear regression fits

is given in Table B8.

Data was extracted from these results to observe the change in

roughness with particle size of powder sprayed and the general trend

observed was that roughness increased with increase in particle size

(see Table 3.3 and Figures 3.28 to 3.31). The <lO~m size range was

not used in the linear regression fits of the last two graphs due

to the poor qualities of film. (See Figures 3.30 and 3.31 and the

data given in Table B9).


Figure 3.26. Roughness Wavelength vs. Film Thickness for Various Powders

Linear Regression Fits

0.9

0.8
~

~
~ 0.7 lO\lm
...00
.c
<=
..... 0.6 40-50\lm
Q)

'"
>OJ
Commercial
;3 0.5 +50\lm
III
III
Q)
<= 0.4
.c00
"
~ 30r40\lm
0.3
10-20\lm

0.2 20-30\lm

0.1

0
0 10 20 30 40 50 60 70 80 90 100
Film Thickness (\lm)
Figure 3.27. Roughness Average vs. Film Thickness for Various Powders

Linear Regression Fits

1.6
'.1:..: ';; ~ ,; .. :. ~ I: , ~ '" ::.
~

13
;:l.
~

QI
01) <lOllm
cu 1.2
'"'
Qlo

~
Ul Commercial
Ul
QI 40-501lm
,; 0.8
.c01) +501lm
::0
0
P<::"
30,..401lm
0.4 1O-201lm

20-301lm

0
0 10 20 30 40 50 60 70 80 90 100
Film thickness (Ilm)
Figure 3.29. Roughness Wavelength vs. Size of Powder Sprayed

at Various Coating Thicknesses

.9
.-.
0-
-
Key
30)lm Film
50)lm Film
70)lm Film
.8 0- 100)lm Film
~

~
m
.
.c
00
p .6

•q
Q)
.-<
Q)
;.
C1I
;3:
III
III
Q)
p •4 0
• ~
.c

~
00

" •
0
0
0

.2 •
0 •

0
0
0 10 20 30 40 50 60 70
Average Particle Size ()lm)
Figure 3.30. Roughness Average vs. Particle Size Sprayed

Various Film Thicknesses

1.0

a 0.8 30\lm
;:I.
~

<Il 50\lm
00
'"
~
<Il
0.6
~
III 70\lm
III
<Il
c::
..c: 100\lm
00
::>
0
p::
0.4

0.2

o
10 20 30 40 SO 60
Particle Size (\lm)
Figure 3.31. Roughness Wavelength vs. Particle Size Sprayed

.Various Film Thicknesses

0.6

,~ O.S 30\lm
'-'
..c:.... SO\lm
on 70\lm
.::QI
...
. QJ
0.4
:> 100\lm
~
U)
U)
QJ 0.3
.::
..c:on
0
lXi'"
0.2

O. 1

o
10 20 30 40 o o
Particle Size (\lm)
Table 3.3. Surface Roughness Results for Different Sized
Powders at Various Film Thicknesses

Film Thickness (Jlm)

Powder Type Test Parameter 30 50 70 100

Roughness
0.81 0.66 0.52 0.30
Average
Commercial
Roughness 0.48 0.42 0.35 0.26
Wavelength
Roughness 1.18 1.12 1.06 0.97
Average
<lOJlm
Roughness
0.73 0.74 0.75 0.77
Wavelength
Roughness 0.36 0.31 0.26 0.18
Average
lo-20Jlm
Roughness 0.23 0.18 0.13 0.07
Wavelength
Roughness 0.32 0.36 0.48
0 .• 41
Average
20-30Jlm
Roughness 0.23 0.28 0.33 0.41
Wavelength
Roughness 0.45
0.51 0.39 0.31
Average
30-40Jlm Roughness 0.29 0.27 0.26 0.23
Wavelength
Roughness 0.77 0.66 0.35 0.39
Average
40-S0Jlm
Roughness 0.53 0.46 0.40 0.31
Wavelength
Roughness 0.76 0.67 0.44
0.57
Average
+50Jlm
Roughness
0.46 0.43 0.40 0.35
Wavelength

/
- 68 -

The Talysurf 10 has enabled the measurement of the surface

roughness of coatings in a simple manner. Both the values that have

been measured, Ra and Aa, complement each other and should be used

together to describe a surface profile.

The results in the figures and tables have shown that in

general the average roughness decreases with thicker films but

increases with larger particle sizes used in the spraying. Comparison

of these results with figures for other surfaces reveal that there

is little change in these qualities with the parameters being

observed here.

As for the gloss tests, the commercial powder shows more

definite trends for the change in roughness average and wavelength

with film'thickness. The linear regression fits, although only

shown for comparison, appear quite good correlations in these cases.

This could be due to having a less random and loose packing than

the narrow size fraction packings.

3.7 Adhesion

The cross hatch test, as described in BS 3900 : Part E6, was

employed as it was a test that is commonly used and is simple in

nature.

A series of spacing edges were made as described in the British

Standard. A block of wood,with a rubber base, of suitable size to

be held on the workpiece by hand, was used to hold the eleven spacing

edges (See Figure 3.32). Due to the small size of the sample plates

difficulty was encountered in keeping the block and spacers steady

on the plate whilst the grooves were cut. Therefore a large plate

with a 2 in square recess was used to house the plate whilst cutting

proceeded. In this way the spacers were held flush against the

sample and housing plate.


1

10 Spacers Wooden 1
lmm thick Block

1
o 1

Rubber Base 1

1
Figure 3.32. Series of Spacing EdgeS for Cross-Cut Test

1
- 69 -

3.7.1 Experimental Method

A sample plate was placed in the housing plate and the

block of spacers suitably positioned over the plate •. Using a new

bladed scalpel, 11 grooves were cut through the film to the metal

substrate. The cuts were made at a uniform and slow speed across

the surface of the coating. The spacers were then rotated by 90 0

and a similar set of cuts were made to produce a cross hatch on the

coating.

The number of small 1 mm squares of coating removed by this

action were then compared to the classification chart in the British

Standard. (See Figure 3.33).

A further test was carried out, similar to that described in

ASTM Standard 03359-74, in which a piece of adhesive tape was placed

over the cross hatching. The tape was then removed and the remaining

coating was again compared to the classification chart.

Due to the destructive nature of this test (and the corrosion

test) only one plate was used to test the adhesion of the coatings

for any particular thickness and powder type.

3.7.2 Results and Discussion

Five different thickness ranges were tested for each of

the powder types. Table 3.4 shows the results from both these tests

in terms of the characterisation number for each of the powder types

and film thicknesses. Difficulty was found in cutting completelY

through the coating to the metal for films that were thicker than

BOum.
Figure 3.33. British Standard Classification Chart (BS3900)

Appearance of surface or crou-cut


Clusin· .area from which n.aking has occurred
calion Description (example for lix panllel cuta)

o The edges of the cuts are completely smooth; none of the


squares of the lattice is detached.

Detachment of smallnakes of the coating at the intersections


of the cuts. A cross·cut area not distinctly greater than 5 %
is affected.

2 111e coatingllas naked along the edges and/or at the inter.·


sections of the cuts. A cross·cut area distinctly greater than
5 %, but not distinctly greater than 15 % is affected. '.

3 The coating has naked along the edges of the cuts partly or
wholly in large ribbons, and/or it has naked partly or wholly
on different parts of the squares. A cross·cut area distinctly
greater than 15 %, but not distinctly greater than 35 % is
affected.

4 111e coating has naked along the edges of the cuts in large
ribbons and/or some squares have detached partly or wholly:
A cross-cut area distinctly greater than 35 %. but not dis:
tinctly greater than 65 % is affected.

5 Any degree of naking that cannot even be classified by


classification 4.
Table 3.4. Results of Adhesion Tests - Cross hatch and Tape Test

<lOIl.m . lD-2011m 20-3011m 30-4011m 40-5011m +5011m Commercial

t t t t t t t
C T C T C T C T C T C T C T
(llm) (llm) (llm) (llm) (llm) (\lm) (\lm)

15 0 0 18 0 0 29 0 0 28 0 0 33 0 0 37 0 0 21 0 0

25 0 0 27 0 0 38 0 1 36 0 0 39 2 3 49 0 0 33 0 0

38 1 2 39 0 0 41 1 3 49 0 0 40· 1 2 58 0 0 43 0 0

60 1 2 46 0 0 48 0 0 55 0 0 54 0 0 65 0 0 52 0 0

84 1 3 SS 0 0 59 0 0 66 0 0 65 1 2 72 0 0 60 0 0

t film thickness (average of 5 measurements)

C - Cross hatch adhesion test

T Cross hatch test followed by tape test


- 70 -

As expected, these coatings typically gave very good adhesion

results. Even the <lO~m range powder type coatin~s gave good results.

These results reflect the fact that the coatings had been stoved

correctly and that the substrate surface was in good condition.

Some brittleness was observed in a few of the coatings but this

could be due to slight variations in the oven temperature. However,

over 75% of all the coatings tested showed no signs of loss of

adhesion due to the cutting of the crosshatch or pulling on the

film during the removal of the tape.

These results are expected since a commercially formulated

powder has been used for the experiments and it should exhibit good

flow and adhesion properties after a correct stoving cycle. This

test also shows that the powder film has been completely fused. If

this was not so and a dry, unfused powder monolayer was left close

to the substrate then large areas would be lifted when the upper fused

film had been corrupted.

Following the tape test, only 3 samples had a classification

greater than 2. In these cases the results show that less than 35%

of the film was removed, with no squares being wholly removed.

3.8 Corrosion

ASTM Standard Bl17 was followed for the assessment of corrosion

resistance of coatings using the salt spray test. The apparatus

(See Plate 3.8) consisted of a fog chamber, a salt solution reservoir

compressed air supply and atomising nozzles. The sample plates were
0
held on larger tin plates inclined at 30 to the vertical and parallel

to the principal direction of the flow of the salt spray through the

chamber. The coated plates were held by double sided sellotape and
Plate 3 . 8. Salt- Spray Test Chamber
- 71 - .

the edges of the samples were painted using a zinc based paint to

stop the steel edges from corroding and hence keep the. main sample

area clean from corrosion stains.

3.8.1 Experimental Method

Salt solution of 5 parts by weight sodium chloride in

95 parts distilled water was used. The salt spray chamber was kept

at approximately 37 0 C throughout the period of the test. The

condensed fog was collected so that drops of solution did not

continually fall onto the specimens.

The samples were split into two halves. One half were scribed

with an 'X' using a scalpel blade. It was ensured that the cuts

were made right through the coating to the metal substrate. The

other half were left untouched. The plates were left in the spray

chamber for 48 hours and then removed. Their conditions after testing

were recorded.

3.8.2 Results and Discussion

Two plates were taken from the same thickness ranges as

used for the adhesion tests. In all cases there was no detachment of

coating from the plates scribed with an 'X'. Corrosion had taken

place on the metal and was seen by the red/brown stains on the cuts.

There was no evidence of pin holes in any of the surfaces of

the coating; no brown spots were apparent.

This test shows the resistance of the film to a corrosive

atmosphere but also gives an indication of the adhesion of the

coating, especially in the case of the X-cut coating. All powder

size fractions of this commercial powder gave excellent results, as

expected with epoxy resin powder coatings. No detachment of the

coatings again confirmed that the stoving cycle had been correctly
- 72 -

carried out and that adhesion was representative of normal industrially

produced coatings.

3.9 Porosity

The measurement of porosity has been carried out in two ways.

The first test involved the removal of coatings from the substrate.

This was achieved by first coating the substrate with a very thin

film of aluminium produced by a vacuum coater. This aluminium

coating was then dissolved in nitric acid after the powder coating

had been stoved. The coating could then be removed from the plate.

A sample of the coating was then placed in an electron micro-

scope. The coating was subjected to X-rays and a photograph taken

of those X-rays permeating through the coating. Where absorbance

of X-rays took place due to the high atomic weight pigment atoms the

picture was darker. Any holes through the coating or air pockets

would be shown as light grey areas as some .absorbance took place.

Thus the porosity of the film could be seen over a sample area.

The corrosion test previously described also acts as a

porosity test. Any pores through to the substrate act for sites of

corrosion and red/brown rust spots would appear in the surface of the

coating where this has occurred.

3.9.1 Results and Discussion

Analysis of the plates from the corrosion experiments

have already shown that no pin holes could be seen in the coating.

Electron microscope photography also showed that there were no air

pockets in the coatings. These tests were dcne over a random sample

of coatings throughcut the various powder types.


- 73 -

The effectiveness of the stoving cycle is again confirmed and

suggests that change in particle size does not unduly effect the

nature of the coating in this respect.

3.10 Conclusions

A series of tests have been developed such that the physical

qualities, and in particular surface properties, of a set of coatings

have been examined. A commercial standard epoxy resin powder was

fractionated into 6 size ranges. Each fraction was used as a

separate powder for coating a set of sample plates. A reproducible

method of coating was employed using a commercial type spray gun

system and over 2500 plates were produced using this method.

The results from these tests have given a comprehensive study

of the effect of particle size range on the final film properties

of stoved powder coatings. It has been clearly observed that although

different particle sized powders have been used to form a coating,

the films still retain excellent properties of corrosion resistance

and adhesion. This is true for any thickness of film between 20

and 100 ~ms.

However, it has been observed that larger sized particles do

produce rougher coatings which have a surface profile which has a

close structure with very fine superimposed roughness.

Associated with this increase in roughness, larger particle

sizes also gave slightly lower gloss results than finer particles.

Thicker films on the whole gave films with better gloss characteristics

and smoother surfaces than thinner films.

The <lO~m range showed poorer surface characteristics than

any of the other powders used. This is due to the very poor flow

properties of the powder. The powder tended to be very cohesive and


- 74 -

hence gave rise to the spraying of 'clumps' which consequently gave

poorer finishes. It would not be recommended that a powder of this

type be used for the production of powder coatings. This powder type

was not included in the linear regression analysis of results for

different powder size ranges.

In general the results of the gloss tests can be considered

to show no significant differences when the results are compared

to those obtained from other finishes, such as a mirror or matt

surface. The coatings appear to be very similar to each other. Visual

inspection confirms this and all coatings have very good appearances.

The roughness results are very similar in this respect but perhaps

show marginally more significant changes in roughness with particle

size.

The adhesion, corrosion and porosity tests confirm the excellent

qualities of the coatings and that the stoving cycle has been

correctly carried out.

These results appear to agree with the findings of Nix and

Wolpert, as described in Chapter 1, in that the effects due to

particle size on the levelling of coatings are small. In these

experiments all the powders came from the same powder batch and

therefore no effects due to formulation need be considered. The

type of clusters of particles mentioned by Wolpert (S. Wolpert et aI,

1972) have given noticeable results for the <lO~m range powder.

It can therefore be concluded, on the basis of the results

for this powder and its size ranges, that particle size has negligible

effect on film properties. Therefore any increase in deposition

efficiency gained by changing particle size of the powder can be made


- 75 -

without fear of losing physical qualities of the final film. This

conclusion also appears true for coatings of any thicknesses between

20 and 100~m.

3.11 Summary

The apparatus and experimental methods used in the assessment

of the physical qualities of electrostatic powder coatings have been

described.

Particular emphasis has been placed on the measurement of film

thickness, gloss and related properties, and surface roughness.

Experimental results, presented as a three way comparison, have been

discussed. It has been found that there is no significant difference

in the properties of films with size of powder sprayed or thickness

of film produced.

The following chapter describes investigations performed in

order to discover and understand some of the reasons for these

results.
CHAPTER 4

PARTICLE STOVING AND PACKING OBSERVATIONS

4.1 Introduction

4.2 Stoving Observations

4.2.1 Experimental Observations

4.2.2 Comparison of Different Types of Powder

4.2.3 Determination of Spreading Factor

4.2.4 Experimental Results

4.2.5 Discussion of Results

4.3 Particle Packing Considerations

4.3.1 Effect of Formulation on Powder Characteristics

4.3.2 Experimental Observations

4.3.3 Quantification of Observed Effect

4.3.4 Experimental Results

4.3.5 Discussion of Results

4.4 Summary
-------

- 76 -

4.1 Introduction

The examination of the coatings produced from different sized

particles has shown that there is very little difference in the

physical properties of the coatings for any of the sizes used. The

differences in surface roughness and gloss were greater than in other

properties but the physical significance was negligible. However, it

has been observed that in all the gloss and surface roughness

measurements the coatings of commercial powder exhibit more definite

trends than any other powder. This is the only powder with a wide

size distribution encompassing all the sizes used in the other

powders.

To try and gain some understanding of the reasons why this lack

of any apparent effect is observed further investigations into the

way in which particles pack and stove have been carried out.

Investigations were on a very exploratory basis but some of the

findings gave rise to more formal types of investigations which in

turn produced very interesting results.

4.2 Stoving Observations

Experiments of a very simple nature were conducted to observe

the way in which powders flow out. This was accomplished by observing

particles melting on a microscope hot stage. Some of these experiments

were recorded by using a cine camera placed above the microscope

eyepiece (having removed the objective lens). A diagram of a similar

apparatus employed is shown in Figure 4.1.

A diagram showing the way in which particles flow has been already

given in Chapter 1.

Various stoving experiments, inVOlving sprayed plates and single

particles, were carried out. A few 2" square plates were coated

using the commercial equipment and then placed on the hot stage.
h
.--.....J ,'--_ _ _ _ _ _..,
c
Cine Camera

I I

Microscope

To Temperature
Powder Sample Controller
Heating Element

Glass Slide

~.
Light
Beam
Mirror

Figure 4. L Microscope Hot-stage Apparatus


- 77 -

For this case it was found that the thicker layer of particles

gave focussing problems. When the packed particles started to melt

the level of the coating dropped substantially. Due to the small

depth of field of view the coating then moved out of focus and

therefore manual ·adjustment was required whilst filming took place.

Also, as the coating levelled out to its characteristic smoothness,

the surface could not be clearly seen due to the lack of detail

making focussing increasingly difficult.

It should be noted that this method of heating is different to

that encountered by a coating in the stoving oven. Heat is only

being applied from the base through the metal, and hence, up through

the coating.

In a stoving oven the heat, for thin plates, will come from

both sides. For large bulky objects that have a large amount of base

metal the heat for melting will mainly come through the surface and

hence the powder layer. It is more likely in these caSes that unstoved

powder could be left near the substrate giving poor adhesion qualities.

As expected, in the case of the thin sample plates used in the previous

experiments, this was not observed and both adhesion and corrosion

characteristics were to a high standard.

4.2.1 Experimental Observations

From the cine films that were taken of single and groups

of particles melting, it was found surprising the extent to which the

powder spread during the flow out stage. When an irregularly shaped

particle was melted it initially tended to form a ball. This is shown

in Plate 4.1. The particle lost all of its irregular shape charac-

teristics. After this stage the powder then flowed out and covered
Plate 4.1. Melting of an Irregular Shaped Particle
- 78 -

nearly three times of its initial substrate area.

When a f ~w particles plac !( together on the same plate were

stoved it could be clearly seen how the whole of the surface can be

coated with a relatively small amount of powder. This is shown in

Plate 4. 2 . In the first picture the irregular particles are seen to

cover a small area of the plate. As the time of heating progresses

some of the particles contract into balls and then begin to spread .

They then flow into other particles and the flow out continues.

This is also shown on the next plate (Plate 4.3) in which the

latter stages of powder flow is shown in silhouette form . These

photographs were taken using transmission microscopy . A few separate

particles fuse together to form a coating.

However, although this tends to show that the powder from

individual particles mixes in with others, the actual mixing between

particles is limited . This was confirmed by spraying a sample plate

with a mixture of coloured particles. Three different colours were

used, equal amounts being mixed together, and these were sprayed

onto a glass slide so that the underneath could also be seen .

Spraying of glass was effected by first spraying the plate with an

anti-static spray agent which allowed the plate to be grounded to a

better extent than with a normal glass plate.

The results from this experiment, after stoving of the coating,

are shown in Plates 4.4(a) and 4.4(b). It was found very difficult

to focus since there was no surface detail. The boundaries between

different coloured particles are very hazy and do not help in this

matter.
Plate 4.2 . Melting and Fusing of Powder Particles
• • • o •
• • •


\
.t

00

• -: •

• •
Plate 4. 3 . Silhouettes of Melting and Fusing of Powder Particles
Plate 4.4. Staved plate coated with coloured particles viewed
from (a) above and (b) underneath (lower picture)
- 79 -

It can be seen that there is a small amount of mixing at the

boundaries but the bulk of the particle, although having flowed,

keeps its basic colour.

4.2.2 COmparison of Different Types of Powder - Spreading vs Finish

The previous simple experiments have shown that, on

melting, this powder flows to a great extent and this could be the

overriding factor when comparing coating properties with particle

size sprayed. Due to the excellent flow out characteristics of this

powder any effect of particle size could be nullified thus giving

only small changes in physical properties observed.

The effect of flow out characteristics was investigated to see

if there was any correlation between the flow out properties of powder

and the final finish'that is produced. A spreading factor, in terms

of the increase in diameter of a particle being heated on a hot plate,

was used to represent the flow out of the powder. The'final surface

finish was assessed by visual observation of gloss and surface

smoothness.

4.2.3 Determination of Spreading Factor

Five different types of powder, which were known to

produce different final film finishes, were used. The apparatus

employed is shown in Figure 4.1. Several particles were placed on

a glass slide and placed between the heating elements. The temperature
0
was rapidly raised to l80 C and the melting of particles recorded by

a cine camera at 12 frames/sec.

Spray coatings of the powders were produced using the standard

technique and then stoved in the oven at the manufacturers recommended

conditions. The coatings were visually assessed and placed in order

of merit.
- 80 -

The films of particles melting were analysed using a Vanguard

image analyser. This device allows a frame of film to be enlarged

onto a screen (0.5 x 0.3m) where, using an electronic cursor,

positions on the screen can be logged by computer. The initial

diameters of the particles were recorded by logging the leftmost,

rightmost, upper and lower points of the particle when it had

contracted to its smallest size. This was repeated when the particle

had spread on melting to its fullest extent. The change in size of

a typical particle as it flows can be seen ~n Plate 4.5. From this

data the vertical and horizontal diameters were averaged for 10

particles of each powder. The data was analysed by a computer program

which gave a spreading factor based on the ratio of the initial and

final particle diameters.

4.2.4 Experimental Results

A summary of all the results obtained from the determination

of spreading factor and visual assessments for the five different

powders is shown in Table 4.1.

The recommended time and temperature of stoving for each powder

is also given. The spreading factor is given as the ratio of the final

and initial melted particle diameters. The ratio of areas is also

given.

The best three coatings were difficult to visually assess in that

they were all very similar. The polyester powder which was the worst

gave a mottled type finish which was very uneven and obviously had

not flowed out very well. In this case levelling had not taken place

to the same extent as the others.



• "



• •
,

~.~------~~.~~.----~,~

.,
,~-- ----.--------~

I i6 ••• •


• • •


... •
0
'"

J;. -. •

I
0
!
Cl
a
I •
('>
'~~
Change ~n Spreaded Area of Particle during Coating
Table 5.1. Stoving of Five Different Powders - Comparison of Spreading Factors

Cure Visual Inspection t Measured Diameter Diameter


Time Area
Powder Type Temp Ratio Grading t
(mins) Ratio
(OC) Gloss Smoothness Before After A/B

Slow Cure Amine


180 20 2 1 1.03 1. 82 1.77 3.13 1
Epoxy

Polyester 180 15 5 5 0.97 1.27 1. 31 1.72 5

Epoxy Polyester 180 10 1 2 0.99 1.61 1.62 2.62 3

Fast Cure
180 10 3 3 1.01 1. 76 1. 74 3.03 2
Anhydride Epoxy

Fast Cure Amine


180 10 4 4 1.09 1.67 1.52 2.31 4
Epoxy

tGradings 1 = best, 5 worst.


- 81 -

4.2.5 Discussion of Results

The determination of a spreading factor using this method

can be seen to show a good indication of the type of finish that might

be expected from a powder. The higher the spreading factor, the

better the flow characteristics of the powder and hence the higher the

quality of the coating produced by electrostatic spraying.

The standard powder used in the earlier experimental tests

obviously had excellent flow characteristics which resulted in very

good films being produced with respect to adhesion, corrosion

resistance and porosity. The effects seen in these stoving experiments

agree with the work carried out by Wolpert et al and Nix et al.

(S.M.Wolpert et aI, 1972; V.G.Nix et aI, 1973). Wolpert stated that

the levelling of a coating is very dependent on the gel time which

occurs when there is a drastic change in viscosity. Viscosity is

critical to the levelling but below a certain amplitude surface tension

effects are important. Nix agreed with this but found that other

factors also have a bearing. Particle clusters were observed to be

one source of orange peel. After individual particles had melted and

flowed the radius of curvature of imperfections will be due to either

the clusters or to orange peel 'hills and valleys'. Trapped air in

the dry powder layer also appeared to have an effect on the flow of

powder coatings. Spitz concluded that the ripple in powder coatings

originates from the uneven deposition of powder due to particle size

and particle clusters. Again the clusters were observed to have more

effect than particle size. The ripple wavelength increased with

levelling according to Orchards theory (G.T. Spitz, 1973; S.E. Orchard,

1962).
- 82 -

In this study the powder sizes investigated originate from the

same powder formulation. Therefore any changes in surface property

will be due to particle size, packing orientation, porosity (trapped

air), and particle clusters; as opposed to different melt viscosities,

gel times or surface tensions. For this powder, however, little

difference has been observed with particle size or film thickness.

This is probably due to the excellent flow properties given by the

formulation. Slight increases in roughness with particle size were

observed which seem to agree with the findings of the previously

mentioned workers.

The commercial powder showed more definite trends than any

other powder sprayed and so the observations of packing of this

distribution on a substrate is now described.

4.3 Particle Packing Considerations.

In line with general observations of particles stoving on a

substrate, experiments were carried out to investigate the packing

of different sized particles. To achieve this different coloured

powders were used to represent size fractions of the powder

distribution.

Various colour combinations could then be used to observe where

certain sizes had landed on the substrate or Ln the packed layer.

4.3.1 Effect of Formulation on Powder Characteristics

Since it is possible that different powders have different

characteristics with regards charging,· stoving etc., it was necessary

to look at the characteristics of the powders being used.

It was understood that apart from the obvious difference of

pigmentation, the four different coloured powders were basically


- 83 -

chosen for their similar formulation. However to confirm that they

did· not behave substantially different they were subjected to two

tests; namely determination of spreading factor and measurement of

charging characteristics.

Films taken of the red, orange, white and blue powders melting

were analysed in the manner previously described. The results are

given in Table 4.2. It can be seen that the standard deviations about

the average of the results is very small. There is very little

difference in the spreading factors of the powders although it was

observed that the red powder was found to be much slower in the time

it took to start melting.

Since the ratio of the diameters on spreading will be dependent

on the size of particles used, particles of similar size were chosen

for analysis. This can also be seen from the table of results. The

orange, white and blue powders were chosen for further experiments

since they provided a contrasting mixture of colours and the red

powder was different in that it had a langer time to melt start.

The charging characteristics measurements were carried out at

the Electrical Engineering Department.of Southampton University by

Dr. S. Singh. An airless gun was used to charge the powder and a

Faraday cup used to collect the charged powder. A grid was placed

in front of the Faraday cup to remove excess ions. The weight of

powder collected and the charge leaked away is recorded hence giving

a charge to mass ratio measurement. (S. Singh, 1979).

The results of these measurements are given for the three powders

in Table 4.3. It waS not possible to make measurements for the <20~m

range as the powder would not flow well. Slightly higher values of

Q/M were obtained for the orange and blue powders but it was considered
Table 4.2. Spreading Factor Results of Four Coloured Powders

White Orange Red Blue.

Before After A/B Before After A/B Before After A/B Before After A/B

1217 2050 1.684 1270 2188 1.723 1225 1986 1.621 1076 1811 1.683

1337 2287 1.711 1110 1846 1.663 1207 1942 1. 609 1489 2360 1.585

1310 2177 1.662 1219 2137 1.753 1011 1721 1.702 1184 2180 1.841

1057 1763 1.668 1512 2601 1. 720 1289 2211 1. 715 1072 1838 1.714

1205 1770 1.469 1145 1973 1. 723 1553 2440 1.571 1384 1969 1.422

1480 2488 1.681 1275 2345 1.839 1772 1988 1.696 1349 2293 1.699

1146 2264 1. 975 1169 1883 1.610 988 1748 1. 769 1118 1842 1.647

977 1713 1. 753 1061 1895 1.786 1136 1813 1.595 1007 1634 1.622

1179 1977 1.676 1076 1897 1. 763 1146 1821 1.589 1007 1945 1.931

1209 2084 1.723 1188 2027 1. 706 1155 1940 1.679 1241 2020 1.628

1350 2190 1.622 1217 2180 1.791 1168 1903 1. 629 1442 2443 1.694

1250 1868 1.494 1421 2035 1.432 994 1596 1.606

Average 1.677 1.709 1.652 1.673

Standard Dev. 0.122 0.101 0.060 0.122

All measurements are in terms of diameter of melting particles


- 84 -

that these were insignificant differences. (S. Singh, 1980a).

None of the powders reached the predicted saturation charge, as shown

in the table, and it was noticeable that the small particles had

effectively the same charge as large ones.

4.3.2 Experimental Observations

The average size distribution of the coloured powders

was found by determining the distribution of each of the powders

using a Coulter Counter. Dispersion of the powder in the electrolyte

was found difficult, but was aided by addition of 'Tergitol'. The

average distribution is shown in Figure 4.2. Each of the powders were

classified using the Alpine zig-zag classifier into three fractions,

namely <20~m, 20-40~m and >40~m. These fractions were then mixed

together in appropriate quantities as given in the table below to

form the three different mixtures.

Size Range <20~m 20-40 >40~m

Wt % in Range 16.0· 35.0 49.0

Mixture 1 Orange White Blue

Mixture 2 Blue Orange White

Mixture 3 White Blue Orange

Each of the mixtures were sprayed onto the large 10" x 8" plate

which was covered with twenty 2" square sample plates. After spraying,

the plates were compared to the original powder fed to the spray gun,

and this is shown for the three mixtures in plates 4.6, 4.7 and 4.8.

A leitz microscope and associated equipment was used to photograph

the powder coating in closer detail. The apparatus is shown in

Plate 4.10.
Figure 4.2. Average Size Distribution of Coloured Powders

Coulter Analysis

100

.
....
N
80
"'H
0
'"
:>
N
.... 60
....~.
.
.:3

........:> 40
<11
.....
§
u
20

O;------.------,------.------,------,------,------.------,------.r------y
o 10 20 30 40 50 70 80 90 100
Particle Size
Table 4.3. Charge to Mass Ratio Measurements of Coloured Powders

Measured Predicted Fraction of


Powder Size Fraction
Q/M Q/Mt Predicted

-3
Orange 20-40 5.25xlO-4 1.17xlO 44%

+40 -4 -4
Orange 5.30xlO 7.07xlO 75%

Blue 20-40 5.65xlO-4 1.17xlO-3 48%

Blue +40 5.05xlO-4 7.07xlO -4 71%


-4
White 20-40 4.50xlO 1.17xlO-3 38%

+40 -4 -4
White 4.47xlO 7.07xlO 63%

tassuming saturation charge, where qs

and where K = 4
£0 = 8.85 x 10- 12
Eo = 4 x 10 5 V
Plate 4.6. Comparison of coating and feed material for mixture 1

Plate 4.7. " Comparison of coating and feed material for mixture 2
Plate 4.8. Comparison of coating and feed material for mixture 3

Plate 4.9. Microphotograph of upper layers of coating of mixture 1


Plate 4 . 10. Leitz microscope and apparatus


- 85 -

Interesting effects were observed in each case. The uppermost

layers predominantly consisted of particles of the largest size

fraction. The comparison with the feed material clearly shows the

difference in colours. Further microscopic observations of the

individual 2" plates confirmed this and an example is shown in

Plate 4.9. Here the largest fraction, blue, is seen to be predominant

in the upper layers. The smallest fraction, orange, appears to be

closest to the substrate. Two other points from this photograph

should be noted. Firstly the packing can be seen to be very open;

large crevasses are present between particles and the layers are

fairly tall compared to the particle size. Secondly, it is observed

that very fine particles, typically <lO~m, stick to larger ones and

in this case very fine orange particles are seen around the large blue

pa~ticles.

In Plates 4.11 and 4.12 the surface layers are viewed at

slightly different angles . Depth of field is very small but similar

observations are made and tree-type structures can be seen.

Hence, it appears that a size segregation effect takes place

through the packing in that large particles are preferentially

deposited in the upper layers.

4. 3. 3 Quantification of Observed Effect

In order to quantify the effects that have been observed,

to obtain data rather than consider only pictures, the size distributions

of powders deposited on the plate were measured and compared to that

in the feed hopper.


Plate 4 . 11. Microphotograph of a powder coating viewed at an angle

Plate 4.12. Microphotograph of a powder coating


- 86 -

Two of the coloured powder mixtures were initially used for this

with size distributions being measured by the coulter counter.

However, problems of dispersion and sampling made the use of this

instrument unacceptable. Only a very small sample from the plates

can be placed in the coulter beaker and hence difficulty is encountered

in obtaining representative results. Dispersion of large particles

was difficult without using high stirrer speeds which then gave

problems with air bubbles.

The Microscal photosedimentometer was chosen as a device that

would measure size distributions in the range required and which gave

reproducible results because of the amount of sample used.

The experiments previously described were therefore repeated

using four different powders, two black and two white. These were:-

Powder 1 Drynamels White Epoxy Anhydride (2......-02i..C)

Powder 2 Ault and Wiborg White Epoxy/Polyester (966-651-015)

Powder 3 Drynamels Black Epoxy Anhydride (LFA 6319)

Powder 4 Teodur Black Epoxy Polyester (SP 926/18/1)

Each of these powders had Q/M measurements taken at Southampton

University. Their charging characteristics are shown in Table 4.4.

(5. Singh, 1980b). Again there appeared to be very little difference

l.n the results.

The powders were sprayed onto twenty 2" square sample plates

held onto the larger 10" x 8" plate. One of the samples was then

washed into 500 ccs of distilled water. The Microscal wide angle

scanning photosedimentometer sensitivity control was adjusted so that

an intensity of 100% was obtained, when the light beam through the
Table 4.4. Charge to Mass Ratio Measurements for Four
Different Powders

Powder Type Colour Size Fraction Q/M x 10- 4

20-40 4.1
White
>40 3.6
Epoxy
Anhydride
20-40 5.3
Black
>40 4.8

20-40 4.9
White
>40 3.8
Epoxy
Polyester
20-40 5.0
Black
>40 3.8
- 87 -

cell of clean liquid was at its maximum. A photograph of the apparatus

LS shown in Plate 4.13.

The sample was then placed in the glass cell and the suspension

was well stirred. At this time the chart recorder was switched on.

The scan is started after a suitable length of time, this usually

being when the chart recorder falls to below 30% of full scale. The

cell is then automatically lowered 50 that the light beam 'scans' the

sedimentation tank and reduces the height of fall of particles in

the sampling zone. The chart recording is analysed and the optical

density of the suspension (as a ratio) is measured at various times

and recorded. A typical chart is shown in Figure 4.3.

The data is then used in a computer program to obtain the size

distribution of the sample. This is based on the settling rates of

different sized particles as given by Stokes Law:-

where U is obtained by knowing the height dropped in time t (U = hit).


Three samples were used for each powder sprayed, and each feed

material. This method has the advantage of analysing all the powder

on a sprayed plate and hence reduces sampling errors.


Plate 4.13. Microscal wide angle scanning photosedimentometer
o ·2 '. .J ·5
OPTICAL DENSITY (D)
.-----~-- -i\
·
·
)'\
l\
· "-
~-~\-- -----------------.
\\
-r--
l\
1\

-----'--.---~-.::.\..,-:.\.\- -----.:.-----'--"--- -----

. - - - . ----:--7---;
j

i
---\.----
I
---;--

:
--,-~
i
i
.

:

-7 ~8 ,9 I-a \.
iI '"
.....
i : ""...c:
'"
....""
o
""
'"
(l
..,
'"
::: '<
.....
(l

....
'"
(')
:r
...'"
~

...0
,."

. , 3
,
!-~';"

i i '"
:r
o_____~·~~___~-~i.___~3~ ___.~~_~~~~~:6 0

~PT1CAL DENSITjr (D)


'" ""0
rn
I \\ '....."
0-
3
:::'"
'"30
I, ,.,'"
'..."
- 88 -

4.3.4 Experimental Results

The results from each of the scans are given, for all

of the samples and for each powder, in the tables in Appendix C

(Tables Cl to C8). Each of the distributions were drawn and the

cumulative weight percentages under given sizes were collected and

averaged to form the distributions shown at the right hand side of

these tables. These have been determined for each powder, before

and after spraying.

The size distributions are summarised in Table 4.5 for the four

different powders. The comparison of size distributions are shown

in Figures 4.4 to 4.7. Only distributions representing the smallest

50% of the powders have been shown in each case since it is the fines

that are of particular interest. In each case the amount of fines

is considerably lower in the powder after spraying. On average there

is between 7 and 15% by weight less material smaller than 30~m in

the powder on the sprayed plates. Therefore there is a definite

change in size distribution of the powder on spraying, and involves

a loss of small particles. It appears from the graphs that there is

equal amounts of particles <lO~m. Less particles of this size are

lost probably because many of them stick to large particles and hence

still land on the plate.

4.3.5 Discussion of Results

The analysis of size distributions of several powders

and mixtures before and after spraying have agreed with photographed

observations shown in earlier plates. When powders are electro-

statically sprayed for a reasonable length of time a general loss

of fines takes place. This is quantitatively confirmed by the changes


Table 4.5. Comparison of Size Distributions of Four Different Powders Before and After Spraying

Cumulative Weight % < Size

Powder 1 Powder 2 Powder 3 Powder 4

Size After Before After Before After Before After


Before
\lm

10 5.7 4.9 5.8 6.8 5.4 3.5 6.5 3.8

15 7.3 5.6 7.5 6.9 6.5 4.0 9.1 5.1

20 10.9 7.3 10.0 7.6 9.3 6.7 14~1 6.6

25 19.1 10.9 15.5 10.5 14.9 11.0 20.9 9.7

30 29.5 16.8 24.5 18.0 23.0 16.5 29.0 14.5

35 42.0 24.5. 35.5 28.3 32.8 22.5 37.9 21.3

40 57.1 34.7 45.8 39.3 43.4 29.5 46.7 29.2

45 70.2 47.3 55.5 50.0 54.2 36.5 56.5 37.2

50 82.1 62.3 - 60.0 - 44.0 - 47.4


Figure 4.4. Size Distribution of Powder 1 Before and After Spraying

Microscal.Analysis

Key
100 0- Before
.
.....N
o - After
....
-0
U)

80
§

60

40

20

o
1 o 3 4 5 6
Particle Size (~m)
Figure 4.5. Size Distribution of Powder 2 Before and After Spraying
Microscal AnalYllis ..
Key
70 o - Before
o - After
QJ
N
·M 60
...
Cl)

QJ
."
:§ so
I>t

40

30

20

10

O~--------~----------r---------.----------r---------'---------.
o 2 3 40 so 60
Particle Size (~m)
Figure 4.6. Size Distribution of Powder 3 Before and After Spraying
Microscal Analysis
Key
o - Before
70 o - After
60

50

40

30

20

10

01----------r---------,----------r---------.---------~--------~
o o 20 30 40 50 60
Particle Size (~m)
Figure 4.7. Size Distribution of Powder 4 Before and After Spraying
Microscal Analysis
Key

80
o - Before
o - After
QJ

.....N
en
I< 60
QJ
"t:I
I'l
P
w
...
..c
.....
OJ)

40
~
QJ

.......:>
....'"
u
~ 20

O~--------r---------r--------'---------'---------'---------'
o 10 20 30 40 so 60
Particle Size (~m)
- 89 -

of size distribution shown in the last four figures. For each of four

powders there is a definite loss of fines, and some medium sized

particles, from the powder. The photographs suggest that this

increasingly occurs as spraying progresses such that on the upper

layers of the powder packing large particles predominate. It has been

observed that near the substrate small particles are present in the

powder layer.

However, at this time, it is not clearly understood why these

observations take place. These effects could be due to the balance

of forces on different sized particles changing with size, velocity

etc. 6harge on the particles, the depth of the coating could also be

controlling factors. It is also possible that there is an optimum

size distribution where, for a given thickness of film.or spray rate

etc, a loss of fines does not occur during spraying.

In the next two chapters the forces acting on the particles will

be considered in a theoretical manner with the aid of a computer

simulation program. The effects of particle size, air velocity and

film thickness will be considered. The effect of monosized distri-

butions on the packing of particles will also be investigated to try

to gain a better understanding of the way in which a powder packing

is formed. By so doing it is hoped that suggestions can be made for

the optimisation of the process.

4. 4 Swinnary

The determination of a spreading factor by measuring the change

in size of a melting powder particle has been shown to give an

indication of the type of quality of finish to be expected of a powder

coating. The results suggest that, for the powder used in the experiments

of Chapter 3, its excellent floW characteristics outweigh any effects


- 90 -

caused by the size of powder sprayed on the physical properties of the

coating. Plate 4.14 shows how the powder coating can level out.

From macro and micro observations of mixtures of coloured

particles, representing different size ranges, a size segregation

effect through the powder coating is noticed. A predominance of large

particles is observed in the upper layers of the coatings. This effect

has been quantitatively investigated by the determination of size

distributions of powders before and after spraying. For four different

powders a loss of fines was determined in each case.

Although this effect has been observed for powders with a wide

distribution it is not clear what occurs for monosized powders. The

structure of the packings obtained has been seen to be of a very

open nature although comparisons of packings have not been made with

those obtained from the narrow size range powders.

It is possible that the observed effects are due to the balance

of forces acting on the particles close to the substrate which changes

as coating progresses. Therefore a study of the forces acting on

particles under different operating conditions has been carried out by

computer simulation. The setting up of a force balance equation is

described in the following chapter.


Plate 4.14. The Surface of a Powder Coating during Stoving
CHAPTER 5

TRAJECTORY EQUATIONS

5.1 Introduction

5.2 The Force Balance

5.3 Charge on a Particle

5.4 Electric Field

5.5 The Forces

5.5.1 Drag Force

5.5.2 Flow Past a Flat Plate

5.5.3 Field Force

5.5.4 Interparticulate Forces

5.5.5 Image Force

5.5.6 Repulsive Forces

5.5.7 Attractive Forces

5.6 Other Representations of Forces

5.7 Comparison with 3-D

5.8 Summary
- 91 -

5.1 Introduction

In the previous chapter investigations into the way particles

of a size distribution pack onto a substrate were described. It was

found that there was a size segregation effect taking place. Large

particles were preferentially deposited on the upper layers of a

packing and there was an overall loss of fines. It was suggested

that these effects were possibly due to the balance of forces (electro-

static and aerodynamic). acting in the close proximity of the substrate.

With an increase in film thickness it could be put forward that these

effects accounted for some of the small changes in surface properties

for the commercial powder.

Although trends and effects have been observed for this wide

size distribution powder it is not clear what differences occur

between packings of various monosized powders. In order to gain an

insight into the type of packings and variations therein, produced by

the spraying of monosized distributions, a computer simulation

technique has been employed. The intention of this method was to

investigate the variations in orientation and structure of packings

obtained for powders of different distributions for various operating

conditions. The formation of a computer program to do this would

allow many conditions to be examined with relative ease compared to an

extensive experimental program. This method also allows comparison

of the forces on each particle within a packing. It then might be

possible to predict what type of powder size distribution would be

most effective for a given situation.

In order to simulate any process mathematically it is first

necessary to look at the problem and suggest a model that describes


- 92 -

the system. To enable a study of the spraying of particles onto a

substrate the flight of the particles from the gun to the substrate

must be described and this can be achieved by using a set of trajectory

equations.

By considering a force balance on any particular particle in

flight, taking into consideration all the forces acting on that

particle, differential equations are obtained. These describe the

variation of the position, velocity and acceleration of the particle

in terms of the system variables. Integration of these equations then

enables the movement of a particle to be followed until such time

that it lands onto the substrate, hits another particle or is over-

sprayed.

When dealing with any theoretical simulation technique one of

the biggest problems is to establish and recognise all the assumptions

that are required to enable a solution to be obtained. At the same

time analogy to the real system must be kept where possible in order

for the results to be meaningful. The initial stages of formation of

the simulation concept are kept as simple as possible but as investi-

gations proceed the basic model is continually developed to reflect the

real situation. '

This chapter describes the trajectory equations used to represent

the flight of particles onto a substrate and the assumptions employed.

5.2 Force Balance

To obtain the trajectory equation a force balance is applied to

a particle in flight.

Consider a particle heading towards the substrate as shown in

Figure 5.1. It can be seen that the forces acting on the particle

are:-
Figure 5.1. Forces Acting on a Particle in Flight

Field
Force
I Drag Force

Image Repulsive Forces

, .- , , .-
,
",.
/
" \ " \ / \ , " Earthed
-
I I \
I Ef); I .$ : I $;",.- I
$: Substrate
\
, \
I \ , ",.,/
~
I

,.. -
",., I
\
\. '
I
...,...~-
,
"- '"
Image Layer
'-- I
\
(f),'---
I

\

EfJ!, " ,
-- "
"- , _ /
,
\
I
\

EB,
\
, I
- 93 -

(i) Drag force due to relative motion with air

(H) Field force due to a charged particle being in an electric

field between the gun and target

(iii) Inertial force due to momentum of the particle

(iv) Gravity force (dependent on direction of spray and position of

substrate)

(v) Image force (attractive) due to induced image of the charged

particle itself in the substrate

(vi) Interparticulate electrostatic forces (repulsive and attractive)

due to charges in packing and their induced images.

These are all the forces that act on the particle and these must

be adequately described so that the mathematical model will represent

the system.

Initially the following assumptions will be. made for the

construction of the force balance:-

(i) The spraying of particles is such that they fall onto the

substrate with no effect due to gravity. Only a small area of

packing can be considered due to the number of particles

involved. Gravity effects have been initially ignored because

the significance of this force on the packing is dependent on

the direction of the plane of the substrate.

(ii) The simulation is carried out in 2 dimensions only. This enables

a pictorial representation of the packing to be obtained. Apart

from the fact that 3-dimensional analysis is more complex, there

is no easy method for representing the packing pictorially.

Reading a set of coordinates and figures does not readily give

an indication of the type of packing obtained and it is even


- 94 -

more difficult to detect minor differences and especially errors.

Calculation of porosity and analysis of size distribution through

the thickness of packing would be one of the ways an impression

could be gained.

(iii) The particles are spherical. Predicting the drag forces on a

particle taking into account the velocity profile of the air

stream is difficult in itself. Trying to account for the

various shapes of .the particles gives an extra degree of

difficulty. It would also create problems when deciding

which surface or edge first touched another particle or the

substrate. Considering spherical particles makes this much

simpler.

Other assumptions will be described as the individual components

of the force balance are considered. In the x and y directions,

therefore, a force balance can be applied to a particle in flight:-

Inertial force = Drag force + Image force + Field force + Attractive + Repulsive·
of particle Force due force due
(zero in y) to images to packed
of packing charged
particles

5.3 Charge on a Particle

As previously mentioned, the charge on a particle, with regards

to this process, plays a very important role with regards to the

packing of particles. Therefore, it is worth considering how the

particles are charged and the value of that charge. For the purpose

of this work only corona charging will be considered since this type

has been used in the experiments and is more common than tribocharging.

Tribocharging does not normally interfere with corona charging unless


- 95 -

the tribocharges produced are very high or are of opposite sign to

the corona electrode, whence they will quench the corona.

The field strength at which a corona begins to form is subject

to various parameters:-

(i) radius of wire of electrode, a

(ii) roughness factor for electrode, m

(iii) relative air density, p

This field strength is given by Peek to be

E = 3 x 10 6 m (p + 3.0/p/a) volts/m

Hence, the sharper the point of the electrode the less voltage is

required (F.W. Peek, 1929).

Particle charging takes place between the corona glow boundary

and the earthed substrate. The particles are charged by the electrons

and negative ions in that region. There are two types of charging

mechanism that take place, field and diffusion charging, each being

significant for a particular size range. For particles greater than

a radius of 0.5~m field charging is the dominant mechanism, hence

only this mechanism will be discussed. (S. Oglesby, 1978).

Ions will tend to move along the electric field lines and hence

collide with particles in the air stream. The charges are maintained

on the surface by the image force of the induced dipoles within the

particle. A particle will continue to be charged until the electro-

static field produced by these charges repel further ions. This is

known as the saturation charge. The electric field lines are shown

in figures 5.2 and 5.3 and show how the field lines intersect the

dielectric particles and hence will attract the ions.

If a conducting sphere of radius a, with a charge q, is

suspended in a uniform field Eo the field near the particle will be


-
----~. ------~----~ -------

---- --=::::-.---~---
Figure 5.2. Electric Field in the Vicinity of an Uncharged Particle

-------~---========
Figure 5.3. Electric Field in the Vicinity of a Charged Particle
- 96 -

due to two components:-

(i) Self field caused by the charge on the particle

(ii) Applied field, modified by the conducting particle

and is described as

where E = resultant field

e = angle between point on particle and field

£
o
= permittivity of free space

Since the electric field in the vicinity of the particle (E) is zero

when the saturation charge, qs' is reached (external field = field

due to charge) then

= l21Ta 2 £ E
o 0

e equals zero as the particle charge increases, hence cos e =1


(M.M. Pauthenier, 1932).

For a non conducting particle the equation is

=
12 [£!2 ]

where £ = dielectric constant of partiCle.


\
The charge on the particle can be related as a function of time

by the expression

q (t) =


o
where T = time constant described by T = p.b
1

where Pi = ionic charge density

and b = ionic mobility


- 97 -

-3 -4
The time constant is usually of the order of 10 or 10 second

and so saturation charge is normally reached within this time. The

velocity of the powder from the gun is approximately 2-5 m/sec and

the spray distance is normally about 20-30 cms. Hence the time of

flight is about 0.1 sec and thus there is ample time available to

reach saturation charge.

The dielectric constant of epoxy powder is normally about 4

and the electric field is usually 600 kV/m.: (Handbook of Chemistry

and Physics, 1980). From the equation it can be seen that a large

change in dielectric constant will only cause a small difference in

the overall charge. Other physical variables though can effect the

charge on a particle.

Climatic conditions, such as humidity and heat can cause

clustering of particles and hence effectively give lower charging

due to the increase in particle size. This could be responsible for

the small changes in measured Q/M with particle size. The charging

voltage, gun to target distance and powder feed rate can all affect

the acquisition of charge. The charging voltage and gun target

distance obviously alter the strength of the charging field thus

large spraying distances give lower charge. An increase in powder

feed rate can also have the same effect as the powder cloud can shield

the electric field or possibly quench the corona by space charge.

These ,effects are shown in Figure 5.4 (S. Wu, 1976).

5.4 Electric Field

The electric field plays a very important part in the e1ectro-

static coating process. It is the mechanism that enables the particles

to be charged and aids transport of the particles to the substrate.


Figure 5.4. Effects of Operational Variables on Particle Charge
Thermosetting Acrylic Powder (S. Wu)

Spray ChaFging
Powder feed Distance
rate 1 /min Voltage kV
cn
1.0 20 100
Powder Feed Rate

Spray
0.8 ·80
15

Charging
0.6 60 Voltage

10

0.4 40

5
0.2 2

o 0t-------~------~~-----.--------r-------,-------~
-.5 -1 -1.5 -2.5 -3

Particle Charge per mass (Coulomb/Gram x 10- 6 )


- 98 -

The magnitude of this latter effect is not certain at this time,

compared to the drag and other electrostatic forces acting on the

particle. However it is obvious that if the particle is not charged

it will not be retained if it reaches the target.

To obtain an equation which describes the electric field between

the gun and target it is necessary to consider the two main components:-

(i) Applied voltage at the gun electrode

(ii) Space charge effect due to the charged powder cloud.

The effect of the space charge is to suppress the effect of the field

near the corona but enhance it near the substrate. The magnitude

of this effect is dependent on the charge on the particles and the

powder output rate.

The system is considered to be a spherical coordinate system

with the origin at the electrode of the gun, as shown in Figure 5.5.

If we consider a substrate of radius R at a distance L from the

gun and that the substrate is effectively a spherical cap of radius L

then all quantities will be symmetrical about the axis in the conical

space bounded by the cap and electrode.

Hence the electric field is dependent on the space charge

density, Ps' and may be found by solving Poisson's equation which is

given by

+
I av =
r ar

where €o is the permittivity of free space

v is the voltage at r
and r is the radial distance from the electrode

Wu has described a solution to this equation. The effect of

ions that go to make up the space charge can be neglected since the

charged particles contribute 75% to 95% of the total.


Spray gun

~-
z

L
r

Earthed
Target
R«L)

--

.. Figure 5.5. Coordinate System for Analysing Field Between


Gun and Target
- 99 -

Integration gives

v =

Since the effect due to a layer of charged particles on the field is

greater than that due to space charge the system is considered to have

no space charge i.e. (ps = 0).

Therefore the boundary conditions are V(L) = Vp


V(O) = Vo
where Vp = voltage of powder layer

Vo = voltage on electrode of radius b

Hence V = Vo In (r/L) - Vp In (rib)


In (b/L)
-av
and since E = a.r
vo - V
P
E = r In (Llb)

The field at the electrode is obtained by letting r = b giving

V - V
=
o P
Eo
b In (Llb)

and thus bEo


E =
r

The variation of electric field, given by this expression, is shown

in Figure 5.6 (S. Wu, 1976).

5.5 The Forces

5.5.1 Drag Force

The force acting on a particle due to the difference in

velocities between itself and the fluid around it is dependent on

the Reynolds number of the particle:-


Figure 5.6. Variation of Electric Field with Distance from Spray Gun

Using WU Expression

2000
Spray distance = 0.2m
Applied voltage = -60 kV
Electrode radius = 0.005m

1600

1200

--...
>
El

...
."

....'"
'"
.~ ~800
......
(J

...'"
'"

400

O~--------~--------r-------~--------~---------r
0.04 0.08 0.12 0.16 0.2
Distance from gun (m)
- 100 -

->
V
reI I dp
Re =
p lJ

where IV->ll is the magnitude of the relative velocity between the


re
fluid and particle. The force on the particle can be obtained from

Newtons law:-

~D
=

where CD is the drag coefficient

If CD is plotted against the particle Reynolds number a graph

as in Figure 5.7 is obtained (J.M. Coulson et al., 1968). It can be

seen that for Reynolds numbers of less than 1 the graph is a straight

line. This region is described by Stokes Law where

24
C = Re
D

and from the equation for the force on a particle we have

24 lIdp2 ~
= Re 4 2
p

which becomes (substituting for Re )


P

For Reynolds numbers between 1 and 1000 the drag coefficient has been

described by Schiller and Naumann by the following equ,,~ion (L. Schiller

et al., 1933).

= 24 (1 + 0.15 R 0.687)
Rep ep

whence the equation for the drag force becomes

= 3111ldp r re l (1 + 0.15 Re 0.687)


p
Figure 5.7. Relationship Between CD and Rep for a Sphere

Region A Region B Region C Region D


C = 24 C = 24 (1+0.1.5Re 0.0/') CD.= 0.44 CD = 0.1
D - D -- p I
Re p Rep I

log (C D/2)

C /2 = 0.221 ---
D

C /2 = 0.051 --
D

5
Rep 0.2 Rep = 500 Rep = 2 x 10
log Rep
- 101 -

These are the two areas of interest since it is likely that the
-
particles in flight in the EPC process will have Reynolds numbers in

one of these two regions.

From experiments with the computer program, as described in

Chapter 6, the Reynolds number was typically found to be less than 1.

Hence Stokes Law can be applied and so the components of drag

force in the x and y direction are:-

=
dx
FDx 31TJldp (U: - dt)

and
= 31Tlldp (v - ~)
dt

The velocity components of the fluid (in this case air) will

be dependent on the position relative to the gun and substrate.

Therefore a model must be used that describes the variation of

these components with radial distance from the gun.

Initially the x component of the air velocity (perpendicular

to the substrate) was taken to be a constant and the y component

equal to zero. Therefore the air velocity component did not change

with position, and in effect the air flowed straight through the

substrate. This model was found unsatisfactory since it gave

particle velocities near the substrate far in" excess of those

expec~ed in r€ality.

Therefore the model representing flow around a flat plate has

been used. It is assumed that, after the jet of air has left the

gun, at some point between the gun and substrate the air flows

perpendicular to the target for a short period (i.e. parallel flow).

The air is then perpendicular to a flat plate and the flow is modelled
- 102.-

accordingly. This assumption is of course dependent on the size of

the substrate and the magnitude of the errors will be discussed

later.

5.5.2 Flow Model

To obtain the equations of the model for flow past a

flat plate there is a succession of basic transformations that have

to be carried out. The initial model is described by the wand z

planes. The transformations are shown in Figure 5.8.

Consider the flow past a circle of radius a of a parallel air

stream and transform this by a rotating transformation so that the

circle becomes a straight line and the flow is no: longer uniform

and parallel to it. (H.R. Valentine, 1967).


2
The flow past a cylinder is given by W = z1 + a and this is
Zl
then rotated by 90 0 by the transformation z2 = -iz l • The final
2
transformation Z = z2 + a gives the flow past a flat plate where
z2
the length of the plate is 4a. The z plane is then rotated again

through 90 0 by the transformation z': = iz.

The combination of these transformations gives:-


-z2
zl = i = l.Z2 (elimination of zl)

2 2
... + a
W = l.Z2 = i (z2 - ;....) (flow past a circle)
iZ2 z2

Z
w = - [z Z - azr
Zz

also
zz = [zz + : : ] z

therefore wz + zz = 4aZ

hence W = i1z 2-4a Z


Zl-plane
..-plane

Yz
Y
Z2-plane

4a

Overall, w =

Figure 5.8. Transformations used for Flow Model


- 103 -

If the half width, 1, is equal to 2a

Then

,
The final rotation back to the horizontal gives (z" = it)
w = 1z;.2 + 12

When the original uniform velocity (at infinite) of the fluid is U

then the transformation is

From streamline and potential theory we have that

w = ~ + iljJ

where ljJ = stream function (ljJ = f(x,y»

with ti = dljJ v =
-dljJ
dy dx

and $ = potential function

= d$ d~
with u y: = dy
dx

Since Z = x + iy

dw =
dw
di dx

dw a ($ + iljJ) ~ + ialjJ
and so
dz
=
ax
=
ax ax
= u - iv

= u - i~~"

By manipulation of this equation and the real and imaginary parts

the conversion can take place as follows :-

dW.' =
Uz
d"z
(z2 + 12)!
- 104 -

subs ti tuting z = x + iy
dw U(x + iy)
dz =
«x+iy)2+ l2)!

= U(x+iy) (x 2_y2 + 12 - 2i X)!


«x 2_y2 + 12)2 + 4x2y2) 1
i9
Putting in the form of Re

dw ( .)( i9)!
dz = U x + 1.:/: Re
«x2_y2+l2)2 + 4x2y2)!

where R = [(X 2 _ y2 + 12)2 + (4x2y2) J!

and 9 = tan
-1
[ x 2 -2:/:x
+ 12 - yJ

dw (x + iy)e i9 / 2
Hence
dz
= .;:rU

Substituting for e i9/2 = (cos 9/2 + i sin 9/2)

dw U
dz = ;:r (x + iy)(cos 9/2 + i sin 9/2)

hence dw U
dz = ;:r [x cos 9/2 - Y sin 9/2 + i ( x sin 9/2 + y cos 9/2) J

Since dw = u - iv
dz

the velocity components can be·o.seen to be

u = ~ [x cos 9/2 - y sin 9/2J


R!
and

v = -~ [x sin 9/2 + Y cos 9/2J


R!

where 9 and R are as described earlier and U is the uniform velocity.

u and v can now be substituted into the drag term equation.


- 105 -

Typical values of the x and y velocity components (u and v)

are shown at different distances away from the plate in Figures 5.9

and 5.10. It can be seen that a distance of O.lm away from the plate

the x component of the velocity is 45% of the uniform velocity. It

should be noted that this solution gives the air velocity as infinite

at the edge of the plate. However in these investigations the edge

of the plate can be considered as being at infinite since the area

of the plate considered is near the origin and is vety small. It

should also be mentioned that in these calculations the origin is

at the plate but in the main trajectory calculations the origin is

at the gun.

5.5.3 Field Force

The field force acting on a particle is dependent on the

position of the particle in the field and the charge on the particle.

i.e.
-+
= qE

The strength of the field, without space charge, at a radial

distance r has been seen to be

vo - Vp
Ex = r lrL(L/b)
cos e

in the.x direction.
x
Since cos e =
r
and r =

V
q • o
=
1n(L/b)

since Vp is assumed to be zero

and q is given by a previous equation.

The equation is similar for the field force in the y direction.


Figure 5.9. Air Velocity Model for Flow Past a Flat Plate

Uniform Velocity = I m/s

1.0

--
<Il

!! O.B

.......g
o
QJ

-a" 0.6
><
.....:
t;-
.... 0.4
o
o
.-<
~
...."
..: 0.2

o
o 0.1 0.2 0.3 0.4 0.5
Distance from Plate-X (m)
Figure 5.10. Air Velocity Model of Flow Past a Flat Plate

uniform Velocity = 1 m/s

0.5

a'"
~ 0.4

.....,"
0

......
A
QJ

0.3
Y 0.06

I><

..."
...t-
u
o.
....
0
y
~
...<... o.

Distance from Plate - X (m)


- 106 -

5.5.4 Interparticulate Forces

There are three types of forces that have to be

considered here:-

(i) The image force of the oncoming particle which only has a

component in the x direction. The image is due to the induced

dipole charge formed in the substrate.

(ii) The repulsive forces due to the packed particles on the substrate

- x and y components. The repulsive force is due to the field

exerted by the particles of the·same charge sign and the force

is dependent on the distance between the particles and the

charge on each of them.

(iii) The attractive forces due to the attraction of the oncoming

particle and the images of the packed particles on the substrate,

again induced.

The convention used for the summation of these forces is shown

in Figure 5.11. The equations formed are based on the cartesian

coordinate system in 2-dimensions using x as the direction from the

gun to the substrate and y being the direction that the target lies

in. The origin of the two axes is the needle point of the electrode.

(Note: different coordinate system to flow model).

An oncoming particle, Po is considered approaching the substrate


on which there is a packing of particles already present. The

particles on the substrate, Pi' the images, Pj' and the image of the

oncoming particle, Pk' must all be considered.

The force between two particles is given by the equation

F'=
Figure 5.11. Convention used for Calculation of Inter-particulate Forces

,-- ... ,
,. - ... ',
,
.... -I
'-
. .'
I

Ro. ,

, ,
I \
., ,-'I

Spray Gun


L
.. Substrate
- 107 -

and therefore the component in the x direction is given by

F =
x
(S. Oglesby, 1978)

For a monosized distribution it would be assumed that the

charge on all particles was the same. Therefore for the force due

to a particle in the packed layer (where q1 = qZ)


Z
-qo
=
4TTE: R • Z
o 0],

but for the force due to an induced image (where q1 = -qZ)

=
4TTE: R ~
o oJ

5.5.6 Image Force

The image force due to the induced image of the oncoming

particle is given by

= (no y component)
2
4TTE: o R0 k

where (L - xo ) is the distance between the

particle and the substrate

Therefore =

and it is assumed that q =-q


o k

5.5.7 Repulsive Force

The repulsive force due to the packed particles is given

in a similar way. It is assumed that the charge on the packed

particles does not leak away to any appreciable extent. In all

these calculations it is assumed that the charge is concentrated at

the centre of the particle. The charge is therefore, in the case of


- 108 -

monosized particles, exactly equal to that of the oncoming particle.

Hence
r ~
=

and in the x direction

-qoqi cos Si
F·~x =
41lE R •
o o~

From Figure 5.11 it can be seen that

(x. - x )
~ 0
x.~ - x 0
cos 8. = =
~ R.
o~
[ (x.
~
- x )
o
2+ (y. - y )
~ 0
2J !

Therefore
(x.~ - x o )
F·~x =
+ (Yi -

The·tota1 force acting on the particle in flight will therefore be the

sum of all such individual forces

i
= E . 3/
1 + (y. - Y )2
~ 0
J 2

and similarly for the y direction.

5.5.8 Attractive Forces

The attractive forces due to the induced images of the

packed particles on the oncoming particle is given by a similar

expression.

In the x direction

j
= E cos 8
1

which simplifies to
,

- 109 -

(x.-x )
= ~ J 0
1 (R .)3/2
OJ

Combining all these forces gives the final differential equations

describing the trajectory of the particle in flight. These are

summarised in Table 5.1.

5.6 Other Representations of Forces

The calculation of the repulsive and attractive forces due to

the charged particle packing and the equations used have been

described. It is found that the calculation of all such individual

forces associated with a packing is very time consuming for packings

in excess of 50 particles. An upper limit of 300 particles considered


is imposed due to the available computing time. These calculations

are necessary so that a representation of the packing can -be obtained

and an insight into the effects of particle size, packing thickness

etc on the overall packing can be gained.

However similar investigations can be made by considering

particles heading towards a uniform and even packing. A theoretical

layer of charge, with a given charge density, can be considered to

exist on the substrate.

By calculating the charge on a particle of a certain size and

assuming a porosity of packing the charge per unit length and unit

thickness (2 dimensional case) can be calculated. In effect this

acts as a layer of particles of a given thickness but with a width

of one particle diameter (in the 3rd dimension). The charge density

in this case has units of coulomb/square metre. The convention used

for these theoretical considerations is shown in Figure 5.12.


Table 5.1. Summary of Forces on Particle

Force x-direction . y-direction


x
o
Field FE x = q
0 1n(L/b) • R 312 = qo 1n (Llb)'
o

Drag Re < 1

FDx = 311\ld o( (j-~~ )

Re > 1

FDx = 31I\l d0 ..u- dK\[l 68i1


( d~ +. ° l5Re°.
P J FDy = 311\ldO(v-adVVt)
=- [l+O.l5Re°p.687]

where

u=U (x cosS/2-y sinS/2)


:r
R
0 0 :r
v=U (x sinS/2+y cosS/2)
0 . 0
·R

where R and S are given in the text


q 2
F = _ _0,,-_ _
Image
I 16 E (L-x )2
o 0
°
• 1
Repulsive FR =~-q q. Ix.-x I FR =E-q q.
x 1 0 1 . 1 0 Y1 0 1 . IYcYol
411EO R. 3/2 411E 0
R. 3/2
01 01

FA =~-q q.
J
Attractive I xj-xo I FAy=i-qoqj .IYj-Yol
x 1 ..:£..:J..
411E
.
0 Roj 3/2 4111;
0 R 3/2
oj
Spray Gun

if
0

I
I
Xo

GT
I

0-
Yo
0
o!
I

I
0

I
I

t I Charged Layer
I
(

----- -------

• •

Figure 5.12. Convention used for Theoretical Charged Layer Calculations


.,

- 110 -

The force between two charged particles (of the same sign) is

and in the x direction the force component is given by

FRx =

In considering a uniform charged layer the total force on the

oncoming particle is the sum of the forces due to all points in the

charged layer. If the layer is on a plate of thickness t and width

z ~. the overall force is

i GT (x-x )
= f
-i GT-t
f [ z 0 J
3 2 dxdy
«x-x) + (y-y )Z) /
o 0

where GT is the distance between gun and target, and Qa 1S the charge

density in terms of coulombs per unit area.

Putting
c

and substituting
x' = x - x0
y' = y - Yo

and xl = GT-t-x 0
X
z = GT-x0
YI = -(Hy 0 )
yz = i-y 0

the equation becomes


- 111 -

. Y2 X2[
= - C f f x' 3/2]dx'dY '
Yl xl + y,2)

Hence

= -
-1
C f Y2 dy , [ ---;;---=--';;" 1 +
2 ,2 2
Yl (X
2 + Y )

= - C [ - log(y' +!x~ + y,2) + log(y' + !xi + y' 2JY2


Yl

= - C

which gives

(Y2 + !xi 2
+ Y2) (Yl + !x~ yi) ]
+
FRx = - Clog [
(Y2 + !x~+ 2
Y2) (Y l + !xi yi)
+

substituting back

= - C

If x o = GT - O.5t .. FRx =0
Similarly in the y direction

i GT
= f f
-i, GT-t
- 112 -

Thus
Y2 x2
FR
Y
= -C f f
Yl xl
[ (X ,2 +Yy'2)3/2
I ] dx'dy'

using the same subs ti tu tions, and this becomes

(xl+~~+xi) (x2+4i+x22)
FRy = -Clog [ ]
x2+~~+X~ (x l +.0i+ x l
2

Substituting back again the equation for the force on a particle in

the y direction due to a charged layer is:-

(GT-x )+~i+y )2+(GT-x )2 J


FRy = -Clog [ _0 ----;::o==:::::=o~
. /22
(GT-xo -t)+{(i+y 0 ) +(GT-x 0 -t)

Hence if Yo = 0, i.e. the particle is travelling along the x axis,

FRy = 0.

The force on the particle in the x and y directions due to the

induced image layer is obtained by exactly the same derivation except

that

= GT-xo

and = GT-xo +t

and therefore similar equations are obtained.

Also, for the image force the sign of the constant is reversed

due to the image charge layer being of opposite sign

hence C =

These equations therefore allow the simple calculation of the

force on a particle in any position due to a charged layer of chosen

thickness, charge density and size. This can then replace the section
- 113 -

of the computer program that calculates interpartic1e forces and can

show the effects of different thicknesses of layers on the flight

and packing of particles. Other effects, such as that due to changes

in transport air velocity and particle charge can also be observed.

5.7 Comparison with the 3 Dimensional Case

The equations described to represent forces acting on a particle

in flight consider the 2 dimensional case only. This simulation

could be considered as particles being sprayed from a gun in one

plane only and landing on a very narrow substrate, extending in the

same plane. This in itself does not seem an unreasonable qualification

but even if the particles were sprayed in a thin plane the electro-

static forces would give rise to a 3 dimensional effect in the case

of different sized particles. This is due to their centres being in

slightly different positions. In the computer program the charge on

the particles is considered to be concentrated at the centre of each

particle. In reality, of course, they are spread over the surface of

the particle, hence the dependence of saturation charge on surface

area.

If the corresponding equations previously derived for the

repulsive and attractive forces due to a charge layer are extended

for a plate of infinite width a relationship is obtained for 2

dimensions. For 3 dimensions the total force on a particle can be

shown to be zero if an infinite charge layer, and related image layer

are considered.

Although the forces due to drag and the electric field can be

symmetric (in 3 dimensions) about the x axis, the electrostatic forces

cannot. The random landing of particles and subsequent effect on


- 114 -

trajectories due to packed particles leaves no symmetry about the

x axis. Therefore any particle trajectory is dependent on all the

particles around it and will therefore be forced in one direction

or another because of the resultant of all the individual forces.

In this respect the 2-dimensional case must be considered as

a large simplification of the true situation. However it is hoped

that the simulations gained by this technique do show some of the

effects in packing .·that .can occur, albeit only quaE tatively.

5. 8 Summary.

The equations of motion have been developed (in 2 dimensions)

for the calculation of the trajectories of individual particles in

flight. These equations take into account the electrostatic forces

due to all the particles packed on the substrate. Although the

summation of these forces requires a.vast amount of calculation when

more than 50 particles have landed, it is necessary so that effects

close to the substrate can be observed.

Various assumptions have been made to arrive at the form of

these equations and although they greatly simplify the problem it is

thought that this treatment will still provide valuable information.

In the next chapter the computer program, the way in which it

works and the experiments carried out will be described.


CHAPTER 6

COMPUTER SIMULATION

6.1 Introduction

6.2 Nag Routines

6.3 Program Description

6.3.1 Problems encountered

6.3.2 Main subroutine

6.3.3 Function subroutine

6.4 Testing of Program

6.4.1 Drag coefficient

6.4.2 ·Starting velocities

6.4.3 Error testing

6.4.4 Particle splaying

6.4.5 Further effects

6.5 Experimental Results and Discussion

6.5.1 No field force

6.5.2 Monosized powder

6.5.3 Size distributed powders

6.5.4 Theoretical charged layer

6.5.5 Effect of gravity

6.6 Conclusions

6.7 Summary
- 115 -

6.1 Introduction

In the previous chapter- 'the derivation of the trajectory

equations via the force balance on a particle was discussed. The

complex nature of the equations is such that straightforward

integration by analytical methods is not possible. Numerical

methods have therefore to be used. A computer program has been

written and employed to enable the calculation of solutions quickly

and efficiently.

For this purpose a computer library routine which solves

ordinary differential equations has been included in the program

that allows control of the integration procedure and will minimise

errors to desired levels. There are various routines available from

the NAG library which integrate differential equations in various

,ways. Three such routines have been used, the final one being most

suitable for this problem. Restrictions of computing time, accuracy,

and type of equations involved led to this choice of routine.

During the development of the program four different computers,

of varying size and location, were employed, all of them using the

Fortran language. The basis of the program was built around a set

of simplifying assumptions, separate to those used in the formation

of trajectory equations. These assumptions were:-

(1) The particles are spherical

(2) Only one particle is in flight at any instant in time

(3) Each particle, for a given run, starts at a random point at

the same distance from the substrate with identical velocity

components.
- 116 -

(4) No effect due to gravity is considered

(5) When a particle touches another particle or the substrate it

remains stationary.

(6) Each particle carries its equilibrium charge (according to

its size).

(7) The particles do not lose any of their charge on contact with

the packing.

6.2 NAG Routines

One of the simplest NAG routines available for numerical

integration of ordinary differential equations (Type D01ABF) was

first used to solve the trajectory equations (Nottingham Algorithms'

Group, 1973). This is based on Merson' s form of the Runge-Kutta

method of numerical integration. It was discovered that no solution

could be obtained using this routine since to work within prescribed

error limits an infinitely small step length was required. This is

characteristic of a stiff system of equations. Stiff systems of

equations are those which have rapidly decaying transient solutions

(or whose Jacobian matrices have eigenvalues with large negative

real parts). The degree of stiffness (given as the ratio of the

largest to smallest real parts, in magnitude) may vary across the

range of integration.

Gear's method of integration is more efficient for stiff systems

and therefore the NAG routine D02AJF was employed (NAGFLIB, 1976).

Although solutions were achieved with this routine, it was found that

the program became very time consuming (in terms of c.p.u. time) when

dealing with large numbers of particles. A newer, more efficient

routine, D02QBF, was therefore used and it was found that this worked
- 117 -

much quicker than its predecessor. This routine integrates a stiff

system of ordinary differential equations also using Gear's method

but has a greater variety of facilities for interrupting the

calculation. The routine, and its use, is fairly complex (NAGFLIB,

1978).

D02QBF advances the solution of equations

z.1 ' =

where i = 1,2 . t- • • N

over an integration range using a number of steps of a variable-

order variable-step Gear method. The stepsize and order are chosen

so that the local error estimate is kept within a prescribed error

bound. The way in which the size of the error bound was chosen is

described later.

The equations are contained in a subroutine which is called

by the NAG routine after initial values for the variables have been

set. Interrupt procedures allow for the access of information during

the course of the integration.

The integration can also be stopped at any point in time by

such interruptions, a ·.facility which was not available in earlier

routines.

The information obtained also provides enough data for accurate

interpolation between integration points using other suitable NAG

routines. Since the routine only solves first order differential

equations it is necessary to describe the trajectory by four equations.

The subjects of the equations are given as follows:


- 118 -

dZ dZx
l
dt = Z = f l {t,Zl,ZZ,Z3,Z4)
dt

dZ
Z dx
dt = dt = f l {t,Zl,ZZ,Z3,Z4)

dZ
3
dt
= .izZ = f l {t,Zl,ZZ,Z3,Z4)
dt

dZ
4
dt = ~
dt = f l {t,Zl,ZZ,Z3,Z4)

hence Zz = x, Z4 = Y and Zl and Z3 represent the velocity of the

particle in the x and y directions respectively.

The program is initially supplied with:-

(l) initial values of t and Z.


~

(Z) the length of the integration range

(3) the number of differential equations

(4) a set of control variables

(5 ) the Error bound

(6) interrupt conditions

(7) a failure parameter and type of error test parameter

and (8) working space (dimensioned spare variables)

The routine chooses a step size which just keeps within the

error bound and then evaluates the derivatives in terms of T, Zl' ZZ'

Z3 and Z4. The trajectory is calculated and continues until the

particle is close to the substrate. The routine is then interrupted

after each integration step to compare the particle position relative

to other stationary packed particles and the substrate.


- 119 -

The NAG routine has not been internally reprogrammed in any

way and its working will not be described here. Literature is

available'on the program and the method of 'integration (NAGFLIB,

1978; G. Hall et aI, 1976; A.C. Hindmarsh, 1974; C.W. Gear, 1971).

If the program fails due to an error in the NAG routine

(IFAIL > 0 on exit) a set of information variables are printed which

describe the current status of the routine. This includes the type

of error encountered. Common errors found were:-

(1) integration range finished before particle was found to have

landed.

(2) start parameters set wrongly

(3) stepsize too small for solution to be obtained.

The number of successful/unsuccessful integration steps, maximum/

minimum stepsize used, last integration point calculated and the

maximum values of the differentials are just some of the outputs

available. Although the routine is complex and cons1derable time is

required to understand it, it is very versatile when handled

efficiently.

The mode of operation of the program will now be described.

6'.3 Program Description

Apart from the NAG routine (which is called from the Library)

the program is split into two main sections:-

(1) Main subroutine which handles the setting of variables, initial

conditions and determination of when the flight of particle

is over (i.e. it has hit another particle or the substrate).

(2) Function subroutine which calculates the individual force

components which are combined to form the differential equations.


- 120 -

The overall scheme of the program is shown in Figure 6.1 which

gives a simplified flow chart.

6.3.1 Problems Encountered

The computation time for calculating the sum of all

individual interparticu1ate forces (electrostatic, including images)

is not in itself too large, but when this procedure is carried out

many thousands of times the restraints on computer time becomes

important. This requires that the program is written in such a way

that there is no wasteful computation. Careful use of variables

and arrays and avoidance of repeated operations are just some of the

considerations.

In one of its earliest forms using D02AJF the program took

just under one hour of c.p.u. time on the Loughborough University

ICL 19048 computer to calculate the trajectories of 50 particles.

For the next fifty particles the c.p.u. time was nearly four times

as long and since this was nearing the available computer users time.

alternative methods were sought.

Using a computer link from Loughborough via Nottingham a

program file was created and used to run programs on the CDC 7600 at

the Manchester computer centre. This very large main frame machine

allowed a much larger number of trajectories to be considered.

However, due to the unavoidable low priority status being used

development was extremely slow. The computer plots took approximately

two weeks from the day of program execution to return to Loughborough.

The latest version of the NAG routine, D02QBF, was written into

a program initially developed on the DEC 10 at ICI plc, Runcorn with

the cooperation of the Process Technology Group. The program was

streamlined with particular attention being paid to the assignment of


Figure 6.1. Simplified Flow Chart of Program Procedure

MAIN subroutine
!
Dimension Variables
!
Input Data
1
Call plotter, draw axes and label
!
Set Constants
!
Print initial data

...----1 Calculate random start point


1
1---- Call NAG G1i!15CAF
Calculate particle diameter ______________~~
1· Call NAG GI)I5CAF
Set boundary conditions
l
Calculate constants

1
Set flags for NAG routine
!
Call NAG routine~________________,
, . - - - - -..., DI)I2QBF

plot & print trajectory 14------------------{~N~A~G~r]o;u~tji~n~e~~------------~


Check particle position Call FUNCTION subroutine
relevant to other
particles 1
1 Calculate all forces
~----------Compare to substrate
1
!
Calculate final
Set up differential
------"
equations
position .....----'
1
~------------------- Plot final position
l
STOP
- 121 -

variables for calculation constants where parts of equations were

continually repeated. The combination of this routine and programming

allowed 150 trajectories to be calculated in 1 hour and 350 in

approximately three and a half hours.

Moving the program onto ICI's VAX 11/780 computer (manufactured

by OEC) also enabled the use of the BENSON electrostatic plotter which

reduced plotting time to one quarter and gave 'on the spot' print-

outs. VAX stands· for Virtual Address eXtension and is an extension

of the POP-ll computers (O.R. Mason et aI, 1978). It is a very

powerful minicomputer which is suited to having a small number of

users with large programs.

In terms of software, the determination of the final position

of particles caused many problems. As a particle approaches the

substrate it is possible that it can collide with any other stationary

particle or the substrate. Each time a check is made on its position

relative to all others many operations have to be carried out. Careful

programming was necessary in order to keep this to a minimum and a

variety of methods were considered. Finally, the most simple method

was chosen. Once a particle in flight reaches the start of the

powder layer the distance between the centres of particles are compared

to the addition of the two radii. This is carried out after each

integration step. More precise details are given later.

Error testing is also important. Not only must the accuracy of

the results be checked but invisible errors (i.e. typing and input

errors) must be detected by the user. Hand computations are therefore

carried out to check the magnitude of forces, velocities, correct

signs etc. such that the trends observed agree to an extent with those

expected.
- 122 -

6.3.2 Main Subroutine

From Figure 6.1 it can be seen that the main subroutine

is responsible for all of the operations apart from calculating the

differential equations and the integration procedure.

The input of data comes from a data file unique to each run

condition. Basically the initial velocity, type of powder and start

positions are read in via this file. This command file also controls

output and assigns the name of the data.file for the plotter.

The plotter file is opened and data is sent for the drawing

of axes, labelling and printing of conditions. The GIND-F language

is used for drawing (GINOF, 1976). Following input of the data on

file, all of the initial conditions, constants and printouts are

completed before setting the NAG routine control variables. A random

number generator, NAG G0SCAS/G0SCBF, is used to give random numbers

for the calculation of start position and particle size (if not

monosize distribution) (NAGFLIB, 1977a; NAGFLIB, 1977b). If a size

distribution is being considered the random number is used in connection

with a number size distribution to obtain a particle diameter which,

together with all other particle sizes, is representative of the

dis tribu tion.

The mass, charge and other calculation constants can then be

obtained.

The NAG controls are set such that initiallY the integration

will run without interruption until a specified value in the x

direction is reached.

Interruption then occurs at a suitable distance away from the

substrate where the plotting of the trajectory commences. Once this

point has been reached the integration is interrupted after each


- 123 -

integration step. Control variables are therefore readjusted

accordingly.

Each time the integration is interrupted control is given back

to the main subroutine from NAG. If the particle has not reached

a distance equal to that of the highest particle in the packing the

particles position is simply recorded, plotted and the program returns

to NAG. The trajectory progress is printed if desired. Individual

component forces printout is also available if required.

If the particle is within the layer then it follows a program

seqeunce as summarised in Figure 6.2.

If a failure is detected by the NAG routine then a controlled

stop .is carried out. This means that with an IFAIL > 0 on exit from

the NAG routine the program prints the control variables which gives

valuable' information for the assessment of the cause' _of :the 'error.

6.3.3 Function Subroutine

This routine formulates the differential equationsc.and

therefore calculates the forces acting on a particle in the x and y

directions. All of the statements in this routine have been


',-
simplified as far as possible so that the c.p.u. time is reduced.

For each trajectory approximately 250 integration steps are used and

for each call of this subroutine approximately 600 operations are

carried out for each comparison with a 100 particle packing. Therefore
9
for 350 particles at least 10 operations are performed in this

subroutine alone. Wherever calculation constants can be used for

each new particle, calculation is done in the main subroutine. A

summary of the routine is given in Figure 6.3.


Figure 6.2. Simplified Flow Chart for MAIN Subroutine (Stopping Section)

New coordinates from NAG routine


!
Check for failure in NAG

Check
!
time of integration and y position ______________-,

J
Print trajectory or forces
!
Compare particle position with that
of uppermost particle if out of

r-________~Compare
l
position with other stationary
bounds

J particles in reverse order


if collision
compare I
and record -
highest
~
check rest to see
1
Check with Record
t
Yes -
if another one
~ sUbstrate~ if not dummy
has b een h·l.t reached position
layer

1 if collision
record as old position..J
No l
plot trajectory
J
calculate position
by interpolation
~
recall NAG routine

calculate position
on substrate

~l
interpolate other variables
~
plot final position and finish trajectory
l
check to see if last particle is highest
~
Print position and if hit, particle number

!
RESTART LOOP _ -_ _ _ _ _ _ _ _ _ _...J
Figure 6.3. Simplified Flow Chart of FUNCTION Subroutine

From NAG routine

1
Calculate x and yair velocities
using model

!
Calculate Particle Reynolds number

!
Calculate x and y components of drag force

!
Calculate image force

1
Calculate x and y components of field force

1
...---Calculate distance between oncoming and stationary
particle and its image

1
Calculate angles between image and particle
if
1st
particle
next
particle
1
Calculate attractive and repulsive forces

t,_ _ _ _ _ _ _ _
!
Sum with those already calculated

~
1
__________________~Develop 4 differential equations

l
Return to NAG routine
,
- 124 -

The use of arrays is purposely limited. On entry to this

subroutine the x and y coordinates and velocities are transferred

to single variables since they are used many times and this reduces

computing time. The air velocities at the new coordinates are

calculated using the flow model and the calculation of forces f').1To'(Ys.

Eventually the"differential equations are produced:-

In x direction

F(l) = (Drag force + field force + Image force + Particulate forces)/

mass of particle = d 2x
dt 2

F(2) = Velocity of particle = dx


dt

In y direction

F(3) = (Drag force + field force + particulate forces)/Mass of

particle = .62
dt

F (4) = Velocity of particle = 2z


dt
The program is shown in Figures Dl (a-p) with explanations of

each variable used in the program given in Appendix D. For calculations

considering a layer of charge the appropriate equations are substituted

into the FUNCTION subroutine for the particle summation loop. The

relevant section of program is shown in Figures D2(a-b). A typical

printout from a run of 10 particles,all 25~m diameter, is given in

Figures D3(a-e) with suitable explanations.


- 125 -

6.4 Testing of Program

In order to check that some of the assumptions and method of

programming gives meaningful results the program has to be tested ~

in several ways.

6.4.1 Drag Coefficient

The drag coefficient used in the determination of the

drag force is dependent on the size of the particle Reynold's number.

If Rep < 1 then Stokes Law applies but if it, is greater than 1

modification to the drag coefficient is required. Therefore, in the

trajectory printout, the particle Reyno1ds number is given after

each integration step. Typically the Reyno1ds number for a 10~m

particle is less than 0.3, but it does depend on the size of the

particle and the relative velocity. However, as the particle

approaches the plate the Reyno1ds number is often less than 0.001.

Hence the assumption that

24
= appears reasonable.
Re

Typical values of Re obtained from trajectories of partic1es·,during


p
which no field is applied are shown in Table 6.1.

The maximum Reyno1ds numbers occur at the start of the trajectory

when the velocities of the particle are set. Therefore the relative

velocity is maximum at this time. The minimum values are usually

obtained when the particle is close to the substrate. These figures

show that, although the initial starting conditions are an approxi-

mation, in the area of interest (close to the substrate) Stokes

Law wi 11 app 1y •
Table 6.1. Reyno1ds Numbers of Particles During Flight Without
Applied Field

Particle Particle Particle


Diameter = Diameter = Diameter =
lO)lm 25)lm 50)lm

Particle Re t Re. t
p p Re t
Velocity m/s p

1 <0.0001 0.0007 0.0061

3 0.0004 0.0068 0.0668

5 0.0012 0.0205 0.2636

7 0.0022 0.0395 1. 377*

10 0.0047 0.0930 6.4410

t Re taken for particle at a distance of 1 cm from substrate.


p

No particles. landed on substrate except

*where 8 landed

o
where all 10 landed
- 126 -

6.4.2 Starting Velocities

Due to the very complex nature of the air flow exit the

spray gun between the electrode and target, whose representation

is out of the scope of this work, a simplified model is used. In

the last chapter formulae were developed for representing the air

flow around a flat plate. For the calculation of trajectories

leading up to the packing of particles on a plate it is necessary

to use some initial starting conditions.

On assuming a uniform gas velocity heading towards the plate

it is possible using the equations of Chapter 5 to calculate the

component air velocities at any position. It has been assumed that

the trajectories start mid-way between gun and target. The initial

velocities of the particles are assumed to be equal to the uniform

gas velocity, as used in the flow model, in the x direction and zero

in the y direction. For a uniform velocity of 1 m/s this would mean

a discrepancy in the x direction air and particle velocities of

approximately 0.5 m/so

In order that these starting conditions did not cause severe

errors in the trajectory calculation a computer program was formed

in which only the drag and inertia forces on the particles were

considered. A run using varying sized particles then showed how the

particle velocity settled after the initial start.

This experiment can also be considered in a theoretical

manner. Considering the forces acting on the particle in the x

direction (neglecting gravity) and assuming that the air velocity

is not dependent on x:-

M dV
rel
dt
= -3~~d V
p reI
(assuming Stokes' Law applies)
- 127 -

Integration gives
= e
-tiT

M
where T =
31TlIdp

The time taken for the relative velocity to be halved is given by

t = Tln2.

For a 10um diameter particle in air this time would be

2.3 x 10- 5 secs. To·'reduce the relative velocity to within 1% of

the initial relative velocity would therefore take approximately

seven of these times (i.e. 0.2 msecs). The distance travelled in this

time (assuming average V = 0.75 m/s) would be 120um.


rel
This agrees well with results from the computer program. For ten

different particles,. from slightly different y starting positions,

the distance taken for the particle to settle to within 1% of the air

velocity was between 200um and 2900um (16um diam). The times taken

were between 0.4 msecs and 3.5 msecs. The results are shown in

Table 6.2.

It therefore appears that these starting conditions are not

unreasonable since the particles reach a steady state generally within

0.5 - 3% of their total flight distance. It does depend though on the

velocity and size of particles as can be seen from the table. Since

the area of interest is close to the substrate these results suggest

that any approximate starting velocity conditions could be used whilst

still maintaining accuracy in the observation area.


Table 6.2. Effect of Initial Starting Conditions on Particle Trajectory
Without Field

Diameter of Flight time to particle velocity Distance


Particle within 1% of air velocity Travelled
(\lm) (mi 11 iseconds) (\lm)

6.7 0.675 379

6.7 0.674 378

5.4 0.363 211

8.2 0.860 500


-
9.3 1.14 757

9.7 1.21 697

10.4 1.36 892

16.1 .3.43 2629

8.7 0.96 657

5.8 0.42 246

TOL = 10- 7

Uniform Air Velocity =1 m/s


- 128 -

6.4.3 Error Testing - Setting of Tolerance Bound

The NAG routine works in such a way that it chooses the

largest integration step possible which just keeps the estimated errors

within a tolerance bound provided by the user. The larger the

tolerance the larger the integration steps since less accuracy is

required. Although this means faster computation the results may not

represent the true system. In severe cases the particles can 'jump'

across the whole of the particle layer and the substrate by a

considerable amount. This effect is caused by the very large

integration steps being used in the initial, smoother running part of

the integration range. The influence on the particles of the charged

layer, in this situation, is not obserVed due to the very slow

response of the system to its presence. The stopping section of the

program is also reliant on the fact that the step size is small

enough that particles cannot be 'jumped over' within the packing.

A series of computer experiments were carried out in which the

effects of changing tolerance error were observed. For each run the

start conditions remained the same but the final positions were

recorded for each of 10 particles. The tolerance error chosen for

further experiments was that employed when the final positions of

particles had not altered significantlY compared to the next largest

tolerance error.

The convergence of solutions to a constant value with changes

in tolerance bound can be clearly seen in Figures 6.4 to 6.6. The


7
tolerance error chosen on this basis was 10- • The figures show how

a consistent solution is produced as the tolerance bound is decreased.

In general it appears that larger particles do not need as small

tolerance errors as small ones. Figure 6.7 compares the plotted


Figure 6.4. Calculated Y Position vs. Tolerance Bound

Uniform Velocity =1 mts

.-.-
Key
3000
+- Particle No 2 (6.7)lm)
Particle No 3 (5.4)lm)
Particle No 7 (10.4)lm)
0- Particle No 8 (16.l)lm)
0 0- Particle No 10 (5.8)lm)

2500

2000

e
~
'"
.....,"0
.... 1500
"0
,..'"

1000

500
-"'-- -.--~.. ----.----.;----.
o

.~
-1 -2
". i-.-. . . .
-3
---0

O+-~~~--~----~---.----'-----r----r----'
o -4 -5 -6 -7 -8
0 8===g---g---g

Log (Tol)
lO
Figure 6.5. Calculated Y Position vs. Tolerance Bound

Uniform Ve loci ty = 5~Ji1/. s

-r - Particle Number 2
11 - Particle Number 3
3000 4t-- Particle Number 7
Cl - Particle Number 8
C> - Particle Number 10

2500

Il--·~ 11----11----11----11----11
2000

6

~
"-

1500
...."0
...."" 4t
'"0
"",..

1000 .-..I_C>--C>----O-- C>---C>-- C> ---- C>


C>

0---'<-
+
500
4t _ _ 4t - - 4 t - - - 4 t - - 4 t - - - . •

O;----,-----~----~---~----r_---~----~--~
o -1 -3 -5 -6 -8
Log (Tol)
lO
Figure 6.6. Calculated Y Position vs. Tolerance Bound

Uniform Velocity = 10 m/s

Key
600
;- - Particle Number 2
11 - Particle Number 3
e - :-Particle Number 7
Cl - Particle Number 8
e> - Particle Number 10

500

400 ----------._------.-----.
=
~

~
""
300
......."0
....
'"
0
e
0-
><

2000 e>............
O-----e>--e>--e>--O--e>

+----
1000 -+--+--~--+--+--+
0---\1 - - - 0 0 - - - - - 0 - - 0 0 C c
--e--e e • • e

0-r-----,------r-----~----~----~----~r_----~----~
-3 -4 -5 -6 -7
Log (Tol)
lO
Figure 6.7. Comparison of Particle Trajectories tor Differenc
Tolerance Bounds (U = 5 m/s)

" "

,
'.

Tolerance
Bound = 0.01

Tolerance
Bound = 1 x-lO- 8
- 129 -

trajectories for three of the tolerances. For large bounds the

trajectory consists of long straight line steps due to the limited

accuracy of integration. For a bound of 10- 8 the trajectories are

extremely smooth and integration steps are in the order of 10- 5

seconds, which typically represents a movement of <lOum. (This is

dependent on the size of particle and velocities but it is found that

the distance between steps decreases on nearing the substrate or

packing).

6.4.4 Particle 'Splaying'

From a series of test runs it was found that changes in

air transport velocity caused large changes in the area covered. This

meant that comparison of packings was difficult since a different

number/unit area of particles were being sprayed for each situation.

The widths of the spray close to the substrate were therefore measured

and compared (as shown in Table',6.3) for various conditions. It was

observed that the width increased as velocity increased. However,

for very large particles (50um) at high velocities their inertia

caused the particles not to follow the streamlines and hence in this

case the width of spray actually decreased.

The start lengths were rescaled so that the width of spray was

approximately the same in each experiment. The predicted start

widths, shown in Table 6.4, were therefore used to obtain equal areas

of spraying so that better comparisons could be made.

6.4.5 Further Effects

When the packing on the substrate is built up it is

sometimes found that particles, particularly small ones, can either

(i) be repelled a long distance away from the spray zone or (ii) reach

an equilibrium position in space above the packing and hence have

very long flight times (in excess of 0.5 s). Since in both these
Table 6.3. Effect of Air Velocity on Splaying of Particles of
Different Sizes Starting at Identical Positions

Particle size, Ilm 10 17.5 25 25 50 Dist

Start length * Ilm 300 400 500 1000 1000 700

Velocity m/s Width of band t , Ilms

1.0 1350 1350 1480 2960 2210 4420

3.0 2610 2280 2120 4240 2270 <6000

5.0 3580 2970 2600 5200 2090 >6000

7.0 4364 3450 2820 - 1760 >6000

10.0 5450 4090 3060 - 1520 >6000

t measured at a distance of l500l!m from substrate

* width of start band O.lm from substrate


Table 6.4. Predicted Start Widths for Equal Areas of Coverage
with Variation in Operating·Conditions

Velocity
Start length t , llm
m/s

1.0 430 580 680 890 310

3.0 230 350 460 870 200

5.0 170 270 380 940 130

7.0 140 230 350 1100 100*

10.0 110 190 320 130 100 *

Particle Dist.
10 17.5 25 50
Size, llm

t calcaulated on basis of results from Table 6 to cover


area of 200011m wide

* approximated values
- 130 -

cases any. further computation for the given particle gives no

additional information the integration is controllably stopped.

For case (i), when the particle is beyond a distance of

0.009m from the area of interest the integration is stopped and the

particle position is logged at (-999, -999). In this way the particle

is then too far away to contribute to the interparticulate forces but

the coordinates indicate in the printout what has occurred. In

case (ii) once the particle has been in flight for longer than 0.45s

its final position is logged as the last current position. In the

printout it is shown to have hit particle number 999 which again

indicates what has happened. Both these precautions avoid wasting

computer time.

6.5 Experimental Results and Discussion

The computer simulation program has been used to primarily

observe the effects of particle size, air velocity, charged layers on

the packing of particles on a substrate. A comparison is also made

between monosized and size distributed powders.

Computer plots are used to pictorially represent the packing,

with each particle trajectory also being plotted. The forces acting

on a particle at any stage in its flight can be known from the print-

out and a few examples are given in the text.

Nearly 100 experimental runs have been carried out. It is

impossible to describe in detail each one and so several runs which

are considered representative are discussed in fuller detail.

However, the plots for all runs, if not given in the main text, can

be found in Appendix E.
- 131 -

6.5.1 No Field Force

In determining the flow regime that must be described

for the calculation of dra'g force (drag coefficient equation)

experiments were carried out in which no field forces and no particle

charges were considered. Figures 6.8 to 6.14 show some of the

results of these runs. In each figure an inset shows the same

experimental conditions but with field and electrostatic forces

acting on the particles. The scale of the inset is exactly the same

as the main figure. In general, without an applied field, the

particles do not reach the substrate but follow the air streamlines

around the plate. Hence deposition does not occur.

As particle size increases the spread of the spray becomes

narrower and, in fact, for 50~m particles starting with a velocity of

5 m/s contact with the substrate is made. However, since there is no

electrostatic forces, the particles would be unlikely to remain on

the substrate.

Figures 6.11 to 6.14 show the effect of uniform velocity on

the spread of particles. As velocity increases inertia causes the

particles to deviate from the flow of air. At a velocity of 10 m/s

the particle traj.ectories appear very similar to those where field

forces are acting.

These experiments clearly show the importance of the electro-

static forces ,for the deposition of particles on the substrate. This

appears particularly true for small particles which tend to follow

the air flow.


Figure 6. B. Trajectories of Particles With and Without Electrostatic Forces - Particle diameter = lO~m.
Uniform Velocity 5 m/s

With Field

Without Field

-
Figure 6.9. Tra'ectories of Particles With and Without Electrostatic Forces - Particle diameter ~ 25~m.
Uniform Velocity - 5 m s

--.,--

~~ -"~
...
er pockjno
: . 11

M."!1I"d'!mi~'\"""

With Field

Without Field

-
Figure 6.10. Trajectories of Particles With and Without Electrostatic Forces - Particle diameter = 50~m.
Uniform Velocity = 5 m/s

With Field

Without Field
Figure 6.11. Tra'ectories of Particles With and Without Electrostatic Forces - Particle diameter = 50~m.
Uniform Velocity = 1 m .s

ing :

With Field

Without Field
Figure 6.12. Trajectories of Particles With and Without Electrostatic Forces - Particle diameter>= 50~m.
Uniform Velocity = 3 m/s

With Field
-.."

Without Field
Trajectories of Particles With and Without Electrostatic Forces - Particle diameter SOllm,
Figure 6.13.
Uniform Velocity - 7 m/s

i
With Field
I
~.l:

n Without Field

I
I
Figure 6,14. Tra'ectories of Particles With and Without Electrostatic Forces l - Particle diameter = 50jJm,
Uniform Velocity = 10 m s

With Field

I Without Field

- I
- 132 -

6.S.2 Monosized Powders

Earlier it was discussed how changes in particle size

and velocity caused varying amounts of 'splaying' of particles and

hence gave different amounts of coverage on the substrate. Experiments

were therefore split into two sections:-

(i) those in which various start widths were employed to achieve

equal areas of coverage over the range of conditions

(ii) those in which the start widths were kept constant.

In order to compare the effects of size and velocity on

particle packing it is primarily the first section that will be

discussed in detail. Due to the large extent of splaying in the

second case for the smaller particles comparison of the effects of

packing on particle trajectories is difficult since the coating is

very thin.

Experiments were conducted for four different particle sizes

(lOllm, l7\im, 2S11m and SOllm diameters) and five uniform air velocities

(1,3,S,7 and 10 m/s).

The variation in conditions was used so that effects of these

parameters could be observed to give indications of the dominant

forces acting on the particles. In general those experiments us~ng

a uniform air velocity of S m/ s will be discussed to a fuller extent

since this relates more to the real system (M.L. Ang, 1981). However

the other results are valuable in showing the changes in observed

effects. Also it can be seen from Figure 6.lS that in a typical

commercial powder 90% by number of the particles are less than l711m

diameter. Therefore the experiments involving 101lm particles are more

relevant than others.


Figure 6.15. Number Size Distribution of Epoxy Powder

100

80
QJ

....
N

""
QJ
."
C
:=> 60
IN!

"
QJ

~z 40

20

04-~~,,~--,,-----r-----r-----T-----T~---'------r-----r-----'
o 10 20 30 40 50 60 70 80 90 100
Particle Diameter (~m)
- 133 -

Figures 6.16 to 6.19 show packings of 300 particles of the four

different sizes at a uniform gas velocity of 5 m/so Figures El to

E16 show results at other velocities. From general observations of

these plots the following remarks can be made:-

(i) Small particles deviate from streamline trajectories more than

large particles. Their movements are much more 'jerky' than

larger particles.

(ii) When particles of larger sizes are repelled from a section of

coating, the particles take longer before they start returning

to the substrate. There is more evidence of 25~m particles being

'shot' away from the coating and landing in uncovered regions.

(iii) With small particles it can be observed that some reach an

equilibrium position from which they do not subsequently move.

However, once this position has been calculated it is assumed

that the particle remains stationary and therefore is not

affected by any other particle entering the near vicinity. In

practice this would not occur and the presence of other

particles coming close to the one in equilibrium would cause

this situation to be disturbed.

(iv) Dendrite type formations are formed which for large particles

can be very tall (>10 particle diameters). The upper.:layers of.

the packings appear more open and this phenomena agrees with

microscopic observations.

Deep crevasses between the dendrite structures are observed in

the plots. It must be stated, however, that the computer

simulation assumes that on touching particles remain stationary.

No bounce-off or collapse of such formations occurs. However,

the simulated effect does appear very similar to that observed

in spray coatings.
Figure 6.16. Tra'ectories of Particles of a Monosized Powder - Particle diameter =
Start Width - 0.00017 m.
10~m, Uniform Velocit = 5 mLs,
Figure 6.17. Tra'ectories of Particles of a Monosized Powder - Particle diameter = l7~m. Uniform Velocit
Start Width = 0.00027 m
Figure 6.18. Trajectories of Particles of a Monosized Powder - Particle diameter = 25~m, Uniform Velocity =5 m!s,
Start Width - 0.00038 m


- - - - - - ----_.

Figure 6.19. Trajectories of Particles of a Monosized Powder - Particle diameter = 50\Jm, Uniform Velocity = 5 m/s,
Start Width - 0.00094 m
- 134 -

(v) SOum particles appear relatively unaffected by packings of

particles up to 1000um in height. Their trajectories deviate

by only small amounts, and this is more typical at the edge of

the area considered.

The individual component forces have been recorded at points

along a particle trajectory for three different particle sizes.

This allows comparison of the magnitude of each of the forces acting

on the particle and gives an indication of which is the dominant

force or forces. Tables 6.S to 6.10 show the results of this analysis

for particle sizes of 10um, 2Sum and SOum using a uniform air

velocity of S m/so Tables 6.S to 6.7 consider the ten!=h'.p:arti:i:le

of each packing whilst the other three tables consider the 200th

particle. The figures in these tables were obtained from the force

and trajectory printouts for each run.

In considering the tenth. particle of each packing and comparing

for the three different sizes the following effects are observed:-

(i) For small particles the drag and field forces during the initial

stages of the trajectory are very similar in magnitude (but

opposite in direction).

(ii) Since only a few particles have already landed, the image force

is the largest force just before the particle lands. For larger

particles the packing is much closer and so the interparticulate

forces are correspondingly larger. The SOum particle has

actually landed on top of two others and so the interparticulate

force in the x direction is large and negative.

(iii) For large particles the drag term is dominant but the field

force increases relative to it as the particle approaches the

substrate. This is due to the relative velocity between the


Table 6.5. Individual Forces Acting on the 10th Particle in Flight

Particle Air
Velocities· Velocities
Xt Y Time in UI' VI' U V
air air Re
(m) (IIm) flight (m7s) (m7s) p
(m/s) (m/s)
(secs)

0.100 -39 - 5.00 0.00 - - -


0.110 -42 0.004 2.37 0.00 2.06 0.00 0.174

0.150 -68 0.025 1.46 0.00 1.23 0.00 0.128

0.190 -235 0.074 0.41 -0.01 0.24 -0.01 0.096

0.199 -492 0.103 0.20 -0.01 0.03 -0.01 0.096

* -588 0.110 0.20 -0.02 0.00 -0.01 0.112

DRAG FIELD IMAGE PARTICULATE


Xt FDx FDy FE x FE FI ! (FR,c+F ~ ! (FRy+F Ay
(m) Y
(N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -5.2E-10 1.8E-13 4.9E-10 -1. 9E-13 3.1E-18 -2.5E-20 -4.8E-24

0.150 -3.8E-10 1. 4E-13 3.6E-10 -1. 7E-13 9.7E-18 -1. 4E-19 -1. 5E-22

0.190 -2.9E-1O 2.6E-13 2.9E-1O -3.4E-13 2.2E-16 -1.5E-17 -3.7E-19

0.199 -2.8E-lO 5.1E-lO 2.7E-10 -6.8E-13 1.6E-14 -4.0E-15 8.9E-17

* -3.4E~lO 9.1E-13 2.7E-10 -8.0E-13 9.6E-10 6.0E-13 -5.5E-14

Particle size = 101lm, Uniform Velocity, U = 5 m/s, Particle Number 10

Start Width = 0.0002m ,


* last but one caluclated x position, distance from substrate = 511m
t approximate x position from spray gun
Table 6.6. Individual Forces Acting on the 10th'Particle in Flight

Particle Air
Velocities ., Velocities
Xt Y Time in UP. VP. U v. ,Re
air a~r p
(m) (jJm) Flight (m7s) (m7s)
(m/s) (m/s)
(secs)

0.100 -39 - 5.00 0.00 - - -


0.110 -40 0.002 3.52 0.00 2.06 0.00 2.049
.
0.150 -59 0.018 1.88 0.00 1. 23 0.00 0.913

0.190 -140 0.052 0.73 0.00 0.26 0.00 0.660

0.199 -211 0.068 0.46 -0.01 0.03 -0.01 0.604


';' .

* -225 0.070 0.45 -0.02 0.00 -0.01 0.632

DRAG FIELD IMAGE PARTICULATE

X t FDx FDy FEx FEy FI l:(FR,c+FAx) L (FRy+FAy)


(m) (N) (N) (N) (N) (N)
(N) (N)

0.100 - - - - - - -
0.110 -6.lE-9 -1.9E-13 3.lE-9 -1.lE-12 1. 2E-16 -2.4E-18 -1.lE-21

0.150 -2.7E-9 5.7E-13 2.3E-9 -9.lE-13 3.8E-16 -1.3E-17 -1. 9E-20

0.190 -2.0E-9 5.6E-13 1.8E-9 -1.3E-12 9.lE-15 -1. 6E-14 -2.9E-17

0.199 -1. 8E-9 7.3E-13 1.7E-9 -1.8E-12 8.9E-13 -1.lE-12 -1. 7E-13

* -1. 9E-9 4.4E-11 1. 7E-9 -1. 9E-12 5.9E-9 B.lE-lO -4.9E-lO

Particle size = 25jJm, Uniform Velocity, U =5 m/s, Particle Number 10

Start Width = 0.0002m


* last but one calculated x position, distance from substrate = l2.5jJm
t approximate x position from spray gun
Table 6.7. Individual Forces Acting on the 10th Particle in Flight

Particle Air
Velocities Velocities
Xt Y Time in Up Vp U V . Re
(m/s) air an p
(m) (Ilm) Flight (m?s)
(m/s) (m/s)
(secs)

0.100 -39 - 5.00 0.00 - - -


0.110 -39 0.002 4.72 0.00 2.07 0.00 7.45

0.150 -47 0.012 3.2S 0.00 1.22 0.00 5.79

0.190 -70 0.02S 1. 70 0.00 0.2B 0.00 3.99

0.199 -S2 0.035 1.26 0.00 0.03 0.00 3.46

* -S4 0.036 1.20 0.00 0.00 0.00 3.37

DRAG FIELD IMAGE PARTICULATE

Xt Fnx FDy FEx FEy r(FRx+FAx) r(FRy+FAy)


Fr
(m) (N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -2.2E-B -3.9E-12 1. 2E-B -4.5E-12 1. 9E-15 -1.4E-16 -1. 5E-19

0.150 -1. 7E-B -2.4E-13 9.lE-9 -2. BE-12 6.2E-15 -B.SE-16 -1. SE-IS

0.190 -1. 2E-S -1. 3E-13 7.2E-9 -2.6E-12 1. 3E-13 -7.7E-14 -9.4E-16

0.199 -1.0E-S 1. 2E-13 6.SE-9 -2.SE-12 1. 4E-ll -8. lE-ll -1.2E-ll

* -1.0E-B 9.0E-12 6.SE-9 -2.9E-12 1.1E-9 -3.5E-S 6.3E-10

Particle size = 501lm, Uniform Velocity, U = 5 m/s, Particle Number 10

Start width = O.oooZm


* last but one calculated x position, distance from substrate = 120!lm
t approximate x position from:spray gun
- 135 -

particle and air decreasing. The interparticle electrostatic

forces, although increasing, cannot slow down the particle

enough to stop it from landing.

(iv) In the y direction, the forces are generally much smaller than

those in the x direction, but the field force is generally the

largest until the particles are close to the packing when

electrostatic forces increase and cause the particle to move

laterally, whence the drag force increases.

(v) In general it is observed that the field and drag forces do not

change by any appreciable extent over the range of the trajectory.

The electrostatic forces however can increase by 10 orders of

magnitude between the start point (O.lm from substrate) and the

substrate.

(vi) The particle Reynolds number (Rep) for 10~m particles is less

than one throughout the trajectory and it is only for 50~m

that Re than one for any length of time.


p is greater
When the 200th particle is considered it is observed that the

magnitude of some of the forces has changed, especially near to the

substrate. From Tables 6.8 to 6.10 the following observations are

made:-

(i) Small particles are heavily influenced by the electrostatic

forces near the substrate. The repulsive forces are 2-4 times

larger than the drag or field forces. The drag and field

forces are of similar size throughout the trajectory and hence

the balance of forces is controlled by the electrostatic

forces. It is noticeable that in the example of the 10~m particle

the particulate electrostatic force in the x direction was positive

on landing. This was because it found a path between the

packing to the substrate and therefore had many similar charged


Table 6.8. Forces Acting on the 200th Particle in Flight

Particle Air
Velocities Velocities
X t Y Time in Up Vp U V . Re
air ur p
(m) (Ilm) Flight (m/s) (m/s)
(m/s) (m/s)
(secs)

0.100 30 - 5.0 - - - -
0.110 33 0.004 2.37 0.00 2.06 0.0 0.17

0.150 53 0.025 1.46 0.00 1.23 0.00 0.13

0.190 171 0.072 0.44 0.00 0.27 0.00


. 0.10
0.199 390 0.104 0.19 0.01 0.03 0.01 0.09

* 558 0.112 0.18 -0.04 0.00 0.01 0.10

DRAG FIELD IMAGE PARTlCULATE

X t FDx FEx FE . 1: (F Rx +FRy) J:(FRy+FA )


Fny y FI y
(m)
(N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -S.2E-I0 -1. 4E-13 4.9E-1O 1. 4E-13 3.1E-18 -9.6E-1O -6.7E-22
0.150 -3.8E-1O -1.1E-13 3.6E-1O 1.3E-13 9.7E-18 -5.4E-18 -3.4E-21

0.190 -2.9E-1O -2.0E-13 2.9E-1O 2.6E-13 2. ZE-16 -5.8E-16 7.6E-18

0.199 -2.8E-1O -4.1E-13 2.7E-1O 5.3E-13 2.1E-14 -2.8E-13 1.1E-14

* -3.0E-I0 8.9E-ll 2.7E-I0 7.6E-13 1.0E-9 1.1E-9 1. 6E-I0

Particle size = 10llm, Uniform Velocity, U = 5 mls, Particle Number 200

* last but one calculated x position, distance from substrate = 51lID


t approximate x position from spray gun
Table 6.9. Forces Acting on the 200th Particle in Flight

Particle Air
Velocities Velocities
X t Y Time in Up Vp Uair V. Rep
(m) (Ilm) Flight (m/s) (m/s) ur
(m/s) (m/s)
(secs)

0.100 68 - 5.0 0.00 - - -


0.110 70 0.002 3.52 0.00 2.06 0.00 2.04

0.150 103 0.018 1.88 0.00 1.23 0.00 0.91

0.190 242 0.052 0.73 0.01 0.26 0.01 0.66

0.199 367 0.068 0.45 0.01 0.03 0.01 0.59

* 1325 0.075 0.29 0.17 0.00 0.03 0.45

DRAG FIELD IMAGE PARTICULATE

X t FDx FEy FI ! (FR,.; +FRy) !(FRy+FAy)


Fny FEx
(m)
(N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -6.1E-9 3.3E-13 3.1E-9 2.0E-12 1. 2E-16 -2.4E-16 -5.6E-20

0.150 -2.7E-9 -9.9E-13 2.3E-9 1. 6E-12 3.8E-16 -1.3E-15 7.5E-19

0.190 -2.0E-9 -9.6E-13 1. 8E-9 2.3E-12 9.2E-15 -1. 6E-13 3.6E-15

0.199 -1. 8E-9 -4.1E-12 1. 7E-9 3.1E-12 8.8E-13 -8.6E-ll 5.9E-12

* -1. 2E-9 -5.8E-1O 1. 7E-9 1.lE-ll 6.1E-9 3.5E-10 3.6E-ll

Particle size = 25)lm, Uniform Velocity, U = 5 m/s, Particle Number 200

* last but one calculated x position, distance from substrate = l31lm


t approximate x position from spray gun
Table 6.10. Forces Acting on the 200th Particle in Flight·

Particle Air
Velocities Velocities
X t Y Time in Up Vp Uair V Re p
(m) (Ilm) Flight (m/s) (m/ s) air
(secs) (m/s) (m/s)

0.100 169 0.000 5.00 0.00 - - -


0.110 170 0.002 4.72 0.00 2.07 0.00 7.45
0.150 202 0.012 3.2B 0.00 1.22 0.00 5.79
0.190 300 0.029 1.67 0.01 0.26 0.01 3.96
0.199 357 0.035 1.21 0.02 0.03 0.01 3.32

* 362 0.035 1.13 0.03 0.02 0.01 3.1B

DRAG FIELD IMAGE PARTlCULATE

X t E(FRx +FRy ) E (FRy +FA;)


FDx Fny FEx FEy FI
(m)
(N) (N) (N) (N) (N) (N) (N)
0.100 - - - - - - -
0.110 -2.2E-B 1.8E-11 1.2E-8 L9E-ll 1. 9E-15 -1. 7E-14 4.5E-17
0.150 -1. 7E-8 LOE-12 9.lE-9 1.2E-ll 6.2E-15 -9.9E-14 5.8E-16
0.190 -1. 2E-8 5. ZE-13 7.2E-9 1.1E-11 1. 4E-13 -1.lE-ll 4.5E-13
0.199 -9.9E-9 -6.2E-ll 6.8E-9 L2E-ll 1. 5E-ll -B.9E-9 LOE-9

* -9.3E-9 -2.0E-10 6.8E-9 L2E-ll 2.8E-ll -4.7E-B 2.BE-B

Particle size = 501lm, Uniform Velocity, U =5 m/s, Particle Number 200

* last but one calculated x position, distance from substrate = 7531lm


t approximate x position from spray gun
- 136 -

particles above it (see Figure 6.21). The resultant force,

in the x direction, was therefore towards the substrate.

(ii) Large particles have consistently high drag forces throughout

the flight showing that the particle has very high inertia.

Although the electrostatic forces become much larger as the

particle approaches the packing the inertia cannot be overcome

and the 50~m hits the packing at a velocity of over 1 m/so

At this time the electrostatic force in the x direction is

five times the size of the drag force.

(iii) On comparing electrostatic forces of the 200th and 10th

particles it can be clearly seen that the forces are an order

of magnitude larger at 1 cm away from the substrate for the

200th particle, due to the charged layer that has been built.

Since effects due to inertia were noticed for the SOum

particles when no field is applied, the forces acting on a SO~m

particle at different velocities have been tabulated in Tables 6.11

to 6.13.

From these tables it can be seen that:-

(i) At low velocities (1 m/s) the field forces acting on the

charged particle cause an increase in particle velocity.

Initially the field force is larger than the drag force and

so the particle actually accelerates. This is one of the few

occasions that, at the start.of the trajectory, the field

force is larger than the drag force.

As the field force decreases with increasing distance from the

gun the drag force again dominates. At higher velocities the

drag force is always greater than the field force (in the x

direction) •
Table 6.11. Forces Acting on a 50~m Particle in Flight

Particle Air
Veloci ties Velocities
X t Y Time in Up Vp U V . . ·Re
(m/s) air al.r p
(m) (llm) Flight (m/s) (m/s) (m/s)
(secs)

0.100 -92 - 1.00 0.00 - - -


0.110 -96 0.007 1. 59 0.00 0.41 0.00 3.330

0.150 -130 0.031 1.47 0.00 0.25 0.00 3.428

0.190 -182 0.064 1.00 0.00 0.06 0.00 2.641

0.199 -201 0.075 0.89 0.00 0.00 0.00 2.501

* -205 0.076 0.89 -0.02 0.00 0.00 2.507

DRAG FIELD IMAGE PART ICULATE

X t FDx Fny FEx FEy FI l: (FRxc-+:FRy) r(FRy +FAy )


(m) (N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -1.0E-8 6.0E-12 1.2E-8 -1.lE-lO 2.0E-15 -4.7E-17 -1.2E-18

0.150 -1.0E-8 7.8E-12 9.lE-9 -8.0E-12 6.0E-15 -2.5E-16 -1. 4E-18

0.190 -7.9E-9 6.5E-12 7.2E-9 -7.0E-12 1.lE-12 -2.lE-14 -6.5E-16

0.199 -7.4E-9 6.8E-12 6.8E-9 -6.9E-12 1. 5E-11 -2.5E-11 -9.2E-12

* -7.4E-9 1. 3E-lO 6.8E-9 -6.9E-12 2.4E-8 6.9E-9 -8.9E-9

Particle size = 5011m, Uniform Velocity, U = 1.0 m/s, Particle Number 5

* last but one calculated x position, distance from substrate = 25~m

t approximate x position from spray gun


Table 6.12. Forces Acting on a sO~m Particle in Flight

Particle Air
Velocities Velocities
X t Y Time in Up V U V. Re
p air aH p
(m) ( m) Flight (m/s) (m/s) (m/s) (m/s)
(secs)

0.100 -92 - 5.00 0.00 - - -


0.110 -93 0.002 4.72 0.00 2.07 0.00 7.447

0.150 -110 0.012 3.2S 0.00 1.22 0.00 S.7S9

0.190 -162 0.02S 1. 70 0.00 0.2S 0.00 3.990

0.199 -192 0.035 1. 26 0.00 0.025 0.00 3.470

* -197 0.036 1.22 -0.02 0.00 0.00 3.456

DRAG FIELD IMAGE PARTICULATE

X t FD x FD y FE y E(FRx+FRy) E(FR +FA )


FEx Fr y Y
(m)
(N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -2.2E-S -9.1E-12 1.2E-S -1.1E-ll 1. 9E",ls -4.sE-17 -1.lE-19

0.150 -1. 7E-S -s.sE-13 9.1E-9 6.7E-12 6.2E-ls -2.7E-16 -1. 4E-lS

0.190 -1. 2E-S -3.1E-13 7.2E-9 -6.2E-12 1.3E-13 -2.4E-14 -7.2E-16

0.199 -1.0E-8 -3.6E-13 6.SE-9 -6.6E-12 1.sE-ll -2.6E-ll -9.6E-12

* -1.0E-S 9.2E-ll 6.SE-9 -6.7E-12 2.4E-S 7.5E-9 -1.0E-S

Particle size = 50~m, Uniform Velocity, U =5 m/s, Particle Number 5

* last but one calculated x position, distance from substrate = 2s~m


t approximate x position from spray gun
Table 6.13. Forces Acting on a 50~m Particle in Flight

Particle Air
Velocities Velocities
X t Y Time in Up Vp U V. Re
air al.r p
(m) (~m) Flight (m/s) (m/ s) (m/s) (m/s)
(secs)

0.100 -91 - 10.00 0.00 - - -


0.110 -92 0.001 9.51 0.00 4.12 0.00 15.146

0.150 -100 0.006 7.23 0.00 2.49 0.00 13.319

0.190 -126 0.013 4.25 0.00 0.55 0.00 10.397

0.199 -139 0.015 3.34 -0.01 0.05 -0.01 9.245

* -141 0.016 3,25 -0.01 0.00 -0.01 9.133

DRAG FIELD IMAGE PARTICULATE

X t FDx FE y E(FR,,:+Fay) E(FRy+FAy)


FDy FEx FI
(m) (N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -4.5E-8 -2.5E-IO 1. 2E-8 -1.0E-ll 1. 9E-15 -4.8E-17 -1.1E-19

0.150 -4.0E-8 -1. 7E-1O 9.lE-9 -6.lE-12 5.9E-15 -2.6E-16 -1. lE-18

0.190 -3.1E-8 -1.3E-ll 7.2E-9 -4.8E-12 1.3E-13 -2.6E-14 -6.1E-16

0.199 -2.8E-8 -1. 3E-ll 6.8E-9 -4.7E-12 1. 5E-ll -3.0E-ll -8.2E-12

* -2.7E-8 2.2E-ll 6.8E-9 -4.8E-12 5.9E-9 -4.2E-9 -2.0E-8

Particle size = 50~m, Uniform Velocity, U = 10 m/s, Particle Number 5

* last but one calculated x position, distance from substrate = 50~m


t approximate x position from spray gun
- 137 -

(ii) At 10 m/s uniform velocity the drag force is dominant for the

whole of the trajectory and the particle lands at a velocity of

over 3 m/so

(iii) Although the electrostatic forces increase by several orders

of magnitude they are still 3 orders of magnitude less than the

drag force at a distance of only 1000pm from the substrate.

Hence for normal thicknesses of coatings repulsion of these

particles will not occur (except perhaps after having landed).

From the computer plots it is shown that small particles (lO~m)

are strongly influenced by the electrostatic forces due to packed

particles on the substrate. These particles are subject to a great

deal of movement close to the packing before they eventually land.

In Tables 6.5 and 6.8 it is shown how, close to the substrate, all

forces acting on the particle are of similar size. Since small

particles approach the packing at very low velocities (0.2 m/s) they

can respond to any sudden increase in force acting on the particle.

This is usually due to it approaching some packed particles on the

substrate.

Figures 6.20 to 6.22 show the final stages of the 200th 10~m

particle for velocities of 1 m/s, 5 m/s and 10 m/so These correspond

to small sections of packings shown in Figures 6.16, El and E4.

The trajectories are subject to 180 0 turns in the direction of travel and

are obviously related to the packing that is present in the close

vicinity to them.

Also, (i) as velocity increases it is observed that the

particles generally travel closer to the packing than at lower velocities.


. re 6.20.
Figu Trajectory of Par tiC Substrate
. le Number 200 Close to the Par tiC . r
. I e diamete
=

Uniform velocity = I m/s

350
Figure 6.21. Trajectory of Particle Number 200 Close to the Substrate

Particle diameter = 10~m

Uniform velocity = 5 m/s

500
Figure 6.22. Trajectory of Particle Number 200 Close to the Substrate

Particle diameter 10pm

Uniform velocity 10 m/s

500 550 600


- 138 -

(ii) it is less likely that the particle finds an equilibrium

position at higher velocities since the electrostatic forces

do not have such a marked effect as at low velocities.

(iii) Thicker packings are obtained when using higher uniform air

velocities.

In Figures 6.20 and 6.22 the final particle position is in

'mid-air' since its time of flight is already greater than 0.45

seconds. The computer printout shows that for each case the x and y

velocities of the particle at ,these positions are zero and hence all

the forces are balanced. In Figure 6.20 (1 m/s) there are other

particles which are also stationary above the packing. Unfortunately

due to limitations of the program the particles are not moved once

their trajectory has finished. In reality the presence of particle

200 would probably cause all of them to move to new positions, some

possibly landing on the substrate or packing.

Typically, these figures show that small particles will 'search'

for areas of low population (less repulsive force) and then try to

land. In Figure 6.21 this was achieved but although the particle

turned back in Figure 6.22 particles 189 and 145 were on the substrate

which stopped it from landing. Particle 200 had therefore committed

itself to entering in a 'valley' (bounded by Nos 106, 60, 172, 33)

where eventually it reached a balanced position.

By observing these examples in greater detail it can be

understood how small particles are able to cover the substrate to a

larger extent than can larger ones. Although in the property testing

experiments no difference in adhesion qualities was noticed it is

possible, for poorer flowing powders, that small particles are needed

to ensure full contact and wetting of the substrate.


- 139 -

Similar results are observed for experiments in which the

start width was constant (Figures El7 to E36). In these plots the

splaying of particles is clearly shown and comparisons show how

particle size and air velocity affect this. The very small areas of

coverage of sOum particles enhance the dendrite effect and show how

very thick packings could be obtained if these formations did not

collapse.

In general these plots show similar effects to the first set

of plots and therefore the conclusions based on these results are

not affected by the splaying effect.

6.5.3 Size Distributed Powders

It was found that simulating a mixture of particle sizes

was much more difficult than monosized powders. Therefore, results

have to be unfortunately limited in their scope. The splaying of

particles of different sizes caused problems since the majority of

large particles landed in a small area in the centre of the plate.

However, as mentioned in Chapter 3, this is a real effect and

segregation of a powder does occur with a static gun.

Nevertheless, interesting effects are observed by considering

a distribution. The splaying effect is enhanced because such a small

number of particles are considered and the start width at x = 0.1


is the same for all sizes. This procedure was adopted since any

alteration of start width with size would suggest preferential

deposition of particles (especially large ones) and the number

distribution representing a powder would not be kept constant. The

plots are shown in Figures 6.23 to 6.27 for five different operating

velocities.
Figure 6.23'. Trajectories of Particles of a Size Distributed Powder - Uniform Velocity = 1 m/s,
Start Width - 0.00031 m ..
Figure 6.24. Trajectories of Particles of a Size Distributed Powder - Uniform Velocity ~ 3 m/s,
Start Width - 0.00020 m
Figure 6.25. Trajectories of Particles of a Size Distributed Powder - Uniform Velocity =5 m/s,
Start Width - 0.00013 m
Figure 6.26. Trajectories of Particles of a Size Distributed Powder - Uniform Velocity =7 mLs,
Start Width - 0.00010
Figure 6.27. Trajectories of Particles of a Size Distributed Powder - Uniform Velocity = 10 mls,
Start Width = 0.00010
- 140 -

The following effects are observed:-

(i) As velocity increases the severity of the segregation of

particle sizes across the substrate increases.

(ii) There is less meandering of the particles with increase in

uniform velocity.

(iii) At higher velocities the large dendrite formation of particles

in the centre of the substrate causes a shadowing effect since

small particles are repelled away from it. The tall dendrite

therefore has a larger overall effect compared to small

dendrites formed at low velocities and alters the trajectories

of small particles at an earlier stage (compare Figure 6.24 -

3 m/s with Figure 6.27 - 10 m/s).

(iv) Similarly to the monosized results, small particles are

responsible for covering more of the substrate whilst large

particles tend to form tree formations.

Although these results can only represent a fraction of the

packing that is usually produced, they show that small particles will

not land in areas where there is a thickness of coating due to large

particles. This agrees with the experimental findings in that fines

are lost when a commercial size distribution is sprayed.

Small particles are therefore repelled from the dendrite

structures to thinner layers of coating. As a surface becomes

completely covered and layers are built the fine particles will not

be able to land and so will be over sprayed. From these results and

those of the monosized experiments it is suggested that only a fixed

amount of fines can be coated onto a substrate over and above a given

thickness of coating. Therefore, if thick films are required, . small


- 141 -

particles are wasted once the first few layers have been completed.

Similar-effects are found when the start width is constant for

various transport air velocities (see Figures E37 to E41).

6.5.4 Theoretical Charged Layer

In the last Chapter equations were derived for the

calculation of force on a particle due to a charged layer of a given

thickness across the width of the plate. Experiments, using these

equations, have been carried out for three particle sizes, three

veloci ties and three layer-_,thicknesses (50Jlm, 100Jlm, 250Jlm thick).

In each situation the charge density (charge/unit area) has

been taken as 2.22 x 10- 5 cou10mb/m 2 . Some of the results obtained

are shown in Figures 6.28 to 6.33 and the remaining plots to complete

the series are shown in Figures E42 to E62.

The results can be summarised as follows:-

(i) For velocities of 1 to 10 m/s no 10Jlm particles landed on a

layer of charge of any thickness employed here.

(ii) 25Jlm particles landed on 50Jlm and 100Jlm thick layers at any

velocity but none landed on the 250Jlm thick layer.

(iii) All 50Jlm particles landed under all conditions.

(iv) For conditions in which the particles did not land an equilibrium

distance from the charged layer was reached and the particles

travelled away from the centre of the plate. All particles had

the same trajectory after having reached this equilibrium

distance.

(v) As the thickness of charge layer was increased so the equilibrium

distance increased.
Figure 6.28. Trajectories of Particles Approaching a Charged Layer 250llm thick - Particle diameter = 101lm,
Uniform Velocity - 5 m/s
Figure 6.29. = 2Sj.1m,
Uniform Velocity
Figure 6.30. Trajectories of Particles Approaching a Charged Layer 250pm thick - Particle diameter = 25pm,
Uniform Velocity - 1 m/s
" I
Figure 6.32. Trajectories of Particles Approaching a Charged Layer 250~m thick - Particle diameter = 25~m,
Uniform Velocity - 10 m/s

I~
~ 'X..\ v

CH

-1
Figure 6.33. Trajectories of Particles Approaching a Charged Layer 250~m thick - Particle diameter = 50~m,
Uniform Velocity = 5 m/s


- 142 -

(vi) Uniform air velocity was found to have little effect on the

equilibrium distance compared to the effect of layer thickness.

Table 6.14 shows the distance of the equilibrium trajectory from

the charge layer for each of the conditions considered and shows the

effects of remarks (v) and (vi). It is also observed that, as

velocity increases, the closest distance of the trajectory to the

charge layer is smaller.

Since the 25~m particles are deposited at low velocities

(1 and 5 m/s) but not at 10 m/s, the forces acting on a 25~m particle

have been recorded in Tables 6.15 and 6.16 for various positions

during the trajectory.

Table 6.15 shows the forces from the start point until the

particle reaches its closest position to the charge layer.

In the early stages of flight the drag force is dominant. As

the particle approaches the substrate the electrostatic forces increase

and ~l drops and so therefore the drag force correspondingly


re
decreases. Eventually the repulsive forces cause the particle to

turn around and at this point the repulsive forces are greater than

the addition of the drag and field forces. Table 6.16 shows results

of the trajectory after this point and'Figure 6.34 is a pictorial

representation of this situation.

As the particle turns and heads back towards the gun (Point A)

the repulsive force drops proportional to 1/d 2 , where d is the distance

between particle and layer. The field force, however, changes

negligibly over this distance and so then becomes the dominant force.

The relative velocity of the particle slowly approaches zero and then
Table 6.14. Variation of Distance of Equilibrium Trajectories from
Charge Layer with Velocity and Thickness of Layer

Particle Uniform Distance from Layer


Diameter Veloci ty of Equilibrium Trajectory
()lm) m/s )lm

1 52 (43) 105(101) 259(258)


10 5 51(43) 102 (98) 247(246)
10 51 (42) 100(96) 233(233)

1 landed landed 261(223)


25 5 landed landed 257(180)
10 landed landed 249(57)

1 landed landed landed


50 5 landed landed landed
10 landed landed landed

50 100 250

Thickness of charge layer ()lm)

* Figures in brackets represent closest position of trajectory

from charge layer.


Table 6.15. Forces Acting on a Particle in the Vicinity of a Layer
of Charge

Particle Air
Velocities Velocities
X t Y Time in Up Vp U V Re
air air P
(m) (Ilm) Flight (m/s) (m/s) (m/s) (m/s)
(secs)
0.100 -95 0.000 5.00 0.00 - - -
0.110 -98 0.003 3.45 0.00 Z.04 0.00 1.98
0: 150 -143 0.018 1.88 0.00 l.Z3 0.00 0.91

0.190 -337 0.05Z 0.73 -0.01 0.Z6 -0.01 0.66

0.199 -513 0.068 0.4Z -0.01 0.03 -0.01 0.55

* -545 0.070 0.00 -0.01 0.01 -0.01 0.01

DRAG FIELD llIAGE PARTICULATE

X t FDx FE,c FE y l:(FR;,+FRy) E(FRy +FAy)


FDy Fr
(m)
(N) (N) (N) (N) (N) (N) (N)

0.100 - - - - - - -
0.110 -5.8E-9 -Z.OE-13 3.1E-9 -Z.7E-1Z 1. ZE-16 -5.ZE-14 -Z.4E-17

0.150 -2.7E-9 1. 4E-1Z 2.3E-9 -Z.ZE-1Z 3.8E-Hi -1. 6E-13 -4.8E-17

0.190 -Z.OE-9 1.3E-1Z 1.8E-9 -3.ZE-1Z 9.0E-15 -3.6E-1Z -4.0E-17

0.199 -1. 6E-9 1. 6E-lZ 1. 7E;"9 -4.4E-1Z 9.0E-13 -3.7E-10 -6.3E-18

* 4.6E-ll 1. 7E-1Z 1. 7E-9 -4.6E-1Z 9.0E-1Z -5.4E-9 -Z.lE-18

Particle size = Z51lm, Uniform Velocity, U =5 m/s, Particle Number Z5

* closest calculated x position to layer, distance from substrate = 33Qjlm

t approximate x position from spray gun


Table 6.16. Forces Acting on a Particle Aoproaching a Layer of Charge

Particle Air
Velocities Velocities
Xt Y Time in Up Vp U . V . Re
a1r a1r p
(llm) (llm) Flight (m/s) (m/ s) (m/s) (m/s)
(secs)

630 -587 0.073 0.00 -0.02 0.02 -0.01 0.03

462 (A) -643 0.077 0.00 -0.02 0.01 -0.02 0.01


523(B) -699 0.080 0.00 -0.02 0.01 -0.02 0.01

498(C) -758 0.083 0.00 -0.02 0.01 -0.02 0.01

509 (D) -S2S 0.OS7 0.00 -0.02 0.01 -0.02 0.01

506 (E) -970S 0.lS2 0.00 -0.25 0.01 -0.24 0.03

DRAG FIELD nlAGE PART ICULATE

X t L(FR,.:+FRy) L(FRy+FAy)
FDx FDy FEx FEy FI
(llm) (N) (N) (N) (N) (N) (N) (N)

630 5.9E-ll 1. SE-12 1. 7E-9 -5.0E-12 2.5E-12 -1.1E-9 -4.3E-lS

462(A) 4.1E-ll 2.0E-12 1.7E-9 -5.5E-12 4.6E-12 -2.2E-9 -3.5E-lS

523 (B) 5.3E-ll 2.2E-12 1. 7E-9 -6.0E-12 3.5E-12 -1.6E-9 -4.3E-lS

498(C) 4.9E-ll 2.4E-12 1. 7E-9 -6.5E-12 4.0E-12 -1. SE-9 -4.4E-18

509 (D) 5.2E-ll 2.6E-12 1. 7E-9 -7.0E-12 3.SE-12 -1. 7E-9 -4.9E-1S

506 (E) 5.3E-ll 3.0E-ll 1. 7E-9 -S.3E-ll 3.SE-12 -1.8E-9 -6.9E-17

t approX1mate
. "
x pos1t10n f rom su b strate

Particle size = 2511m, Uniform Velocity, U =5 m/s, Particle Number 25


Figure 6.34. Trajectory of Particle Number 25 Close to the Charged Layer

Particle Size = 25~m


Charged layer thickness = 250~m

Uniform velocity = 5 m/s

650
B

550
D

C
450

350

250

I
-800
I
-700
I
-600
i
-500
- 143 -

the field pushes it back towards the substrate (Point B). Whilst

this occurs, the y components of the field and electrostatic forces

slowly push the particle in the y direction (in this case negatively)

such that it is moving away from the centre of the plate.

The particle oscillates several times being pushed in· ,either x

direction by the field and electrostatic forces (Points C & D).

Eventually a situation very close to an equilibrium balance of forces

in the x direction is reached (Point E). As the particle travels

away in the y direction the field (x) will decrease slightly, compared

to the electrostatic force, and hence the x position of the particle

slowly increases away· from the substrate. Since all particles, for

a given set of conditions, are the same size they follow the same

trajectory, irrespective of their initial y positions.

The trajectories of lO~m particles approaching the. substrate

are not as oscillatory since they approach the equilibrium distance

at much lower velocities and their stopping distances are much shorter.

The inertia of the 25~m particles allows them to come closer to

the substrate before they are turned away.

Overall these experiments have been extremely valuable. They

have shown that, for identical coating thicknesses and air velocities,

the landing of particles is dependent on particle size. Also, these

results agree with the quantitative analysis of size distributions

before and after spraying. For a given set of operating conditions,

at some juncture small particles will be repelled whilst larger

particles wilt continue to be deposited. Therefore a change in size

distribution will occur through the particle packing, with small

particles closest to the substrate.


- 144 -

6.5.5 Effect of Gravity

The force due to gravity acting on the particles has not

been included in these experiments since its effects are dependent

on the plane in which the target is positioned. Obviously the gravity

force is greater for larger particles since it is proportional to

mass.

For comparison of the magnitude of forces acting on a particle,

the size of the forces due to gravity are

10\lm particle 5.14 x lO-lON

25\lm particle B.03 x 10-9 N

50\lm particle 6.43 x 10-BN

For particles' of the order of 10\lm in size the force is

approximately equal to the drag and field forces acting on the particle

throughout its trajectory. When packings are present, the inter-

particulate electrostatic forces will be greater than the force due

to gravity close to the substrate and hence similar effects would be

expected as have been already observed. However, gravity would cause

any overall displacement of landing position compared to those

obtained here. Again its effects close to the substrate are reduced

compared to other sized particles because of the short stopping distances

of very small particles.

For large'particles the effect due to gravity will be more

dependent on the location of the substrate. If vertical, the large

particles would land at lower positions on the substrate, or possibly

miss completely. Higher operating air velocities would help overcome

this as the particles leave the gun with more inertia.

For horizontal surfaces, depending on their position relative

to the gun, gravity would give the particles more inertia towards the

plate and hence thicker coatings would be expected.


- 145 _.

6.6 Conclusions

The writing of a computer program to simulate the trajectories

of particles sprayed towards a substrate has given an insight into

the mechanism of packing of particles by this process in a qualitative

manner.

The method and working of the computer program has been described

together with the necessary tests that are required to ensure correct

computation and calculation. Many stated assumptions have been used

in order to simplify the system such that a reasonably large number of

particles can be considered. Preliminary experiments showed that

(a) the tolerance bound required to give representative results


7
was 10-

(b) that Re is generally less than 1 close to the subs~rate and


p
hence Stokes Law can be applied

and

(c) that the particle velocities, although initially approximated,

reach a tangible solution in the first few centimetres of the

trajectory.

The monosized experiments have shown that each powder size

range will tend to have a maximum thickness that can be coated onto

the substrate. Small particles are strongly influenced by electro-

static forces close to the substrate and due to their mobility are

able to find uncoated areas of substrate. Large particles are

dominated by their own inertia and in general the forces due to the

powder packing are not strong enough to cause any appreciable

deviation in their trajectories. These particles form dendrite type

formations which agrees with experimental evidence.


- 146 -

In. these computer experiments particles were found to reach an

equilibrium position where the resultant of all forces acting on· the

particle was zero. Close up studies of the movement of 10~m

particles. have shown how this can occur and also how particles follow

the contours of the particle packing.

Throughout the computer experiments it has been noticeable

that both particle size and air velocity affect the splaying of

the particles. Decreasing particle size and increasing velocity

enhances this effect. The size distributed powder results confirmed

the observations of size segregation across a coating. Large

particles were prolific in the centre whilst small particles landed

towards the outside of the plate. Large particles tend to cause a

shadowing effect in that small particles are deflected far away from

them and hence parts of the substrate are left less populated close

to a large particle dendrite tree.

Since three hundred particles took approximately 3 hours to

compute, a wide, thick coating could not be simulated by this method.

The experiments involving a charged layer allowed an advanced

coating CL e. thicker) to be considered but required less computing

time. The plots showed that for a coating of 50~m or thicker no

10~m particles would be deposited for velocities between land 10 m/s.

25~m particles were repelled from a 25~m thick layer but 50~m

particles always landed. Hence for identical coatings and conditions

the plots have shown that deposition is related to particle size.

Large particles will continue to be deposited as the packing increases

in thickness and small particles are repelled.


- 147 -

6:7 Summary

By using a computer program to simulate the packing of particles

on a substrate, a valuable insight has been gained into understanding

which forces act dominantly on particles in different circumstances.

Results of computer analysis have shown good agreement with

experimental results and have given indications of the reasons behind

observed effects. By considering the effects of a charged layer and

different air velocities on monosized powder distributions the

size segregation effect that was quantified in Chapter 4 has been

confirmed. Although such simu1ations can only be considered

qualitative they have endorsed experimental trends.


CHAPTER 7

FINAL DISCUSSION AND SUMMARY

7.1 Aims of the Investigations

7.2 Summary of Results of Investigations

7.2.1 Testing of Properties

7.2.2 Packing and Stoving

7.2.3 Trajectory Equation

7.2.4 Computer Simulation

7.3 Limitations of this Work

7.3.1 Experimental

7.3.2 Theoretical

7.4 Future Work

7.S Final Concluding Remarks


- 148 -

7.1 Aims of the Investigation

The electrostatic powder coating industry has many advantages

over other industries in the finishing markets. However its growth

has been curtailed to approximately 10% of the world market due to

the limitations of use by major coating producers. Problems

associated with colour change and control of film thickness have

been paramount in causing this curtailment.

The aim of these investigations has been to observe the effects

produced by changing the particle size distribution of the powder

used to produce spray coatings. In particular, changes in physical

properties of the coatings have been monitored for a variety of

different sized powders. This has been combined with studies of

particles packing onto substrates so that improved operating

procedures (e.g. optimisation of size distribution etc) can be

suggested based on the results of such investigations. It is

important that if the size distribution of a powder is altered to

gain (i) higher deposition efficiency, (ii) control of film thickness,

or (iii) reduction of problems associated with colour change,. that

there is no undue deterioration in physical properties of the film.

Although several workers have carried out fundamental

investigations into the role of operational parameters on deposition

efficiency surprisingly little attention has been given to the effect

of physical characteristics of the powder on the properties of

coatings. Previous research has been summarised in the text and

clearly shows that a great deal of information is available with

regards to obtaining high deposition efficiencies by controlling

operational variables, such as charging voltage, substrate width


- 149 -

etc etc. However individual forces acting on each particle (due to

the influence of such parameters) has not been considered and it has

been one of the aims of this work to obtain a clearer understanding

of which forces are most important in the electrostatic spraying

process.

Other researchers, who have dealt with the levelling of powder

coatings, acknowledge that the irregularities in the powder layer

can affect the final film produced i.e. extent of orange peel.

Similarly, although the fusing and flowing of the powder layer and

the effect of formulation has been investigated, little insight into

the way in which a packing is actually formed has been obtained.

A literature survey revealed numerous methods which are

frequently used in industry for analysing powder coatings. The

primary difficulty associated with such testing techniques is that

no single property can be measured without encompassing another. This

is especially true for mechanical type tests where a variety of

inter-related properties are measured. The analysis of data from

quality testing techniques also presents problems. Comparisons to

standards are often required and even then the significance of

differences obtained must be questioned.

A series of tests were adopted and used in conjunction with a

developed experimental procedure for providing reproducible test

samples for the different powders used. Coatings were prepared,

tested and collated with thickness, powder type and physical property

measured. By considering only one powder in these experiments the

effects of powder formulation could be neglected. Similarly effects

due to stoving were neglected since identical conditions were used

throughout the production period.


- 150 -

The results from these tests were indicative of the fact that

a commercial powder had been chosen. The coatings, as expected, had

generally excellent physical properties and hence identification of

any significant differences in their physical properties was made

all the more difficult.

The mechanisms and effects associated with the packing of

particles on a substrate were investigated both experimentally and

theoretically.

Although practical investigations were very simple in nature

they permitted observations of how different sized particles are

distributed in a packing. Results of these exploratory experiments

led to quantitative analysis in which the powder size distributions

were determined before and after spraying.

Theoretical investigations were carried out in order to

understand some of the observed phenomena. A fundamental approach was

taken to investigate the role of forces acting on particles of

different size in flight. Although such approach has to be necessarily

qualitative, the aim was to gain an overall insight into the forces

that primarily act on particles in flight, and especially when particles

come close to the substrate. It was then the intention to try and

apply the results, with reservations on applicability, to the general

system.

Information obtained from computer experiments enabled the forces

acting on a particle at any point in its trajectory to be considered

and compared to other positions, different sized particles and various

operating conditions. Although the simu1ations were restricted to

2 dimensions several interesting effects were observed.


- 151 -

Therefore the overall scheme of research was:-

(i) Development of a reproducible coating technique.

(ii) Production of test samples using narrow size fractions of

a commercial powder.

(iii) Assessment of the physical properties of coatings by a suitable

test procedure ..

(iv) Observations of particles packing and stoving.

(v) Development of trajectory equations to describe the flight of

particles.

(vi) Simulations of particles packing under different operating

conditions, including studying

(a) monosized powders

(b) size distributed powder

(c) layer of charge" on'substrate

for changes in particle size (a, c) , transport air velocity

(a,b,c) and thickness of charge layer (c) •

7.2 Summary of Results of Investigations

7.2.1 Testing of Properties

The results of tests on these coatings for properties

of gloss, roughness, adhesion etc showed that there is no significant

difference for changes in particle size sprayed or thickness of

coating. Narrow size fractioned powders showed no trends, with the

results being scattered about an average value. The commercial powder

did show more marked trends for roughness and gloss with changes in

thickness. When comparing the spread of results from these tests with

low and high quality finishes it could be concluded that the variations

were very small.


- 152 -

This.work suggested that any change made to the size distri-

bution of the powder sprayed, in order to increase or control

deposition, would not appreciably effect the surface finish. This

is an important conclusion.

7.2.2 Packing and Stoving

Simple stoving experiments suggested that the excellent

flow. characteristics of the epoxy resin powder were responsible for

the small variations in testing results.

From microscopic observations of packings (unfused) a variation

of size distribution was observed through the packing depth and an

overall change in distribution between the powder sprayed and deposited

was quantified. A loss of fines was determined and it appeared from

observations that this was primarily from the upper layers of the

coatings.

7.2.3 Trajectory Equation

Using a force balance, a set of trajectory equations were

developed to describe the flight of particles to the substrate.

Several simplifying assumptions were used to allow easier handling

of these equations. Although straightforward models were used for the

air flow, electric field and calculation of drag coefficient,

individual interparticulate electrostatic forces were incorporated.

The approach was kept 2 dimensional to permit pictorial representation

of the results, where 3-dimensional analysis would have given many

problems.
- 153 -

7.2.4 Computer Simulation

The trajectory equations were found to be of a stiff

nature. A suitable NAG library routine was employed to effect

numerical integration over the flight range. Many problems were

encountered. in handling the 'physical' situation of particles landing

on a substrate and in minimising computing time. Summation of all

interparticulate electrostatic forces at each step of the integration

limited the number of particles that could be considered due to the

restraints on computer time.

Although several assumptions in the program have been used the

overall representations appear very reasonable. The flow model

appeared adequate and when no electrostatic forces were applied it

was shown that deposition would not occur in the majority of cases.

Starting the trajectories mid-way between gun and target with

approximated initial velocities was also shown to be reasonable.

The results obtained show clearly similar effects to those

observed in the practical experiments. Particle size and transport

air velocity were both found to be very important in determining the

type of packings produced.

From simulations of the spraying of mono sized particles it has

been shown that smaller particles tend to produce thin packings

covering a higher percentage of area of the substrate. These

particles are seen to be strongly influenced by electrostatic

forces in the vicinity of the packing and substrate. Their movements

are jerky, frequently change direction, whilst large particles

trajectories are more streamline being less influenced by the

charged packing structures. Transport air has been observed to


- 154 -

influence the thickness of packings. This is due to the splaying

effect and also the change in the balance between aerodynamic and

electrostatic forces.

Results from studies with size distributed powders indicate

that large particles produce 'shadow' areas in which there is an

absence of fines. As the packing becomes thicker so more fines are

repelled and large particles become abundant in central and upper

regions of the coating. Both these effects show good agreement with

experimental findings.

Large particles formed tall dendrite type structures which did

not occur for small particles. These large particle tree formations

gave rise to the formation of large crevasses and valleys in the

upper layers of the packing which is not so noticeable close to the

substrate due to the presence of the small particles.

Consideration of a theoretical charged layer showed how particles

of different size behave in different ways. This set of experiments

allowed particles approaching much thicker coatings to be studied.

Results indicated that particle size was more important than air·velocity in

determining whether particles landed on a coating. It was clearly

shown that as the thickness of a coating increases deposition is

subject to size restrictions. Small particles are not deposited with

thick coatings and in such cases the balance of forces close to the

substrate is of paramount importance.

Although the flow pattern, surroundings and plate size in a

coating system will undoubtedly be different for each industrial

application, these results have outlined the important factors in

particle deposition. In the production of thick coatings a wide size


- 155 -

distributed powder will produce a fines rich overspray and will

drastically reduce deposition efficiency as coating thickness

increases. For thinner films better control would be obtained if

less large particles were used. Increased movement of spray guns

will cause less size segregation and hence produce more uniform

films .

7.3 Limitations of this Work

In carrying out any piece of research limitations of time and

resources means that certain areas are not as fully investigated as

they could be and that simplifying assumptions have to be applied.

This work is no exception and it is important to recognise not only

the achievements but the limitations of the work.

7.3.1 Experimental

Narrow size distributed powders and a commercial powder

were chosen for tests on film properties with the intended exclusion

of effects due to formulation. It appears from the results of these

tests and stoving observations that the results could be strongly

related to the powder formulation used, whereas more definite trends

might be observed for other types of powders.

7.3.2 Theoretical

There are many assumptions that have been incorporated

into the computer program to allow for a relatively easy method of

solution to be followed.

Also excluded are some effects that would be readily observed

in the real situation but are very difficult to simulate. Together

these are summarised below:-


- 156 -

(1) 2 dimensional only

(2) Simple air model

(3) No gravity force

(4) No bounce-off or movement of particles after landing

(5) Small area considered

(6) One particle in flight at a time

(7) Spherical particles

(8) No space charge

(9) No leakage of charge from packing

However, the aim has been to gain an insight into the important

parameters that effect the process and this has been achieved to a

considerable extent.

Development could be continued so that the above list is

shortened as far as is reasonably possible. This would serve to

make the simulation closer to reality than at present.

In its present form, therefore, the computer simulation technique

can only be considered as a qualitative method of observing the

effects of forces acting on particles, even though the trends that

have been already observed do agree qualitatively with those obtained

experimentally.

7.4 Future Work

The investigations reported here have provided important

conclusions with regards to the electrostatic coating process. Both

experimental and theoretical work has shown how particles behave in

different ways according to their size and the operating conditions

employed.

It has been shown that for the commercial type powder little

change in physical properties of the coating occurs with ch8nge in


- 157 -

particle size or film thickness. However, staving experiments suggest

that this could be due to the powder's excellent flow properties.

Hence further experimental work could be performed to investigate

whether these observations are true for other powder formulations.

This work has purposely avoided changes in formulation since the

experimental program was already very extensive. Future work could

therefore be restricted to the testing of fewer properties i.e. just

adhesion and gloss. Comparisons could be made between poorer flowing

powders and excellent flowing powders. The effect of size ranges

could also be investigated in this way with regards the levelling

of the powder and in turn related to the porosity of the packing.

Investigations into particle packing have shown how various

sizes behave differently with regards deposition. This area could

be more fully investigated by

(i) Measurement of size distributions of powder on a plate at

different coating times.

(ii) Collection and size analysis of overspray.

(iii) High speed colour cine filming of a spray of particles landing

on a substrate using a powder with coloured size fractions.

The computer program has enabled effects of changing air

velocity and particle size to be studied. Trends showing the way

in which the magnitude of forces acting on the particles change have

been described and this itself has proved very valuable for under-

standing the coating system. However the program is subject to

many simplifying assumptions and therefore there is plenty of scope

for changes to be made in an attempt to simulate the real system

more closely:-
- 158 -

(i) Air flow. The model for air flow is currently based on flow

past a flat plate and the trajectory is started half way

between the gun and target. Development of equations to

describe the air flow from the gun (including taking into

account the type of deflector, nozzle etc) would help to optimise

the system With regards deposition efficiency, gun design etc.

(ii) At present only 2-dimensional analysis is made. Although this

does give indications of trends, the effect of a packing on

a particle in flight would be different for the three dimensional

case. The setting up of the equations is not in itself

difficult but representation of 3-D results would require

considerable thought.

(iii) During computer experiments it was found that for large numbers

of· particles (~300) solving the trajectory equations came close

to the user time limit. The time for computation increases

rapidly as the number of particles landed increases since each

particle in the packing is used for the calculation of electrostatic

forces and is checked for collision. The use of a theoretical

model to simulate a layer of uniform charge aided this situation,

but reduces the amount of information that can be gained e.g.

packing orientation, porosity, size segregation etc. If a

packing model could be developed to describe the force acting

on the particle due to the packing this could reduce computation

substantially.

(iv) No attention is g~ven in respect of space charge and only one

particle is considered in flight at any instant. The program

could be adjusted so that, say, 300 particles were in flight at


- 159 -

once and so the. effect of each particle in flight on the others

would be taken into account. It would then be less likely that

a particle would reach an equilibrium condition as shown in the

plots here. This technique would require more computation time

since 300 particles have to be compared and accounted for

through each of the 300 particle trajectories.

(v) In trying to accomplish (iv) particle bounce would need to be

considered since it would be very probable that particles

would collide in flight.

(vi) As gravity is ignored in the experiments reported in this work,

the positional direction of the substrate is of little consequence.

The packing of particles onto a 3-dimensional object (i.e. a

cube) would possibly produce very interesting information on

how ,particles are deposited on surfaces facing different

directions. The air flow model would require adjustment·, but

if accomplished the wrap around effect could be investigated.

(vii) The problems due to splaying were very marked due to the program

being based on a static type gun. The comparison of packings

obtained by using a gun which traverses across the object could

show the type of coatings produced by robot guns in an industrial

plant, and also· reduce the problems due to the effects of

splaying.

7.5 Final Concluding Remarks

The intention of the investigations reported here was to

ascertain whether changes in the particle size distribution of electro-

statically sprayed powders effect the final finish of the coatings

produced. A positive conclusion formed on the basis of data obtained


- 160 -

from the testing of several hundred sample plates was that particle

size or film thickness did not affect the physical properties of

coatings for the epoxy powder employed: This suggests that any change

in the size distribution of the powder can be made to' increase

deposition efficiency or to control film thickness without fear of

losing' the excellent physical properties associated with powder

coating.

The use of a computer',simulation technique has given an under-

standing of real effects observed in the laboratory, and although

simplified, has been extremely useful in indicating the important

operating parameters of the electrostatic powder coating process.

It has enabled the process to be studied from a 'novel fundamental

standpoint rather than in the more usual experimental manner. Further

development of this technique should therefore be instigated to widen

the range of applicability.


REFERENCES

Listed in alphabetical order of firs!: named authors


REFERENCES

1.0. Ai tken , 1968, Proc. lnt. Sym. Powder Coatings, Paper No V/18,

13-15 Feb.

R. AlIen, 1973, lnd. Fin. Surf. Coatings, 25 (301), 20-27.

M.L. Ang, 1981, Ph.O. Trajectories of Charged Particles in Electro-

static Powder Coating Systems, Loughborough University.

Anon, 1974, Prod. Fin., 39 (3), 24-25.

Anon, 1976, lnd. Fin. Surf. Coatings, 28 (332), 17-23.

Anon, 1976, lnd. Fin. Surf. Coatings, 28 (331), 10-16.

Anon, 1974, Prod. Fin., 39 (2), 20-21.

R.A. Ashdown, 1974, Prod. Fin., ~ (2), 33-36.

ASTM Recommended Practices for Testing Polymeric Powders and Powder

Coatings, 1973, 03451-75.

ASTM Handbook of ASTM Standards, 1975, Part 27.

J.O. Bassett, R.P. Corbett, J. Cross, 1975, Inst. Phys. Conf. Ser.,

No 27.

O.R. Birchenough, M.J. Marsha11, 1972, lnd. Fin. Surface Coatings,

24 (288), 4-7.

R.R. B1ake1y, 1977, lnd. Fin. Surf. Coatings, 29 (344), 28-36.

J. Boxa11, J.A. Fraunhofer, 1974, J. Paint Manufacture, 44 (5),

26-29.

A.W. Bright, 1977, J.O.C.C.A., 60(1), 23-27.

British Standard Methods of Tests for Paints, 1974, B.S. 3900,

Part E7.

British Standard Method of Tests for Paints, 1974, B.S. 3900,

Part E6.

British Standards Method of Tests for Paints, 1967, B.S. 3900,

Part 02.
w. Brushwe11, 1973, Ind. Fin. Surface Coating, 25 (297), 24-30.

G.D. Cheever, 1975, J. App1. Po1ym. Sci., 19, 147-163.

A.E. C1axton, 1975, Preprints Oil & Colour Chemists Assoc. Biennial

Confs., Scarborough, June, 28-34.

J.H. Co11ing, W.E. Craker, J. Dunderda1e, 1968, J.O.C.C.A., 51,

514-546.

M. Cow1ey, 1974, Ind. Fin. Surf. Coatings, 25 (308), 20-21-

W. Crisp, 1973, Ind. Fin. Surf. Coatings, 24, 299.

J. Cross, 1975, Static Electrification, Inst. Phys. Conf. Ser.,

No 27, 202-213.

J.A. Cross, S. Singh, A.B. Ahmed, 1980, J.O.C.C.A., 63, 326-329.

V. K. Croutch,. 1976, 4th American Conf. on Powder Coating, Toronto,

Canada.

E. W. Drew, 1977 , Prod. Fin ... 41 (11) , 8-10.

M.J. Drury, 1974, Prod. Fin, 39(2), 37-38.

J .A. Fraunhofer (Part 1), 1971, Prod. Fin. , 35 (8) , 22-25.

J .A. Fraunhofer (Part 6), 1972a, Prod. Fin. , ~(9), 47-52.

J.A. Fraunhofer (Part 7), 1972b, Prod. Fin. , ~(11) , 12-18.

J.A. Fraunhofer (Part 8) , 1972c, Prod. Fin. E (1) , 20-25.

S. Gabrie1, 1975, J .. O.C.C.A., 58, 52-61-

N.l. Gaynes, 1977, Testing of Organic· Coatings, Noyes Data Corp.

GINOF User Manual, 1976, Issue 2, Computer Aided Design Centre,

Cambridge.

A. Go1ovoy, 1973a, J. Paint Tech. , 45(580) , 42-48.

A. Go1ovoy, 1973b, J. Paint Tech·.· .. 45 (585) , 74-79.

A. Go1ovoy, 1973c, J. Paint Tech. , 45(585), 68-73.

A. Go1ovoy, A.D. Co1vin, 1975, J. Paint Tech., £(604), 65-70.

R. Goudie, 1974, Prod. Fin., 12.(3) , 14-17.


H.K. Hammond Ill, 1974, Official Digest, April, 360-365.

Handbook of Chemistry and Physics, 1977, 58th Ed., CRC Press.

M.J. Hannon, D. Rhum, K.F. Wissburn, 1976, J. Coatings Tech., 48

(621), 42-54.

G.F. Hardy, 1974, J. Paint Tech., 46(599), 73-82.

S.T. Harris, 1976, The Technology of Powder Coatings, Portcullis

Press.

S. Huey, 1964, Symp. Gloss Measurement, Official Digest, April.

D. Jarvis, 1974, Prod. Fin., 11(3), 12-13.

M.B. Kilcullin, 1975, Preprints Oil Col. Chem. Assoc. Biennal Conf.,

Scarborough, 50-61.

C.H.J. Klaren, 1977, J.O.C.C.A., 60, 205-213.

K. Koffmann, T. Kosbahn, 1966, Farbe. U. Lacke, 21(2), 119-127.

C. Korf, 1976, Powder Coating, R.A. Chandler Ltd.

T. Kosbahn, 1964, Farbe. U. Lacke, 70 (9) , 693-702.

S. Kut, 1972, Prod. Fin. , 36(4), 21-29.

S. Kut, 1971a, Prod. Fin. , ~(3), 32-38.

S. Kut, 1971b, Prod. Fin., ~(2), 28-30.

S. Kut, 1971c, Prod. Fin. , 36 (1) , 31-37.

S. Kut, 1974, Prod. Fin. , 39 (2) , 25-3l.

P. G. DeLange, 1978, Fin. Ind., l(3), 35-38.

T.D.T. Latter, 1976, Prod. Fin. , 40 (8) , 46-47.

R. Lever, 1978, Fin. Ind., l(2), 41-44.

B. Lindberg, 1975, Preprints Oil Col. Chem. Assoc., Biennial Conf.,

Scarborough, 105-119.

M.J. Marshall, D.R. Birchenough, 1972, Prod. Fin., lL(2), 40-41.

D.R. Mason, J.L. Stevens, 1978, Introductory Course to VAX 11/780,

ICI Corporate Laboratory.


S.L. Mason, 1974, Prod. Fin., ~(2), 22-24.

W.J. McGill, 1977, J.O.C.C.A., 60, 121-126.

E. Miller, 1974, Fundamentals of Powder Coating, S.M.E.

C.A. Mitton, 1968, Accelerated Environmental Testing, Paper No V/17,

Proc. Int., Symp., Powder Coatings.

A.D. Moore, 1973, Electrostatics and its Applications, Wiley

Interscience.

Myers and Long (Ed.), 1968·, Characterisation of Coatings, Physical

Techniques Part 1, Marcel Dekker.

NAGFLIB, 1976, NAG library routine D02AJF, Integration of a system

of first order ordinary differential equations over a range using a

divided difference formulation of Gear's method for stiff systems

of equations 1387/0 : Mk 5.

NAGFLIB, 1977a, NAG library routine GOSCAF, Generation of a pseudo-

random real number, uniformly distributed between 0 and 1, 1443/0 :

Mk 6.

NAGFLIB, 1977b, NAG library routine G05CBF, Setting the state of the

generation storage to a repeatable value, 1440/0 : Mk 6.

NAGFLIB, 1978, NAG library routine D02QBF - Integration of a stiff

system of first order differential equations using a variable-order,

variable-step Gear method, 1640/0 : Mk 7.

K.C. O'Neill, 1977, Aust. O.C.C.A. Proc. News, 14, 6-12.

V.G. Nix, J.S. Dodge, 1973, J. Paint Tech., 45(586), 59-63.

Nottingham Algoriths Group, 1973, NAG Library routine D02ABF,

Integration of a system of ordinary differential equations over a

range using Merson's Method. Doc. No. 564.

s. Oglesby, Jnr, G.B. Nichols, 1978, Electrostatic precipitation,

Marcel Dekker Inc, New York.


P.H. Ong et al., 1975, Static Electrification, Inst. Physics Conf.

Ser" No 27.

S.E. Orchard, 1962, J. App1. Sci. Research, l!(A), 451-465.

D. Payne, 1973, Ind. Fin. Surf. Coatings, 24, 299.

M.M. Pauthenier, M. Moreau-Hanot, 1932, J. Phys. Radium, l(7),

590-615.

F.W. Peek, 1929, Dielectric Phenomena in High Voltage Engineering,

3rd Ed. McGra~Hi11.

P.P.I. Company, 1963, Brice Phoenix Light Scattering Photometer

Instruction Manual Operation Manual OM-2000, Philadelphia, USA.

A. Quach, 1973, Ind. Eng. Chem. Prod. Res., Dev., 11(2),110-116.

P.S. Quinney, N.J. Tighe, 1971, Br. Po1ym. J., l(10), 274-278.

L. Schi11er, A. Naumann, 1933, Z. Ver. deut. Ing. 22, 318.

S. Singh, 1979, Coordinated Powder Research Programme. Southampton

University, Progress Report, June.

S. Singh, 1980a, C.P.R.P. Southampton University Progress Report,

Jan.

S. Singh, 1980b, C.P.R.P. Southampton University Progress Report,

March.

J. Smarsh, 1972, J. Paint Tech., 44 (565), 30-37.

S. Spinde1, 1973, Am. Paint. Tech., ~, 33-38.

G.T. Spitz, 1973, ACS Div. Org. Coatings Plastics Chem. Papers, 33(2)

502-507.

R.C. Spragg, D.J. Whitehouse, 1972, Measurement and Control, ~,

95-101.

H. Stein, 1972, Ind.-Fin. Surf. Coatings,- 24(294),32-33.


B. Svoboda, o. Ko1ar, 1962, Chem. Prumysl, 12(7), 281-384.

M. Tahan, 1974, J. Paint Tech., 46(59)), 35-47.

M. Tahan, B.J. Tighe, 1974, J. Paint Tech., 46(590), 48-57.

J.D. Tiernay, 1975, Prod. Fin., ~(11), 14-15.

Y.C. Ting, 1978, M.Sc. Thesis, Back Ionisation in Electrostatically

deposited layers, Southampton University E1ec. Eng. Dept.

J.D. Toff, 1979, Prod. Fin., il(ll) , 10-12.

H.R. Valentine, 1967, Applied Hydrodynamics, 2nd Ed. Butterworths.

K. Ven1et, 1973, Ind. Fin. Surf. Coatings, 25(299), 30-41.

B. Whiting, J.K. Dennis, 1977, Report, Dept. Metallurgy and Materials,

University Aston in Birmingham.

W.G. Willows, 1968, Powder Coatings Int. Symp., 13-15 Feb, London.

S.M. Wo1pert, 1973, ACS Div. Org. Coatings, P1ast. Chem. Papers,

33 (2), 485-492.

S. Wo1pert, J.J. Wojtkowiak, 1972, ACS Div. Org. Coatings, P1ast.

Chem. Papers, ~(2), 456-461.

S. Wu, 1976, Po1ym. P1ast. Tech. Eng., l(2), 119-220.

U. Zor11, 1972, Progr. Org. Coatings, 1, 113-154.


U. Zor11, 1963, Deutsche Farben-Zeitschrift, 1I(1), 6-12.
NOMENCLATURE

a radius of wire of electrode; radius of particle; half width

of plate

b ionic mobility, radius of electrode

C calculation constant in interparticulate electrostatic force

CD drag coefficient

d diameter

dp diameter of particle

E field strength, resultant field

Eo Uniform field

Ec critical electrical breakdown strength

F Force

FAx force between particles of opposite charge in x direction

FD drag force in x direction


x
FE field force acting on particle ~n x direction
x
Fr image force in x direction
x
FRx force between particles of like charge in x direction

GT distance between gun and target

1 half width of plate

L distance between gun and target

m roughness factor

q charge on particle

qs saturation charge

Qa charge density of theoretical layer of charge

r radial distance from electrode

r radius of curvature of corona glow region


c
R calculation variable (define in Section S.S:.?); .. distance, between

two particles

Re particle Reynold's Number


p
t time; thickness of theoretical layer of charge

u velocity of air in x direction

U uniform air velocity

v velocity of air in y direction

VreI relative velocity between air and particle

V voltage potential

Vo voltage potential at electrode

x displacement in X-direction

y displacement in Y-direction

z displacement in Z-direction

e Angle to horizontal; calculation variable as defined in section

E dielectric constant of powder

EO permittivity of free space

~ viscosity of fluid

P density of fluid

Pf density of fluid

p.
~
ionic charge density

T time constant

q, potentfal function

.p steam function
Subscripts for use with x, y, q, d and R

i refers to a packed particle

j refers to the image of a packed particle

k refers to the image of the oncoming particle

o refers to the oncoming particle


APPENDIX A

GLOSS TEST RESULTS


Table Al. Gloss Test Results at Various Film Thicknesses for Commercial Powder

Film Thickness ()lm)

22 27 33 37 42 47 52 56 62 67 74 85 100

Average 20.68 22.04 26.30 27.19 26.43 28.07 28.04 29.95 27.49 30.45 29.34 30.91 30.99
Gloss
Factor Standard Deviation 1. 36 0.92 1. 79 1.11 4.44 0.80 1.60 1.84 1.64 1.25 1.71 2.23 1. 26

Probable Error 0.60E 0.411 0.80' 0.496 1.986 0.35E 0.7It 0.823 0.73" 0.559 0.76' 0.997 0.563

Average 99.4 01.2 102.6 ,03.8 ~02.8 05.6 104.6 106.0 ,05.2 108.8 1,03.2 08.2 ,05.8
Peak
Height Standard Deviation 2.94 2.14 1.96 1.72 6.55 2.42 3.14 3.16 3.25 3.66 2.79 3.92 3.31
(mV)
Probable Error 1. 31' 0.957 0.87 ' 0.769 2.929 1.082 1.40l 1.413 1.45 1.637 1.248 1. 753 1.480

Width at Average 4.82 4.60 3.9.2 3.82 3.89 3.76 3.74 3.55 3.83 3.57 3.53 3.51 3.42
Half Peak
Height Standard Deviation 0.24 0.14 0.21 0.10 0.32 0.06 0.13 0.16 0.14 0.09 0.24 0.17 0.07
(DEG)
Probable Error 0.107 0.063 0.094 0.045 0.143 0.027 0.05E 0.072 0.06" 0.040 0.107 0.076 0.031
Table A2. Gloss Test Results at Various Film Thicknesses for <10~m Size Range Powder

Film Thickness (~m)

16 22 28 32 37 44 56 62 67 74 84 95

Average 21.48 17.36 8.25 17.71 22.35 18.56 25.31 18.38 25.28 22.91 25.10 24.46
Gloss
Standard Deviation 3.24 0.46 3.20 3.30 1. 93 2.68 4.49 2.71 3.36 2.23 2.54 2.82
Factor
Probable Error 1.449 0.206 1.431 1.476 0.863 1.199 2.008 1.212 1.503 0.997 1.136 1. 261

Average 99.8 90.2 92.8 90.8 94.4 90.8 97.2 88.0 97.2 96.2 95.2 96.4
Peak
Height Standard Deviation 5.78 2.32 4.78 4.12 3.14 2.93 5.19 3.52 5.74 3.71 3.54 4.03
(mV)
Probable Error 2.584 1.038 2.138 1. 843 1.404 1.310 2.321 1.574 2.567 1.659 1.583 1.802

Average 4.73 5.20 5.20 5.29 4.25 4.96 3.94 4.86 3.89 4.23 3.82 3.98
Width at
Half Peak
Standard Deviation 0.56 0.12 0.64 0.86 0.29 0.51 0.58 0.47 0.34 0.32 0.24 0.33
Height
(DEG)
Probable Error 0.250 0.053 0.286 0.385 0.130 0.228 0.259 0.210 0.152 0.143 0.107 0.148
Table A3. Gloss Test Results at Various Film Thicknesses for 10-20~m Size Range Powder

Film Thickness (~m)

18 22 27 32 37 42 46 56 62

Average 29.88 28.57 29.75 31.59 31.21 32.24 29.25 30.55 31.99
Gloss 1. 96 2.43
Standard Deviation 2.92 0.87 1.53 0.79 1.04 1.71 0.73
Factor
Probable Error 1.306 0.389 0.684 0.877 0.353 0.465 1.087 0.765 0.326

Average 110.4 105.4 108.6 111.2 109.0 110.0 106.8 107.2 108.6
Peak
Height Standard Deviation 2.42 2.33 2.58 2.04 1. 79 2.76 2.14 4.79 1.85
(mV)
Probable Error 1.082 1.042 1.154 0.912 0.801 1.234 0.957 2.142 0.827

Width at Average 3.73 3.69 3.66 3.53 3.49 3.41 3.67 3.51 3.40
Half Peak
Standard Deviation 0.34 0.09 0.12 0.17 0.07 0.08 0.27 0.09 0.09
Height
(DEG)
Probable Error 0.152 0.040 0.055 0.076 0.031 0.036 0.121 0.040 0.040
Table A4. Gloss Test Results at Various Film Thicknesses for 20-30~m Size Range Powder

Film Thickness (~m)


-
27 32 37 42 48 52 57 62 67 71 77 84 100

Average 28.83 29.59 30.18 30.42 31.08 30.35 31.02 29.64 29.24 31.43 31.30 29.34 29.37
Gloss Standard
1.60 0.74 0.91 1. 70 1.58 1.34 1.65 0.86 1.86 1.87 3.29 4.22 4.75
Factor Deviation
Probable
0.716 0.331 0.407 0.760 0.707 0.599 0.738 0.385 0.832 0.836 1.471 1.887 2.124
Error

Average 104.2 106.8 107.0 107.2 107.8 107.8 106.0 ~03.0 104.8 107.6 105.0 103.8 103.6
Peak
Standard 2.93
Height 0.98 1.10 3.66 1.17 1.83 2.19 3.74 2.79 3.44 4.52 5.31 3.50
Deviation
(mV)
Probable
1.310 0.438 0.492 1.637 0.523 0.818 0.979 1.673 1.248 1.538 2.021 2.375 1.565
Error

Width at Average 3.62 3.61 3.55 3.53 3.48 3.56 3.42 3.48 3.59 3.43 3.38 3.75 3.59
Half
Standard
Peak 0.11 0.08 0.09 0.13 0.16 0.13 0.12 0.14 0.15 0.16 0.23 0.48 0.42
Deviation
Height
Probable
(DEG) 0.049 0.036 0.040 0.058 0.072 0.058 0.053 0.062 0.067 0.072 0.103 0.215 0.189
Error
Table AS. Gloss Test Results at Various Film Thicknesses for 30-40~m Size Range Powder

Film Thickness (~m)

29 33 37 42 48 52 57 62 67 72 " 77 82 87 93 104

Average 27.09 28.23 27.38 29.63 28.51 28.25 30.59 28.58 28.00 26.73 26.91 26.21 27.30 27.06 28.60
"Gloss Standard
1.32 1.49 0.99 2.29 1.98 2.30 1.36 0.87 1.37 0.69 1.55 1.48 1.01 2.49 2.18
Factor Deviation
Probable 0.93
0.59 0.67 0.44 1.02 0.89 1.03 0.61 0.39 0.61 0.31 0.69 0.66 0.45 1.11
Error

Average 106.4 107.8 106.4 107.6 105.4 105.2 110.25 106.2 106.6 105.8 104.6 102.4 104.8 104.8 108.0
Peak
Standard
Height 2.33 2.48 3.56 2.06 2.15 4.79 3.34 1.94 1. 36 0.98 1.62 2.87 1.33 3.82 1.9
Deviation
(mV)
Probable 0.85
0.91 1.11 1.59 0.92 0.96 2.14 1.49 0.87 0.61 0.44 0.72 1. 28 0.60 1.71
Error

Width at Average 3.93 3.83 3.89 3.65 3.71 3.74 3.61 3.73 3.82 3.96 3.90 3.92 3.84 3.90 3.80
Half
Standard
Peak Deviation 0.12 0.14 0.16 0.21 0.20 0;16 0.07 0.12 0.20 0.10 0.18 0.16 0.15 0.30 0.31
Height
Probable
(DEG) 0.05 0.06 0.07 0.09 0.09 0.07 0.03 0.05 0.09 0.05 0.08 0.07 0.07 0.13 0.14
Error
Table A6. Gloss Test Results at Various Film Thicknesses for 4o-50~m Size Range Powder

Film Thickness (pm)

33 37 42 47 52 57 63 75 84 92 97 106;, 135

Average 25.97 25.85 28.58 28.38 27.46 26.95 26.92 28.86 30.37 27.97 28.00 27.42 28.61
Gloss Standard
0.83 3.72 2.19 1. 58 0.97 1. 73 1.63 3.05 0.94 2.68 1.43 2.36 2.59
Factor Deviation
Probable
0.371 1.664 0.979 0.707 0.434 0.774 0.729 1.364 0.420 1.199 0.640 1.055 1.158
Error

Average 104.4 104.4 105.0 106.8 103.0 104.6 103.0 106.0 108.2 104.6 105.4 105.6 107.0
Peak
Standard
Height 1.96 2.28 2.10 2.79 1.10 4.36 2.37 4.15 1.60 3.72 1.96 3.01 2.76
(mV) Deviation
Probable 0.88 1.02 0.94 1.25 0.49 1.95 1.06 1.86 0.72 1.66 0.88 0.35 1. 23
Error

Width at Average 4.02 4.10 3.69 3.77 3.75 3.89 3.83 3.70 3.57 3.76 3oT] 3.87 3.76
Half
Peak Standard 0.10 0.17 .
Deviation 0.56 0.24 0.18 0.09 0.14 0.27 0.10 0.25 0.15 0.23 0.26
Height
Probable
(DEG) 0.05 0.25 0.11 0.08 0.04 0.06 0.08 0.12 0.05 0.12 0.07 0.10 0.12
Error
Table.A7. Gloss Test Results at Various Film Thicknesses for +50pm Size Range Powder

Film Thickness (pm)

38 43 47 52 57 62 67 71 76 81 87 92 97 104 122

Average 27.10 28.15 28.71 30.99 28.92 29.32 30.00 30.89 29.72 29.09 28.99 29.37 29.00 29.59 30.70
Gloss
Factor Standard 2.46 2.33 0.83 1.22 1.37 1.46 3.01 2.27 3.26 1.22 1.16 1.59 2.04 1.91
1.10
Deviation
Probable 0.55 0.61 0.49 0.65 1.35 1.02 0.55 0.52 0.71 0.91 0.85
1.10 1.04 0.37 1.46
Error

Average 103.8 106.2 105.2 108.2 105.6 105.8 105.4 106.8 105.0 101.8 104.8 106.2 104.8 107.00 107.20
Peak
Standard 2.73
Height 3.19 2.04 2.48 2.04 3.71 1.02 1.94 2.45 8.35 1.72 2.14 3.19 2.68 1.47
Deviation
(mV)
Probable
1.43 0.91 1.11 0.91 1.22 1.66 0.46 0.87 1.10 3.73 0.77 0.96 1. 43 1.20 0.66
Error

Width at Average 3.85 3.80 3.66 3.49 3.66 3.61 3.52 3.49 3.55 3.51 3.62 3.62 3.62 3.63 3.50
Half
Standard
Peak 0.25 0.33 0.04 0.09 0.10 0.11 0.16 0.29 0.21 0.17 0.10 0.11 0.13 0.20 0 •. 18
Deviation
Height
Probable
(DEG) 0.11 0.15 0.02 0.04 0.05 0.05 0.07 0.13 0.09 0.08 0.05 0.05 0.06 0.09 0.08
Error
Table A8. Results of Linear Regression Fit to Experimental
Data from Gloss Test

Powder Measured y Correlation


Slope ·t = 15 t = 100
Type Quantity Intercept Coefficient

Gloss Factor 0.117 21.17 0.850 22.92 32.87


Commer-
Peak Height 0.0829 99.91 0.726 101.2 108.2
cia I
Width -0.0145 4.63 0.802 4.41 3.18

Gloss Factor 0.084 17.13 0.669 18.4 25.5

<10Jlm Peak Height 0.0292 92.58 0.210 93.0 95.5

Width -0.0167 5.39 0.750 5.14 3.72

Gloss Factor 0.043 28.93 0.505 29.6 33.2

1O-20Jlm Peak Height -0.0193 109.31 0.157 109.0 107.4

Width -0.006 3.79 0.723 3.70 3.19

Gloss Factor 0.0027 29.98 0.065 30.0 30.3

2o-30Jlm Peak Height -0.038 107.97 0.473 107.4 104.2

Width 0.00 3.54. 0.017 3.54 3.54

Gloss Factor -0.014 28.81 0.274 38.6 27.4

3O-40Jlm Peak Height -0.0268 107.8 , 0.330 107.4 105.1

Width 0.0041 3.75 0.234 3.81 4.16

Gloss Factor 0.018 26.53 0.450 26.8 28.3

4o-50Jlm Peak Height 0.024 103.5 0.480 103.9 105.9

Width -0.0018 3.93 0.395 3.90 3.75

Gloss Factor 0.019 27.98 0.448 28.3 29.9

+50Jlm Peak Height 0.0083 104.98 0.131 105.1 105.8

Width -0.0022 3.77 0.495 3.74 3.55


Table A9. Linear Regression Results. for Gloss Test at
Various Film Thicknesses
Using data from Table 3.2

Film Thickness -- 30um 50um 70um 100um

Slope -0.066 -0.069 -0.077 -0.086


Intercept 31.15 31.56 32.16 32.87
Gloss
Factor d = l5um 30.16 30.52 30.96 31.58
d = 55um 27.52 27.76 27.88 28.14

Slope -0.097 -0.077 -0.053 -0.012


Peak Intercept 109.8 108.9 107.8 106.1
Height d = l5um 108.3 107.7 107.0 105.9
d = 55um 104.5 104.7 104.9 105.5

Slope 0.0052 0.0065 0.0079 0.0073


Peak Intercept 3.52 3.44 3.36 3.34
Width
d = l5um 3.60 3.54 3.48 3.45
d = 55um 3.81 3.80 3.79 3.74

Calculations do not include <lOum size range


FIGURE Al GLOSS FACTOR VS. FILM THICKNESS
POWDER SIZE RRNGE : COMMERCIRL
KEY
40 m- EXrEKIHENTAL DATA

38 - - - BEST fIT

36

n:= 34
Cl
t-
u 32
a:
lL
30
(f)
(f)
Cl 28
-1
t=I
26

24

22 m
m
20
0 10 20 30 40 50 60 70 80 90 100

FILM THICKNESS lfM.)


FIGURE RZ GLOSS FRCTOR VS. FILM THICKNESS
KEJ rOWDER SIZE RANGE: <10fM.
~- £Xr£RIH£NTAL DATA

30 --- - 6EST fIT

28

26

a:: 24
0
. t-
Ll 22
er:
LL
20
en
en 18
C)
-1
<.:J
16

14

12

10
0 10 20 30 40 50 60 70 80 90 100

FILM THICKNESS (~M. )


FIGURE A3 GLOSS FACTOR VS. FILM THICKNESS
KEY rOWDER SIlE RRNGE : lo-zofM.
m- EXrERIHENTAL DATA

35 - - - _ BEST FIT

34

33

~ 32
Cl
t-
Ll 31
Cl:
tL. 30
(i) [!]
en 29
Cl [!]
---l
Cl
28

27

26

25
0 10 20 30 40 50 60 70 80 90 100

FILM THICKNESS ()-1N_J


FIGURE R4 GLOSS FRCTOR VS. FILM THICKNESS
rOWDER SIlE RRNGE : ZO-30pN.
KEY
35 ~- EXrEftlHEHTAL DATA

34 - BEST fIT

33

~ 32
Cl
t- ~
~
u 31 ~ m
Cl: m ~
LL !I!
30
en ~ ~
en 29 ~ ~
Cl ~
--I
t;J
28

27

26

25
0 10 20 30 40 50 60 70 80 90 100
FILM THICKNESS (fM. )
FIGURE AS GLOSS FACTOR VS. FILM THICKNESS
rOWOER SIIE RRNGE : 30-40fM.
~EY

34

33 --- - BEST I' JT

3Z

Cl::: 31
Cl I!J
t-
l...) 30
I!J
a:
u.. Z9
(J') I!J I!l I!J
t!l [']


(J')
Cl
za I!l
...J I!l I!J
t.:J I!J
Z7 I!l
I!J

I!J
Z6

Z5

Z4
la zo 30 40 50 60 70 80 90 100 110

FILM THICKNESS ~M.)


FIGURE AS GLOSS FACTOR VS. FILM THICKNESS
rOWDER SIIE RRNGE : 40-S0fM.
t<EY
34 ~- EXrE"I"E"T~L D~T~

33. - - - - BEST fIT

32.

a:: 3.1
0 I!l
~
u 3.D
Cl:
u.. 2.9 I!l
I!ll!l I!l
ch
en 2.8 I!JI!l
Cl
-....l I!J
l.:)
V I!J I!J

26

2S

24
0 20 40 60 ao 100 120 140 HiO 180 zoo
FILM THICKNESS
FIGURE A7 GLOSS FACTOR VS. FILM THICKNESS
rOWDER SIIE RRNGE : >50fM.
KEY
35 m- EXrEnl"EHTPL DPTP

34 - - - - BEST fIT

33

~ 32
0
t- I!I
LJ 31 I!I
m
cr·
lJ...
30 m
!!J m
(F) !!J m
(F)
I!l !!J !!J !!J
Cl
29 t!I
--1
t:J I!I
28
!!J
27

26

25
25 35 45 55 65 75 85 95 105 115 125

F"ILM THICKNESS ~M. J


FIGURE RS PERh HEIGHT VS. FILM TH I C~\NESS
rOWDER SIIE RRNGE .. COMMERCIRL
KEI
115 m- £~r£"IM£NT~L D~T~

113 - B£ST r IT

1 11
>
2: 109 I!l
I!l

107
t- I!l
I!l
:r: 105
CJ
--.
Lu
:r: 103

~
IT
101
LU I!I
a... 99

97

95
0 10 20 30 40 50 60 70 ao 90 100

FILM THICKNESS ()J N - J


FIGURE A9 PEAK HEIGHT VS. FILM THICKNESS
fOWDER SIlE RANGE; <lO)lN.
j(EY
105

103 --- - BEST rn

~
101
• I!I
>-
:::L 99
m m
97 m
I!I
t-
I:
t..:J 95 m
--.
lU 93 ID
I:
y;
a: 91
lU
G- 89

87

85
0 10 zo 30 40 50 60· 70 80 90 100

FILM THICKNESS ljlM. )


FIGURE AIO PEAK HEIGHT VS. FILM THICKNESS
rOWDER SIlE R~NGE : 10-Z0fM.
KEY
115

113 - - - - BEST I'IT

I!l
~
111
• I!l
> I!l
~ 109 I!l
I!l
107 I!l
t-
::c I!l
c.::J 105
.....
LW 103
::c
I!l
:s:: 101
cc
LW
Il.. 99

97

95
0 10 ZO 30 40 50 SO 70 80 90 100

FILM THICKNESS. [fM.)


FIGURE All rEAK HEIGHT VS. FILM THICKNESS
rOWDER SIZE RRNGE : ZO-30fM.
~EY

115 m- EXrEftInEHTRL DRTR

113 - - - - BEST I'IT

~
III

>
~ 109
[!) [!)
[!)
I!I
107
m
t-
I [!) I!I
t!J 105
...... I!I
I!l I!l
L&..I 103 I!l
::J:

:x:: 101
CC
L&..I
CL 99

97

95
0 10 ZO 30 40 50 60 70 80 90 lOO
FILM THICKNESS (~M.)
FIGURE AIZ FEAK HEIGHT VS. FILM THICKNESS
rcrwDER SIlE RRNGE : 30-40~M.

KE'l'
115 ID- EXrEnINENTAL DATA

113 - - - - BEST FH

III

109
[!) [!)
[!)
107 [!) [!) [!)
l!l
[!) . l!l
105 [!) [!) [!)

103
:s::: 101
a:
w I.
£L 99

97

95
10 20 30 40 50 60 70 80 90 100 110

FILM THICKNESS (}JM. )


FIGURE R13 PERK HEIGHT VS. FILM THICKNESS
rOVOER SIZE RRNGE : 40-S0JH.
t<EY
115 I!I- EXrEftlnEHTAL DATA

113 - - - - BEST rJJ

~
111

>
-
::E 109

107 m
I!I
m
I- m
:c 105 m m
m ill I!I
Cl I!I
...... Iil
w I!I I!I
:c 103
~
a: 101
w
0... 99

97

95
0 ZO 40 60 80 100 1Z0 140 160 180 ZOO

FILM THICKNESS lfM- J


FIGURE A14 PEAK HEIGHT VS. FILM THICKNESS
rOVOER SIIE R~NGE : >50}JN.
J'(EY
115 I!I- Dr'EnJ"EItTJI\. DPTA

113 - BEST rn

111

::t:
:::J 109
I!J

107 m m III

t- m m m
rn e
:::c
l:) 105
I!I I!I m I!I
...... I!I
W
:::c 103 I!I
~
a: 101
w
0... 99

97

95
25 35 45 55 S5 75 85 95 105 115 125

fILM THICKNESS ~M. J


FIGURE A15 WIDTH AT HALF PEAK HEIGHT VS. FILM THICKNESS
rOWOER SIlE RRNGE : COHHERCIRL
f'(E'I'

5.5
t- --- - BEST r IT
I:
t:J
o-t 5
l.I-I I!l
I:
I!l
:);:::
a: 4.5
l.I-I
a...
1L 4
....J
a: I!l
I:

t- 3.5
a:
I:
t-
q 3
o-t
~

2.5
0 10 20 30 40 50 so 70 80 90 100

FILM THICKNESS (?M.J


FIGURE R16 WIDTH RT HRLF PERK HEIGHT VS. FILM THICKNESS
rOWOER SIZE RRNGE : <IO~M.

KEY
6
t- - - - - 8EST rn
:c
t..:I
.-. 5.5
w m
:c
:x:: 5
a:
w
lL

lL. 4.5
-'
a:
:c
t- 4
a:
:c
t-
Cl 3.5
.-.
3:

3
0 10 20 :SO 40 50 60 70 80 90 100

FILM THICKNESS (jJM_l


FIGURE A17 WIDTH AT HALF PEAK HEIGHT VS. FILM THICKNESS
rOWDER 5 IlE RFlNGE : 1 O-ZO)JI't.

KEY
4.5
t- --- - BE5T r IT
:r: 4.3
t:J
...... 4. 1
lI-l
:r:
:x: 3.9
a:
lI-l 3.7 !!I III
LL
I!I
l.L. 3.5 I!I
---l
a:
:r: 3.3
I!I
-
t-
a: 3.1
:r:
t- 2.9
O
...... 2.7
~

2.5
0 la 20 30 40 50 so 70 80 90 100

FILM THICKNESS YJM. )


FIGURE Rl8 WIDTH RT HRLF PERK HEIGHT VS. FILM THICKNESS
rOWDER SIlE RRNGE : 20-30~M.

t<EY
4.5
t- - - - - BEST r IT
:::c 4.3
c..::J
..... 4. 1
LW
:::c
:::.::: 3.9
a: m
LW 3.7
a... I!l I!l m
m
LL 3.5
.....J
a:
:::c 3.3
t-
a: 3.1
:z: 2.9
t-
O
....,
~
2.7

2.5
0 10 20 30 40 50 so 70 80 90 100

fILM THICKNESS [~M.J


FIGURE R19 WIDTH RT HRLF PERK HEIGHT VS. FILM THICKNESS
rOWDER SIlE RRNGE : 30-40~M.

f'(EI
5
I- --- - BEST rn
I 4.8
t::J
.....,
lI-J
4.5
I
~
4.4
a:
lI-J4.2
a..
lL. 4 m
.....J I!I
m I!I m m
a: I!I
I!I
I 3.8 m m m
I- I!I
I!I
a: 3.5
I
l- 3.4
q
.....,
~
3.2

3
10 20 30 40 50 so 70 80 90 100 110

FILM THICKNESS £)JM_)


FIGURE A20 WIDTH AT HALF PEAK HEIGHT VS. FILM THICKNESS
rOWDER SIIE RRNGE : 40-50fM.
t<E'1'
4.5 ~- EXrE"INENT~L D~T~

r- 4.3 --~\ - BEST r IT


:r:
(.:J
....., ID
LW 4. 1
:r:
:::.::: 3.9 ID ID
a: ml!J mm
LW 3.7
a..
lL. 3.5
...J
a:
:r: 3.3
r-
a: 3. 1
:r: 2.9
t-
q
.....,
;3; 2.7

2.5
0 20 40 60 80 100 lZO 140 160 180 ZOO

FILM THICKNESS ~M. J


FIGURE R21 WIDTH RT HRLF PERK HEIGHT VS. FILM THICKNESS
rOWDER SIZE RRNGE ; >50UM.
KO
-4.5
t- --- - BEST r IT
:r 4.3
CJ
....,
w 4. 1
:r
::.:::: 3.9
Cl:
l~ 3.7 ID
B ID
£L l!I ID l!I l!I
[!]
l!I I!l l!I l!I
1L 3.5 I!l
_.J
Cl:
::c 3.3
t-
Cl: 3. 1
::c
t- 2.9
q
....,
3=
2.7

2.5
25 35 45 55 65 75 95 105 115 125

FILM TH I CKNE55 [UN. )


APPENDIX B

ROUGHNESS TEST RESULTS


Table Bl. Surface Roughness Results at Various Film Thicknesses for Commercial Powder

Film Thickness (\lm)

22 27 33 37 42 47 52 56 62 67 74 85 100

Average 1.03 0.99 0.73 0.76 0.70 0.53 0.56 0.50 0.54 0.51· 0.40 0.42 0.45
Roughness
Standard 0.25 0.16 0.09 0.08 0.11 0.11 0.08 0.12
Average 0.21 0.29 0.22 0.21 0.14
Deviation
(\lm)
Probable 0.06 0.04 0.03 0.02 0.02 0.03 0.03 0.03 0.02 0.03
0.05 0.06 0.05
Error

Average 0.56 0.53 0.46 0.50 0.15 0.34 0.37 0.36 0.38 0.33 0.36 0.35 0.28
Roughness
Standard
Wavelength 0.09 0.10 0.09 0.15 0.06 0.08 0.10 0.08 0.07 0.08 0.16 0.17 0.09
(nun)
Deviation

Probable 0.03
0.02 0.02 0.02 0.03 0.01 0.02 0.02 0.02 0.02 0.02 0.04 0.05
Error
Table B2 .. Surface Roughness Results at Various Film Thicknesses for <10~m Size Range Powder

Film Thickness (~m)

16 22 28 32 37 44 56 62 67 74 84 95

Average 0.86 1.09 1.33 1.29 1.21 1.44 1.01 1.34 1.00 1.12 0.97 0.78
Roughness
Standard
Average 0.20 0.29 0.35 0.39 0.39 0.28 0.42 0.27 0.36 0.40 0.33 0.23
Deviation
(~m)

Probable 0.06 0.08 0.09 0.09 0.15 0.10


0.07 0.09 0.11 0.08 0.06
Error

Average 0.47 0.67 0.77 0.79 0.78 0.92 0.72 0.91 0.74 0.83 0.71 0.59
Roughness
Standard
Wavelength 0.07 0.15 0.16 0.19 0.19 0.16 0.25 0.16 0.22 0.27 0.17 0.14
. Deviation
(mm)
Probable
0.03 0.04 0.04 0.04 0,04 0.09 0.05 0.06 0.07 0.04 0.04
Error
Table B3. Surface Roughness Results at Various Film Thicknesses for 10-20~m Size Range Powder

Film Thickness (~m)

18 22 27 32 37 42 46 56 62

Average 0.33 0.35 0.40 0.37 0.38 0.35 0.40 0.24 0.24
Roughness
Average Standard 0.42 0.11 0.12 0.06 0.09 0.10 0.21' 0.05 0.05
Deviation
(~m)

Probable
0.01 0.02 0.03 0.01 0.02 0.02 0.06 0.01 0.01
Error

Average 0.23 0.25 0.22 0.26 0.23 0.21 0.23 0.14 0.15
Roughness
Standard
Wavelength 0.04 0.06 0.06 0.14 0.05 0.06 0.09 0.05 0.05
Deviation
(mm)
Probable
0.01 0.01 0.01 0.03 0.01 0.01 0.03 0.01 0.01
Error
Table B4. Surface Roughness Results at Various Film Thicknesses for 20-30~m Size Range Powder

,
Film Thickness (~m)

27 32 37 42 48 52 57 62 67 71 77 84 100

Average 0.45 0.44 0.39 0.41 0.38 0.37 0.31 0.35 0.38 0.38 0.40 0.45 0.53
Roughness
Standard 0.12
Average 0.11 0.11 0.06 0.07 0.07 0.05 0.09 0.13 0.12 0.10 0.17 0.29
Deviation
(~m)

Probable 0.02 0.03 0.02


0.02 0.01 0.02 0.01 0.02 0.03 0.03 0.03 0.04 0.08
Error

Average 0.30 0.29 0.28 0.29 0.26 0.28 0.21 0.24 0.27 0.29 0.32 0.34 0.56
Roughness
Standard 0.14 . 0.12
Wavelength 0.05 0.08 0.07 0.08 0.04 0.08 0.04 0.06 0.09 0.08 0.42
(mm)
Deviation

Probable
0.01 0.12 0.12 0.02 0.01 0.02 0.01 0.01 0.02 0.02 0.04 0.03 0.12
Error
Table B5. Surface Roughness Results at Various Film Thicknesses for 30-40~m Size Range Powder

Film Thickness (~m)

29 33 37 42 48 52 57 62 67 72 77 82 87 93 102

Average 0.46 0.54 0.56 0.50 0.43 0.43 0.37 0.46 0.39 0.33 0.38 0.37 0.40 0.33 0.31
Roughness
Standard
Average 0.62 0.16 0.09 0.13 0.10 0.12 0.11 0.08 0.10 0.09 0.08 0.09 0.11 0.09 0.06
Deviation
(~m)

Probable 0.02
0.03 0.04 0.02 0.03 0.03 0.03 0.03 0.02 0.03 0.03 0.02 0.03 0.02 0.02
Error

Average 0.25 0.31 0.32 0.32 0.26 0.26 0.27 0.29 0.24 0.22 0.23 0.22 0.25 0.28 0.22
Roughness
Standard
Wavelength 0.04 0.07 0.06 0.07 0.04 0.06 0.10 0.06 0.06 0.06 0.06 0.05 0.06 0.13 0.08
Deviation
(nun)
Probable
0.02 0.02 0.01 0.02 0.01 0.01 0.02 0.03 0.01 0.02 0.02 0.01 0.01 0.04 0.03
Error
Table B6. Surface Roughness Results at Various Film Thicknesses for 4o-50~m Size Range Powder

Film Thickness (~m)

33 37 42 47 52 57 63 75 84 92 97 106 135

Average 0.96 0.83 0.68 0.65 0.48 0.50 0.55 0.55 .0.44 . 0.41 0.39 0.37 0.28
Roughness
Standard 0.22 0.16 0.08 0.10 0.08 0.04
Average 0.30 0.25 0.22 0.16 0.18 0.11 0.06
Deviation
(~m)

Probable 0.07 0.06 0.05 0.05 0.04 0.04 0.04 0.03 0.03 0.02 0.03 0.02 0.01
Error

Average 0.61 0.54 0.49 0.51 0.34 0.41 0.39 0.37 0.36 0.33 0.30 0.34 0.23
Roughness
Wavelength Standard 0.12 0.08 0.10 0.09 0.10 0.19 0.12
0.16 0.13 0.22 0.15 0.06 0.07
(mm)
Deviation

Probable 0.03 0.02 0.02 0.03 0.02 0.03 0.03 0.06 0.03
0.04 0.03 0.05 0.02
Error
Table B7. Surface Roughness Results at Various Film Thicknesses for +50~m Size Range Powder

Film Thickness {~m}

38 43 47 52 57 62 67 71 76 81 87 92 97 104 122

Average 0.80 0.75 0.74 0.62 0.57 0.62 0.53 0.55 0.46 0.48 0.45 0.53 0.45 0.46 0.39
Roughness
Standard
Average 0.14 0.12 0.13 0.17 0.15 0.19 0.14 0.18 0.13 0.13 0.13 0.10 0.18 0.15 0.13
Deviation
{~m}

Probable
0.04 0.03 0.03 0.04 0.04 0.04 0.03 0.04 0.04 0.04 0.03 0.02 0.06 0.04 0.04
Error

Average 0.45 0.46 0.49 0.45 0.37 0.41 0.37 0.41 0.33 0.33 0.33 0.38 0.40 0.33 0.34
Roughness
Standard
Wavelength 0.07 0.08 0.18 0.11 0.08 0.11 0.08 0.11 0.09 0.07 0.06 0.12 0.08 0.13 0,10
Deviation
{mm}
Probable
0.02 0.02 0.04 0.03 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.03 0.03 0.03 0.03
Error
Table B8. Linear Regression fits of Roughness Average and Wavelength Results

~_Measured: y Correlation
Powder Type Slope t = 15 t = 100
Quantity Intercept Coefficient

Roughness --0.0076 1.037 0.860 0.923 0.277


Average
Commercial
Roughness -0.0032 0.576 0.868 0.528 0.256
Wavelength

Roughness
-0.0003 1.272 0.367 1.268 1.271
Average
<lO;Jm
Roughness 0.711 0.118
0.0006 0.72 0.771
Wavelength

Roughness
-0.0025 0.434 0.613 0.397 0.184
Average
l0-20Jlm
Roughness 0.299 0.808
-0.0023 0.265 0.069
Wavelength

Roughness 0.221. 0.595


0.0027 0.262 0.491
Average
20-30Jlm
. Roughness
0.0024 0.164 0.609 0.2 0.404
Wavelength

Roughness 0.593 0.842


-0.0028 0.551 0.313
Average
30-40Jlm
Roughness -0.0009 0.317 0.601 0.227
0.304
Wavelength

Roughness 0.929
-0.0054 0.867 0.848 0.389
Average
40-50Jlm
Roughness -0.0031 0.618 0.880 0.308
0.572
Wavelength

Roughness 0.889 0.889


-0.0045 0.822 0.439
Average
+50Jlm
Roughness
-0.0017 0.513 0.755 0~488 0.343
Wavelength
Table B9. Linear Regression Results for Roughness Average and
Roughness Wavelength at Various Film Thicknesses

Using data from Table 3.3

Film Thickness 30\lm 50\lm 70\lm 100\lm

Slope . 0.0125 0.0102 0.0056 0.0043


Intercept 0: 11 . 0: 13 0:20· 0.21
Roughness
Average
d = l5\lm 0.29 0.29 0.28 0.27
d = 55\lm 0.79 0.69 0.51 0.45

Slope 0.0076 0.0068 0.0061 0.0046


Intercept 0.082 0.086 0.09 0.11
Roughness
Wavelength
d = l5\lm 0.20 0.19 0.18 0.18
d = 55\lm 0.50 0.46 0.43 0.37

Calculations do not include <lO\lm size range


FIGURE 81 ROUGHNESS RVERRGE VS. FILM THICKNESS
r~WDER SIIE RRNGE : CoMMERCIPL
~EY

1.1
[!]
~
1 ---", - BEST I'IT

:E
3- .9
.8
1+1
t.:l
a: .7
0:::
1+1
:> .s
a:
U)
.5
U)
1+1 .4
:z
:::r:
t.:l .3
::J
Cl
0::: .Z

. 1
a 10 Zo 30 40 50 sa 70 80 90 100

FILM THICKNESS ()J M. )


FIGURE 82 ROUGHNESS RVERRGE VS. FILM THICKNESS
r~WDER SIIE RRNGE : <l0fM.
t(E'I'
1.5
I!I
~
1.4 --- - BEST r IT
• ID
::E ID
::J.... 1. 3 I!I

1.2
4J
=
a: 1.1 ID
I!I

a:::
4J [!J [!J
> 1 [!J
a:
U")
.9
[!J
U")
4J .B
:z [!J
:c
t.:l
::J
. "}
C)
a::: .6

.5
0 10 20 30 40 50 60 70 80 90 100

FILM THICKNESS (JlM. 1


FIGURE 83 ROUGHNESS AVERAGE VS. FILM THICKNESS
rOWDER SIZE R~NGE : lO-ZO~M.

KEY
.6

--- - BEST rn

:0: .5
3-
LU .4 I!l
t.:] 6 I!l
C± !!l I!l
~ I!l
LU
> _3
CC
I!I I!l
If)
Cl')
tU
.2
~
I
d
::l _1
0
a::

0
0 10 20 30 40 50 60 70 80 90 100

FILM THICKNESS ()JM. )


FIGURE 84 ROUGHNESS RVERRGE VS. FILN THICKNESS
KEY rOWDER SIZE R~NGE : 20-30pM.
m- EXrEftlHENTPL DPTP

.6 - - - - - BEST fIT

I!l

w •4
t.:I
a:
~
w
>- .3
a:
en
en .2
w
::z
:r::
t.:I
::> . 1
Cl
~

o ~--------------~-----------------------~~---~---~----~
o la 20 30 40 50 60 70 110 90 100

FILM THICKNESS [~M.)


FIGURE 85 ROUGHNESS RVERRGE VS. FILM THIC~NESS

rOWDER SIZE RRNGE : 30-40~M.

t<E'J'
.s
~ I!l - - - - BEST rn
•. 55 I!l
::E

3-. .5 I!l

b.J I!l I!l


c..:j. 45
~ I!l [!J

W I!l
>- .4
a: I!l

~.3S
I!l
. lL.l
:z
I: .3
t.:l
:::l
~.2S

.2 ~----------------~------~-----+-----~------------~----------~

10 20 30 40 50 60 70 80 90 100 llO

FILM THIC~NESS ~M. )


FIGURE 86 ROUGHNESS RVERRGE VS. FILM THICKNESS
rOWDER SIZE R~MGE : 40-50~~.

t(EY

1
m
~.9 --- - BEST r IT

~ I!l
B
:3c..
.7
LW
t.::J
a:. S
a::
LW
>.5 I!ll!l
a:
U).4
U)

~.3
I:
d .2
:J
Cl
a::. 1 I!l
I!l
0
0 20 40 60 BO 100 120 140 160 HlO ZOO

FILM THICKNESS
FIGURE 87 ROUGHNESS RVERRGE VS. FILM THICKNESS
rOwDER SIZE RR~GE .. >SO}JM .

~EY

.9 m_ EXrE"I"E"T~L D~T~

~
I!I
- BEST rn

.8
::£ I!I I!l
~.7
I!I
LW .6 I!I
f..:]
et: ID
~ .5
w I!l ID
>-
a: .4
(I)
(I)
w .3
:z
:::z:: .2
f..:]
=>
Cl
~ .1

----'
0
25 35 45 55 65 75 85 95 105 115 1Z5

FILM THICKNESS r}.IN. )


FIGURE Ba ROUGHNESS WAVELENGTH VS. FILM THICKNESS
POWDER SIIE RRNGE: COHHERCIRL
KEY
.7

• --- - BEST r IT
~ .6

[!]
;: .5
c..:J
z
4J
....J .4
4J
>
a:
3= .3
if)
en
4J .2
:z
I
t.:J
::J • 1
Cl
0:::
o ~----------~------~----~------------------~------~-------------
o 10 20 30 40 50 60 70 80 90 100

FTLM THICKNESS
FIGURE 89 ROUGHNESS WAVELENGTH VS. FILM THICKNESS
I"ClWDER 5 IlE RRNCE •• < 1 OPI1 •

~EY

1 m- EXrE"'HE"TAL DATA
~

::E I!l
- BEST I'll
::E. 9
I!l
:c. S I!l I!l I!l
t-- I!I.
t:J I!l
I!l
2.7
w I!l
...J
~
>.e I!I
a:
~
.5
en I!I
(J')
w
z-lI
:r:
t:J
~3
0-
~

.2
0 10 20 30 40 50 60 70 90 100

FILM THICKNESS ~M. 1


FIGURE BID ROUGHNESS WAVELENGTH VS. FILM THICKNESS
rOWDEH SIZE HRNCE : lO-ZOpM.
t(EY
.3

:£ --- - BEST r IT

-.25

:c
r-
~ .2
:z
lI.I
...J
lI.I
>.15
er:
:;.:

~ .1
lI.I
:z:
:::r:
~.05
:::J
o
a:::
o ~----------~------------~----~----~----~~----~------------
o la 20 30 40 50 so 70 80 90 100

FILM THICKNESS (~M.l


FIGURE 811 ROUGHNESS WAVELENGTH VS. FILM THICKNESS
rOWDER SIZE RRNGE : ZO-30fM.
f'(EY
.7
:::E - - - - BEST r IT
:::E • 6
III

:c
t- .5
t.:J
:z:
lI-I
~ .4
lI-I
::> III
et: III
3: .3 [!) [!) l!l 15- III III
[!)
Cl')
Cl') l!l
lI-I • 2
:z
:c
d
:::> • 1
o
a:::
o ~-------------------~-----------------~--~---~
o 10 ZO 30 40 50 so 70 80 90 100

FT LM TH I CKNESS ()-JM. )
FIGURE B12 ROUGHNESS WAVELENGTH VS. FILM THICKNESS
r~WDER SIZE RRNGE : 30-40fM.
KEY
.4
_. BEST fIT
:::E
:::E. 35

[!J [!J
[!J

!!l
!!l
[!J
[!J

!!l

~ .2
LW
:z
I:
l.:J.15
:::l
o
0::

•1
10 20 30 40 50 60 70 130 90 100 110

FILM THICKNESS fjJ M. )


FIGURE 813 ROUGHNESS WRVELENGTH VS. FILM THICKNESS
rOWDER SIlE R~NGE : 40-50}1I1.
KEY
.7 m- EXrE"IHEHTRL ORTR
~

~ - BEST I'n
~ [!]
.6

:r:: .5
t-
<.:J
:z:
4J
-1 .4
w
>
a:
::;.: .3
(f')
(f')
4J .2
Z
::c
l:l
=> . 1
Cl
t:t::
0
0 20 . 40 60 BO 100 120 140 160 IBO ZOO

FILM THICKNESS (jJM. J


FIGURE 814 ROUGHNESS WAVELENGTH VS. FILM THICKNESS
rOWDER SllE RRNGE : >50fM.
KEY
.55 ID- EXrEnJ"EHTAL DATA
- - - - - BEST FIT

.5 !!l

::c
t-
c..::J .45
z
LW
....J
w .4
>
a:
~

V)
V)
.35
w
z
::c
c..::J .3
::l
Cl
~

.25 ~---------------~~---+-----~-----------~-----------------
25 35 45 55 65 75 85 95 105 115 125

FILM THICKNESS
l3:
H
n
~
'"n
F:
'"
H
~
t::l
H

'"
~
~
i'd
>tj

~
H
gJ t::l
~ H
H I><
!i1l n

~
><
'"
H
'"
f!l
'"fl
~

'"
Table Cl. Mierosea1 Photosedimentometer Analysis of Powder 1. Before Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 16.4 Ht. of Scan (ems) : 17.1 Ht. of Scan (ems) : 17.2
Averages t
Time to Scan (mins ) : 24.2 Time to Scan (mins): 17.8 Time to Scan(mins): 18.0

Optics1 Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) (\lm) undersize (mins) (\lm) undersize (mins) (\lm) undersize (\lm) undersize

- - - - - - - - - - 10 5.7
(0.96)
- - - - - - - 3.4 55.5 100.0 15 7.3

0.8 - - - 3.0 55.7 100.0 5.3 46.2 69.7 20 10.9


(0.68)
0.7 1.2 78.4 100.0 6.2 41.5 71.6 7.0 40.1 53.9 25 19.1

0.6 4.3 48.8 72 .1 9.3 34.2 50.4 9.3 34.7 40.2 30 29.5

0.50 6.7 41.4 50.5 13.3 29.5 32.9 12.5 30.9 28.4 35 42.0

0.45 8.2 37.2 41.3 16.8 26.0 25.4 14.8 28.0 23.1 40 57.1
0.40 10.2 33.1 33.0 22.0 22.8 18.7 18.6 25.0 18.3 45 70.2

0.35 13.1 29.2 25.7 26.0 20.1 12.9 23.0 23.2 14.1 50 82.1

0.30 17.0 25.3 19.2 28.4 17.6 7.8 25.2 21.7 10.1 60

0.25 20.5 21.5 13.6 31.0 9.1 3.3 27.2 15.8 6.4

0.20 23.2 9.9 8.8 42.0 0.9 1.0 37.0 5.5 3.7

0.15 - - - - - - - - -
t values obtained from size distribution curves
Table C2. Microscal Photosedimentometer Analysis of Powder 1. After Spraying

Sample 1 Sample 2 Sample 3


Ht. of Scan (ems) . 16.4 Ht. of Scan (cms) : 17.1 Ht. of Scan (cms) : 17.7
Averages t
Time 1:0 Scan{odns): 24.2 Time to Scan (mi ns ) : 17 .8 Time to Scan (mins): 18.0

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) (pm) undersize (mins) (pm) undersize (mins) (pm) undersize (pm) undersize

- 10 4.9

- 15 5.6
(O. 75) (O. 75)
0.8 1.0 89.6 100.0 3.4 "58.8 100.0 1.3 90.1 100.0 20 7.3

0.7 2.5 61.5 72.6 3.9 54.2 89.2 1.9 69.6 85.3 25 10.9

0.6 4.2 49.2 53.8 4.7 48.9 69.3 3.8 52.4 62.6 30 16.8

0.50 6.0 43.0 38.8 5.9 44.7 51.4 5.7 45.3 45.6 35 24.5

0.45 7.1 39.5 32.2 6.7 41.9 43.2 6.7 42.1 38.2 40 34.7

0.40 8.4 36.0 26.1 7.7 39.0 35.5 7.6 39.1 31.3 45 47.3

0.35 10.4 31.9 20.6 8.9 36.2 28.3 9.0 36.0 25.0 50 62.3

0.30 13.6 27.9 15.8 10.4 32.9 21. 7 10.6 32.4 19.1 • 60 82.3

0.25 17 .9 24.4 11.5 13 .1 28.8 15.7 13.8 28.1 13.8

0.20 22.9 21.3 7.8 17.8 24.1 10.4 18.4 23.9 9.2

0.15 26.6 9.8 4.5 21.3 10.8 6.0 21. 3 11.0 5.4

t values obtained from size distribution curves


Table C3. Mieroseal Photosedimentometer Analysis of Powder 2. Before Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17 .5
Averages
Time to Scan (mins) : 17.0 Time to Scan(mins): 21.8 Time to Scan(mins): 19.2

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%.
Density (mins) (\lm) undersize (mins) (\lm) undersize (mins) (\lm) undersize (\lm) undersize
(1.0) (0.94)
- - - - 4.3 50.1 100.0 1.3 89.8 100.0 10 5.8
(0.9) (0.9)
- - - - 6.2 42.0 80.3 2.0 68.5 88.1 15 7.5
(0.73)
0.8 0.9 104.6 100.0 8.7 35.8 63.7 4.0 50.8 65.3 20 10.0

0.7 1.6 71.6 87.6 11. 7 31.0 49.6 6.4 40.8 48.5 25 15.5

0.6 4.5 46.7 59.3 15.4 27.1 37.4 9.5 33.8 34.9 30 24.5

0.50 8.1 37.7 40.9 20.0 24.6 26.7 13.5 29.5 23.7 35 35.5

0.45 10.2 33.8 33.5 22.1 23.4 21. 9 16.2 27.0 18.8 40 45.8

0.40 12.5 30.7 26.8 22.9 22.4 17.3 19.2 24.5 14.3 45 55.5

0.35 15.0 27.8 20.7 23.8 21.1 12.9 21. 2 22.1 10.3 50

0.30 17.7 25.4 15.2 25.2 15.8 8.7 23.0 15.3 6.6 60

0.25 19.1 20.1 10.2 33.4 5.7 5.6 32.5 4.9 4.0
,

0.20 25.7 7.9 6.3 - - - - - -


0.15 - - - - - - - - -
t values obtained from size distribution curves
Table C4. Microscal Photosedimentometer Analysis of Powder 2. After Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.2 Ht. of Scan (ems) : 17.6
Averages t
Time to Scan(mins): 15.0 Time to Scan(mins): 12.8 Time to Scan(mins): 14.6 .

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) (\lm) undersize (mins) (\lm) undersize (mins) (\lm) undersize (\lm) undersize

- 10 6.8
(0.95) (0.93)
- 1.5 77 .8 100.0 1.8 74.2 100.0 15 6.9

0.8 1.6 77.0 100.0 3.1 57.9 68.9 3.2 58.3 73.5 20 7.6

0.7 3.1 57.6 74.8 4.7 48.0 53.5 4.6 49.0 57.5 25 10.5

0.6 5.0 46.5 55.9 6.5 40.9 40.7 6.4 41.9 44.0 30 18.0

0.50 7.2 40.6 40.7 8.9 36.4 29.8 8.6 37.5 32.5 35 28.3

0.45 8.5 37.4 34.1 10.2 34.0 24.9 9.9 35.0 27.4 40 39.3

0.40 10.0 34.5 28.0 11.8 31.3 20.4 n.3 32.3 22.6 45 50.0

0.35 11. 7 . 31. 7 22.3 13.6 28.6 16.2 13.6 29.6 18.2 50 60.0

0.30 14.1 28.7 17.1 14.8 26.6 12.4 15.4 27.5 14.1 60

0.25 16.2 26.1 12.4 15.8 23.6 8.9 16.3 25.7 10.3

0.20 17 .6 12.5 8.2 18.4 10.8 5.7 17.6 12.3 6.8

0.15 - - - - - - - - -
t values obtained from size distribution curves
Table C5. Microsca1 Photosedimentometer Analysis of Powder 3. Before Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.5
Averages t
Time to Scan(mins): 22.0 Time to Scan (mins): 15.7 Time to Scan(mins): 27.2

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt% .
Density (mins) (!lm) undersize (mins) (!lm) undersize (mins) (!lm) undersize (!lm) undersize
(1.0)
- - - - - - - 5.8 43.7 100.0 10 5.4
(0.83)
0.9 1.0 104.6 100.0 - - - 7.9 37.7 80.0 15 6.5

0.8 1.4 79.1 89.2 - - - 10.5 32.9 62.7 20 9.3

0.7 3.3 54.4 62.0 1.2 79.4 100.0 13.6 28.9 47.7 25 14.9

0.6 6.0 41.3 43.3 4.2 47.4 62.7 17.6 25.4 34.4 30 23.0

0.50 9.8 34.5 29.1 8.3 37.1 40.5 22.7 23.0 22.8 35 32.8

0.45 12.0 31.2 23.2 10.6 32.9 31.8 26.2 20.5 17.5 40 43.4

0.40 14.7 28.0 17.8 13.4 29.4 24.1 30.2 16.6 12.8 45 54.2

0.35 18.3 24.5 13.0 16.2 26.2 17. 2 34.8 11.8 9.0 50

0.30 23.4 17.9 8.8 18.2 22.4 11.1 40.1 4.6 6.3 60

0.25 31. 8 6.7 5.7 21.3 13.1 5.8 - - -


0.20 - - - 31.7 2.9 2.8 - - -
0.15 - - - - - - - - -
t values obtained from size distribution curves
Table e6. Microsca1 Photosedimentometer Analysis of Powder 3. After Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17 .5 Ht. of Scan (ems) : 17 .5
Averages t
Time to Scan (mins) : 21.5 Time to Scan(mins): 18.5 Time to Scan(mins): 11.2

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) (j.Im) undersize (mins) (j.Im) undersize (mins) (j.Im) undersize (j.Im) undersize
(1.0) (1.0) (0.1)
- 1.7 76.7 100.0 1.7 80.1 100.0 18.8 9.8 4.0 10 3.5

0.9 2.6 65.1 69.8 2.4 67.8 80.5 - - - 15 4.0

0.8 3.6 56.0 55.1 3.3 58.2 64.0 - - - 20 6.7

0.7 4.7 48.4 42.4 4.4 50.4 49.9 - - - 25 11.0


(0.56 )
0.6 6.5 40.7 31.5 5.9 43.1 37.6 1.1 95.2 100.0 30 16.5

0.50 9.4 34.9 22.3 8.2 38.0 27.2 1.9 75.0 76.9 35 22.5

0.45 12.0 31. 2 18.3 9.7 34.7 22.5 2.8 62.6 61. 7 40 29.5

0.40 14.6 28.4 14.8 11.8 31. 5 18.3 3.9 53.8 49.1 45 36.5

0.35 17.4 25.9 11.6 14.3 28.4 14.5 5.1 46.7 38.2 50 44.0

0.30 21.1 23.1 8.6 17.8 25.1 11.0 6.9 39.8 28.7 60

0.25 23.9 19.7 6.0 20.8 21. 7 8.0 9.7 33.5 20.7

0.20 27.2 15.1 3.8 23.3 18.6 5.4 12.6 27.3 13.9

0.15 32.3 6.2 2.1 25.9 8.5 3.1 16.1 21.8 8.4

t values obtained from size distribution curves


Table C7. Microsca1 Photosedimentometer Analysis of Powder 4. Before Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.5 Ht. of Scan (ems) : 17.5
Averages t
Time to Scan(mins): 18.5 Time to Scan(mins): 21. 7 Time to Scan (mins): 19.2

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) ()Im) undersize (mins) ()Im) undersize (mins) ()Im) undersize ()Im) undersize
- (1. 2)
- 1.6 65.3 100.0 10 6.5

- 7.7 37.8 46.8 15 9.1

0.8 10.7 32.2 36.6 20 14.1

0.7 14.5 27.9 27.8 25 20.9


(0.56)
0.6 0.7 91.5 100.0 19.0 23.5 20.3 0.8 97.9 100.0 30 29.0

0.50 5.7 43.2 53.7 24.5 19.8 13.9 2.7 65.0 70.3 35 37.9

0.45 8.5 36.0 42.8 26.7 17.4 11.2 . 4.6 48.9 54.9 40 46.7

0.40 11.8 31.0 33.7 28.9 15.2 8.8 6.4 41.6 . 42.5 45 56.5

0.35 15.4 26.2 25.8 30.8 13.2 6.8 8.7 35.9 32.0 50

0.30 20.6 21.5 19.2 32.6 11. 2 5.0 11.6 30.5 22.9 60

0.25 23.7 18.7 13.8 34.3 5.1 3.5 16.8 24.7 15.2

0.20 25.1 8.9 9.0 22.4 16.9 8.9

0.15 30.5 6.1 4.6

t values obtained from size distribution curves


Table C8. Microsca1 Photosedimentometer Analysis of Powder 4. After Spraying

Sample 1 Sample 2 Sample 3

Ht. of Scan (ems) ; 17.5 Ht. of Scan (ems) ; 17.5 Ht. of Scan (ems) ; 17.5
Averages t
Time to Scan(mins}; 15.4 Time to Scan (mi ns) ; 13.8 Time to Scan(mins}; 15.0

Optical Time Size cum. wt% Time Size cum. wt% Time Size cum. wt% Size cum. wt%
Density (mins) ()lm) undersize (mins) ()lm) undersize (mins) ()lm) undersize (j1m) undersize
(0.98)
- 1.1 98.9 100.0 10 3.8
(0.9) (0.82)
- 1.6 80.0 80.5 1.2 99.7 100.0 15 5.1
(0.73)
0.8 2.6 63.2 60.7 1.0 103.0 100.0 1.4 83.1 94.1 20 6.6

0.7 4.1 51.2 45.1 1.5 78.5 89.5 2.6 63.2 69.6 25 9.7>

0.6 6.0 42.7 32.5 3.1 57.1 62.9 4.1 51.0 50.9 30 14.5

0.50 8.4 37.4 21.9 5.2 47.1 43.6 6.1 43.9 35.9 35 21.3

0.45 10.1 33.9 17.3 6.5 42.4 35.6 7.3 40.2 29.4 40 29.2

0.40 12.5 30.1 13.1 7.9 30.4 28.4 8.7 36.6 23.5 45 37.2

0.35 15.9 26.2 9.4 9.7 34.2 21.9 10.7 32.2 18.1 50 47.4

0.30 18.2 22.9 6.2 12.5 29.8 16.1 14.6 27.4 13.3 60

0.25 20.3 13.9 3.3 15.4 25.3 11.1 17.5 22.6 9.3

0.20 31.1 3.3 1.6 18.2 18.6 6.8 20.9 10.1 6.0

0.15 24.5 7.1 3.6

t values obtained from size distribution curves


- - - - - - -----------

APPENDIX D

COMPUTER PROGRAM LISTING AND PRINTOUT


Figure Dl (a-p). Listing of Computer Program

C ELEX:TOOS'lATIC PCMDER CXlI\TING - PARTICLE PACKThG

C C ' C X . ' l I .

C 'llITS PRCX;R¥-I IS AN ATI'EMPT 'ID SIMULA'IE PACKThG OF ELEX:TOOS'lATICALLY


C SPRAYED PARTICLES am::> AN EM.'lHED SUBSTAAlE 'lAKThG IN'IO ACCOONT
C ALL '!HE FORCES ACTnl> CN '!HE PARTICLES I.E. DRAG,INERTIA,GRAVIlY,
C IMAGE, AND IN'IER-PARTIOJLAlE ATl'RACTICN AND REPULSICN roRCES.
C '!HE DIFFEREm'IAL mJATICNS WHICH DESCRIBE '!HE PARTICLE 'mAJEX:T-
C ORIES ARE IN'lffiRAmD USThG A NlI.G LmRARY ROUTINE 'ID CAI.CUIAlE '!HE
C PA'lH OF A PARTICLE AND FINDITS FINAL PACKED rosITICN.
C '!HE X-axlRDINlI.lE S'lART rosITICN IS ARBI'lARILY rnosm 9:MEWHERE
C BE'lWEEN '!HE GUN AND 'lARGET AND A RANIXM NUMBER ROUTINE IS USED
C R)R '!HE S'lARTING POSITICN IN '!HE Y-DIREI:TICN. IT IS ASSUMED 'lHAT
C '!HE PARTICLES ARE SPHERICAL AND 'lHAT WHEN '!HEY TCO::H '!HEY REMAIN
C IN 'lHAT POSITICN.
C 9JLVIN3 '!HE O.D.E. 's IS ACHIEVED BY NUMERICAL IN'IE:iRATICN OVER
C A RAN:;E USThG GEM.'S ME'lHClD roR STIFF mJATICNS. '!HE <XJTPUT IS
C CXJNTR)LLED BY CCNTOOL PARAMEmRS AND '!HE IN'l'ffiRATION BY A 'IOLER-
C AOCE ERroR, BUT '!HE OPTIMUM Sm? IENG'lH IS IJE'IERoIINED BY '!HE
C ROUTINE. '!HE FINAL rosITION °OF '!HE PARTICLE IS CALCUIAmD BY
C IN'IEAAJLATICN OF '!HE LAST WJ <rMPUTED OXlRDINlI.'IES OF '!HE ~
C PARTICLE, '!HE LAST S'lEP BEING '!HE CNE 'lHAT PASSES IN'IO A PACKED
C PARTICLE OR '!HE SUBSTAAlE.

C '!HE CCMPU'IER PRCGRAM IS SPUT IN'IO '!HE R)LIruIN:; SEX:TICNS .-

C (1) RE'A!)ThG OF M'lA IN'IO PRCGRI\M

C (2) SETTING SCALES AND AXES roR PIDT

C (3) INITIAL CCNDITICNS AND SIZE DIS'llUBUTICN

C ( 4) S'lARTING OF 'mAJEX:'IORY CALCULATICNS & CALLING NAG ROUTINE

C (5) CH:EI:KThG PARTICLE rosITICN & PRINTING 'mAJEX:'IDRY

C (6) CCMPARISON WI'IH S'lATICNARY PARTICLES & 'rnAJEX:'IORY PIDTTING

C (7) FINDING COLLISICN PARTICLE & CALCULATING FINAL rosITICN

C (8) roSI'l'ICN CN SUBSTAAlE

C (9) IN'IERPOLATION OF 0'lHER VARIABLES

C (10) OUTPUT OF FINAL rosITICN

C (11) CAICULATICN OF roRCES - IN SUBROUTINE

C (12) CALCULATICN OF IN'IERPARTICULAlE FORCES


Page 2

C SECTION 1

IN'IH>ER RANIXl,)(sTEPS, YSTEPS ,NEX>STP ,CHEl:K ,JK , FORCEP


IN'IH>ER IN!) ,NUMBER, N, IW, IW1,MPED ,CXMENT, PID'IN) , IFAIL,K ,JJ , JW
lNT8GER ID1,ID2,l,J,KK,M

RmL*8 CHXJJ ,CHYJJ ,CHDSQ ,AlRVX ,AlFNY ,CHFAD,CHRSQ


RmL*8 FP1,FP2 ,FP3 ,FP4, IN'1'X1,INTX2,CSQ,P ,Rl,R2, 'lYPE
RmL*8 S'lARTL, P'IST , XLST, ONE' , HEIGHT, UP, VP , 'ffiAJPR,CHAR:;E
RmL*8 UNI'ISP,XPAPER,YPAPER,AXISST,«L'IH,YL'IH,A,B
REA.L*8 Q,D1RIB(ll) ,RFmlT,PI,CAXlS,VISC,ELSTR,DlEI:ST,~
REA.L*8 GUN'lAR,~T, TIPRAD,PHI ,DENS ,CELEX:,SQH
REA.L*8 Cl3UN ,CIMGJN ,CCRGE ,CMASS ,ClDAAG , 'lEND , IDL, R,CEFLEx::
REA.L*8 FRAC,FRAC10,DIAM,D(500),RADIUS,X1(500) ,Y1(500)
REA.L*8 PC(500),REYOON,REY,FLIGHT(500),S~(500)
REA.L*8 S'lART,MI\SS ,CFOR:E,CDRAG, lEG,RAD ,NUX,NUY ,SIZE,T, Z (4)
REA.L*8 CIN(7) ,COMM(5) ,CON(5),OOUT(lG),PW(4,4),W(4,22)
REA.L*8 PRINTX, PRIN'tY , PRlN'lH ,OIDX ,OIDY ,OVELX ,OVELY ,aY ,OX
REA.L*8 VELX,VELY ,X,Y ,XJJ ,YJJ ,DTPSQ,RADJJ ,RADSQ,CX,cy
REA.L*8 GRAn, INCEPT,C ,A1 ,B1, INTX, IN'lY ,IEVER, IN'IU , IN'lV ,OIDT, PRIN'lD
REA.L*8 PRIN'IS, INTT, 'ffiAJST, XAXIS , E ,D1RIB0 ,Gf/AV

REA.L*8 DRAGX ,DRAGY , IMAGE ,FIErDX ,FIEIDY ,F'IOTX,F'IO'IY


DIMENSION pID'IN) (7) ,CXMENT(7)

c:aM:N/BL3/DRAGX,DRAGY, IMAGE,FIEIDX ,FIEIDY ,F'lO'IX,F'ID'IY ,AIRVX,AlFNY


C~/BL1/UNF ,HEIGHT ,GUN'lAR,CEIE: , SQH,CGUN,CE:FJ:.OC,NUMBER
c:aM:N/BL2/CIMGJN ,0, PC,CHAR:;E,MASS ,CFOR:E ,CDRAG ,Xl, Y1,GRAV, I ,CSQ
CCl+m/BL4/REYOON ,REY

EX'lEmAL FCN ,D02QBF


EX'1ERNAL SE'ICPU

C
C - SET UP CALL ID MAlLOOX • USED ID KILL JOB AF'IER
C - DP.Y SHIFT HAS BEEN SET BY '!HE OPERA'lOR
C
CALL ~lIULBX(SETCPU)
C INITIAL a::NDITIONS
C

C NUMBER OF PARTICLES
CALL X04AAF (1, 5)
CALL X04ABF(1,5)
REA.D (5, 309G) NUMBER
C GUN ID 'lMGET IYmNCE
GUN'lAR=O • 200
C AANlXM NUMBER GENERA'lOR S'lART NU-!BER
RI\NOO=O
C START UNG'IH FOR PARTICLES IN Y DIREl:TION
CALL INruT (5)
CALL INPF (STARTL)
Page 3

JW=6
C S'mRT D'D'INCE OF S'lART LENG'IH FRCM Y AXIS
P'IST=S'mRTL/-2. ODO
C 'lOLERANCE ERroR
CALL INroT (5)
CALL INPF ('lOL)
C S'mRT UNG'IH m x DIREJ:TION
XLST=O.lDO
C AIR Gl'.S VEIOCITi WHEN m NJRM1'.L Ii'IJ:M
CALL INroT (5)
CALL INPF (UNF)
CALL INroT(5)
CALL INPF (TiPE)
CALL INroT(5)
CALL INPF (D'lRIBO)
IF(D'IRIBO.a;).O)<D 'lO 1
00 1 1=1,11
D'IRIB(I)=DTRIBO
1 ClJNTINUE
REND (5,55)FOHCEP
55 FOlM'.T(I3)
REND(5,122) (PLOTNO(I) ,1=1,7)
REND(5,122) (~(I) ,1=1,7)
C HEIGHT OF 'lARGET PIA'lE
HEIGHT=O.200
C mITIAL PARTICLE VEIOCITi IN X DIREJ:TIOO
UP=UNF
C INITIAL PARTICLE VEIOCITi IN Y DIREJ:TICN
VP=O.ODO
C S'mRT OF 'mAJEX:'IORY PRrnT
TR~-0.0999999DO


C SECTIOO 2


C PIDT CONDITICNS
C

CALL S!\.VOAA
C UNIT OCP..LE FUR PIDTTING
UNI'ISP=0.080ODO
C IEN3'IH OF PAPER m xDIREJ:TICN
XP~10000.ODO
C· IEN3'IH OF PAPER m Y DIREJ:TICN
~2500.0DO
C S'mRT OF AXIS MiiAY FR:M ORIGIN IN X DIREJ:TICN
AXISST=O.O
C NUMBER OF S'IEP UN3'IHS IN X DIREJ:TION
XS'lEPS=40
C NUMBER OF S'IEP IENG'IHS IN Y DIREJ:TION
YS'lEPS=10
C UN3'lH OF E7\CH S'IEP m x
XL'lH=100
C UN3'lH OF E7\CH S'IEP m Y
YL'lH=100
C NUMBER OF NffiI'\TIVE S'IEPS IN X
NEXiSTP=20
CALL UNI'IS (UNI'ISP)
CALL DEVPAP (XPAPER, YPAPER, 1)
CALL WINJ::X:M (2)
CALL CHASIZ(30.0,30.0)
CALL MJV'102 ((XS'lEPS*XL'lH+2000.0) /2-500.0 ,YPI\PER-50.0)
Page 4

CALL aJAHOL(55H E*LLEX::'IR)S'D'.TIC *UC*rD>.TING : *UP*L(W)


lER *UP*IACKING*.)
CALL MDVT02«XSTEPS*XLTH+2000.0)/2-470.0,YFAPER-60.0)
CALL LINT02«XSTEPS*XLTH+2000.0)/2-500.0+ll80,~60.0)
CALL CHASIZ(20.0,20.0)
A= (XSTEPS*XLTH) +1000.0
CALL MDVT02(A,YPAPER-100.0)
CALL aJAHOL(3OH CALCUIATION DI'.'D'. *.)
B=YPAPER-150.0
CALL MDVT02(A,B)
CALL aJAHOL (25H ID OF P'ICIS S'lUDIED = *.)
CALL MDVT02(A,B-25.0)
CALL aJAHOL (25H X <XXJRDINA'tE S'lART = *.)
CALL MDVT02(A,B-50.0)
CALL aJAHOL (25H S'lART LENGTH RAN3E = *.)
CALL MDVT02(A,B-75.0)
CALL OlAHOL (25H RI\NIXM NO VARIABLE = *.)
CALL MDVT02(A,B-100.0)
CALL aJAHOL (25H VEL OF PARTICLE (X) = *.)
CALL MDVT02(A,B-125.0)
CALL aJAHOL(25H UNIFORM AIR VEUX:ITi= *.)
CALL MDVT02(A,B-150.0)
CALL aJAHOL(25H 'IOLERAOCE ERROR = *.)
CALL MDVT02(A,B-200.0)
CALL CHAARR(PLOTNQ,7,4)
CALL MDVT02(A,B-225.0)
CALL CHAARR (CXMENT, 7 , 4)
A=A+400.0
CALL M0V'I02 (A,B)
CALL CHAINT (NUMBER, 11)
CALL MDVT02 (A,B-25.0)
CALL OJAFIX (XIST ,14,6)
CALL MDVT02(A,B-50.0)
CALL OJAFIX(S'lARTL,14,6)
CALL MDVT02(A,B-75.0)
CALL CHAINT (RANOO, 11)
CALL !Ol'I02 (A,B-100.0)
CALL OJAFIX (UP, 14,6)
CALL MDVT02 (A,B-125. 0)
CALL OJAFIX (UNF , 14,6)
CALL !.ov'I02(A+60.0,B-150.0)
CALL CHAFID ('IOL, 10)
CALL !Ol'I02(A-400.0,B-200.0)
C IABELLING OF AXES

C X AXIS

XAXIS=lOOO.O+(XLTH*NEGSTP)
CALL !Ol'I02(1000.0,50.0)
CALL LINT02 «lOOO.O+XSTEPS*XLTH) ,50.0)
CALL LINT02( (lOOO.O+XSTEPS*XLTH) , (YSTEPS*YLTH1+50.0)
CALL !Ol'I02(1000.0,50.0)
CALL LINT02(1000.0,(YSTEPS*YLTH)+50.0)
CALL MDVT02«1000.O+XSTEPS*XLTH)/2-5.0,5.0)
CALL aJAHOL (12H SUBSTRA'tE*.)
CALL CHASIZ(15.0,15.0)
IDl=XSTEPS+l
DO 10 I=l,IDl
M= (I - (NmSTP+1) ) *XLTH+AXISST
Q=(I-l)*XLTH+1000.0
CALL MDVT02(Q,55.0)
CALL LINT02(Q,45.0)
CALL MDVT02 (Q-50. 0,30.0)
CALL CHAINT (M, 5)
10 ClJNTINUE
Page 5


'"C Y AXIS
C'
ID2=YS'lEPS+l
DO 11 I=1,ID2
M= (1-1) *YL'IH
Q=(I-1)*YL'IH+50.0
CALL M0VT02(1005.0,Q)
CALL LINT02(995.0,Q)
CALL M0VT02(930.0,Q-5.0)
CALL C!lIUNT (M, 4)
C=1000.0+ (XS'lEPS*XL'IH)
CALL M0VT02 (C-5.0,Q)
CALL LINT02(C+5.0,Q)
CALL M0VT02(C+10.0,Q)
CALL C!lIUNT (M, 4)
11 OJNTINUE

CALL M0VT02(0.O,50.0)
CALL LINT02 (XPAPER, 50.0)

ccco::ccccccccnxxrcccccca:::cccco:xn:n:x:ccax:co:xx:ccco o::o::ccccccca::ccc
C SOCTION 3

C SETrING a:::tIDITIONS
C
C
C SIZE D'IRIBUTION OF PARTICLES
C
IF(DTRIBO.NE.O)GO 10 11222
DTRIB(1)=0.000005DO
DTRIB(2)=O.00000575DO
DTRIB(3)=0.0000065DO
DTRIB (4) =O.00000725DO
DTRIB (5) =O.OOOOO8DO
DTRIB(6)=O.00000875
DTRIB(7)=0.0000095DO
DTRIB(8)=0.0000105DO
DTRIB (9) =O.0000125DO
DTRIB(10)=O.000017ODO
DTRIB(11)=O.000095DO
11222 OJNTINUE
RFKlNT=O.ODO
PI=4*n~~(1.ODO)
CAXIS=XAXIS-AXISST
C VISCOSITf OF AIR
VISC=O.0000178D0
C EI.EX:'lRJDE STRENi'IH
EISTR=6D05
C DIELEJ:::TRIC a:NS~T
DIEl:::ST=4.0DO
C PER-1ITTIVITf OF FREE SPlICE
PEI1M=8 .850-12
C EI.EX:'lRJDE NEEDLE RADIUS
TIPRAD=0.005DO
C FIElD PO'IDITIAL
PHI=60000.ODO
C DENSITf OF EaoIDER
DENS=1000
~
Page 6

C CAICUlATICN <rNS'llINTS
C
C
CELEl:=4*PI*PERM
S(JI=HEIGHT**2
CGUN=PHI/ (DrDG (GUN'mR/TIPRAD) )
CIlQJN=2*GUN'mR .
CCRGE=3*PI*(DIECST/(DIECST+2.0))*PERM*ELSTR
CMl\SS=PI*DENS/6
ClDRAG=3*PI*VISC
C
C TIME OF IN'lE3RATICN FOR E7\CH PARTICLE (M1\X)
TmlO=5.000
WRI'lE (JW ,1000)
WRI'lE (JW ,1020)
WRI'lE (JW ,1021)
WRI'lE(JW,123) (PLOTNO(I),I=l,7)
WRI'lE(JW,122) (CXXoIENT(I) ,I=l, 7)
WRI'lE(JW,1023)NUMBER
WRI'lE (JW ,1024) RANDC)
WRI'lE(JW,1025)XLST
WRI'lE(JW,1026)S~TL
WRI'lE (JW ,1027)UNF
WRI'lE(JW,1028)HEIGHT
WRI'lE(JW,1029)UP
WRI'lE(JW,1030)VP
WRI'lE(JW,1031)TOL
WRI'lE(JW,1032)TEND
WRI'lE(JW,1033)TRAJST
122 FORMI'.T(7A4)
123 FORolAT(- PIDT FILE NAME IS -, 7A4)
1020 FORolAT(2 (lX/) ,30H B:XlNIl/>.RY <rNDITIOOS)
1021 ~T(30H )
1023 FORolAT(2OH 00 OF PARTICLES = ,nO)
1024 ~T(2OH RANlXM NUMBER GEN= ,no)
1025 ~T(2OH X S~T FROM GUN = ,E20.10)
1026 ~T(2OH S~T WID'IH IN Y = ,E20.10)
1027 ~T(2OH UNIFORM AIR VEL = ,E20.1O)

1028 ~T(2OH HEIGHT OF 'll\R:2ET = ,E20.10)


1029 FORolAT(2OH INITIAL X VEL. = ,E20.10)
1030 roRolAT(2OH INITIAL Y-VEL. = ,E20.10)
1031 FORolAT(2OH TOIERAN:E ERroR = ,E20.10)
1032 FORolAT(2OH IN'lE:;RATICN TIME = ,E20.10)
1033 ~T(2OH TRAJECTORY S~T = ,E20.10)
WRI'lE (JW ,1000)

ccaxx:ccco::o x:cccoX'cccccccccoX'cccco::OXCO)))X'ccax'OX'CC'CCCCO:CC
C sa:::TIOO 4

C CAIJ:lJIATICN OF SIZE I'.ND S~T rosITICN


CALL GOSCBF (RIINDO)
00 14 I=l,NUMBER
R=GOSCAF (R)
FRAC=GOSCAF (FRIIC)
FRIIC10=FRIIC*10
IND=O
00 16 3=1,10
E=J
IF (FRIIC 10 • LT. E) GO TO 11115
16 CCNTINUE
Page 7

11115 DIAM=DTRIB(J+1)-((J-~10)*(DTRIB(J+1)-DTRIB(J»)
D (I)=DIAM
RADIUS=DIIIM/2
S'mRT=S'mRTL*R+P'IST
S'mRTl(I)=S'mRT
C CAICUIATION CXNS'll'IN'IS

CHARiE=CCRGE* (DIAM**2)
CSQ=CHAR;E**2
PC (I) =CHAR;E
MASS= (DIAM**3) *01ASS
REYCXJN=DIAM*1.0/VISC
GRAV=O
CEFLEJ::=CHAR;E*CGUN
CFOO::E=CIlMiE/CELEJ::
CDRAG=CLDRAG*DIAM
IE6=LD06
RAD=RADIUS*IE6
NUX=RAD/3
NUY=RAD/6
SIZE=RAD/2.5
CALL CHASIZ(SIZE,SIZE)
C B:lUNIlARY CXNDITICNS FOR EACH PARTICLE
'1'=0.000
Z(l)=UP
Z (2)=XIST
Z(3)=VP
Z(4)=S'mRT
N=4

C SETTING OF Nl\G CXNTroL PAR/IME'lERS

CIN (1) =1. 000


CIN (2) =0 .000
CIN(3)=0.000
CIN (4) =0 .000
CIN(5)=0.000
CIN(6) =0.000
CCM-1 (1) =0.000
CCM-1(2) =0.000
CCM-1 (3) =1.000
W(1,21)~999
W(2,21) =TAAJPR
W(3,21)~999
W(4,21)~999
CCM-1(4)=O.000
CCM-1 (5) =0.000
X=O.OOO
Y=O.OOO
VELX=O.OOO
VEU'=O.OOO
OIDX=O.OOO
OIDY=O.OOO
OVELX=O.OOO
0VEI.:i=0 • 000
OIDT=O.OOO
MPED=O
IFAIL=l
IW=4
I'tl1=22
C PRINT IID'.DING FOR 'Il'AJEX:'IDRY 'll'J3LE

PRnm{= (GUN'lM.-·Z (2) ) *IE6


Page 8

PRIN'!Y=Z (4) *IE6


PRIN'IH=CIN (5)
IF(I.NE.FOHCEP)GO IQ 11333
WHI'lE (JW, 3008)
WHI'lE (JW ,30311I
WHI'lE (JW, 3008)
WHI'IE (JW, 3030)
WHI'lE (JW, 3008)
11333 <XlNTINUE
IF ('lYPE.ElJ.O)GO IQ 10000
WHI'lE (JW, 3008)
WHI'lE (JW, 3043)I
WHI'lE (JW, 3044)
WHI'lE (JW, 3008)
WHI'lE(JW,3060)D~
WHI'lE(JW,3061)S~RT
WHI'lE (JW,3062) PTST
WHI'lE(JW,3063)~
WHI'lE(JW,3064)MASS
WHI'lE (JW, 3008)
WHI'lE (JW, 3045)
WHI'lE(JW,312) T,PRINTK,PRIN'lY,Z (1) ,Z(3) ,AIRVX,AIRVY,PRINTH

C CALL IQ NAG !UJTINE

10000 CALL D02QBF(T,TEND,N,Z,CIN,IQL,FCN,COMM,CQN,COUT,MPED,


lPEDERV ,m ,W ,IW ,IW1,IFAIL)

C sa::TIOO 5

C a:MES OUT OF D02QBF WHEN REACHING FIDIT OR EIIERY S'lEP ~

IF(IFAIL.GT.O)GO IQ 99999
IF(IND.ElJ.l)GO IQ 10005
W(2,21)~999
COMM(4)=1.
INO=l
OY=(~-Z(2»*IE6+50.0
OX=(Z(4)*IE6)+CAXIS
CALL MO\I'I02 (OX ,OY)
GO IQ 10006
10005 VELX=Z(l)
VELY=Z(3)
X=Z (2)
Y=Z (4)

C am::K IF PARTICLE IS OOT OF BJUNDS OR IN LAYER REXiIOO

IF(T.GT.TEND-0.5)GO IQ 22333
IF(Y.LT.-O.OlDO)GO IQ 22333
IF(Y.GT.O.OlDO)GO IQ 22333
PRINTX=(~-X)*IE6
PRIN'!Y=Y*IE6
C S'lEPrENG'lll
IF(I.NE.EOHCEP)GO IQ 11334
Page 9

WRITE(JW,313)T,PRINT.K,~,D~GX,DRNGY,FIELDX
1 ,FIEIDY , IMAGE ,F'lDT.K,.E"IOT.{
11334 CONTINUE
IF(~.EQ.O)GO 10 11111
PRIN'lH~T (4) -<XXJT (5)
WRITE (JW, 312) T, PRINT.K,~ , VELX, VEIX ,AIRVX ,AIINY ,PRIN'lH,REY

C SEI:TIClN 6

ccccc:cccccccccr::cccccrn::a::axrCcccfXxxrcn::u:o 0 CCCCCCCCCCCCCCC
c· am:::K ON rosITION REIATIVE 10 SUBSTRATE & 0'lHER PARTICLES

11111 CONTINUE
IF(X+~IUS+RFRJNT.LT.m::JNT)GO 10 10001
IF(I.EQ.l)GO 10 10001
K=I-l
00 31 J=l,K
JJ=I-J
XJJ=Xl(JJ)
YJJ=Yl(JJ)
DTPSQ=(X-XJJ) **2+ (Y-YJJ) **2
~J=D(JJ)/2
RADSQ= (~IUS+~J) **2
IF (D'WSQ.LT.~SQ)GO 10 10222
31 CONTINUE
10001 IF(X+~IUS.GT.~)GO 10 10003

C PARTICLE H1IS PASSED 'lllROOGH ALL CHEX:KS ruT MUST BE IN


C 'mI\JEl::1ORY PRINT REGICN AND SO 'IHEREFORE PIDT STEP
C X AND Y ARE REVERSED FOR PID'I'l'II'G

CX=(Y*IE6)~IS
CY=(~-X)*IE6+50.0
CALL LIN102 (CX ,CY)
OX=CX
OY=CY
10006 OIDX=X
OIDY=Y.
OIDT=T
OVELX=VELX
OVELY=VEIX
IFAIL=l
GO 1010000
C PARTICLE IS roJND 10 !UT 1IN)'lEER PARTICLE
C CALCUIATE IN'IERSEI:TIClN OF LINE BE'lWEEN 'I'M:) 'ffiAJEI:'IDRY POINTS
C AND CIRCLE OF ~IUS (~IUS+~J)

C am:::K IS In'l CARRIED CXJT 10 SEE IF AN'! 0'lHER PARTICLE

cm.
C IS lilT. SEARCH CARRIES CN 'IHROUGH REST OF S'mTIClNARY PARTICLES

C S&::TICN 7

10222 IF(JJ.EQ.l)GO 10 10002


CHEI:K=JJ-l
00 32 JK=l,CHEJ:K
Page 10

JJ=CHECK+l-JK
CHXJJ=xl (JJ)
CHYJJ=yl (JJ)
CHDSQ=(X-cHXJJ) **2+ (Y-cHYJJ) **2
CHRl\D=D(JJ)/2
CHRSQ= (RADIUS~) **2
IF(CHDSQ.LT.CHRSQ)GO 'ID 10022
32 CXNrINUE
JJ=CHECK+l
GO 'ID 10002

C 1INO'IHER PAATICLE H!\S BEEN roJND - tn.. FIND OOT WHICH mE


C H!\S 'lHE IDST OIIERSHCOT - BICl;ER 'lHE DIE EEREl.'O: 'lHE MJRE IT
C H!\S OVER SHOT

10022 IF ( (CHRSQ-OlDSQ) • LT. (FADSQ-DTPSQ) ) GO 'ID 32


RADSQ=CHRSQ
DTPOO=CHDSQ
XJJ=CHXJJ
YJJ=CHYJJ
GO 'ID 10222

C 'lHE ~T PARTICLE 'ID 'lHE !AST POINT CN 'lHE 'mAJEX:'IORY


C IS Im REJ:ORDED AND SO IN'IERPOrATION RXlTINE <XM-lEN:ES

10002 GRAO= (y-oIDY) / (X-{)IDX)


INCEPT=Y-GRAD*X

C Im USING x**2+Y**2=R**2 C!\I.CUIAlE INTElCEPT

P=GRAD*XJJ+INCEPT-YJJ
A=HGRAD**2
B=2*GRAD*P
C=P**2-RADSQ
IF (B.LT.O)GO 'ID 888
Rl=(-B-DSQRT(B**2-4*A*C»/(2*A)
GO 'ID 8888
888 Rl=(-B+DSQRT(B**2-4*A*C»/(2*A)
8888 a:JNTINUE
R2=C/ (A*Rl)
INTXl=XJJ+Rl
INTX2=XJJ+R2
IF (INTXl.GT.INTX2)GO 'ID 10010
INTX=INTXl
GO 'ID 10011
10010 INTX=INTX2
10011 INT'{=GRAD*INTX+ IOCEPT

C Im GO AND CA.ICUIAlE 'LEVER' ,O'IHER VARIABLES AND


C PIDT FINAL POINT

GO 'ID 10007

C sa::TICN 8

(XXX'CO)))XX'CCCCCCCCCC(,(X'CCO:CCOJ::COX'O:'CXX'C'CX'CCXl a:x:o:xJ)))xxx))))::C

C OUOJrATION OF FINAL POSITION CN SUBS'ffiA.lE USm:; EOJATION


C OF LINE BE'lWEEN OID & NEW POSITION
Page 11

C WE KNOW '!HAT X MUST EX:UQIU. 'GtlN'mR-RADIUS'

10003 JJ=O
GRAD= (Y-0rDY) / (X-0IDX)
INCEPT=Y-GRI\D*X
rn'IX=GUN'lAR-RADIUS
INT'f=GRAD*rn'IX+:m::EPT

C SEX:TICN 9

C N:M 00 AND CAICUIA'lE 'LEVER' AND 0lHER VARIABLES

10007 LEVER=(X-rn'IX)/(X-0IDX)
Xl (I) =rn'IX
Yl(I)=INT'f
IN'lV=VELY- LEVER* (VELY-ovELY)
rn'lU=VELX-LEVER* (VELX-0VELX)
INTl'=T-LEVER* (T-OIDT)
FLIGHT(I) =INTT

C PIDT ro8ITICN OF mRTICLE I\ND 'rnAJOC'IDRY


C
CX=(INT'f*IE6)~IS
CY=(~-rn'IX)*IE6+50.0
au.r. MJI1'I02 (CX ,CY-RAD)
au.r. AHCBY2(0.0,RAD,0.0,0.0,0)
CAL!MO\1'102 (CX-NUX ,CY-NUY)
au.r. CHAINT (I , 3)
au.r. MJI1'I02 (CX ,CY)
au.r. LIN'I02 (OX ,OY)
C '!HEN PRINT FINAL ro8ITICN
C a:MPARE FOSITICN OF mRTICLE 10 FroNT CNE

IF (!N'n{.GT.FROOT)GO 'IQ 10008


F'OCN'l'- rn'IX
RFImT=RADIUS
10008 CONTINUE
IF (T'iPE.EXl.O)GO 'IQ 21111

C SEX:TICN 10

ccccccccccoX'C'O cco:n::CCXX))X'(l ))X'C'O))))X'C'O:CCCCCO O))X'C'():X:CCCCCCO:CC

WRI'lE(JW,3045)
WRI'lE (JW, 3000)
WRI'lE (JW, 3006) I
WRI'lE (JW ,3007)
WRI'lE(JW,3001)T,OIDT,rnTT
WRI'lE (JW, 3002) VELX ,OVELX , IN'IU
WRI'lE(JW,3003)X,OIDX,rn'IX
WRI'lE(JW,3004)VELY,QVELY,rnTV
WRI'lE(JW,3005)Y,OrDY,INT'f
Page 12

WRI'lE (JW ,3008)


WRI'lE (JW ,3000)
C CONTINUING FINAL PRIN'IOJT

WRI'lE (JW, 3008)


PRIN'lD=RADIUS*IE6
PRIlm(= (GUN'mR-INTX) *IE6
PRIN'lY=INT':{*IE6
PRINTS=S'D'.RT*IE6
WRI'lE (JW , 3009) PRIN'ID
WRI'lE(JW,3010)~
WRI'lE(JW,3011)PRIN'lY
WRI'lE(JW,3012)INTT
WRI'lE (JW,3013) PRINTS
WRI'lE (JW, 3008)
WRI'lE (JW, 3000)
WRI'lE (JW, 3008)

QJ 'ID 21111

22333 INTY=999.99
JJ=O
INTX=999.999
C PRINT FINI\L PARTICLE roSITION & HIT NUMBER

21111 CONTINUE
IF(I.NE.FORCEP)GO 'ID 11335
WRI'lE (JW, 3008)
WRI'lE (JW , 3030)
11335 CONTINUE
WRI'lE(JW,307S)I,INTK,IN~,JJ,DIAM
14 CONTINUE

WRI'lE (JW, 1001)


WRI'lE (JW ,1002)
WRI'lE (JW ,1003)
DO 18 J=l,NUMBER
FPl=(GUN'mR-Xl(J»*IE6
FP2=S'D'.RTl(J)*IE6
FP3=Yl (J) *IE6
FP4=D(J)*IE6
WRI'lE (JW ,1004)J ,FLIGHT(J) ,FP1,FP2,FP3,FP4
18 CONTINUE
WRI'lE (JW ,1005)
WRI'lE (JW, 1000)

WRI'lE (JW, 3071) FRJNT, RFRJNT


WRI'lE (JW, 3008)
WRI'lE(JW,3072)AIRVX
WRI'lE(JW,3073)AIRVY
WRI'lE(JW,3080)DRAGX
WRI'lE(JW,3081)DRAGY
WRI'lE(~N,3082)IMAGE
WRI'lE(JW,3083)FIELDX
WRI'lE(JW,3084)FIELDY
WRI'lE(JW,308S)FTOTK
WRI'lE(JW,3086)~

1000 FORMAT(///'****************************************************
1*******************##**#*************#*#****#****##****')
Page 13

1001 roRMI\T (//,'


1 FINAL PARTICLE roSITION')
1002 roRMI\T('
1 ')
1003 roRMI\T(/I I 1'- FLIGHT TIME HEIGHT FRJ
lM PIA'IE Y S'lART Y <rORDINA'IE PARTICLE 0
2IAME'lER' )
1004 roRMI\T(2 (lX/) ,13H PARTICLE ID ,I3,3H ,5F20.10)
1005 rom>.T (4 (lX/) )
3000 FORMAT(//~ *******************************************************
1***********************************************************~)
3001 roRMI\T(2OH NEW TIME = ,E20.10,2OH OL
10 TIME = ,E20.10,2OH INTPL'O TIME = ,E20.10)
3002 rom>.T(2OH NEW X VEL = ,E20.10,2OH OID X VE
LL = ,E20.10,2OH INTPL'O X VEL = ,E20.10)
3003 roRMI\T(2OH NEW X <rORD = ,E20.10,2OH OID X <rO
=
lRD ,E20.10,2OH INTPL'O X <rORO =,E20.10)
3004 roRMI\T(2OH NEW Y VEL = ,E20.10,20H OID Y VE
LL = ,E20.10,2OH INTPL'O Y VEL = ,E20.10)
3005 rom>.T(2OH NEW Y <rORD = ,E20.10,2OH OID Y <rOR
10 = ,E20.10,2OH INTPL'O Y <rORD = ,E20.10)
3006 roRMI\T(2(lX/) ,33H FINAL PARTICLE EOSITION - NUMBER,13)
3007 roRMI\T (38H )
3008 roRMI\T(3 (lX/))
3009 roRMI\T(2OH PARTICLE SIZE = ,F20 .10)
3010 roRMI\T(2OH X <rORDINA'IE = ,F20.10)
30ll rom>.T(2OH Y <rORDINA'IE = ,F20.10)
3012 rom>.T(2OH TIME IN FLIGHT = ,F20.10)
3013 rom>.T(2OH S'lART Y <rORD = ,F20.10)
3043 rom>.T(' EUCTroS'll\.TIC POWDER <X\1\.TING 'ffiAJEC'IORY - PARTI
lCLE NUMBER ',13)

3044 roRMAT(' - - - - - - - - - - - - - - - - - -
1 ')
3031 roRMI\T (' PARTICLE 'ffiAJEC'IORY - roOCES CliI PARTICLE ID.
1 ',13)
3030 rom>.T(' TIME X <rORDINA'IE Y <rORD DRAG X DRAG Y
1 FIEID X FIEID Y IMAGE 'lOT IN'IER X 'lOT IN'IER Y')
3045 roRMAT(' TIME X <rORDINA'IE
1 Y <rOROINA'IE X VEIJ:x:I'I'l Y VEL
2CCI'I'l AIR VEL X /I.IR VEL Y S'IEP IlNi'IH REYNOIDS ID ' , I)
313 rom>.T(F8.5,F12.4,FI0.4,7E12.4)
312 roRMI\T(F12.6,2F12.4,4F12.6,E14.4,F12.4)
3071 rom>.T(//,' X <rOROINA'IE OF HIGHEST PARTICLE = ',E20.10,
1/ " RADIUS OF '!HIS PARTICLE = ' ,E20 .10 , 11)
3072 rom>.T(' Am VELX= ',E20.10)
3073 rom>.T(' Am VELY= ' ,E20.10)
3075 roRMI\T(/' PARTICLE' ,13,' IANDED , X = ',F14. 7,' , Y =
I' ,F14. 7,' HIT PARTICLE ID. ',I3,' PARTICLE OIAME'IER = ',F14. 7)
3080 rom>.T(' ORAGX =' ,E20.10)
3081 roRMI\T(' ORAGY =' ,E20.10)
3082 FORMAT(' IMAGE = ',E20.10)
3083 FORMAT(' FIEIDX = ',E20.10)
3084 rom>.T (' FIEIDY =' ,E20 .10)
3085 FORMAT(' FTOT% = ',E20.10)
3086 FORMAT(' ~ = ',E20.10,111)
3096 FORMAT (13)

C ERroR S'll\.'lUS CCliI'lroL - CXJTruT OF NAG PARAME'IERS

99999 00 91 KK=1,4
Page 14

wru:m(JW,3015)KK,CCUT(KK) ,KK,CIN(KK) ,KK,CON(KK),


1 KK ,CCt-M (KK)
91 ClJNTINUE
KK=5
wru:m(JW,3016)KK,CCUT(KK) ,KK,CIN(KK) ,KK,CON(KK)
00 92 KK=6,7
wru:m (JW ,3017)KK,CCUT(KK) ,KK,CIN (KK)
92 CXJNTINUE
00 93 KK=8,16
wru:m (JW, 3018) KK ,CCUT(KK)
93 CXJNTINUE
WRIm(JW,3333)~IL
CALL IlE\IEl'ID
3333 roJM\.T(' ~IL=' ,13)
3060 roJM\.T(' DIAME'lER OF PARTICLE = ',E20.10)
3061 FO!Ml'.T(' S'lART roSITION IN Y = ' ,E20.10)
3062 roJM\.T(' DIS'1l'.NCE FOCM ORIGIN = ' ,E20.10)
3063 roJM\.T(' CllARGE ON PARTICLE =' ,E20.10)
3064 roJM\.T(' w.ss OF PARTICLE = ',E20.10)
3015 roJM\.T(' CCUT NO.',I2,2X,E13.6,' CIN NO.',I2
1,2X,E13.6,' CON NO.',I2,2X,E13.6,' CCt-M NO.',I2
1,2X,E13.6)
3016 roJM\.T(' CCUT NO.',I2,2X,E13.6,' CIN NO.',I2,2X
1,E13.6,' CON NO.',I2,2X,E13.6)
3017 roJM\.T(' CCUT NO.',I3,2X,El3.6,' CIN NO.',I2,2X,E
113.6)
3018 roJM\.T(' CCUT NO.' ,I2,2X,E13.6)

S'lOP
END

SUBRXJ'I'INE R:N(T,Z,F)

C SEJ:TION 11


~*8 X,Y,Xl(500),Yl(500),T,Z(4),F(4),VELX,VELY,CSQ
~*8 SQX,~,NEWX,GUN'lAR,S(l<X ,RCCNS ,CELEX: ,HEIGHT
~*8 SQH ,SQRRC, '!HE'D>. ,SINT ,COST ,AIRVX,AIIN'f ,UNE' ,DRAGX
~*8 DRAG'{ ,CDRAG ,IMAGE,CHAR:iE,ELEX:,CGUN ,FIEIDX ,FIEIDY
~*8 F'IO'IX,F'IO'l'l,XI,YI ,IMX,IMY ,Cn-GlN,CEFLEX:
~*8 DIFIMX ,DIFIMY ,DIFXI ,DIFYI,DTSQJ ,D'ISQI ,SQRDI
~*8 SQRDJ,OQSI,CQSJ,SINI,SINJ,ATTFX,ATTFY,REPFX,REPFY,CONST
~*8 CFORCE,PC (500) ,PAR'IFX,PAR'lFY ,w.ss,GRAV,D(500)
~*8 DIFVX,DIFVY ,VREt"REY,REY<XN

IN'lfXiliR I,!N03, II

c:cMoICN/BLl/tJNF ,HEIGHT ,GUN'lAR,CELa:: ,SQH ,CGUN,CEFLEC,NUMBER


c:cMoICN/BL2/Cn-GlN ,D,PC,CllARGE,MASS ,CI'OICE ,CDRAG ,Xl, Yl,GRAV , I ,CSQ
. c:cMoICN/BL3/DRAGX ,DRAG'{ , IMAGE ,FIEIDX,FIEIDY ,F'l'O'I'X,F'ID'IY ,AIRVX ,AIIWY
Page 15

X=Z (2)
Y=Z(4)
VELX=Z (1)
VELY=Z(3)
SQX=X**2
S\1i=Y**2

C VEIOCI'lY PROFILE CAICUIATIONS

NEWX=X-GUN'Il'.R
~=NEWX**2
RCCNS=OSQRT( (~-S\1i+SCfI) **2+ (2*NEMt*Y) **2)
SQROC=DSQRT(BCCNS)
'IHE'I!'.=m'lM ( (- 2*NEWX*Y) / (~-S\1i+SQH) )
SINT=OSIN ('IHE'I!'./2)
CDST=OCQS ('IHE'I!'./2)
AIRVX=-UNF* (NEWX*CDS'!'-Y*SINT) /SQROC
AIRW=UNF* (NEWX*SINT+Y*CDST) /SQROC
DIFIJX=AIRVX-VELX
DIFVY'=AIRW-VELY
VREL=OSQRT (DIFIJX**2+OIFVY**2)
REY=REYCON*VREL

C DRAG FOR:E

DRIIGX=CDRAG* (AIRVX-VELX)
DRAGY=CDRAG* (AIRVY-VELY)

C IM!'.GE 'FORCE

IMAGE=CSQ/ (CELEX:* (2* (GUN'D'.R-X» **2)


C FIEW FORCES

ELOC=CEFLEx::/ (SQX+S\1i)
FIEWX=ELEx::*X
FIEIDY=ELEx::*Y

C SEI:TION 12

C IN'lER PARTIOJIA'lE EDR:ES

F'IO'D{=O • 000
F'IOTi=O.OOO
IF(I.EQ.1)GO 10 22222
IND3=I-1
00 21 II=1,IND3
XI=X1(II)
YI=Y1(II)
IMX=C]l.GJN-XI
IMY=YI
DIFIMX=IMX-X
DIFIMY=IMY-Y
DIFXI=XI-X
DIFYI=YI-Y
D'ISQ1=DIFIMX** 2+DIFIMY** 2
DTSQI=DIFXI**2+DIFYI**2
Page 16

SQRDI=DSQRT (DTSQI)
SQRDJ=OSQRT (DTSQJ)
<XlSI=DIFXI/SQRDI
OJSJ=DIFIMX/SQRD,J
SINI=DIFYI/SQRDI
SmJ=DIFIMY/SQRDJ

I'.'l'.lFX~SJ/D'ISOJ
I'.TlFY=SmJ/D'lSQJ
REPFX~I/DTSQI
REPFY=SINI/DTSQI

a;wS'J.'=CroR:E*PC (II)
PM'IFX= (1'.'l'.lFX-REPFX) *a:wST
PM'lFY= (ATlFY-REPFY) *a:wST

F'lO'lX=F'lO'1X+PI'.R'IFX
F'lOT'f=F'lO'l'l+PM'lFY

21 CXJNTINUE

22222 F(l)=(DRI'.GX+FIELDX+IMI'.GE+F'lO'lX)/MASS
F (2)=VELX
F(3)=(DRAGY+FIELDY-+GAAV+F'lOTi)/MASS
F(4)=VELY
RE'lURN
END

$
Figure D2(a-b). Listing of Segment from Charged Layer Program

C SEG!ENT OF PRCXiRAM 'IAYER.FOR'

C USED FOR 'lHE CALCUIATICN OF 'mAJOC'IDRY OF PARTICLES 'lLWARDS A


C SUBS'ffiA'IE CO\lEREI) WI'IH A 'lHEDRETICAL LAYER OF ~.

C SEx::TION 11

C CALCUIATICN OF INTERPARTIOJIA'IE FO!CFS ACTIN:; CN CN:XMING PARTICLE.


C A 'lHEDRETICAL LAYER OF CHAR;E OF '!HICKNESS ''!HICK' IS I\SSUMED 'ID
C EXIST ACIDSS 'lHE WHOrE WID'IH OF 'lHE PIA'IE. AN OI.IERAIL CHAR;E OENSI'lY
C IS CALCUIA'IED {PER UNIT AREA} ON 'lHE BASIS OF 'lHE SIZE OF· PARTICLES
C AND A rorosI'lY OF PACKIN3 FAC'IDR.
C '!HIS APPROACH SAVES ALL 'lHE SlMlATIONS OF INDIVIIXlAL PARTICrE roocES.
C A MCDEL IS USID FOR BO'IH 'lHE R&l'.L & IMAGE LAYERS

C IN'IER PARTIOJIA'IE FORCES

F'IO'IX=O.OOO
F'ID'lY=O.OOO
WID=HEIGHT/2.000
WIDN=WID-Y
AGUN'IH=GUN'11\R-X+'!HICK
GUN'lII=GUN'11\R-'!HICK-X
GUNN=GUN'n\R-X
WIOP=WID+Y
NWIOO=-WID-Y

Tl=WIDN-H)SQRT {GUN'IH**2-tWIDN**2}
T2=NWIDN-H)SQRT{GUNN**2-tWIOP**2}
Bl=WIDN+DSQRT{GUNN**2-tWIDN**2}
B2=NWIDN+DSQRT{GUN'IH**2-tWIDP**2}

REPFX=-CFORCE*C!'.Rm*OUX; {Tl*T2/ {Bl*B2} }

Tl=GUN'IH+DSQRT{WIDN**2-+GUN'rn**2}
T2,q)NN-H)SQRT{WIDP**2+GUNN**2}
Bl,q)NN+DSQRT {WIDN**2+GUNN** 2}
B2=GUN'IH-H)SQRT{WIDP**2-+GUN'rn**2}

REPFY=-CFORCE*C!'.Rm*DUX; {Tl*T2/ {Bl*B2} }

Tl=WIDN+DSQRT{GUNN**2-tWIDN**2}
T2=NWIDN-H)SQRT {AGUN'IH** 2-tWIDP** 2}
Bl=WIDN-H)SQRT{AGUN'IH**2-tWIOO**2}
B2=NWIDN-H)SQRT{GUNN**2-tWIDP**2}

ATlFX=CFORCE*C!'.Rm*DUX; {Tl*T2/ {Bl*B2}}

Tl,q)NN-H)SQRT{WIDN**2+GUNN**2}
T2=AGllN'lH+DSQRT {WIDP** 2+AGUN'IH**2}
Bl=AGUN'IH-H)SQRT {WIDN** 2+AGUN'IH** 2}
Page 2

B2~+DSQRT (WIDP**2-1QJNN**2)

ATlFY=CFOOCE*C'ARE\*OIro (Tl*T2/ (B1*B2) )

F'lO'IX=REPFX+ATlFX
F'IO'IY=REPFY+A'ITFY

F (1) =' (DRAGX+FIEIDX+IMAGE+F'IOTX) /MASS


F (2)=VELX
F(3)=(DRAGY+FIELDY~V+F'IO'IY)/MASS
F(4)=VELY
RE'IURN
END

$
Explanation of Computer Program Variables

Page of
Definition

A Constant used in plotting of axes 4

AIRVX Air velocity in x direction 15

AIRVY Air velocity in y direction 15

ATTFX Attractive force in x direction between oncoming 16

and an image of a stationary particle

ATTFY As above, in y direction 16

AXISST Start of x axis away from origin on plot 3

B Constant used in plotting of axes 4

C Calculation variable 10

C1DRAG Calculation constant for drag force 6

CCRGE Calculation constant for calculating charge on 6

particle

CDRAG Calculation constant for drag force 7

CEFLEC Calculation constant for field force 7

CELEC Calculation constant for interparticulate forces 6

CFORCE Calculation constant for interparticulate forces 7

CGUN Calculation constant for field force 6

CHARGE Charge on oncoming particle 7

CHDSQ Square of distance between oncoming and collision 10

particle

CHECK Number of last particle in co'llision 9

CHRAD Radius of collision particle 10

CHRSQ Square of the sum of radii of oncoming and collision 10

particles
CHXJJ X coordinate of collision particle 10

CHYJJ Y coordinate of collision particle 10

CIMGUN Calculation constant for image force 6

CIN Control parameter (array) for NAG routine 7

CMASS Calculation constant for mass of particle 6

COMENT Filename of particles being studied FILE

COMM Output variable for NAG routine, specifying 7

interrupts (array)

CON Control of stepsize during integration (array) 7

CONST Calculation constant for interparticulate forces 16

COSI Cosine of angle defined in Figure 5.11 16

COSJ Cosine of angle defined in Figure 5.11 16

COST Cosine of e from flow model equation 15

COUT Used for output information from NAG routine (array) NAG

CSQ Square of particle charge 7

CX Plotter position of centre of particle in X direction 9/11

CY As above, Y direction 9/11

D Diameter of particle used in storage array 7

DENS Density of powder· 5

DIAM Diameter of oncoming particle 7

DIEST Dielectric constant of powder 5

DIFIMX Distance in x direction between oncoming particle 15

and an image

DIFVX Relative velocity in x direction 15

DIFVY Relative velocity in y direction 15

DIFXI Distance in x direction between oncoming and 15

stationary particle

DIFYI As above, y direction 15

DRAGX Drag force in x direction 15


DRAGY Drag force in y direction 15
DTPSQ Same as CHDSQ 9

DTRIB Storage of size distribution of particles (array) 3/5

DTRIBO Size of monosized distribution, or flag for FILE

distribution

DTSQI Square of distance between oncoming particle and 15

stationary one

DTSQJ As above, image of stationary particle 15

E NAG work variable (array); Same value as J 6

ELEC Calculation constant used for field force 15

ELSTR Electrode field strength 5

F Value of differentials (array) 16

FIELDX Field force in x direction 15

FIELDY As above, y direction 15

FLIGHT Stored times of· particle flights (array) 11

FPl Final output variable for x position 12

FP2 Final output variable for y start position 12

FP3 Final output variable for final y position 12

FP4 Final output variable for diameter of particle 12

FRAC Generated random number used for calculation of 6

particle diameter

FRAClO Ten times FRAC 6

FRONT x coordinate of highest particle in packing 11

FTOTX Total of interparticu1ate forces in x direction 16

FTOTY As above, y direction 16

GRAD Gradient of line betwen centres of two particles 10


GRAV Force due to gravity 7

GUNTAR Distance between gun and target 2

HEIGHr Length of target 3

I Do loop constant; current particle number in flight 6

II Do loop constant 15

IDl Do loop constant 4

ID2 Do loop constant 5

IE6 Calculation constant 7

!FAIL Failure identifier variable 7

IMAGE Image force on particle 15


IMX X coordinate of image of a packed particle 15
!MY As above, y coordinate 15

INCEPT Intercept of line with gradient GRAD 10

IND Flag for first particle calculation 6

IND3 Do loop set constant 15

INTT Interpolated flight time 11

INTU Interpolated x velocity of particle 11

INTV As above, y velocity 11

INTX Calculated final x position of particle 10

INTXl 1st solution of x position of particle 10

INTX2 As above, 2nd solution 10

INTY Interpolated y position of particle 10

IW Used in dimensioning W 7

IWl As above 7

J Do loop variable 6/9

JJ Do loop variable 9

JW Output device number 3

K Do loop variable 9
KK Do loop,variable 13

LEVER Calculation constant in interpolation calculation 11

M Calculation variable for plotting of axes 4

MASS Mass of particle 7

MPED Flag for calculation of Jacobian by NAG routine 7

N Number of differential equations 7

NEGSTP Number of negative steps on x axis of plotter 3

NEWX Converted x coordinate for flow model 15

NUMBER Total number of trajectories calculated FILE


NUX Siting of particle number on plot, x direction 7

NUY As above, y direction 7

OLDT Stored time of particle old position 7/9

OLDX As above, x position 7/9

OLDY As above, y position 7/9


OVELX As above, x velocity 7/9
OVELY As above, y velocity 7/9
OX Last x coordinate used on plot 8/9
OY As above, y coordinate 8/9
P Calculation variable used in stopping section 10
PARTFX Sum of attractive and repulsive interparticulate 16
forces in x direction

PARTFY As above, y direction 16


PC Storage of particle charges (array) 7

PEDERV NAG routine variable NAG


PERM Permittivity of free space 5

PHI Potential at gun 5

PI 11 5

PLOTNO Name of plot file used FILE


PRINTD Output variable for diameter of particle 12

PRINTH As above, stepsize 8

PRINTS· As above, y start coordinate 12

PRINTX. As above, x coordinate 7/8

PRINTY . As above, y coordinate 8

PTST Distance of start length from x axis 3

PW NAG routine variable NAG

Q Calculation variable for plotting of axes 4

R Random number generated for start position 6

calculation

RI Root of quadratic equation 10

R2 As above 10

RAD Radius of particle in microns 7

RADIUS Radius of particle 7

RADJJ Radius of stationary particle 9

RADSQ Equal to CHRSQ 9

RANDO Number which determines start of random number 2

RCONS Calculation variable in flow model IS

REPFX Repulsive electrostatic interparticulate force in 16

x direction

REPFY As above, y direction 16

REY Particle Reynolds number IS

REYCON Calculation variable for Reynolds number 7

RFRONT Radius of highest particle in packing 11

SINI Sine of angle defined in Figure S.ll 16

SINJ Sine of image angle defined in Figure S.ll 16

SINT Sine of e from flow model IS

SIZE Size of numbers in circles on plot 7


SQH Square of target height 6

SQNX Square of NEWX 15

SQRDI Distance between oncoming and packed particle squared 16

SQRDJ As above, oncoming and image 16


SQRRC Square of RCONS 15

SQX Square of X 15

SQY Square of Y 15
START Start position of particle in y direction 7

STARTl As above, storage array 7

STARTL Length used in calculation of start position FILE

T Time 7

TEND Range of integration 6

THETA e in flow model 15

THICK Thickness of charge layer in microns FILE

TIPRAD Radius of electrode tip 5

TOL Tolerance bound FILE

TRAJPR X position for start of trajectory print 3

TYPE Flag for printout of trajectory FILE

UNF Uniform air velocity FILE

UNITSP Plot scale value 3

UP Initial x velocity of particle 3

VELX Velocity of particle in x direction 7/8/15

VELY As above, y direction 7/8/15

VISC Viscosity of air 5

VREL Relative velocity of particle and air 15

VP As UP, y direction 3

W NAG routine variable 7

X X coordinate 7/8/15
Xl Stored final x position (array) 11

XAXIS Distance of origin from plotter origin on plot 4

XI X coordinate of packed particle 15

XJJ As above, image 9

XLST Initial start position of particle in x direction 3

XLTH Length of x axis steps on plot 3

XPAPER Total length of x axis on plot 3

XSTEPS Number of y axis steps 3

Y Y coordinate 7/8/15

Yl As Xl, y position 11

YI As XI, y position 15

YJJ As XJJ, y position 9

YLTH Length of steps on Y axis of plot 3

YPAPER Length of y axis on plot 3

YSTEPS Number of steps on y axis of plot 3

Z Variable in differential equations (array) 7


Figure D3(a-e). A Typical Computer Printout
••......... "..................... , .......•.....................•.....•.........•.•.••..............••••...
BUUNDARY CONDITIONS
-------------------
PLOT fILE NAME IS PLOT NUMBER fIt::LD25V5
25 UH - WITH fIELD
NO Of PARTICLES = 10
RANDOM NUMBER GEN" o
'X START fROM GUN " 0.1000000000E+OO
START WIDTH IN Y = 0.2000000000E-0]
~NIfORM AIR VEL = 0.5000000000E+Ol
,HEIGHT Of TARGET = 0.2000000000E+00
'INITIAL j( Vt::L. " 0.5000000000EtOl
IINITIAL V-VEL. = O.OOOOOOOOOOE+OO
iTOLERANCE ERROR = 0.1000000000E-06
iINTEGRATION TIME" 0.5000000000E+Ol
TRAJECTORY START = O.OOOOOOOOOOE+OO

FIELD25V. LOG 11 28-JUN-1982 11:49:04.10 Page 2

....................................................................................................... ,
,
.. •
PARTICLE 1 LANDED X = 0.1999875 Y
" 0.000]]62 HIT PARTICLE NU, 0 PARTICLE DIAMETER = 0,000025

PARTICLE 2 LANDED X = 0.1999875 , Y


" -0.0001467 HIT PARTICLE NO, 0 PARTICLE DIAMETER
'" 0,000025

PARTICLE 3 LAt;Dt::D , X 0.1999875 Y :: 0.0004320 HIT PARTICLE NO. 0 PARTICLE DIAMETER " 0,000025
"
PARTICLE 4 LANDED X :: 0.1999875 , Y :: -0.0003642 HIT PARTICLE NO. 0 PARTICLE DIAMETER = 0.000025

PARTICLE 5 LANDED X :: 0.1999875 , Y :: -0.0005252 HIT PARTICLE NO. 0 PARTICLE DIAMETER :: 0.000025

PARTICLE: 6 LAI,nt::D x = 0.1'199875 Y


" 0.0005188 HIT PARTICLE:' NO. 0 PARTICLE DI AMETU = O.OOOO2~

PARTICLE 7 LANDED x = 0.1999875 , Y :; 0.0001581 HIT PARTICLE NO. o PARTICLE DIAMETER = 0.00002!
PARTICLE 8 LANDED , X " 0.1999875 Y= -0.0003390 HIT PARTICLE NO. o PARTICLE DIAMETER " 0.00002!
PARTICLE 9 LANDED x " 0.1999875 , Y = -0.0001746 HIT PARTICLE NO. o PARTICLE DIAMETER = 0.00002!

PARTICLE TRAJECTORY - FORCES UN PARTICLE NO. 10

, "R '''T'''''' "K' "\" ,.,,, "i' r. ",,, "r' "m\" " "r '., 'or '\" , \
TOT INTER Y

OO~O~
. o.
/
.140~
9999 9\ 6415 -39
0.000000 99999.6415 -39.4061 4:\99928
~ 9775E~1 ~
-0. H59E-07\ -0. 2956E-ll \ o. 3392E-08\ -0.133 7E-l
0.000000 2.236062
O.
-0.000705 0~J585E~07
-0.1 759E-17 -0.7331E-21
0.00000 99999.4623 -39.4061 -0.1159E-07 -0.2956E~11 0.~392E-08 ~0.1337E~11 0.9775E-16 -0.1759E-17 -0.7331E-21
0.000000 99999.4623 -39.4061 4.999892 0.000000 2.236058 -0.000705 0.3585E~07
0.00001 99931.3334 -39.4061 -0.1154E-07 -0.2927E-l1 0.33B9E-08 -0.1334E-l1 0.9798E~16 -0.1763E-17 -0.7347E~21
0.000014 ' 99931.3334 -39.4061 4.986258 -0.000007 2.234939 -0.000705 0.3585E-07
0.00003 99963.3899 -39.4063 -0.1149E~07 -0.2899[-11 0.3387E-09 -0.1333[-11 0.9801E-16 -0.1767E-17 -0.7371E-21
0.000027 99863.3899 -39.4063 4.972706 -0.000014 2.233623 -0.000705 0.1364E-04
0.00004 99795.6307 -39.4065 -0.1144E-07 -0.2871E-l1 0.3385E-09 -0.1331E-l1 0.9815E-16 -0.1770E-17 -0.7191E-21
0.000041 99795.6307 -39.4065 4.959236 -0.000021 2.232410 -0.000706 0.1364E-04
0.00005 99728.0548 -39.4069 -0.1139E-07 -0.2843E-l1 0.3382E-08 -0.1329E-ll 0.9828E-16 -0.1774E-17 -0.7411E-21
0;000055 99728.0548 -39.4069 4.945848 -0.000028 2.231199 -0.000706 0.1364E-04
0.00008 99582.7292 ~39.4079 -0.1128E-07 -0.2793E-l1 0.3379E-08 -0.1325E-ll 0.9957E-16 ~0.1792E-17 -0.7455E-21
0.000084 99592.7292 -39.4079 4.917210 -0.000043 2.228594 -0.000707 0.1364E-04
0.00011 99439.2425 -39.4094 -0.1117E-07 -0.2724E-11 0.3373E-08 -0.1322E-11 0.9885E-16 -0.17A9E-17 -0.7498f.-21
0.000114 994J8.2425 -39.4094 4.HRH94'i -o.on:,n~" ?,):.o", -() 00(1'(0' () ·)a,~ .. _ .. ·,
PARTICLE NO 7 0.0702027584 12.5000000000 27.8136610985 158.1201105699 25.00000000

PARTICLE NO 8 0.0702073011 12.5000000000 -59.7520649433 -339.0430918210 25.00000000

PARTICLE NO 9 0.0702074309 12.5000000000 -30.4767370224 -174.6314118500 25.00000000

PARTICLE NO 10 0.0702014229 12.5000000000 -39.4060909748 -225.3507913551 25.00000000

.............................................. , ..•.....•. ,., ................... ,......................... .


X COORDINATE OF HIGHEST PARTICLE = 0.1~99875000E+00
RADIUS OF THIS PARTICLE = 0.1250000000E-04

AIR VELX" 0.3051863 710E-O 3


AIR VELl''' -0.56l4038404E-02
DRAGX = -0.1876140255E-08 Forces at last calculated position in trajectory
ORAGY = 0.4449313203E-10
IMAGE = 0.6559142565E-09
FIELDX = 0.1695927004E-09
FIELDY = -0.1911099020E-l1
FTOTX = 0.8247810659E-09
fTOTl' = -0.46879756~OE-n9
Figure 03(e). Output Parameters for Error Analysis and for Checking Efficiency of Routine

COUT NO. 1 0.359499E-07 CIN NO. 1 0.500000EtOl CON NO. 1 0.100000[+02 COMM NO. 1 O.OOOOOO[tOO
COUT NO. 2 0.261995E-02 CIN NO. 2 O.OOOOOO[tOO CON NO. 2 0.100000[t03 COMM NO. 2 O.OOO·OOO[tOO

'IELD2.5V.LOG;1 29-JUN-1992 11149:04.10 Paqe 10

COUT NO. 3. O.OOOOOOE+OO CIN NO. 3 O.OOOOOO[tOO CON NO. 3 0.120000[t01 COMM NO. 3 O.OOOOOOEtOO
COUT NO. 4 0.702065[-01 ~CIN NO. 4 O.OOOOOOEtOO CON NO. 4 0.130000[+01 COHM NO. 4 O.iOOOOO[tOl
.COUT NO. 5 0.702049[-01 ~NO. 5 0.359499[-07 CON NO. S 0.140000[tOl
:COUT NO. 6 0.500000Et01 CIN-n~0.lSI188E-05
the last integration point (time)
,COUT NO. 7 0.429217EtOO CIN NO. 7 ~Q~~_E+~O~l~~~__
: COUT NO. 9 o.196000[t03 ________ - last but one integration point (COUT +-5) = stepsize)
; COUT NO. 9 0.900000EtOl____ ~
! COUT NO.1 0
O. OOOOOOEtOO __________ - - - the number of successful integration steps used for the last particle
. COUT NO.ll 0.277556E-16 ________
'COUTNO.12 0.293974E-39 the number of unsuccessful integration steps
COUT NO.13 0.1941162E-17
COUT NO.14 0.500000Et01
COUT NO.15 0.400000E+Ol
COUT NO.16 0.320000Et02
IrAIL~ O'~-~-- ________
fORTRAN STOP Failure condition
aUASIMODO jOb terminated at 29-JUN-1982 11149103.91

Accounting information.
Buffered 110 count 1 66 Peak working set s1zel 172
Direct 110 countl 145 Peak virtual sizel 286
Page faults: 1536 Mounted volumes: o
[lapsed CPU time: o 00100148.37 [lapsed timel o 00102127.09
Appendix E

Computer Plots of Particle Trajectories

. -7
For each plot (i) the tolerance bound employed was 10

(H) the start point was O.lm from the substrate

(iii) Gun to target distance was 0.2m


(iv) The random number variable was 0

(v) The number of particles studies was 300 (El-E4l)

For Figures E17 to E62 the start width was 0.0002 m


Figure El. Particle Trajectories for Monosized Powder - Particle diameter = 10~m, Uniform Velocity = 1 m/s,
Start Width - 0.00043 m


Figur·e E2.. Particle Trajectories for Monosized Powder - Particle diame ter ~ 10l'm, Uniform Velocity = 3 m/ s,
Start Width ~ 0.00023 m
Figure E3,· Particle Tra'ectories for Monosized Powder - Particle diameter = 10~m, Uniform Velocit = 7 m/s,
Start Width = 0,00014 m


Particle diameter = 10~m, Uniform Velocity = 10 m/s,
Figure E4. Particle Trajectories for Monosized Powder
Start Width = 0.00011 m
Figure E5.· Particle Trajectories for Monosized Powder - Particle diameter = l7vm, Uniform Velocity = 1 m/s,
Start Width - 0.00058 m
Figure E6. Particle Trajectories for Monosized Powder - Particle diameter l7~m, Uniform Velocity 3 m/s,
Start Width - 0.00035 m
Figure E7. Particle Trajectories for Monosized Powder - Particle diameter = l7~m, Uniform Velocity = 7 m/s,
Start Width = 0.00023 m
Figure E8. Particle Trajectories for Monosized Powder - Particle diameter = l7~m, Uniform Velocity
10 m/s,
Start Width - 0.00019 m
Figure E9 •. Particle Trajectories for Monosized Powder - Particle diameter = 25~m, Uniform Velocity = 1 m/s,
Start Width - 0.00068 m
Figure ElO; Particle TraOectories for Monosized Powder - Particle diameter = 25v m, Uniform Velocit = 3 m/s,
Start Width - 0000046 m
Figure Ell. Particle Tra'ectories for Monosized Powder - Particle diameter = 25~m, Uniform Velocit = 7 m/s,
Start Width - 0.00046 m
10 m/s,
Figur"e E12. Particle Tra"ectories for Monosized powder - Particle diameter Q 2S)Jm, Uniform Velocity
Start Width Q 0.00032 m
Figure E13.' Particle Trajectories for Monosized Powder - Particle diameter = 50~m, Uniform Velocity = I m/s,
Start Width - 0.00089 m
Figure EI4:. Particle Traj ectories for Monosized Powder - Particle diameter = SO\Jm, Uniform Velocity 3 m/ s,
Start Width = 0.00087 m
Figure EIS; Particle Trajectories for Monosized Powder - Particle diameter = SO~m, Uniform Velocity = 7 m/s,
Start Width - 0.00110 m
Figure E16. Particle Tra·ectories for Monosized Powder - Particle diameter = 50~m Uniform Velocit = 10 m/s,
Start Width = 0.00130 m
Figure E17. Trajectories of Particles for Monosized Powder - Particle diameter = lO~m, Uniform Velocity = 1 m/s
3 m/s
Figure E18. Tra'ectories of Particles for Monosized Powder - Particle diameter = lO~m. Uniform Velocity
Figure E19. Trajectories of Particles for Monosized Powder - Particle diameter = IOpm, Uniform Velocity = 5 m/s
Figure E20. Trajectories of Particles for Monosized Powder - Particle diameter lO~m, Uniform Velocity 7 m/s
Figure E21. Trajectories of Particles for Monosized Powder - Particle diameter = 10~rn. Uniform Ve10cit = 10 rn/s

Figure E22. Trajectories of Particles for Monosized Powder - Particle diameter = l7~m, Uniform Velocity = 1 m/s
Figure E23. Trajectories of Particles for Monosized Powder - Particle diameter Q 17~m. Uniform Velocity Q 3 m/s
Figure E24. Trajectories of Particles for Monosized Powder - Particle diameter = 17llm, Uniform Velocity = 5 m/s
Figure E25, Tra'ectories of Particles for Monosized Powder - Particle diameter
l7~m. Uniform Velocit 7 m/s
Figure E26. Trajectories of Particles for Monosized Powder - Particle diameter = 17\lm, Uniform Velocity = 10 m/s
Figure E27. Trajectories of Particles for Monosized Powder - Particle diameter = 25~m, Uniform Velocity = 1 m/s
Figure E28. Trajectories of Particles for Monosized Powder - Particle diameter = 25~m, Uniform Velocity =3 mls
Figure E29. Trajectories of Particles for Monosized Powder - Particle diameter = 25~m, Uniform Velocity = 5 mts
7 m/s
Figure E30. Trajectories of Particles for Monosized Powder - Particle diameter = 25~m, Uniform Velocit

i
Figure E31. Trajectories of Particles for Monosized Powder - Particle diameter = 25~m, Uniform Velocity = 10 m/s
Figure E32. Trajectories of Particles for Monosized Powder - Particle diameter = SO~m, Uniform Velocity = 1 m/s
Figure E33. Trajectories of Particles for Monosized Powder - Particle diameter = SOpm, Uniform Velocity =3 m/s
Figure E34. Trajectories of Particles for Monosized Powder - Particle diameter = 50~m! Uniform Velocity = 5 m/s
Figure E35. Trajectories of Particles for Monosized Powder - Particle diameter 50~m, Uniform Velocity = 7 m/s
Figure E36. Trajectories of Particles for Monosized Powder - Particle diame·ter = SOjJm, Uniform Velocity = 10 m/ s
Figure E37. Particle Trajectories for a Size Distributed Powder - Uniform Velocity 1 rn/s
Figure E38. Particle Trajectories for a Size Distributed Powder - Uniform Velocity = 3 m/s
Figure E39. Particle Trajectories for a Size Distributed·Powder - Uniform Velocity = 5 rn/s
Figure E40~ Particle Trajectories for a Size Distributed Powder - Uniform Velocity = 7 m/s
Figure E41. Particle Trajectories for a Size Distributed Powder - Uniform Velocity = 10 m/s
Figure E42. Trajectories of Particles Approaching a Charged Layer 50~m thick - Particle diameter = lO~m.
Uniform Velocity = 1 mls
Figure E43. Trajectories of Particles Approaching a Charged Layer lOO)lm thick - Particle diameter = IO)lm,
Uniform Velocity - I m/s

----~--~-------.--~-----.----------- - - - - - - - - - - - - - --- .-----


Figure E44. Tra'ectories of Particles e lO\lm,
Uniform Velocity - 1 m s

CHA GEO L
~ -7 -I -, -3 - '"' -I '"' 7 B ," , "
Figure E45. Trajectories of Particles A roaching a Char ed Layer 50~m thick - Particle diameter = lO~m.
Un1form Velocity - 5 m s

- --- ---/I- - \
\

! / I
Figure E46. Trajectories of Particles Approaching a Charged Layer lOOpm thick - Particle diameter = lOpm,
• Uniform Velocity - 5 m/s

II I
Figure E47.· Trajectories of Particles Approaching a Charged Layer 50~m thick - Particle diameter = 10~m.
Uniform Velocity - 10 m/s
Figure E48. Trajectories of Particles Approaching a Charged Layer lOO~m thick - Particle diameter = lO~m.
Uniform Velocity = 10 m/s

I
Figure E49; TraOectories of Particles A roachin ~ 10\lm,
Uniform Velocity - 10 m s

. , •
Figure ESO. Tra'ectories of Particles A 2SIlID,
Uniform Velocity - 1 ID S

' / / / / / / / / / / /./ / / /LLLLL1/ / / / / / / / / /.1 y ~Z''7 'c!"',?, Y,,'77/ / / / / / / / / / / /,.../ /,/ // //7 / /7 / / /, -/
"80" - ,eo -,".. ,~-,,~ -,_ -,~, -" .. -,_ -~. -_ -,... -6", -~ -''''' -~., -'" -,.. '!;U8S'f'RAT~.' ,., ,.. 6.' >e, ••, .' ....- "-,,.. ,-
Figure ESt. Trajectories of Particles Approaching a Charged Layer IOO~m thick - Particle diameter = 2S~m,
Uniform Velocity - I m/s

-1 0 -I 6
Figure E52. Trajectories of Particles Approaching a Charged Layer 50~m thick - Particle diameter = 25~m.
Uniform Velocity - 5 m/s

,~ .L k=
.// / / / / /// / / / / / / / / / /// / / / / / / // / / fI ')"/ i'/,-./ 'L'z";Y / / / / / / / / / / / / / / / / / / / L / / / / / / / / /
Figure E53. Trajectories of Particles Approaching a Charged Layer 50~m thick - Particle diameter = 25~m.
Uniform Velocity - 10 m/s

( (

/ . / / / / / / / / / / / / / / / / / / / / / / / / / / 7 / / 7 'y ~V~~Ji)~/////// /////7777777//////////


Figure E54" Tra"ectories of Particles = 2S]JID,
Uniform Velocity - 10 ID S
Figure E55. Trajectories of Particles Approaching a Charged Layer 50vm thick - Particle diameter = 50vm,
Uniform Velocity - 1 m/s

I .
Figure ES6" Tra"ectories of Particles = SO)JID,
Uniform Velocity = lIDs

-, , - 1

Figure E57". Tra' ectories Qf Particles roaching a Charged La er 250~m thick - Particle diameter = 50~m.
Uniform Velocity - I m s

/
Figure E58" Tra"ectories of Particles = 50jlID,
Uniform Velocity - 5 ID S

-[; a -..
Figure ES90 TraOectories of Particles SO\lm,
Uniform Velocity = 5 m s

!
Figure E60; Trajectories of Particles Approaching a Charged Layer 50~m thick - Particle diameter = 50~m,
Uniform Velocity = 10 m/s

-1 0 -
• ., . •
Figure E61". Trajectories of Particles Approaching a Charged Layer 100llm thick - Particle diameter 501lm.
Uniform Velocity - 10 m/s

(
Figure E62. Trajectories of Particles ~Particle diameter = 50~m,
Uniform Velocity = 10 m s

-, -, • , • J •

You might also like