‘ . ~\ . th 04$” .
‘ 4 . . .VA}. onuvbfifi.
:.:.:. . , . .

.1:
.7...
.

3.523
9.. $3..) u
:0. {Wyn-“Why'd
1.. 9.0... l&v.

.u... h: .
.

0 ‘ J-r‘"
filamwwmf
3- I.

Jr: 5r

n"

‘ .
.mmmm

a
.

L
r.

 

“4 R.
mg
rwwwmu

 

:1. .i
z 6 flit!!! ,3
fawn-ChautfldwUY
. flung?

u 133 .

.43

n
.5 .. _
,1.) .103“.
$1.... . v.9. . ,
.1...
..

.1 5:11.."

g 1?: . w: [T
X“ «3‘..7..i.un.v!

.. 513.15%.

 

 

\D

.5 7’55" I; it

This is to certify that the
dissertation entitled

ULTRAVIOLET RADIATION SURFACE TREATMENT OF
POLYMERS

presented by

ALEKH S. BHURKE

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemical Enflreerm

 

 

Rama/La, // QM‘

Major Professor’s Signature 0

November 26, 2003

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

 

'— v- *- mj__ -fl kc 444A"

 

.fi v Tfi - v

 

 

 

LIBRARY
MIchIgan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c/CIRC/DateDuepGS-p. 15

 

L'LTRAV

 

ULTRAVIOLET RADIATION SURFACE TREATMENT OF POLYMERS
By

Alekh S. Bhurke

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Chemical Engineering and Materials Science

2003

[ITRAVII

L'Itrax‘mlct
pzlymer surt‘it‘cx
Merton of mu
function“ group
farce required fur g

I'\' IICJIIIII
Lid polydienc rui
chemistry ldctcm‘.

irtnd to depend pr

 

rite: of tempera:
3333M [0 that“.
died to describe :~
Meters are rt-
"Ci” be extendctj
Wilmer sun‘m

MIIIIOKL [\2, \

 

ABSTRACT

ULTRAVIOLET RADIATION SURFACE TREATMENT OF POLYMERS
By

Alekh S. Bhurke

Ultraviolet (UV) radiation in the 185 nm — 280 nm (UVC) band can oxidize
polymer surfaces by a combined effect of UV activation of the polymer and the
production of ozone and atomic oxygen from air. UV photo-oxidation creates polar
functional groups that increase surface energy and provide the thermodynamic driving
force required for good wettability and adhesive performance of the polymer.

UV treatment of polycarbonate (PC), polyethylene terephthalate (PET), epoxy,
and polydiene rubber was studied in detail. The changes in wettability and surface
chemistry (determined by XPS) were related to the UV treatment process variables and
found to depend primarily on the net irradiation. In the case Of low Tg polymers, a strong
effect of temperature on the surface properties was also Observed. A process model is
proposed to characterize the UV modification of these polymers. Sensitivity functions are
used to describe the evolution of surface properties as a function of irradiation, and model
parameters are related to physical and chemical properties of the polymers. The model
can be extended to other polymers and ultimately used to predict the properties of

polymer surfaces after irradiation by xenon arc UV lamps.

Keywords: UV, surface modification, surface energy, XPS, polycarbonate.

Copyfightby
ALEKH S. BHURKE

2003

DEDICATION

To my parents who have always

encouraged and supported me.

iv

I would
these tutelage I
Baiting. Nara}
eprtcéate the v.1
(3:13.“th 1';
LNIIJIII \tngc

Craughout the LIL

 

 

 

ACKNOWLEDGEMENTS

I would like to express my gratitude towards Prof. Lawrence T. Drzal under
whose tutelage I have learnt so much over the years. I would also like to thank Professors
Bruening, Narayan, and Jayaraman for their comments and input on this work. I
appreciate the various students, staff, and faculty at the Composite Materials & Structures
Center, whose friendship I have enjoyed during my stay at Michigan State University.
Lastly, my sincere thanks to Dr. Per Askeland who has been a friend, guide, and mentor

throughout the duration of this work.

List of Tablc~

List of Figurc~

Ctepter One —
1.1 Surface
1.2 Own tr
1.3 Compar‘
1.4 Literatur
1.5 Guaix

Retercrzccs

Crapter III 0 — l
3.1 II Lam
3.2 .Ilcasurc
2.3 Meaxum

14 Surface 1

2'5 Adhc‘It'If

2’6 X“) PI}.

2.7 Nam—“Indr-

2'8 summurx

Q

Chm“ Three L

,
3'1 DI‘I‘Cript

 

i” p
~-~ [WC\\}

 

3.2.1 I
3.23 (
3.2.3 l

 

TABLE OF CONTENTS

List of Tables x
List of Figures xiii
Chapter One — Introduction 1
1.1 Surface Treatments 1

1.2 Overview of the UV Surface Treatment 4

1.3 Comparison With Other Surface Treatment Techniques 5

1.4 Literature Review 9

1.5 Goals 31
References 33
Chapter Two — Experimental Methods 37
2.1 UV Lamps and Filters 37

2.2 Measurement of UV Radiation 42

2.3 Measurement Of Ozone Concentration 46

2.4 Surface Energy Measurement 55

2.5 Adhesion Measurements 60

2.6 X—ray Photoelectron Spectroscopy (XPS) Surface Analysis 62

2.7 Nanoindentation Tests 67

2.8 Summary 70
References 71
Chapter Three — Surface Modification of Polycarbonate 72
3.1 Description of Polycarbonate 72

3.2 Process Parameters 73
3.2.1 Flow Rate 73

3.2.2 Ozone Concentration 76

3.2.3 Irradiance 78

vi

3.4.1
5 4 1 I
-‘- '-
35 EPR Sp
3.6 XJDOIDLI
37At1he~tfl
:11 I
27.2 S

‘

3.8 Summarj

et’erences

Chapter Four — S
4.1 Surface C
4.2 Surface C
4.3 [I Treat
4.4 Oxygen 1
4.5 SiIVer Nit
4-6 W Treat
4.7 Whiter} f
4‘3 Analyxi,
4'9 Summar;
Rclfirenees

Ch‘Pli‘r Fire - r

 

5.2 Crosxlir.
5,3 Polytett‘. I
5.4P01_\ttet:
5.5 POIICIIL‘I

3'6 IOanC‘r

 

 

3.3 Wettability and Surface Energy 82

3.4 Molecular Spectroscopy 85
3.4.1 ATR-FTIR Spectroscopy 85
3.4.2 UV Spectroscopy 88

3.5 EPR Spectroscopy 91

3.6 Nanoindentation 96

3.7 Adhesion 98
3.7.1 Tensile Stub-Pull Tests 98
3.7.2 Stub-Shear Tests 101

3.8 Summary 103

References 105

Chapter Four — Surface Anaysis of UV Treated Polycarbonate 107

4.1 Surface Chemistry of Polycarbonate 107

4.2 Surface Chemistry of UV/Ozone Treated Polycarbonate. 111

4.3 UV Treatment in Vacuum 114

4.4 Oxygen Uptake in UV/Vacuum Treated PC 120

4.5 Silver Nitrate Derivatization 123

4.6 UV Treatment in Inert Atmosphere 127

4.7 Washed Films 130

4.8 Analysis of Adhesion Test Fracture Surfaces 134

4.9 Summary 140

References 142

Chapter Five - UV Treatment of Polymers 143

5.1 Epoxy 143

5.2 Crosslinked Tire Rubber Compound 149

5.3 Poly(ethylene terephthalate) 152

5.4 Poly(tetrafluoro ethylene) 154

5.5 Polyolefins 156

5.6 Ionomers (Bexloy®W) 160

vii

 

5.7 Summd-C‘

Reference“

(hi-pier 51x — D1

6.1 Process I

6.1.1 5;
6.1.2 In
61.3 E1
6.1.4 0.
6.1.5 It
6.1.6 Pt
6.1.7 11:

6.3 Procexs .\

 

62.1 L"
6.2.:

62.3 12,
6.2.4 5.

1 . .
6-3 Appltcat;
6-4 Summarx

ReferenceS

Ch‘IPIEI Sewn _
7'1 Abstraet

 

 

7.3.1 (
4 II
/ 3.2 1
7.3.3 1*
7.3.4 (I
1.3.5 I
7-3.6 a

 

5.7 Summary

References

Chapter Six - Discussions and Process Model

6.1 Process Variables
6.1.1 Spectral Output
6.1.2 Irradiance
6.1.3 Exposure Time
6.1.4 Ozone Concentration
6. 1.5 Temperature
6.1.6 Pulse Frequency
6.1.7 Humidity

6.2 Process Model
6.2.1 UVO Surface Treatment Model for Xenon Flashlamps
6.2.2 Determination of Sensitivity Functions
6.2.3 Relation to Physical Properties
6.2.4 Extention to Bulk Polymer

6.3 Application to Unknown Polymers

6.4 Summary

References

C hapter Seven — Market and Economic Analysis
7.1 Abstract
7.2 Introduction
7.3 Surface Treatment Processes
7.3.1 Chemical/Solvent Wash
7.3.2 Flame Treatment
7.3.3 Plasma Treatment
7.3.4 Corona Treatment
7.3.5 Ultraviolet Treatment

7.3.6 Surface Treatment Process Comparison Matrix

viii

162
164

165
165
165
167
169
169
171
172
174
174
177
178
182
191
193
194
198

1 99
199
200
201
201
202
202
203
204
205

7.4 Surface 1
7.5. Potential

_1 111

’J‘

7.

7.5.2 .\1
7.6 Rt‘t'tllllllit
7.7 Summar}

Reference»

Cfiipter Eight - 1
8.1 CIInC1U\11

6.2 Reeummt

.trxndix A — SL
Appendix B - Se
Appendix C ~ C .

 

 

 

7.4 Surface Treatment Cost Comparisons
7.5 Potential Markets
7.5.1 Industry Comparison Matrix
7.5.2 Marketing Considerations
7.6 Recommendations and Implementation Plan
7.7 Summary

References

Chapter Eight — Conclusions and recommendations
8.1 Conclusions

8.2 Recommendations for Future Work
Appendix A — Surface Energy Calculations

Appendix B - Sensitivy Curve Fits
Appendix C — Calculation of Polymer Volumes

ix

207
214
216
217
218
219
221

222
222
229

 

This 1.2
Table 2.1
Table 3.3
Teble 2.3

Tablf 3.1

Table 4.1
Table 4.2

Teblfi 43

Table 4.4

Table 45

Table 4.6

Table 4,7

 

0.
Al
De
Rd
Su:

AT

\lrL

Cir

XPi

\lltl‘

Stir?

VdL‘i

Sun
Ase

Cm]

 

Table 1.1
Table 1.2
Table 2.1
Table 2.2
Table 2.3

Table 3.1

Table 4.1
Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 4.6

Table 4.7

Table 5.1

Table 5.2

Table 5.3

LIST OF TABLES

Comparison of various surface treatment techniques.
Absorption bands for some common organic bonds.
Description of UV lamps.

Radiometric units and descriptions.

Surface energies of liquids used for contact angle measurement.

ATR-FTIR peak ratios of carbonate, hydroxyl, and aromatic
structures in PC after UVO treatment.

O/C ratios in untreated polycarbonate.
Carbon distribution in untreated polycarbonate.

XPS surface chemical composition of UV treated polycarbonate
showing the O/C ratio and C15 curve fit analysis.

Surface chemical composition of polycarbonate UV treated in
vacuum, air and ozone.

Surface chemical composition of UVV treated PC aged in air.
Aged samples oxidized in ozone for 5 minutes prior to analysis.

Comparison of UVV and UVO treatments after water rinse.

Atomic composition and analysis of adhesion test fracture
surfaces (adhesive side) for UVO treated polycarbonate - epoxy
bonds.

Surface chemical composition of UVO treated epoxy.

Temperature of crosslinked tire rubber compound surface after
UV exposure in RC 500 lamp.

Effect of temperature on UVO treated crosslinked tire rubber
compound.

42

59

86

110

110

113

117

119

133

137

145

150

151

Table.“ 5””
511”;

Table 5-5 \th
P01}
mall

Tittle 5‘0 Sufi

1:31:61 9“

133156.] CD”
Tune

Tab-1663 Railt

mill.
(0111

1111:64 0‘"

1.151: 7.1 Star"
Table 7.: CM
llble 7.3 CW1
Table 7.4 Cm
Table 7.5 Cm

Table 7.6 Ci
]\

 

 

Table77 Su:

Iii:
TWTB C
TableSl Pr.
It.

felt C] Re

C11.”

 

Table 5.4

Table 5.5

Table 5.6
Table 6.1

Table 6.2

Table 6.3

Table 6.4

Table 7.1
Table 7.2
Table 7.3
Table 7.4
Table 7.5
Table 7.6

Table 7.7

Table 7.8
Table 8.1

Table C.1

Surface chemistry and atomic composition of UVO treated PET
surface.

Water contact angles of Ultra-High Molecular Weight
Polyethylene after UVO treatments with and without thermal
management..

Surface chemistry of PP and TPO after UVO treatment.
Cooling time and energy per pulse for 2.6 mW pulsed UV output.

Correlation constants for the work of adhesion sensitivity
functions for PC, PET, Epoxy, Rubber, and PTFE.

Ratio of amorphous, glassy, crystalline, and semi-crystalline
molar volumes of polymers to Van der Waals volumes by group
contribution.

Correlation between empirical parameters and physical properties
of UVO treated PET, Epoxy, PC, and crosslinked tire rubber
compound.

Surface treatment process comparison matrix.

Cost analysis of plasma treatment process.

Cost analysis of corona treatment process.

Cost analysis of UV laboratory scale treatment process.
Cost analysis of UV pilot scale treatment process.

Cost analysis of UV production scale treatment process.

Summary of cost analysis based on a production facility treating
200 bumpers (5 ft. x 1 ft.) per hour.

Comparison between the automobile and biomedical industry.
Process model parameters and relation to physical properties.

Reactive volume fraction of polycarbonate from group
contribution theory.

xi

153

158

158

173

180

185

190

206

209

210

211

212

213

214

216

230

238

 

Table C:

Table C .3

1.11116 C4

 

Table C.2

Table C.3

Table C.4

Reactive volume fraction of polybutadiene from group
contribution theory. 239

Reactive volume fraction of PET from group contribution theory. 240

Reactive volume fraction of epoxy (Epon 828 + Jeffamine T403)
from group contribution theory. 241

xii

 

Fgum 11
Figure 1.:
figure 1.3
Figure 1.4

figure 1.)

Figure 1.6

figure 1.7

Figure 1.8
Figure 1.9

Figure 1.10
E1316 1.11
Figure 1.13

Figme 1,13

Flare 1,14

BRETJS

 

 

 

Figure 1.1
Figure 1.2
Figure 1.3
Figure 1.4

Figure 1.5

Figure 1.6

Figure 1.7

Figure 1.8

Figure 1.9

Figure 1.10
Figure 1.11
Figure 1.12

Figure 1.13

Figure 1.14
Figure 1.15

Figure 1.16

LIST OF FIGURES

The adhesive bond interphase.

Schematic of the UV treatment process

Typical outputs of pulsed xenon UV lamps.

Schematic representation of pulsed UV lamp operation at 120112.

The Franck-Condon Principle showing excitation of an electron
from the ground state of the molecule (S0) to a higher
vibrational state of the excited molecule (S1) followed by
vibrational relaxation of the excited state.

Electronic transition in a molecule from the ground state (S0) to a
dissociative state (S 1).

Absorption coefficient as a function of wavelength for
poly(ethylene terephthalate) (PET), poly(tetrafluoro
ethylene)(PT FE), polypropylene (PP) and polyethylene (PE).

Intensity ratio of Cls/Ols bands for LDPE (x) and HDPE (o)
as a function of ozonation time.

Intensity ratio of Cls/Ols bands for LDPE (x) and HDPE (o)
as a function of photo-oxidation time.

Mass transport in gas-solid surface reaction.
VUV photo-oxidation reaction scheme for PE and PP.
Photo-oxidation reaction scheme for PE.

Photo-Fries rearrangement and photolysis reaction scheme for
polycarbonate.

Photo-oxidation scheme for polyethylene terephthalate.
Norrish Type I mechanism for (it-cleavage of carbonyl groups.

Norrish Type H photo-elimination mechanism in polymers
containing carbonyl groups.

xiii

12

14

14

l6

17

19

20

21

23

25

26

27

28

29

29

Figure 1.17

Figure 2.1

Figure 2.2
Figure 2.3
Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 2.9

Figure 2.10

Figure 2.11
1:igure 2.12
Figure 2.13
Fi.gure 2.14

Pi gure 2.15

l2.1gure 2.16

Criegee mechanism for ozone attack on unsaturated bonds.

Xenon pulsed UV lamps (from left) RC 747, RC 500, and RC
740.

Aluminum UV treatment chamber with Suprasil® cover.
Schematic of the experimental setup for UV surface treatment.
UV transmittance of Suprasil®, Quartz, and Pyrex® filters

Lambert’s cosine law showing decrease in irradiance with
increasing angle of incidence.

Irradiance measurements for RC500 lamp with ozone and oxygen
flow.

Irradiance measurements for RC747 lamp with ozone and oxygen
flow.

Absorption spectrum of ozone.

Perkin Elmer Lambda 900 UV-VIS-NIR Spectrometer with
quartz flow cell and sampling nozzle for continuous ozone
measurement.

Ozone concentration in the UV treatment chamber at different
voltages and gas flow rates.

Contact angle (0) at the solid-liquid interface and surface tension.
Kruss Drop Shape Analysis System.

Schematic representation of the tensile stub—pull (PATTI) test
Schematic representation of the stub-shear adhesion test

XPS survey scan showing elemental carbon and oxygen peaks in
polycarbonate.

Deconvolution of the C15 peak envelope in polycarbonate
showing multiple peaks corresponding to carbon atoms with
increasing number of bonds with oxygen.

xiv

30

39

40

40

41

44

47

48

51

52

54

56

59

61

61

64

65

 

EamllT

Hgm3i

Figure 3.3

I),
’aJ

F109?
.-uA
G

FamiJ

Figure 3.5

Figure 3.6

Figure 3.7

Fare 3.8

Figure 3.9

 

Figure 2.17

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Typical nanoindentation load displacement curve showing
loading and unloading behavior.

Chemical structure of polycarbonate.

Contact angles of UV treated PC for ozone flow rates of 10, 20,
and 30 scfl1 indicating no mass transfer limitations.

Contact angles of PC after 30 sec UVO treatment as a function of
ozone concentration. Horizontal line indicates the lowest angles
measured at the lower limit of measured ozone concentrations.

Equilibrium contact angles of deionized water on UV treated
polycarbonate as a function of time at 1 to 5 inches from the UV
source.

Work of adhesion of UV treated polycarbonate with deionized
water as a function of irradiation for data obtained at various
combinations of irradiant powers and exposure times for multiple
lamps.

Polar-Dispersive surface energy of UV treated polycarbonate for
various combinations of irradiance and exposure time.

Acid-Base surface energy of UV treated polycarbonate for
various combinations of irradiance and exposure time.

ATR-FTIR spectra of PC after UV treatment at 2.662 mJ/m2
irradiance and 700 ppm ozone

UV spectra of UVO, UVA, and untreated PC showing UV
induced chain scission and the formation of phenolic groups
(260-278 nm) and phenyl salicylates (245 nm and 315 nm) in the
polymer because of photofries rearrangement.

UV spectra of PC spincoated on quartz after UV Treatment at
2.662 mJ/m2 irradiance and 700 ppm ozone.

EPR spectra of PC exposed to ozone (top), 120 sec UVA
treatment (middle), and 120 sec UVO treatment (bottom).

EPR spectra of epoxy (top), 120 sec UVO treated PC (middle),
and 120 sec UVO treated PC cured in epoxy (bottom).

XV

69

72

75

77

80

81

83

84

87

89

90

94

95

 

figure 3.1.1

Fame3l?

hmmilfi
hamilT
figure 3.18
ham4l
meTl

hgm43

FgmeTA 1
me45 [

hamTfi p

Harem 3

13.11618 (

 

531115 4.9 \

fiflm410

 

Figure 3.13

Figure 3.14
Figure 3.15

Figure 3.16

Figure 3.17

Figure 3.18
Figure 4.1
Figure 4.2
Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8
Figure 4.9

Figure 4.10

Modulus vs depth profiles of UVO treated polycarbonate
measured by nanoindentation tests. Higher modulus in the
surface layers is observed with increasing UVO exposure.
Adhesive strength of UVO treated PC with epoxy adhesive.
Adhesive strength of UVA treated PC with epoxy adhesive.

Adhesive strength of UVO treated PC with polyurethane
adhesive.

Adhesive strength of ozonated and UVO treated polycarbonate
before and after ethanol wash to remove low molecular weight
material.

Slopes of load displacement curves for stub shear adhesion tests.

Cls peak for polycarbonate (a) before and (b) after UV treatment.

O/C ratio of UV treated PC as a function of irradiation.
Schematic of the XPS pre-chamber for UV treatment

O/C ratio of untreated and UVV treated PC exposed to 700 ppm
ozone.

UV/Vacuum treated polycarbonate post oxidized with 700 ppm
ozone.

Reaction mechanism of silver nitrate with carboxylate groups on
UV treated polycarbonate surface. Silver, unlike ozone, is limited
to reactions on the top few monolayers of the surface.

Silver and carboxylate concentration in silver nitrate derivatized
polycarbonate with 60 sec UVV treatment and post-oxidation
with 700 ppm ozone.

O/C Ratio of UVO and UV/Nz + 120 sec 03 treated PC
Work of adhesion of UVO and UV/Nz + 120 sec 03 treated PC

O/C ratio for UVO treated PC before and after rinsing with water
followed by ethanol.

xvi

97

99

99

100

100

102

109

112

118

119

122

125

126

128

129

 

hwfiill

Figure 4.1:

Time «1.13

Figure 5.1

Email

J!

Figure . .3

Figure 5.4

Fzzure
U

Jl
' 1-

Figure 5.6

Flgure 5.7

Eif‘i‘re 5.10

 

51:11:51 1

 

Figure 4.11

Figure 4.12

Figure 4.13

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Figure 5.11

Work of adhesion for UVO treated PC after rinsing with water
and ethanol.

Cls spectra of epoxy adhesive and fracture surfaces (adhesive
side) for 0 sec UVO, 30 sec UVO, 120 sec UVO, and 120 sec
UVO + water rinsed PC.

Cls spectral profile of adhesive fracture surfaces after subtraction
of epoxy Cls signal for 0 sec UVO, 30 sec UVO, 120 sec UVO.
and 120 sec UVO + water rinsed PC.

Chemical structures of Epon 828 (top) and Jeffamine T403
(bottom).

Work of adhesion of UVO treated epoxy with deionized water as
a function of irradiation. Data obtained at various combinations
of irradiance and exposure time.

Acid-Base surface energy of UVO treated epoxy at various
combinations of irradiance and exposure time.

Polar-Dispersive surface energy of UVO treated epoxy at various
combinations of irradiance and exposure time.

Work of adhesion of UVO treated crosslinked tire rubber
compound with deionized water as a function of irradiation. Data
obtained at various combinations of irradiance and exposure
time.

Chemical structure of PET.

Work of adhesion of UVO treated PET with deionized water as a
function of irradiation. Data obtained at various combinations of
irradiance and exposure time.

Chemical structure of PTFE.

Work of adhesion of UVO treated PTFE with deionized water as
a function of irradiation.

Stub pull adhesive bond strength of UVO treated UHMW PE.

Adhesive strength, contact angles and surface temperature of
UVO treated TPO showing no significant change in contact
angles despite the increase in adhesive bond strength.

xvii

132

139

144

146

147

148

150

152

153

154

155

159

159

Figure 5.12

hareél

Bambi

Figure 6.3

hamBl
EamBl
hamB}

hamBA

 

 

Figure 5.12

Figure 6.1

Figure 6.2

Figure 6.3

Figure B.1
Figure B.2
Figure B.3

Figure B.4

O/C ratio, contact angles and surface temperature of UVO treated

Bexloy®W showing surface reorientation effect above the Tg
(50°C).

Work of adhesion sensitivity curves for UVO treated
polycarbonate (PC), epoxy, polytetrafluoroethylene (PTFE),
crosslinked tire rubber compound, and polyethylene terephthalate
(PET) showing the rate of property change as a function of
irradiation.

Comparison of the Van der Waals fractional volume and
fractional area of susceptible sites in PC, PET, Crosslinked Tire
Rubber Compound, and Epoxy with the parameter a calculated
from curve fit to the work of adhesion data. Data for rapid
estimation of a by counting number of sites per mass is also
shown for comparison.

Sensitivity function for surface oxygen content in UVO treated
PC.

Curve fit for work of adhesion of PC.
Curve fit for work of adhesion of epoxy.
Curve fit for work of adhesion of PET.

Curve fit for work of adhesion of crosslinked rubber compound.

xviii

161

181

189

195

234

235

236

237

 

commerciall
including er
112111131068. “I
these materia
carbon based

and inks. The

Surfaee treatme

Thf modilieatit‘

frendly method

 

CHAPTER ONE

INTRODUCTION

“God made solids, but surfaces were the work of the Devil.”

— Wolfgang Pauli (1900- 1958)

Surfaces of solids have been the subject of many scientific investigations over the
years. In the last century, the growing use of polymers has focussed attention on the
surfaces of polymeric materials. Today, hundreds of different types of polymers are
commercially available and find use in almost all conceivable areas of applications
including engineering, medical, transportation, packaging, and commodity goods
industries. With such widespread use arises the need for painting and adhesive bonding of
these materials. Many of the commercially important polymers are low surface energy
carbon based materials with inadequate adhesion and wettability with adhesives, paints,
and inks. The modification of these properties by physical or chemical means is termed
surface treatment. Various methods of surface treatment are available commercially for
the modification of polymers and metals. This work investigates an environmentally

friendly method for the surface modification of polymers with ultraviolet (UV) radiation.

1.1 Surface Treatments
Adhesive bonding of metals, polymers, and polymer composites is an attractive
structural fabrication method. Adhesive bonding can create strong, stable joints with

superior mechanical properties and durability compared to mechanically fastened

 

Structures.
impomm 1'
pie-p15?“ L11
composites.
or limit adhe
painting. req:
arition of C1
Surilre
a the adhesix
tween two C
dimensional 1r
310 bllih Phdfii
Dial [1] “11C:
PTOPOSed to |
[Introstrueture
Filing phaset
lfllfifacial 1711611
We effect of 1
V4001“
OVCTL‘Ome 1116‘

smelt €18 abraxi.

 

0t -
:4111C SOiVer

 

structures. The application of protective coatings and paints to surfaces is also an
important manufacturing process in the durable goods industry. In such processes, the
pre-preparation of the surface is an important step. Processed polymer, polymer
composites, and metallic surfaces contain undesirable compounds or additives that reduce
or limit adhesion. Surface preparation of the adherend, whether for adhesive bonding or
painting, requires the removal of labile organic compounds and contamination, as well as
addition of chemical functionalities that can interact strongly with the adhesive or paint.
Surface treatments are designed to alter the interface between two materials such
as the adhesive and the adherend. The idea of a two dimensional, well-defined interface
between two different materials is necessarily an idealistic one. In most materials, a three
dimensional interphase with unique properties is formed and the boundary between the
two bulk phases is blurred. A model of an adhesive bond interphase was proposed by
Drzal [1] where the interphase between a viscoelastic adhesive and solid substrate is
proposed to be dependent on the surface chemistry, morphology, topography,
microstructure, local chemical composition as well as the bulk properties of the two
adhering phases. Figure 1.1 shows a schematic of such an interphase. The study of any
interfacial phenomenon must take into account the contribution from these factors as well
as the effect of the thermal, chemical and mechanical environments of the interphase.
Various mechanical and chemical surface treatments have been developed to
overcome the problem of weak adhesion in polymers. Mechanical surface treatments
such as abrasion are time consuming, labor intensive and can damage the substrate.
Organic solvents are often used for cleaning surfaces but present various environmental

problems and are being eliminated in order to reduce volatile organic compound (VOC)

emissions.

discharge h:
and used “1
operating CC
shapes. The:
environment;

he manuiaett

1T

f

 

 

 

 

 

 

1.2
0%an

 

 

 

emissions. Other surface treatment techniques such as flame, plasma, and corona
discharge have also been developed [2-10]. While these surface treatments are efficient
and used widely in industry, they suffer from drawbacks such as high cost, hazardous
operating conditions and by-products, and the inability to treat complex geometric
shapes. There is a growing need in industry for a fast, simple, efficient, and
environmentally friendly surface treatment process that can be easily incorporated into

the manufacturing environment [11].

 

  
 
 
  
 
 
 
 
 
 
  
 
 

BULK ADHESIVE

POLYMER of
DIFFERENT
PROPERTIES
ADSORBED
MATERIAL

OXIDE CHEMISTRY

 

.1 .

OXIDE
THERMAL. TOPOGRAPHY
OXIDE
CHEMICAL. MORPHOLOGY
MECHANICAL . BULK ADHEREND
ENVIRONMENTS

ADHESIVE BOND INTERPHASE

O LTD. MSU, 1992

 

 

Figure 1.1 The adhesive bond interphase [1].

1.2 Overview of the UV Surface Treatment

 

Sine
COmb‘il'lLiITOT
This pheno
oxygen erea
lite carbon C
ml U" rail
0‘3"an and t
cFeetii-eiy R.
Phenomenon
‘b 31:10.11 by hi
“FT-15111? to 1‘
Fiapldlt broi
an? Cipogcd 10
SUTT‘JCC gTOUP.
wettabilm an t1
TIT-C radial-w
vapor 1mm isl
recepigye mute;
can yield a SUr

 

ill-sight name
be 5de ”1011-.

 

Since the early 1970's it has been known that Ultraviolet (UV) light in
combination with atmospheric oxygen can clean organic contaminants from surfaces.
This phenomenon occurs when low wavelength UV light interacts with atmospheric
oxygen creating ozone, which oxidizes surface organic compounds to small molecules
like carbon dioxide and water. This reaction occurs in the presence of high-energy (< 185
nm) UV radiation [12, 13]. Ozone absorbs UV radiation at 254nm and dissociates into
oxygen and oxygen radicals creating a very aggressive oxidizing environment which can
effectively remove low molecular weight organic contaminants from the surface. Another
phenomenon that is responsible for cleaning the surface, especially in polymers, is that of
ablation by high energy UV radiation, which can etch and activate surfaces [14-18]. On
exposure to UV light of sufficiently high energy, organic bonds in the surface layer can
be rapidly broken depending on the absorbance of the substrate. When activated surfaces
are exposed to the atmosphere, oxidation takes place with the formation of highly polar
surface groups such as hydroxyl, carbonyl and carboxylic acids that can improve
wettability and adhesion [8,10,13,19-30]. UV light with wavelengths from 184-365nm
(UVC radiation) produced by commercially available xenon and low-pressure mercury
vapor lamps is ideal for the process of surface activation and oxidation. Exposure of a
receptive material to UVC radiation for short times in the presence of oxygen or ozone
can yield a surface with high surface energy, wettability and adhesive strength [31,32].
UV surface treatment also has the ability to treat 3-dimensional surfaces due to the line-
of-sight nature of the process. The by-products of such processes are largely expected to
be small molecules like water and oxides of carbon. The process does not utilize any

solvents and the ozone is dynamically created and dissociated in the treatment

 

em irimmcI
p113“? prOC

ehemlfilh n

1.3 COmP‘
Then
polymers. Ti
treatment. 0'
techniques W
Flame
or organic fue
orig-en radiea
The adhesix'e 1
an increase in
Weak boundar)
10.351'neh1 to r

01 tens of ineht

  
   
  
  
  

Frame treatmer
Treatment is 51
CombUKHOn 01
501136121113

ITCES _

ilFlCiill} Used :

environment. The process is considered environmentally friendly because the dry, gas
phase process does not create VOC emissions and does not use any hazardous wet

chemicals needing elaborate handling and disposal.

1.3 Comparison With Other Surface Treatment Techniques

There are several commercially available techniques for surface modification of
polymers. The most widely accepted techniques today include flame treatment, plasma
treatment, corona treatment, and chemical modification. A brief comparison of these
techniques with the UV treatment process is presented in this section.

Flame treatment involves the rapid oxidation of the polymer surface by hydrogen
or organic fuel flames which can achieve temperatures of 2000 K [33]. The flame creates
oxygen radicals in the air which attack the polymer and form polar groups on the surface.
The adhesive characteristics of the surface are enhanced by a combination of two factors:
an increase in surface energy by oxidation and the physical oxidative degradation of
weak boundary layers and contaminants. The flame head is typically placed very close
(0.25 inch) to the product surface which allows very high treatment speeds on the order
of tens of inches per second. The treated surfaces can remain stable for several weeks.
Flame treatment is not particularly suitable for three-dimensional objects because the
treatment is strongly dependent on the position of the surface within the flame. The
combustion of fuel as the primary source of energy has Obvious environmental
consequences and safety considerations are very important. The process is low cost and

typically used for high speed processing of polymer webs.

 

P1;
or 3111‘ oil
(realt’d ill 1
dTSCl'ltii' :19 l
:31} be C111“
range from
electrons. it
chemical. me
the flame tre.
contaminants
that the surfer
otntrolled at;
messing. A1
10 the extrem
ripensite been

Corona
technique for 11
filmed by the ‘
positioned a \1

and the Mile

 

3111112631116 p.
Heat

IT131115 are

J. ._
$165 1101 um.

Plasma treatment involves the exposure of a material to oxygen, nitrogen, argon,
or any other gas that can be energized to a plasma state [34]. The plasma is usually
created in a partial vacuum. The plasma can be high temperature (energized by electrical
discharge) or cold (energized by radio frequency). The components of the gas mixture
can be chosen to seed required functional groups on the surface and typical exposures
range from a few seconds to tens of minutes [2-3,7-10 ]. The plasma contains energetic
electrons, ions, and UV radiation, which aggressively attack the surface, causing
chemical, morphological, and topographical changes that can improve adhesion. As with
the flame treatment, the surface is also cleaned in the process and organic compounds or
contaminants are removed efficiently. An advantage of plasma over flame treatment is
that the surface of three—dimensional objects can be treated, but the requirement of a
controlled atmosphere and vacuum are disadvantages as they inhibit continuous
processing. Another disadvantage is that the polymer surface can be easily damaged due
to the extremely aggressive environment in the plasma. The process is generally
expensive because of the need for vacuum and batch processing.

Corona treatment, sometimes called ‘non-vacuum plasma’, is a very popular
technique for treatment of polymers in web type applications [3-6,8-10]. The corona is
formed by the application of high voltage (on the order of 10,000 Volts) to an electrode
positioned a short distance from the substrate. The air gap is ionized by the electric field
and the ionized oxygen in the air forms high levels of ozone. The corona and ozone
oxidizes the polymer, resulting in a high energy suitable for printing and bonding. Corona
treatments are less likely to damage treated surfaces than flame or plasma treatments and

does not utilize volatile solvents. The technique is well suited to fast continuous

manut‘aetu
moi-ante
voltages. s
unieh can
Che
These inelt
aid and a.
other appli.
[totally in.
Which min t
issues,
Table
PT'OCCSSCS. I!
CFflt'entionnt
1158 Of [\V In
reductions in
Cllmpm’ng [TIC

mm“ Proee

 

 

manufacturing but generally unsuitable for treatment of convoluted surfaces. The
disadvantages of corona treatment include safety concerns because of the use of high
voltages, static buildup, and the extremely efficient production of high levels of ozone,
which can cause environmental problems.

Chemical treatments are the most widely used processes for surface modification.
These include organic solvent cleaning of surfaces, primer coatings for paint applications,
acid and alkaline baths for metals, detergent washes, surface coatings, and numerous
other applications. These processes are well suited for continuous manufacturing and
typically inexpensive. The disadvantages include the use of solvents and wet chemicals
which can cause VOC emissions, human hazards, effluents, and other waste disposal
issues.

Table 1.1 compares the UV treatment process with the above mentioned
processes. It is observed that UV treatment can address many of the disadvantages of
conventional surface treatment techniques without posing risks to the environment. The
use of UV treatment to replace flame or chemical treatments can result in significant
reductions in VOC and greenhouse gas emissions. Chapter 7 includes an analysis
comparing the economical benefits of these processes and the results indicate that the UV

treatment process is cost competitive with existing processes.

 

Tah

\ _

Current 51.1111

1

1 Environment;
Impact

T

1 Abilit} to
Treat C omple
1 66001lele

1

Treatment
Time

T

Hazards

C051

[

 

, Surtahilit}. 17'
i- 411111.1ka

 

. f

Table 1.1 Comparison of various surface treatment techniques.

 

Conventional Treatments
(Flame, Corona, Plasma,
and Chemical Modification)

UV Surface Treatment

 

Current Status

Flame, Corona and Chemical -
Mainstream technologies.
Plasma — Limited acceptance.

Developing technology -
Equipment manufacturing base
exists.

 

Chemical — VOC emission, waste
disposal.

 

 

 

 

 

 

Environmental Corona — High levels of ozone Low levels of ozone produced in
Impact produced. contained environment.
Flame — Greenhouse gases,
organic fuels consumed.
. . Flame, Corona and Plasma — . .
Ability to Lme-of-srght treatment.
Treatments have severe .
Treat Complex . . . . . Potential to treat convoluted
. limitations in treating convoluted .
Geometries surfaces is excellent.
surfaces.
Corona, Flame — Very fast.
Treatment . .
. Chemical, Plasma — Moderate Fast treatment times.
Time .
treatment times.
Chemical — Human exposure,

Hazards waste disposal. UV protection for humans.
Corona — High levels of ozone Low levels of ozone produced.
production.

Corona, Flame, Chemical —
Cost Inexpensive Inexpensive.
Plasma — Expensive.
Excellent suitabilit for all
Corona, Flame — Web treatment . . y
. . applications - batch and
applications only. Removal of . .
. . . . . . continuous processmg of flat,
Surtability in process gases is requ1red.
. . . convoluted, and large scale
Manufacturing Cheirucal — EnVironmental
. surfaces.
EnVironment concerns, hazardous.

 

Plasma - Unsuitable for large
scale continuous processing.

 

Minimal hazards — UV
protection and removal of low
levels of ozone required.

 

 

 

1.4 Liter:

Ex

the surfaee

degradatio
emission 1.1
In
understandi
wettability a
in the press
mffllllllsm C

010116 18 shot

 

1.4 Literature Review

Exposure of polymers to UV radiation causes chemical and physical changes on
the surface, and depending on the nature, intensity and duration of the radiation, different
phenomenon are observed. High intensity UV lasers can be used to ablate surfaces
rapidly [14,15,17] while long term exposure to low intensity UV radiation leads to photo-
degradation [35-38]. Various types of UV sources such as pulsed and continuous
emission lasers and lamps can be used for these purposes [39].

In recent years, a significant amount of study has been devoted to the
understanding of UV photo-oxidation as a surface treatment process to enhance
wettability and adhesion [19-30]. Deep UV radiation, of wavelengths 185-280 nm (UVC)
in the presence of oxygen oxidizes surfaces and creates polar functional groups. The
mechanism of oxidation with UV radiation in the presence of atmospheric oxygen and/or
ozone is shown below [12,13,40-43 ]:

hv

024036“ (1.1)

0*+02——>03 (1.2)
hv *

03—902+0 (1.3)

UV at 184.9 nm interacts with oxygen to form ozone. Ozone decomposes at 253.7
nm to form singlet molecular oxygen and atomic oxygen. All products of the above
reactions are very reactive and capable of oxidizing the surface. The chemical changes
occurring are often accompanied by changes in the wettability, morphology and

topography of the surface [44-46].

 

1n addtti
surface also oer
eiuse dissoeiutz-
ehiin scission. 1
The presence 11
subsequent reue

Due to t
the photo—aetii.
more suitable 1‘
The production
lamp and come
Exposure to 0/.
often very lon

ottoation PRK‘C

 

Upidkfi 111 [he

1&3QO In 21 T1

1V

treatmenll‘

L'V tre
expoSed subxl
polymers. The
temperature 9
Changes [18.4

Mme 21 Con

In addition to the thermal oxidation by ozone [47], UV initiated photolysis of the
surface also occurs simultaneously [48-49]. UV photons of sufficiently high energies can
cause dissociation of chemical bonds on the surface. Such bond dissociations can lead to
chain scission, molecular rearrangement and creation of free radicals in the polymer [50].
The presence of oxidizing gases near the activated surfaces leads to a wide variety of
subsequent reactions that can form high-energy surface functional groups.

Due to the different wavelengths required for the photo-dissociation of ozone and
the photo-activation of different polymers, UV lamps with a broad spectral output are
more suitable for surface modification than monochromatic UV lasers or excimer lamps.
The production and dissociation of ozone is a cyclic process that occurs under the UV
lamp and combined with the photo-degradation of the surface leads to rapid oxidation.
Exposure to ozone alone can cause significant uptake of oxygen in a polymer surface but
often very long exposure times on the order of hours are required [20,51-53]. The
oxidation process is accelerated in the presence of UV radiation and the levels of oxygen
uptake in the surface achieved by ozonation can be achieved by UV-Ozone (UVO)
treatment in a matter of a few minutes [13, 32]. Increasing the ozone concentration during
UV treatment by using an external ozone generator enhances the rate of reaction.

UV treatment can also cause morphological and topographical changes in the
exposed substrate. Clear evidence is available from UV laser irradiation studies on
polymers. These changes are brought about by ablation and etching, or very high
temperature gradients in the surface regions which can lead to local microstructural
changes [18,44-46]. These physical and chemical changes occur simultaneously and

provide a complex set of variables which must be analyzed in order to understand the

10

 

process. The C
wettability and .
The C01":
exposure deper.
:onditions has 1
there is a large
1
surfaces. 3 5.“
unit‘ersal surf-at
1V 0.11:.
robust method
enhanced adhe.
"dillVfillOIT pmk

iiipttible to tie

for optimum in

 

Figure
l
R1111 provision

factors that car
pulse repetitir
composition. r
luneties. Thest

(V In
studied with t'

5110““ that

process. The effects of these surface changes are realized in terms of changes in
wettability and adhesion.

The complex physical and chemical changes occurring on the surface due to UV
exposure depend on the treatment conditions used and the proper manipulation of these
conditions has the potential to yield tailored surfaces for various applications. Although
there is a large amount of phenomenological data regarding the effect of UV radiation on
surfaces, a systematic understanding of this process can lead to the formulation of
universal surface treatment strategies that can be used for a wide variety of materials.

UV oxidation and surface activation has the potential for creating a low-cost, fast,
robust method for surface preparation of polymer and polymer composite surfaces for
enhanced adhesive bonding. With the right process conditions, UV oxidation and surface
activation process can be: i) capable of cleaning/treating any polymer surface; ii)
adaptable to treat flat or convoluted surfaces; iii) environmentally benign; and iv) tailored
for optimum mechanical performance of adhesive joints and coatings.

Figure 1.2 shows a schematic representation of a typical UV treatment process
with provision of supplemental ozone to accelerate the surface treatment process. Several
factors that can affect the surface treatment process have been identified: UV irradiance,
pulse repetition rate, ozone concentration, temperature, exposure time, chemical
composition, microstructure and morphology of the substrate, mass transfer, and reaction
kinetics. These process variables are discussed briefly.

UV Irradiance. The effect of very strong UV radiation on polymers has been
studied with the use of UV lasers. Many of these studies relate to ablation and it has been

shown that ablation and photo-oxidation are related processes [39]. Ablation of the

11

 

surface occurs
material and e.
the surface mo
pmeesses rexpt
rates of these
radiation is tltu

model.

'0

 

The pt;
specm‘m Land

11“ _
M170115 mate”

 

riii‘zii‘llon in th I

 

surface occurs when the UV radiation fluence exceeds the ablation threshold of the
material and causes gross degradation of the surface. This manifests itself as a change in
the surface morphology and topography. Moreover, the ozone formation and dissociation
processes responsible for the bulk of the surface oxidation are photo-initiated and the
rates of these reactions depend on the photon flux available. The intensity of UV
radiation is thus one of the most important variables to consider in any chemical process

model.

 

  
 

Ozone
Generator

 

 

Substrate

 

 

 

Figure 1.2 Schematic of the UV treatment process

The pulsed xenon arc lamps used in this work have a continuous radiation
spectrum and the identification of key wavelengths and their characteristic effect on
various materials is an important variable. Figure 1.3 shows the typical output of a xenon
filled UV lamp. Preliminary experiments with quartz and Pyrex® filters have shown that

radiation in the 185nm - 280 nm region is necessary for efficient UVO treatment.

 

Pulse I
flexibility in C‘
tt’ lamps 1m
120 Hz. Figure
operating at 1
t'anable. expet‘
can be seen tl
relatit'el)‘ long
surrounding ga
for a gix‘en tot
pulses 10 (1111)“

Ozoner
has found by \
tn the treatme:
Withers: like
“milk 31 £11. 1.
3821‘.QO elthg

 

Ei‘tterator to F

gulf11111665 Illa‘l

10s

 

{lid}. The €11?

 

Pulse repetition rate. One of the advantages of using pulsed UV lamps is the
flexibility in controlling the exposure time and intensity of the radiation. Typical pulsed
UV lamps have pulse widths of 100-200 its and pulse frequencies ranging from 3 Hz to
120 Hz. Figure 1.4 shows a schematic representation of the output of a pulsed UV lamp
operating at 120 Hz. The pulse frequency of lamps is expected to be a significant
variable, especially in the treatment of polymers which are sensitive to temperature. It
can be seen that the emission of UV radiation is accomplished in short bursts with
relatively long dark periods during which heat transfer from the polymer to the
surrounding gas can occur. As the operating frequency is decreased, the energy per pulse
for a given total output can be increased along with a larger cooling period between
pulses to allow dissipation of heat from the irradiated surface.

Ozone concentration. The rate of surface modification by UV photo-oxidation
was found by Walzak et a1. [19] to be proportional to the concentration of atomic oxygen
in the treatment of UV transparent polymers like polypropylene. In the case of other
polymers like poly-ethyleneterephthalate (PET) this dependence was not observed.
Walzak et al. used an external ozone generator to provide a stable source of ozone in the
treatment environment that was independent of the irradiance at 185 nm. A similar
approach has been followed in this study. The motivation for using an external ozone
generator to provide supplemental ozone in the treatment environment is two fold: it
guarantees that the UV treatment occurs at steady state conditions, and it gives the ability
to vary the ozone concentration independently of the UV lamp output which is necessary

to study the effect of ozone concentration on the rate of surface oxidation.

13

100 T

 

 

" I 1
O 0 0 o 0 o T N
0 m 5 4 3 2 I \ U “H— .I
_ E
r
u
.w5
F

 

w 7
0054‘ 3:3 .55. otzzapcd 032:0!

 

1 00
90
80
70

2

30

20

Relative Amplitude Per Unit Area

10

 

0
160 200 220 240 260 280 300 320 240 360
Wavelength (nm)

Figure 1.3 Typical outputs of pulsed xenon UV lamps.
[Source Xenon Corp., Wobum, MA, USA]

 

 

 

 

 

 

 

A
Irradiance
100 us _ L 8.3ms AI
7 [‘ VI

Figure 1.4 Schematic representation of pulsed UV lamp operation at 120Hz.

 

TemP‘
temperature 1
the UV taint“
[R radiation
relaxation fro
energy in The
heating 116C 'JU
higher \‘ibratii
FranchCondo
states that due
msitions occ
electronic trant
the excited ele
Vibrational sta
Moreover. the
[omitted 10 s
Sift”v1" Iemperor
P1336111 and 111*:
the $1355 trans;
Sufi-acts doe 9..

rates of react],

 

between the r'

(1

 

Temperature. One of the important parameters in UV treatment is the substrate
temperature. Substrate temperatures can increase rapidly during UV treatment because
the UV lamps used have a broad spectral output ranging from UVC to the infrared region.
IR radiation is very effective at causing vibrational transitions in polymers and the
relaxation from excited vibrational states is generally accompanied by the release of
energy in the form of heat. In addition, UV radiation can also contribute to surface
heating because the electronic transitions caused by UV radiation often excite bonds to
higher vibrational levels of the excited state (vibronic excitations) according to the
Franck-Condon principle [54]. Figure 1.5 describes the Franck-Condon principle, which
states that due to the mass difference between an electron and the nuclei, electronic
transitions occur at much faster rates than those of nuclear vibrations, and a vertical
electronic transition results in the excitation of the bond to a higher vibrational state of
the excited electronic state. Relaxation from the excited vibrational states to the ground
vibrational state by internal conversion results in the release of thermal energy.
Moreover, the rates of relaxation of vibrational and vibronic excited states are very high
compared to slow thermal diffusion (heat conductivity) in polymers. This can lead to
steep temperature gradients in the surface region even when no infrared radiation is
present and the transient surface temperatures can be very high. Temperatures higher than
the glass transition (Tg) in polymers can have adverse effects on the properties of treated
surfaces due to the higher mobility of the polymer. However, at low temperatures, the
rates of reactions, which are generally temperature dependent, are also slower. Thus, an
optimum range of temperatures may exist for a material where a balance is achieved

between the rates of reactions and adverse effects of high surface temperatures.

15

 

.\l

 

 

Figure I;
from the

excited I!

 

 

Molecular
Potential
Energy

 

 

 

 

F/

 

 

v’=3
v’=
v’=l
b0,—
__ v=2
v:
\V7/V=0

 

 

Internuclear Distance

 

Figure 1.5 The Franck-Condon Principle showing excitation of an electron
from the ground state of the molecule (8..) to a higher vibrational state of the
excited molecule (8;) followed by vibrational relaxation of the excited state.

16

 

Ché’mit
physical and c
modification p
of me incident
transitions in tl
the ground stat
photon. The l

ionization. Elc

 
  
  
  
  

shuttn in HM
repulsive dixso

enough to diw.

i‘

 

 

 

Chemical structure, microstructure and morphology of the substrate. The
physical and chemical characteristics of the substrate play a key role in any surface
modification process. The chemical structure of the substrate determines the absorption
of the incident radiation. Absorption of radiation in the ultraviolet region occurs due to
transitions in the electronic structure of the molecule. The excited electrons can return to
the ground state or escape depending on the energy transferred by the incident exciting
photon. The latter case, where electrons are emitted from the molecule, leads to
ionization. Electronic transitions can also cause bond dissociation in the molecule as
shown in Figure 1.6 if the transition excites the molecule from the ground state to a
repulsive dissociative state. At very low UV wavelengths, the photon energy can be high

enough to dissociate many organic bonds.

 

blr‘
"k .

b‘g‘fiyk’pwt i}...

 
     
 

Bond
Dissociation

S 1
Molecular
Potential

Energy

 

. .. ‘ v w; ’A' we , . z...
t. .Wfififhfizfi 33°} 2' ~

 

 

 

  

 

 

 

 

Intemuclear Distance

 

arse .4 ex'ammammw

A

wevav mswvexnsvM\WW‘<~§*\W\W§M¢WWYM ‘Msmmi‘ 1%“

Figure 1.6 Electronic transition in a molecule from
the ground state (So) to a dissociative state (81).

17

 

The ah
SWCIUTC tchri-
dependent on :
bonds. Table I

!(§]
in .

I1 ' .
b\0fp[10n C(g“

then ~
"’rblno
: p11):

{titre . «‘-

mcroilmctur
L.

h .
flit

per-lies 0f

 

The absorption of radiation by a molecule is a characteristic of the molecular
structure (chromophore) and the photochemical reactions occurring in the material are
dependent on the wavelength of incident radiation and the absorbance of the constituent

bonds. Table 1.2 shows the typical absorption wavelengths for different organic bonds

[55].

Table 1.2 Absorption bands for some common organic bonds.

 

 

 

 

 

 

 

Chromophore hum (nm)
Ether (—O-) 185
Ketone (C=O) 195
Ester (COOR) 205
Aldehyde (CHO) 210

Carboxyl (COOH) ZOO-210
Hydroxyl (O-H) 230

 

 

 

 

The absorption coefficient of a few common polymers is shown in Figure 1.7.
Weakly absorbing polymers such as polypropylene (PP) and polyethylene (PE) have
absorption coefficients about two to three orders of magnitude lower than strongly
absorbing polymers like polyethylene terephthalate (PET) [14] and are consequently
more difficult to modify by photo-oxidative processes. The surface morphology and
microstructure may also play a critical role in UV treatment due to the different physical

properties of amorphous, crystalline and transcrystalline regions in the polymer. In a

18

reactivity and
polyethylene ..

ozonated LDPE

Absorption coefficient [ cm‘1 1

F121.

pt )
”’TFE

 

 

study on UV-ozone treatment of polyolefins, Peeling and Clark [20] found that the
reactivity and oxygen uptake of high density polyethylene (HDPE) and low density
polyethylene (LDPE) varied considerably. Figure 1.8 shows the XPS Cls/Ols ratio of

ozonated LDPE and HDPE measured by Peeling and Clark.

PET

 

Absorption coefficient [ cm~1l

 

10‘ r .
180

U ‘7 T ‘—

200 220 240 260
wavelength [nm] —----

Figure 1.7 Absorption coefficient as a function of wavelength for
poly(ethylene terephthalate) (PET), poly(tetrafluoro ethylene)
(PTFE), polypropylene (PP) and polyethylene (PE) (Kesting et al.[14]).

19

 

Figure

The hi 5
attributed to tl
lDPE leading
HDPE has a
installites. it

measured by l

 

13 W0 treat.

la1108 When it;

me 1:filt'it‘lor .
I

 

9h: '
higher rut,

 

40

SITY RATIO
(.0
O
./

N
O
l

(2,4015 INTEN
6
l
t

 

 

m!-

1‘6 2?. 312 [\I 61.
OZONATION TIME (hr)

Figure 1.8 Intensity ratio of Cls/015 bands for LDPE (x) and HDPE (o)
as a function of ozonation time. (Peeling & Clark [20]).

The higher initial oxidation and uptake of ozone in LDPE compared to HDPE is
attributed to the lower crystallinity of LDPE. Diffusion is faster in the more amorphous
LDPE leading to a higher uptake of ozone/oxygen in the bulk while the highly crystalline
HDPE has a very slow uptake of ozone, but due to the low diffusion through the
crystallites, it is confined to the surface. This accounts for the higher C/O ratios in HDPE
measured by Peeling and Clark at longer exposures. Similar behavior has been reported
in UVO treated LDPE and HDPE. Figure 1.9 shows the corresponding XPS Cls/015
ratios when the polymers were exposed to 254 nm UV radiation and ozone. Apart from
the higher rate of oxidation, HDPE has a higher oxygen content at long exposure times.

The behavior of such systems at short exposure times will be of interest.

20

 

 

9H;

\l

 

Chi/O,s INTENSITY RATIO
0:

ff;
/

 

 

i. 6 1'2 1‘6
PHOTO-OXIDATION TIME (hr) ‘
Figure 1.9 Intensity ratio of Cls/015 bands for LDPE (x) and! HDPE (o)

as a function of photo-oxidation time. (Peeling & Clark [20]).

Besides the influence of crystallinity, other factors may influence the adhesive
performance of UV treated polymers. Crosslinking has been widely reported in the
surface layers of UV irradiated polymers [56-59]. Crosslinking of the polymer chains
can increase the modulus of the surface and increase the adhesive bond strength.
Conversely, excessive surface degradation can lead to formation of weak boundary layers
and low adhesive performance.

Topographical changes in the polymer can occur due to ablation, photo-
degradation, incubation, and other diffusive processes and lead to changes in the surface
roughness. Surface roughness is known to have a strong effect on the wettability and

adhesive properties of solids. Knittel et al. [44-46] report the formation of intricate ring

21

like morphology in UV laser irradiated cold-drawn PET fibers, which they attribute to the
internal stresses present in the fibers. With increasing exposure, the fine rippled surface
of the fibers became coarser. An empirical model has been proposed to relate the
formation of these features to the UV laser exposure. Breuer et. al also report the
formation of rippled surfaces in UV laser irradiated PP films [49]. The appearance of the
surface features was found to be dependent on the surrounding atmosphere. In oxygen
atmosphere, the laser induced surface structures had a lower frequency than in helium.
This may be due to differences in the surface temperature in the two environments. The
release of crystalline or residual stresses at elevated temperatures during UV treatment
can also lead to changes in topography which are important from the point of view of
adhesive performance. Another form of topographical change reported by Walzak et al.
in UVO treated PP is the formation of mound-like structures detected by AFM
measurements [19]. These mounds of polymer are low molecular weight oxidized
fragments which can be removed by water. Similar morphology was observed in corona
treated PP by Strobe] et. a1 [60] and they concluded that the formation of low molecular
weight fragments on the surface did not necessarily lead to the formation of a weak
boundary layer if the fragments could be displaced into a polar solvent such as inks or
adhesives.

Reaction and Transport. Apart from the above factors relating to the chemistry,
microstructure and morphology of the substrate, there are other variables relating to the
transport of the reactive species in the reaction environment that determine the efficiency
of the UV/ozone treatment. The reaction of ozone with the surface is a gas-solid reaction

and the rate of reaction depends on the relative rates of the surface chemical reaction and

22

the mass transfer of reactive species to the surface. The typical gas-solid interphase
consists of the bulk fluid, the boundary layer and the solid surface. The reactive gas
species must diffuse through the boundary layer to adsorb on the surface before the

reaction can occur as illustrated in Figure 1.10.

 

 

>

 

 

Bulk
Fluid Mass transfer across
boundary layer
Boundary . .
Layer Adsorption/Desorption
Surface Reaction
Surface I .

 

Figure 1.10 Mass transport in gas-solid surface reaction.

The chemical reactions occurring in the UV treatment environment can be classified as
gas phase reactions and surface reactions. The gas phase reactions involve the formation
of reactive species, 0* and O3, and photolysis of 03 as described earlier (Equations 1-3).
The mechanisms of these reactions as well as the quantum yield ((1), moles of product
formed per mole of photons consumed), rate and equilibrium constant data are reported in
literature [19]. Photochemical reaction rates are expressed in terms of quantum yields of
the reaction and the photon flux. For the reactions involving dissociation of oxygen and

ozone at 185 nm and 254 nm respectively, the quantum yield for the dissociation of

23

 

 

oxygen is 2 and that for dissociation of ozone is 0.9. The overall reaction kinetics are,
however, dependent on the actual surface reactions that occur during treatment.

Photo-degradation mechanisms of common polymers. UV treatment of polymers
can be considered to be analogous to controlled photo-degradation. A lot of work has
been done in the field of polymer degradation to study the weathering behavior of these
materials. Many of the photo-oxidation and photolysis mechanisms observed in long term
photodegradation are expected to be the initial steps in the UV oxidation process. The
reactions with ozone can yield other products which are not described by these
mechanisms, but they can provide a good reference for the analysis of UVO treated
surfaces.

Reaction mechanisms for various polymers have been proposed in literature.
Figure 1.11 shows one of the proposed reaction schemes for polyethylene (PE) and
polypropylene (PP) under VUV (< 185 nm) irradiation [48,61]. Polyolefins should in
theory be photo-oxidatively stable to radiation above 185 run because of the lack of
chromophores that absorb strongly at longer wavelengths, yet PE and PP polymers can be
oxidized with mercury as well as xenon lamps. This is believed to be a result of
sensitizing impurities in the polymers. Rabek, in his excellent treatise on photochemical
reactions in polymers [50] identifies vinyl and vinylidene impurities in polyethylene as
the starting points for photo-oxidation. The mechanisms of UV photo-oxidation of HDPE
are shown in Figure 1.12. The vinyl impurity in the polyethylene chain leads to the
formation of hydro-peroxy radicals and ultimately aldehydes, ketones and crosslinked

polymer.

24

OH

 

 

 

 

 

 

C|)/
HR
____>+hv ()0 +__> R__C_R +_hV
I o
P]! +O / R (I)
o 0 2
R—C—R +—'O—> R—c—R ——+ R—(ll—R— on ——> R—c—R
R R R o’ R
+‘0H y+°R R_C_R +hv
|
R
+:O
?”i
E. R—C—C—R
--R 1'2 1
<3 o
l I --H/-'CH3 II I
R—c—C—R R—C—C—R
R I I
o
-CH2R u
.__, R—C'I
R

Figure 1.11 VUV photo-oxidation reaction scheme for PE and PP [48].

25

+O 90H 3
2
—CH2-E*CH2-CH2— ——-> —CH2—fi—CH’CH2— ‘_—> _CH2—E_C_CH2—

 

CH2 CH2 CH2
1 Kctonc
‘1’ I?
—CH2—fi—CH-CH2— —> —CH2—fi—CH + CH2
CH2 CH2
Aldehyde
C?
—CH2—fi—CH2-CH3— + —CH2—fi—CH—CH2— ———> —CHz—(l?-CH2-CH2—
CH2 CH1) CH2

.. I

o
—CH3—fi—J‘.H-CH2—

CH2 Crosslinking

Figure 1.12 Photo-oxidation reaction scheme for PE.

The degradation mechanisms of polycarbonate have been the focus of many
studies [35-38]. Figure 1.13 shows the typical photo induced reactions in bisphenol-A
polycarbonate [50]. The most important reaction is the photo-Fries rearrangement of the
carbonate linkage under UV irradiation to form phenyl salicylate, and ultimately, o-
hydroxybenzophenone. Both compounds are UV stabilizers and give polycarbonate an
auto-stabilizing capability. The photo-Fries rearrangement occurs independently of chain
scission mechanisms that cleave the carbonate linkage to liberate C02 and the phenoxy
radical formed can abstract a hydrogen or the methyl group in the bisphenol-A structure.
These products of these basic reactions undergo crosslinking reactions or further

oxidation with ozone to yield a wide range of products.

26

 

 

{iEO—E—O— . ”E: 0.

(Eu, 1 H.

+io-‘3—o— - +$ o ..

Figure 1.13 Photo-Fries rearrangement and
photolysis reaction scheme for polycarbonate.

27

 

Polye:
susceptibilit}
chromophore
direct scissii
mechanism t‘
with the lot-t

Wm lP-O(

 

 

Polyethylene terephthalate (PET) is an aromatic polyester known for its
susceptibility to UV degradation. The aromatic esters are strongly absorbing
chromophores with high absorptivities below 315 nm. The degradation of PET occurs by
direct scission of the ester bonds in the backbone. Figure 1.14 shows the basic
mechanism of chain scission in PET [50]. The radicals formed under decarboxylation
with the loss of CO or C02 and are then oxidized by oxygen or ozone to form reactive

peroxy (P-OO*), alkyloxy (P-O*) or hydro-peroxy (P-OOH) groups.

‘3 Ii
—c—.-—C- + oO—CHz-CHz—O—
it it /
—C‘—C—O-CHz—CHz-O—

I \ it it
—C"C_Oo + oCHz-CH2-0_
I? ‘3.
—C—.o + oC—O—CHz-CHz-O—

Figure 1.14 Photo-oxidation scheme for polyethylene terephthalate.

There are several other important photo and oxidative degradation mechanisms
which are commonly observed in organic molecules. The most important class of these
general mechanisms for polymers are the Norrish Type I and Type II photodegradation
mechanisms in polymers containing carbonyl groups [50,62]. The Norrish Type I

mechanism describes the photo-cleavage of a bond at 0t position in relation to the

28

 

carbonyl SI“
containing l

decarhonyl’dla

RCHgCll

R'Cli

 

Figure

The .‘l
atom on a y .
I. A. ‘.
he {carbon
33mm IO fl"

.\OIT1$1'I mCCt

 

carbonyl group. This type of reaction is responsible for chain scission in polymers
containing ketone and aldehyde groups. At elevated temperatures the acyl radical

decarbonylates with the evolution of carbon monoxide as illustrated in Figure 1.15.

RCHZCHZCHA RCHZCH2CH2' RCH2CH2CH2'
C: O —h—V—» ———> + CO
R'CHZCHZI R'CHZCHzco R'CHZCH2'

Figure 1.15 Norrish Type I mechanism for a-cleavage of carbonyl groups.

The Norrish Type II mechanism is observed in ketones possessing a hydrogen
atom on a 7 carbon atom. The oxygen on the carbonyl group abstracts a hydrogen from
the 7 carbon to form a six membered cyclic intermediate which causes cleavage at the [3

carbon to form end groups with unsaturation as illustrated in Figure 1.16. Both types of

Norrish mechanisms are important in photodegradation of polymers.

0 . O t
R \\C/R R DEC/R R\ ’H"'(“)
H3; (EH A, I i ___, ,c. .c—R'
2 \C/ 2 H2C\C/CH2 H b-—é/
H2 H2 H2 H2
H ,R HO IR'
\C C
II + II
CH2 CH2

Figure 1.16 Norrish Type II photo-elimination mechanism in
polymers containing carbonyl groups.

29

 

The t'
pmposfid b}
attacks the 7:
peroxy radic.
subsequent 1‘

612610111ch a:

Criegee mec:
R\t
H/

 

 

The oxidative attack of ozone on unsaturated bonds in organic molecules was
proposed by Criegee and is known as the Criegee mechanism [63]. The ozone molecule
attacks the 1: bond and forms an ozonide ring. The ozonide ring dissociates to form a
peroxy radical ion and a ketone. The peroxy radical ion which can undergo a variety of
subsequent reactions including chain cleavage in the polymer. The susceptibility of
elastomers and other unsaturated polymers to ozone attack is primarily explained by the

Criegee mechanism.

\ _ / _ _ _
H/C—C\R ___, /c C\

O + O .
(I) o- +3) 0
\ / -——> \\ + H—C—R
7C—C< RA

Figure 1.17 Criegee mechanism for ozone attack on unsaturated bonds.

It can be seen from this brief review of the literature that the photochemistry of
organic molecules, especially polymers, can be very complicated when the effects of UV
irradiation and ozonation are combined. The products of UV photolysis of a molecule can
undergo ozonation, and the products of ozonation can themselves be photoactive. The
determination of the precise reactions occurring in various polymers under conditions of

UV and ozone oxidation is not trivial and beyond the scope of this study.

30

 

15 Goals
The c
surface treatr
different p03
manufacturin
commercial
stabilization .
in the compo
be useful in
This work \V
most. if not a
This
four differen
‘PCI. and c
these materi.
must be cont
The
Changes 0C0
Sifipg Which
Cart bf don,

{hinge in \'

 

totaling”5 a-

Ii -
he gun‘aCe

 

 

1.5 Goals

The objective of this study is to obtain a predictive engineering model for the UV
surface treatment process. The determination of precise mechanisms of UV oxidation for
different polymers is a fascinating field of study, but it has limited use in the
manufacturing environment where pure, repeatable surfaces are rarely available. Most
commercial polymers are processed to various extents and have additives to aid
stabilization and processing. Moreover, there can be considerable batch to batch variation
in the composition and surface quality of commercial polymers. Any model that hopes to
be useful in such environments must be adaptable to these, sometimes large, variations.
This work will focus on developing a model for UV treatment which can be applicable to
most, if not all, polymers.

This will be accomplished by studying the properties of polymers belonging to
four different classes: therrnosets (epoxy), elastomers (rubber), amorphous thermoplastics
(PC), and crystalline thermoplastics (PET). The similarities and differences between
these materials will be used to determine the key UV treatment process parameters which
must be controlled to achieve the required level of surface modification in a material.

The approach will involve the consideration of the sum total of all chemical
changes occurring on the surface during UV treatment instead of modeling the individual
steps which can be different for each polymer. The characterization of surface treatment
can be done with surface energy measurements and surface chemical analysis. The
change in surface functional groups can be analyzed as a function of various treatment
conditions and empirical relationships can be developed between the functionalization of

the surface and process parameters. This macroscopic approach to the problem can

31

I
provide proc

effect of sur
correlated to
ultimate engi

The g-
can account
predictive p:

estimated fro

 

 

provide process optimization guidelines, which can be easily applied in practice. The
effect of surface modification on the adhesive properties of the material can also be
correlated to the functionalization of the surface, providing relationships between the
ultimate engineering properties and treatment process parameters.

The generalization of these observations will be used to develop relationships that
can account for the differences between polymers. This will ultimately lead to a
predictive process model where the effect of UV treatment on the polymers can be

estimated from a minimal set of data and information about the structure of the material.

32

(J)

. Dnfllw

hhddber
.eId/tt'filttl

. GUIOu‘s'b

. Sohn S.

2003.

. Stunner

Tonegro
4165.2t

. SuobeLl

. Sako.§t:

415.199'

   
  

Snobel
.‘ltfllt’jfi in

Kane, .
M‘i‘dmt‘ti

 

 

10.

ll.

12.

13.

14.

References

Drzal L.T. SAMPE Journal, Vol. 19, N0. 5, pp. 7-13, 1983.

Mahlberg, R.; Nieme, H.E.M., Denes, F.,; Rowell, R.M. International Journal of
Adhesion & Adhesives, Vol. 18, pp. 283—297, 1998.

Gutowski, W.S.; Wu, D.Y.; L], S. J. Adhesion, Vol. 43, pp. 139-155, 1993.

Sohn, 8.; Chang, S.; Hwang, I. J. Adhesion Sci. Technol., Vol. 17, No. 3, pp. 453-469,
2003.

Martinez-Garcia, A.; Sanchez-Reche, A.; Gisbert-Soler, S.; Cepeda-Jiminez, C.M.;
Torregrosa-Macia, R.; Martin-Martinez, J. Adhesion Sci. Technol., Vol. 17, No. 1, pp.
47-65, 2003.

Strobe], M.; Lyons, C.S. , J. Adhesion Sci. Technol., Vol. 17, No. 1, pp. 15-23, 2003.

Sako, N.; Matsuoka, T.; Sakaguchi, K.; Composite Interfaces, Vol. 4, No. 6, pp. 40]-
415, 1997.

Strobe], M.; Walzak, M.J.; Hill, J.M.; Lin, A.; Karbashewski, E.; Lyons, C.S. J.
Adhesion Sci. Technol., Vol 9, No 3, pp. 365—383, 1995.

Kruse, A.; Kruger, G.; Baalmann, A.; Hennemann, O.D. Polymer Surface
Modification: Relevance to Adhesion, K.L. Mitta] (Ed.), pp. 291-301, 1995.

Lane, J.M.; Hourston, D.J. Progress in Organic Coatings, Vol. 21, pp. 269-284,
1993.

Bhurke, A.S.; Jensen, M.; UV Treatment — A Low Cost, Environmentally Friendly
Surface Treatment Technique, ML890 Report, Case Center for Computer-Aided
Engineering and Manufacturing, Michigan State University, USA, 1997.

Hashem, T.M.; Zirlewagen, M.; Braun, A.M. Wat. Sci. Tech, Vol. 35, No. 4, pp. 4]-
48, 1997.

Macmanus, L.F.; Walzak, M.J.; Mcintyre, N.S. Journal of Polymer Science: Part A:
Polymer Chemistry, Vol. 37, pp. 2489-2501, 1999.

Kesting, W.; Knittel, D.; Bahmers, T.; Schollmeyer, E. Applied Surface Science, Vol.
54, pp. 330-335, 1992.

33

 

   
  
   
  
  

15, Buchntat
Rch’t'tillt‘

16. Sancaltta
I33. 199

17.130116. .'\1
IS. Krajnovi

l9. \l’alzak.

't‘.
ROmero-
AdheSlmt

T,
Vol. 13

‘9- KdVQ T
1059 3r.
'qmjlciL.
V01. 21 -.

 

15. Buchman, A.; Dodiuk, H.; Rotel, M.; Zahavi, J. Polymer Surface Modification:
Relevance to Adhesion, K.L. Mitta] (Ed.), pp. 199-212, 1995.

16. Sancaktar, E.; Babu, S.V.; Zhang, B; D’Couto, G.C. J. Adhesion, Vol. 50, pp. 103-
133, 1995.

17. Bolle, M.; Lazare, S. Applied Surface Science, vol. 54, pp. 471-476, 1992.

18. Krajnovich, DJ. J. Phys. Chem. A, Vol. 101, pp. 2033-2039, 1997.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Walzak, M.J.; Flynn, S.; Foerch, R.; Hill, J.M.; Karbashewski, B; Lin, A.; Strobe],
M. Polymer Surface Modification; Relevance to Adhesion, K.L. Mitta] (Ed.), pp. 253-
272, 1995.

Peeling, J .; Clark, D.T. Journal of Polymer Science: Polymer Chemistry Edition, Vol.
21, pp. 2047-2055, 1983.

Charbonnier, M.; Romand, M.; Esrom, H.; Seebock, R. J. Adhesion, Vol. 75, pp. 381-
404, 2001.

Praschak, D.; Bahners, T.; Schollmeyer, E. Appl. Phys. A, Vol. 66, pp. 69-75, 1998.
Pasternak, M. Journal of Applied Polymer Science, Vol. 57, pp. 1211-1216, 1995.

Minagawa, M.; Saito, T.; Fujikura, Y.; Watanabe, T.; Iwabuchi, H.; Yoshii, F.;
Sasaki, T. J. Appl. Polym. Sci. Vol. 63, No. 12, pp. 1625-33, 1997.

Yu, J.R.; Chen, Z.L.; Zhu, J.; Liu, Z.F.; Wang, QR. Intern. Polymer Processing,
XIV, 4, pp. 331-335, 1999.

Petit, S.; Laurens, P.; Barthes-Labrousse, M.G.; Amouroux, J .; Arefi-Khonsari, F. J.
Adhesion Sci. Technol., Vol.17, No. 3, pp. 353-368, 2003.

Romero-Sanchez, M.D.; Pastor-Blas, M.M.; Martin-Martinez, J .M.; Walzak, MJ. J.
Adhesion Sci. Technol., Vol. 17, No. 1, pp. 25-45, 2003.

Ouyang, M.; Yuan, C.; Muisener, R.J.; Boulares, A.; Koberstein, J .T. Chem. Mater.,
Vol. 12, No. 6. PP. 1591-1596, 2000.

Kavc, T.; Kern, W.; Ebel, M.; Svagera, R.; Polt, P. Chem. Mater., Vol. 12, pp. 1053-
1059, 2000.

Grujicic, M.; Cao, G.; Rao, A.M.; Tritt, T.M.; Nayak, S. Applied Surface Science,
Vol. 214, pp. 289-303, 2003.

34

 

3]. Dontula.
Society. 1'

32. Bhurke.
106-108.

33. Sullivan.
lnte’mutt

pubhcaur

34. Liston. l
Prilt‘titt?"

35. Webb. J
3.6. Factor.

37. Ramani.
354. 20%

38. Turton. '
I82. 201'

39- Zhang;
All/1651} rr

 

 

40. King X.
1999.

4I.O'Keeffi

Chemim
43‘ GUrol. 3
‘13-Heit, Q:

4'4“ Knittel,
1997_

45. Knlliel‘
1997,

4.6. Knittel.
1998,

31. Dontula, N .; Weitzsacker, C.; Drzal, L.T. Proc. 20lh Annual Meeting of the Adhesion
Society, pp. 97, 1997.

32. Bhurke, A.S.; Drzal, L.T. Proc. 26" Annual Meeting of the Adhesion Society, pp.
106-108, 2003.

33.3ullivan, N.; Branch, M.C.; Ulsh, M.; Strobe], M. Proceedings of the 17th

International Colloquium on Dynamics of Explosion and Reactive Systems, electronic
publication, http://reaflow.iwr.uni-heidelberg.de/~icders99/icders99.html, 1999.

34. Liston, E.M.; Martinu, L.; Wertheimer, M.R. Plasma Surface Modification of

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

Polymers, M. Strobe], C. Lyons, K.L. Mittal (Eds), pp. 3-39, 1994.
Webb, J.D.; Czandema, A.W.; Macromolecules, Vol. 19, pp. 2810-2825, 1986.
Factor, A.; Ligon, W.V.; May, R.J. Macromolecules, Vol. 20, pp. 2461-2468, 1987.

Ramani, R.; Ranganathaiah, C. Polymer Degradation and Stability, Vol. 69, pp. 347-
354, 2000.

Turton, T.J.; White, J .R. Plastics, Rubber and Composites, Vol. 30, No.4, pp. 175-
182, 2001.

Zhang; Esrom; Kogelschatz; Emig Polymer Surface Modification: Relevance to
Adhesion, K.L.Mittal (Ed.), pp. 153-184, 1995.

Ning, X.; Qin, Q. Journal of Computational Chemistry, Vol. 20, No. 6, pp. 623-628,
1999.

O’Keeffe, P.; Ridley, T.; Lawley, K.P.; Maier, R.R.J.; Donovan, R.J. Journal of
Chemical Physics, Vol. 110, No. 22, pp. 10803-10809, 1999.

Gurol, M.D.; Akata, A. AICHE Journal, Vol. 42, No. 11, pp.3283-3292, 1999.

Heit, 6.; Braun, A.M. Wat. Sci. Tech, Vol. 35, No. 4, pp. 25-30, 1997.

. Knittel, D.; Kesting, W.; Schollmeyer E. Polymer International, Vol 43, pp. 231-239,

1997.

Knittel, D.; Kesting, W.; Schollmeyer E. Polymer International, vol 43, pp. 240-250,
1997.

Knittel, D.; Kesting, W.; Schollmeyer E. Polymer International, vol 45, pp. 103-109,
1998.

35

 

47. Watanab.

1“
333-331.

48. HOIILUI‘LIL'
Pt-rlvmer

 

49. Breuer. .‘
Mittal I I;

50. Rahelt. I

Pohnur

51. Fu. X.; I.

 

52. Cheng. l
53. Daseupt.

34. Atkins.
I'niverv.

33. Lothian,

5'6- Qu. B..
1993.

57. Ranby, 1

58' Rmbiu t
Pp. 303-,

59. Stu. vv,
I994.

60. SITObel.

PIUSML'h

6]. “liken.
I993

67
-. HQrSPOt

63.. -
Cnegse.
1313.133--

 

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Watanabe, F.; Shimomura, 0.; Journal of Hygienic Chemistry, Vol. 37, No. 5, pp.
323-331, 1991.

Hollander, A.; Klemberg-Sapieha, J.; Wertheimer, M.; J. Polymer Science Part A:
Polymer Chemistry, Vol. 23, pp. 2013-2025, 1995.

Breuer; Metev; Sepold; Polymer Surface Modification: Relevance to Adhesion, K.L.
Mittal (Ed.), pp. 185-197, 1995.

Rabek, J .F. Mechanisms of Photophysical Processes and Photochemical Reactions in
Polymers, John Wiley, New York, 1987.

Fu, X.; Lu, W.; Chung, D.D.L. Carbon, Vol. 36, No.9, pp. 1337-1345, 1998.
Cheng, F.; Hong, 8.; Ho, C. J. Adhesion, Vol. 67, pp. 123-137, 1998.
Dasgupta, S. Journal of Applied Polymer Science, Vol. 4], pp. 233-248, 1990.

Atkins, P.W.; Friedman, R.S. Molecular Quantum Mechanics, Third Ed., Oxford
University Press, Oxford, 1997.

Lothian, G. Absorption Spectrophotometry, 3rd ed., pp. 101, Hilger, London, 1969.

Qu, B.; Ranby, B. Journal of Applied Polymer Science, Vol. 49, pp. 1799-1807,
1993.

Ranby, B. Polymer Engineering and Science, Vol. 38, N0. 8, pp. 1229 —1243, 1998.

Ranby, B. Polymer Surface Modification: Relevance to Adhesion, K.L. Mittal (ed.),
pp. 303-317, 1995.

Shi, W.; Ranby, B. Journal of Applied Polymer Science, Vol. 51, pp. 1129-1139,
1994.

Strobe], M.; Jones, V.; Lyons, C.S.; Ulsh, M.; Kushner, M.J.; Dorai, R.; Branch, M.C.
Plasmas and Polymers, Vol. 8, No. 1, pp. 61-95, 2003.

Wilken, R.; Hollander, A.; Behnisch, J. Macromolecules, Vol. 3], pp. 7613-7617,
1998.

Horspool, W.M.; Photolysis of Carbonyl Compounds, Photochemistry, Vol. 30, Royal
Society of Chemistry, Cambridge, UK, 1999.

Criegee, R. in: Ozone Chemistry and Technology, American Chemical Society,
pp.133-135, Washington, 1959.

36

 

The equipn
of the treat
Il'obum. .\1
output in th
built alumir
analytical cl
W Spectre»

treated surfa

l] L" La
Three X6110
Bed. The n,
I] The Ia!
Wilde Opti
and all Emit
the heat ger-
litter gum“
dilemma”

 

[RC 747. Rt

 

CHAPTER TWO

EXPERIMENTAL METHODS

The equipment and experimental techniques used for UV surface treatment and analysis
of the treated surfaces are described in this chapter. Xenon UV lamps (Xenon Corp.,
Wobum, MA) were used in this study because of their pulsed operation and enhanced
output in the low wavelength UV-C region. Various polymers were treated in a custom
built aluminum chamber, and treated surfaces were characterized by different surface
analytical characterization techniques including X-ray photoelectron spectroscopy (XPS),
UV spectroscopy, and surface energy measurements. Mechanical characterization of the

treated surfaces was accomplished with nano-indentation and adhesion measurements.

2.1 UV Lamps and Filters

Three Xenon flashlamps operating at frequencies ranging from 3 Hz to 120 Hz were
used. The nominal power, frequency, and shape of the lamps are summarized in Table
2.1. The lamps are equipped with aluminum reflectors and fused quartz windows to
provide optimum transmission of UV radiation. The lamp housings are made of metal
and all emitted radiation is transmitted through the fused quartz windows. A portion of
the heat generated by the bulb is removed from the housing by forced air convection. The
outer surface of the lamp window is considered the area source of UV radiation for all
determinations of irradiant energy and is the reference plane for measurement of the

distance between the lamp and sample surface. Figure 2.1 shows the Xenon UV lamps

(RC 747, RC 740, and RC 500) used in this work. Samples were UV treated in a chamber

37

 

made In
x 1 incl
provide t

DynaxilE

schematic
material h
vicious gl
quartz gla:
1T radiati
but the trar.

ill/7 Ildnxn

to determl' n

 

 

 

 

made from 1/4 inch aluminum plates having exterior dimensions of 8 inch (1) x 8 inch (w)
x 1 inch (h). A 4 inch x 4 inch opening was machined in the top of the chamber to
provide a path for illuminating radiation. The opening was closed with either Suprasil®,
Dynasil®2000, Quartz, or Pyrex® glass plates during treatment. The chamber was
equipped with an inlet and outlet for process gases. The UV treatment chamber and a
schematic of the experimental setup are shown in Figures 2.2 and 2.3. The selection of
material for the treatment chamber window or filter was based on the transmission of the
various glasses as shown in Figure 2.4. Suprasil® and Dynasil® are high purity fused
quartz glasses with very low hydroxyl concentrations, which enhances transmission of
UV radiation below 200 nm. Ordinary quartz has very high transmittance above 230 nm
but the transmittance drops sharply below 200 nm. Pyrex® windows have a cut-off of ca.
10% transmittance at 290 nm. The Pyrex® window was used as a 280 nm high pass filter

to determine the effect of exclusion of low wavelength UV-C radiation on the treatment

 

 

 

 

process.
Table 2.1 Description of UV lamps
Lamp Power (W) Frequency (Hz) Bulb Shape
Xenon RC 500 300 120 5 inches, linear
Xenon RC 740 1500 10 3.5 inches, coil
Xenon RC 747 1500 3, 120 16 inches, linear

 

 

 

 

 

 

38

 

Figure 2.1 Xenon pulsed UV lamps (from left) RC 747, RC 500, and RC 740.

39

 

Figure 2.2 Aluminum UV treatment chamber with Suprasil® cover.

 

 

UV Lamp
—

 

 

 

Quartz (Suprasil®) Window Distance Between

Lamp And Sample

Sample Chamber
\ 1 1

Ozone Inlet \ Sample ! Ozone Outlet

 

 

 

 

Figure 2.3 Schematic of the experimental setup for UV surface treatment.

40

% Transmittance

 

 

 

O # I I I I l I I I I I I I
190 240 290 340 390 440 490 540 590 640 690 740 790

Wavelength (nm)

Figure 2.4 UV transmittance of Suprasil®, Quartz, and Pyrex® filters

4]

2.2 Measurement of UV Radiation

The measurement of radiation is broadly classified into two categories:
radiometry and photometry [1]. Radiometry is the measurement of energy emitted in a
range of wavelengths by a source of radiation while photometry is the measurement of
visible light, more commonly concerned with the effect and response of the human eye to
light. Radiometric measurement at specific wavelengths of the spectrum or spectral bands
is called spectral radiometry. In terms of chemical applications of radiation, spectral
radiometry is widely used since chemical reactions often occur at particular wavelengths
and the measurement of the amount of radiation emitted at those wavelengths can be used
to quantify the processes. The concepts, symbols, and definitions used to describe the
quantitative measurement of radiation are varied and some of the widely used concepts

and symbols are listed in Table 2.2.

Table 2.2 Radiometric units and descriptions.

 

 

 

 

 

 

 

 

Radiometric Unit Symbol Units Description

. Energy transported in the form of
Radiant Energy Q J electromagnetic waves.
Radiant Flux W Radiant energy transmitted by radiation in
(Radiant Power) (I) unit time.
Radiant Emittance M W/m2 Radiant flux emitted from a unit surface.

. . Radiant flux leaving the radiation source
Radiant Intensrty I W/sr per solid angle (steradian, sr).

. 2 Radiant intensity passing through a unit
Radrance L W/sr.m area.
Irradiance E W/m2 Radiant flux received by a unit area.
Irradiation H J /m2 Integrated irradiance over the exposure

time.

 

42

Due to the variety of radiometric units available to quantify radiation, it is
important to choose the right unit to describe the system. In an application such as UV
surface treatment, the amount of radiation reaching the surface is the quantity of interest
and both intensity and irradiance can be used to quantify the radiation. However, it is
important to note that the two units differ in concept: intensity is a unit related to the
source of the radiation while irradiance is a unit related to the receiving surface. Thus,
intensity is more useful in describing the amount of UV radiation emitted by the UV
lamp while irradiance is more useful in describing the amount of emitted radiation
received by a surface exposed to it. Therefore, irradiance will be used to quantify
radiation in this study.

During treatment, the UV lamp is placed above the aluminum treatment chamber with a
UV transparent fused quartz (Suprasil® 300) window sealing the chamber. The UV
radiation emitted by the lamp passes through air, a quartz window, and a layer of ozone,
before reaching the sample. All of these materials absorb radiation to varying extents,
causing attenuation of the UV radiation emitted from the lamp. Another factor to consider
in the measurement of irradiance is the geometric arrangement and nature of the
receiving surface. Lambert’s cosine law [1] states that the irradiance incident on a surface
varies as the cosine of the incident angle, 0, such that E9 = E cos(0). As seen in Figure
2.5, the projected area of the receiving surface orthogonal to the incident radiation
decreases as the incident angle increases. Most real surfaces are Lambertian in nature and
the dependence of the incident irradiance on the angle of incidence must be taken into
account especially since most commercially available lamps are not point sources of

light.

43

 

      

(0°) 100%

 

(30°) 87%

  

 

His-:31 tartar; 5” 9%?

“n-——

I

    

  

-s or-

-----------------
r - . agrvgrearo. we, :._:V,_./,:/Maeaat%7sxe::‘ is age... .3;

 

Figure 2.5 Lambert’s cosine law showing decrease in
irradiance with increasing angle of incidence.

The irradiance at 254 nm wavelength incident on the sample was measured using a
radiometer (International Light, IL17OO Research Radiometer) coupled with a solar blind
photodiode detector (International Light, SED 220), a 254 nm narrow-band filter
(International Light, NS 254), and a quartz cosine-response filter. The setup was
calibrated with NIST traceable standards. The photodetector used has a narrow (8-100)
viewing angle, hence a quartz cosine-response filter was installed over the detector to
allow collection of radiation incident at oblique angles. To obtain an accurate measure of
the amount of radiant energy reaching the sample surface during UV treatment, the entire
detector assembly was mounted inside the chamber at the location where samples were
mounted for treatment. Irradiance was measured with ozone (approximately 750 ppm)
and oxygen gases flowing through the chamber at 30 scfh. Irradiation was measured in
time-integral mode and the irradiance calculated as the average energy reaching the

detector per unit time. Irradiance for RCSOO and RC747 lamps was measured as a

44

 

 

function of varying distance from the treatment chamber. The relationship between
irradiance and distance from a point source is defined by the Inverse Square law which
states that the irradiance decreases as the square of the distance between the detector and
the source [1]. The results of the irradiance measurements for the two lamps are shown in
Figures 2.6 and 2.7. The irradiance increases as the distance from the lamp decreases.
However, the change in irradiance does not follow the Inverse Square law exactly
because the lamps, far from being point sources, are high aspect ratio cylindrical tubes.
Additional deviation is caused by the presence of the chamber and window which limits
the angle of view of the detector assembly. This alleviates the effect of increasing
distance as a greater length of the UV lamp is brought into the limited angle of view of
the detector as the lamp is moved farther away. The irradiance values obtained also take
into account the radiation absorbed by the gas layer (oxygen and ozone) and any
reflection and refraction effects from the quartz window and aluminum chamber. The
data obtained is a measure of the actual irradiance received by a sample placed in the UV
treatment chamber under these specific conditions. Since the irradiance is invariant for a
given set of physical conditions, values from the irradiance—distance calibration curves
were used for calculating the total irradiation (irradiant energy) received by a sample

during treatment such that:

I
H: JEdtzEt (2.1)

0
where H is the irradiant energy or irradiation (J/cmz), E is the irradiant power or

irradiance (W/cmz), and t is the total UV exposure time (seconds).

45

 

 

 

 

The effect of the gas flowing through the chamber has a marked effect on the irradiance.
Ozone has a very strong absorption peak at 254 nm and the incident UV radiation is
absorbed by the layer of gas present between the sample and the quartz window. This
decrease is absent in the case of oxygen (or air), which has no significant absorption at

254 nm, and consequently higher levels of irradiance were obtained.

2.3 Measurement of Ozone Concentration

The UV surface treatment process involves the oxidation of the material surface with
ozone in the presence of UV radiation. Ozone is produced by the interaction of 185 nm
wavelength photons with molecular oxygen and the amount of ozone formed depends on
the output of 185 nm radiation [2]. Xenon lamps have significant output at 185 nm and
with appropriate selection of the equipment, the required ozone can be produced in-situ
during UV treatment. However, reliance on in-situ production of ozone can present
problems during short batch operations such as laboratory experiments because the ozone
produced under the lamp may not reach steady state concentrations immediately on
startup. To avoid this problem, an external ozone generator (Ozotech Inc., OZ6BTU) was
used in experiments requiring ozone to ensure steady—state conditions. Another advantage
of using an external ozone generator is the ability to vary the ozone concentration and
flow rate, independently, to study their effect on the process. The measurement of ozone
concentration can be accomplished in different ways. Iodometric titration and
spectrometric absorption measurement are the two most widely used methods.
Quantitative methods based on absorption measurement are more attractive because of

their simplicity, speed, and the adaptability to continuous sampling.

46

 

 

 

bfl

—e— 02 flow

 

l l 1

l

Irradiance (mW/sq.cm)
-l N O) h 01 O) \I on (O

 

 

+ 03 flow ~-

 

 

 

 

O
O

2

3 4 5 6 7 8 9
Distance from Lamp (inch)

Figure 2.6 Irradiance measurements for RC500 lamp with ozone and oxygen flow.

47

 

 

 

 

16 i -e— 02 flow

 

 

 

 

 

E 14 +03 flow

 

 

 

 

 

 

 

 

 

 

 

012 3 4 5 6 7 8 9101112
Distance from Lamp (inch)

Figure 2.7 Irradiance measurements for RC747 lamp with ozone and oxygen flow.

48

 

 

Beer-Bouguer-Lambert Laws

The principle laws of light absorption of interest in quantitative UV spectroscopy are the
Bouguer-Lambert law and Beer’s law. Bouguer (1729) stated that the proportion of
radiation absorbed by a substance was independent of the intensity of the incident

radiation [3]. This was later formalized by Lambert. The Lambert-Bouguer law states:

I
lnhTO] = ab (2.2)

where 10 is the intensity of the incident light, I is the intensity of transmitted light, b is the
thickness of the transmitting medium, and or is the absorption coefficient characteristic of
the medium. The absorption coefficient 0t contains no concentration factor and is
applicable only to pure materials. When the natural log in Equation 2.2 is converted to
base 10 log, (X is converted to the Bunsen and Roscoe extinction coefficient, K [4].

Beer (1852) related the Bouguer-Lambert law to the concentration of the transmitting
medium by stating that a photon can only be absorbed by a molecule if it collides with
that molecule. Accordingly, the absorption of light through a solution is also proportional
to the concentration of absorbing molecules present in the transparent medium. The

proportionality to concentration can be incorporated in Equation 2.2 to give Beer’s law

[3]:
I0
log T = A = ebc (2.3)

where a is the molar absorptivity, or molar extinction coefficient, c is the concentration,

b is the path length, and A is the absorbance of the medium.

49

 

 

‘2
‘r
W
i
-
.

 

The absorbance can be measured with spectrophotometers and with known values of the
extinction coefficient and path length, the concentration can be calculated. Ozone has a
strong absorption maximum at 253.7 nm, as shown in Figure 2.8, with a molar extinction
coefficient of 0.000308 ppm"cm'l at 0°C and 1 atm pressure [5].

A commercially available ozone generator was used to produce supplemental ozone. The
generator produces ozone by corona discharge through a stream of oxygen source gas.
The concentration of ozone generated can be controlled by adjusting the voltage applied
to the corona tubes. The flow rate of gas through the generator can also be varied. Typical
flow rates used in the ozone generator range from 10 to 30 scfh (std. cubic feet per hour).
As the flow rate of the gas is decreased, the residence time in the corona tubes increases,
leading to higher conversion of oxygen to ozone, and consequently, higher ozone
concentration in the stream.

A Perkin Elmer Lambda 900 UV-VIS-NIR dual beam spectrometer was used to measure
the concentration of ozone. The outlet from the ozone generator was connected to the UV
treatment chamber and gas samples from the treatment chamber were transferred via a
glass nozzle and ozone resistant silicone tubing to a 1mm path length quartz flow cell
mounted in the spectrometer. Continuous flow of the sample gas through the flow cell
was achieved by connecting the outlet of the cell to a small vacuum pump. The

spectrometer and sampling setup is shown in Figure 2.9.

50

Absorbance

 

 

0.12---

I

 

 

 

 

0.00
200

2'50
Wavelength (nm)

Figure 2.8 Absorption spectrum of ozone.

5]

300

 

Sampling Nozzle

Figure 2.9 Perkin Elmer Lambda 900 UV-VIS-NIR Spectrometer with
quartz flow cell and sampling nozzle for continuous ozone measurement.

52

 

Ozo
corn

2111]].

Me
I; C]

Tar

8X15
COIL
111 1;
1111.15
fen.-

flou

Ozone concentrations were measured at room temperature. The concentration can be
corrected for deviation of temperature and pressure from the standard conditions (0°C, 1

atm. pressure) such that :

A T p,
c- X (2.4)

’ 23 f, '17
where

To and P0 = standard temperature and pressure (0°C, 1 atm.)

T and P = experimental temperature and pressure.

The vacuum in the flow cell was found to be less than 1 psi and could not be measured
accurately, hence the small pressure correction term was ignored. Since ozone is a very
reactive gas and cannot be stored without undergoing thermal decomposition [6], no
external ozone standard was available for experimental calibration. Therefore, the ozone
concentrations reported here should be considered approximate values. As will be shown
in later chapters, the exact ozone concentration is not a critical process parameter and
thus the approximate concentrations obtained were considered to be acceptable. Ozone
concentration was measured at different combinations of the generator voltage and gas

flow rates as shown in Figure 2.10.

53

AEf-flfi ‘c-tuhicqcccc RIF-Che

 

Ozone Concentration (ppm)

8000

 

 

 

 

 

 

 

 

 

.J
----- I “
7000 ----~ ——— — -— — — —— — ———',-.————~—" w —————
6000 — - - v -‘ f -—
. __..o
5000 - - 7: jar-2'9
4000 - e .1' ,,mi_ F7...»-
3000 E.-. _ '30" w" 'I o ——
2000 4 # Hf
-x'fi I
1000 —-~—— - xii-'5"
O I I Mih== t t l t t t
0 10 20 30 40 50 60 70 80 90 100
Generator Voltage (% of 110V)
--I--30 scfhflow --o--20 scfhflow "*"10 scfh flow

 

 

 

Figure 2.10 Ozone concentration in the UV treatment
chamber at different voltages and gas flow rates.

54

2.4 Surface
Contact an;
Ilhen a sew
contracts an
drop malte~
angle tilt as
The contact

Young's Er

Where
IS] : 5]

'r. .
ISL : I

III : t
i 2 p
The quaint

Mid Sur

9

“Mod”

2.4 Surface Energy Measurement

Contact angles are a thermodynamic measure of the wettability of a solid-liquid system.
When a sessile drop of liquid is placed on a solid surface, the drop of liquid spreads or
contracts and eventually the solid-liquid-vapor system reaches equilibrium. The angle the
drop makes with the solid at the point of contact at equilibrium is termed the contact
angle (0) as shown in Figure 2.11.

The contact angle is related to the surface and interfacial energy of the solid and liquid by

Young’s Equation [7]:

c056 = MI:- (2.5)
7Lv
where
'st = surface tension of the solid in equilibrium with vapor
731. = interfacial tension of the solid in equilibrium with liquid
YLV = surface tension of the liquid in equilibrium with vapor
0 = contact angle in degrees

The quantity (st-YSL) is the difference in energy between the free solid surface and the
solid surface wetted by the liquid. This provides the driving force for the liquid to cover
a solid surface. The total energy gained by the system when the liquid covers a unit area
of the solid surface is given by the term {yLv + ('st-YSLN which is termed the

thermodynamic work of adhesion (W SLV) [7].

55

adhc

Fig

Young’s equation can be expressed in terms of the thermodynamic work of

adhesion to give the Young-Dupré equation :

WSLV = 71y (1 '1' C05 6) (2.6)
Thus, contact angles can be used as a measure of the thermodynamic work of adhesion
for a given solid-liquid pair. Contact angles can vary from 0° to 180°. For a contact angle
of 0°, the liquid is said to spread on the solid. This represents the ideal state where a
liquid has complete affinity for the solid surface and covers it spontaneously. For contact
angles higher than 90°, the liquid is said to be non-wetting and presents practical
problems for adhesion due to incomplete contact between the two phases. It can be seen
from Young's equation that an increase in the surface free energy of the solid (st) leads

to lower contact angles with the liquid.

 

Vapor (V)

C V ‘1‘ o'-' v -.
waldo-19.7751???

Afi’fifr’fi’h’éV‘fiffih

’YLV

3V

 

'st e ISL>
Solid (S)

 

 

 

 

 

Figure 2.11 Contact angle (0) at the solid-liquid interface and surface tension.

56

l

 

[he surf
interacti
tontpon

respecti‘

The inte

no pha~

The ahov

10 IIIE ddt

The polar

Ctlllldct Ill

ll

“5 eQUt

The surface tension (or surface energy) can be expressed in terms of polar and dispersive
interactions. The total surface tension is expressed in terms of dispersive and polar
components where superscripts d and p refer to dispersive and polar components

respectively :

r = 7" + r” (2.7)
The interfacial tension between two phases can be related to the surface tension of the

two phases by a geometric mean mode] [8] :

1 1
75L = 75V +7’LV "2(7fsj7gIA ”(75704 (2'8)

The above relationship can also be expressed in terms of the work of adhesion according

to the additive model suggested by Fowkes [9],

1+cos6)=Wd +Wp + (2-9)

WSL=7LV( SL SL

The polar-dispersive model in terms of Young's equation gives a relationship between the

contact angle and polar-dispersive components of surface tension [24]:

triathlon)”

7LV

c086=—1+ (2.10)

 

This equation is significant as it allows the measurement of the polar and dispersive
character of the surface. With two or more liquids of known polar and dispersive
components of surface tension, contact angle measurements can be used to determine the

polar and dispersive interactions of the solid surface. Good and van Oss [10] further

suggested that the polar term (79) was not a physically relevant property and the major

57

 

CDI

estin
for a
rneth
cemb

al'eldp

‘IXIUS:
d] tylpi-
c It

components of surface energy can be separated into two categories: i) Lifshitz-van der
Waals forces (7 LW) and ii) forces from Lewis acid (7 +) and Lewis base (7') pair-wise

interactions (electron donor and electron acceptor components) [11] such that:

5L zyLV (1+coso) WSLLW +WALB= 2[,/7§Wyfw +,/yS;/L +,/ySyL] (2.11)

W

Surface treatment of polymers leads to the formation of polar surface functionalities, and
the changes in the polar, dispersive, and acid-base interactions can be determined with
contact angle measurements with three or more liquids. In this study five liquids were
used for surface energy analysis. The liquids used and their surface energies are given in
Table 2.3. The use of five liquids provides five simultaneous equations with three
variables. The parameters were calculated using a sum of least squares approach
(Appendix A), a method analogous to the graphical or linear regression analysis used to
estimate polar and dispersive parameters. This has the advantage of finding the best fit
for all data in three dimensions. This method is recommended by Good [11] over the
method of data reduction involving the calculation of acid-base parameters for multiple
combinations of data for three liquids from the set of five liquids and reporting the
average of all combinations.

Contact angles were measured on a Kruss Drop Shape Analysis System 10 Mk.2
(Kruss, Germany) shown in Figure 2.12. The instrument has video capture capability to
digitize drop shapes. Image analysis was performed using the Drop Shape Analysis
software provided with the system. A manual goniometer (Rame-Hart) was also used to

measure contact angles.

58

Table 2.3 Surface energies of liquids used for contact angle measurement.

 

Liquid ,YLTOtaI ,YLLW ,YL(+) (-)

Yr. YLD YLP

 

Water 72.80 21.80 25.50 25.50 21.80 51.00

 

Glycerol 64.00 34.00 3.92 57.40 34.00 30.00

 

Ethylene Glycol 48.00 29.00 1.92 47.00 29.00 19.00

 

Formamide 58.00 39.00 2.28 39.60 39.00 19.00

 

Diiodomethane 50.80 50.80 0.00 0.00 50.80 0.00

 

 

 

 

 

 

 

 

 

 

Figure 2.12 Kruss Drop Shape Analysis System.

59

he
Ln

Pneu

sunh.
repre-
sunp
tesnn
bendi
the a

peso

W} I
[61152:]:
16515 If
On a I

and q

2.5 Adhesion Measurements
Stub-pull tensile tests (ASTM D4541) were used to measure adhesive bond strength. A
Pneumatic Adhesion Tensile Testing Instrument (PATTI, M.E. Taylor Engineering) was
used to measure the pull-off strength of an aluminum stub adhesively bonded to treated
surfaces with a structural epoxy (Araldite 2015, Vantico). Figure 2.13 is a schematic
representation of the PATTI test configuration. In the case of thin film or flexible
samples, the entire sample is first mounted on a rigid metallic or wooden base prior to
testing to prevent sample bending during tensile testing. This is necessary because any
bending deformation in the sample can lead to the generation of strong peeling forces at
the adhesive interface. Tensile pull-off strength is measured by applying pneumatic
pressure and the failure load is measured. Failure stress is calculated from the peak load.
An alternate stub-shear adhesion test was also developed to measure adhesion on
very thin polycarbonate films which can detach from the stiff base material easily during
tensile testing. The sample configuration is identical to that described for tensile stub-pull
tests but the loading is in pure shear as shown in Figure 2.14. Shear tests were performed
on a UTS mechanical testing machine with a strain rate of 0.05 inch/minute. Failure load
and stress were measured. The tested surfaces were examined to determine the locus of
failure. The failure mode can be either interfacial or substrate. Interfacial failure is seen in
cases where the interface between the sample and the adhesive is weak, while substrate
failure is generally seen when the interfacial adhesion is very high compared to the
strength of the bulk substrate. Examination of the failure surfaces by XPS was used to

determine the locus of failure.

60

 

 

 

 

 

 

Pulling Force

 

 

 

 

 

 

 

 

 

 

 

1 Self-aligning
E. r’ Bearing Plate
E
:1
Pressurized E _
Gas {at 6331‘” Gasket
[ AH Adhesive
I \Test Sample
' \
Pull Stub

Figure 2.13 Schematic representation of the tensile stub-pull (PATTI) test

 

 

Aluminum
Backing

\

 

R\\\\\\\\\\\\\\\\\\\\\\\V

 

 

Shear Load

I

 

Aluminum Stub

 

 

\ Epoxy

Adhesive

\ Polymer Film

 

 

 

Figure 2.14 Schematic representation of the stub-shear adhesion test

61

'Iv')

EqL

2.6 X-ray Photoelectron Spectroscopy (XPS) Surface Analysis
XPS surface analysis is based on the principle of photoemission of electrons to obtain
quantitative and qualitative information about the atomic composition and chemical
structure of the surface of a material [12]. Photons from an X-ray source are bombarded
onto a surface in an ultra high vacuum environment, causing emission of core and
valence photoelectrons which are measured by an analyzer. The process is very surface
sensitive because although the absorption depth of the incident X-rays is large, the
emitted photoelectrons have a short mean free path and the photoelectrons emitted from
the bulk of the material quickly lose their kinetic energy due to inelastic scattering and
cannot escape the material. Only electrons near the surface of the material are able to
escape and reach the detector. The mean free path of emitted photoelectrons in a solid is
given by Equation 2.12.

_ E

_ a(ln E + b)

 

(2.12)

where E is the energy of the incident photon, a and b are parameters related to the
concentrations of valence and core electrons in the material. Ninety five percent of the
emitted photoelectrons reaching the detector emerge from within a depth of 3}» from the
surface. Typically this results in limiting the maximum sampling depth of XPS to 75-100
A from the surface.

The binding energy (E3) of a photoemitted electron is given by the following

equafion.

EB=hV-EK—¢Spec (2.13)

62

_

tt'hcre Bit 5 l
photon. and
emitted elcc'
energy of ti
with t'artttu
compostttttt

spectrum 0
lite area u:

intensity of

Where
L: X~ray fl
‘1: : COnL‘em
6" : phOIOlO
D431) : I

Flint

’7 .

where Ex is the kinetic energy of the emitted electron, hv is the energy of the incident
photon, and (itSpec is the work function of the spectrometer. The kinetic energy of the
emitted electron is measured by the analyzer allowing the determination of the binding
energy of the electrons. Each element has a unique set of binding energies associated
with various electron orbitals which allows accurate determination of the elemental
composition of a surface from its XPS spectrum. Figure 2.15 shows the typical XPS
spectrum of a polycarbonate film. Peaks for carbon (Cls) and oxygen (015) are labeled.
The area under the peak is proportional to the atomic concentration of the element. The

intensity of the signal for an element i (1,) is given by
11 = IoniGiD(8t )Xi (Si) (2.14)
where
10 = X-ray flux,
m = concentration of element 1',
0', = photoionization cross-section of element i,
D(8i) = transmission function of the analyzer, and

kit-Ii) = mean free path of the emitted electron.

It is difficult to measure 10 and D(£,) directly hence absolute atomic
concentrations are seldom measured by XPS. However, if the relative concentration, C,
of two elements (A and B) is calculated using Equation 2.15, the dependence on x-ray

flux is eliminated and only the kinetic energy dependence of D03) is retained.

63

C/S

 

2.2 I I I l l l l I l l

-C1s

1.6—

1.4~

c/s

-01s

 

 

0.8 — WW'WVVMWK, L Mt
. WM“ 1"“ 5E

 

 

 

 

 

0‘4 1 l l l l l l l l l
1 100 1000 900 800 700 600 500 400 300 200 100
Binding Energy (eV)

Figure 2.15 XPS survey scan showing elemental
carbon and oxygen peaks in polycarbonate.

Hydrocarbon

Overall Peak

 

Envelope
0 C — 0
ll (hydroxyl/ether)
O — C—O
(carbonate)

—-=—_€’
wr—

 

 

l l l l l l l
294 292 290 288 286 284 282
Binding Energy (eV)
Figure 2.16 Deconvolution of the Cls peak envelope in polycarbonate showing

multiple peaks corresponding to carbon atoms with increasing number of
bonds with oxygen.

65

 

a _LA_ 03138;“
CB 13 (IA/1,4835 (2.15)

The term in the right bracket in Equation 2.15 is called the sensitivity factor of the
element. Sensitivity factors can be calculated theoretically using cross-sections
determined by Schofield [13] and the mean free paths [14]. The mean free paths are
dependent on the material and the 8 term depends on the instrument. Sensitivity factors
can also be determined experimentally by using standard analytical specimen of known
compositions. Thus the relative atomic concentrations and atomic ratios can be easily
calculated for a given specimen.

Information about the chemical environment of the atom can also be determined
by XPS. As the electronegativity of the neighbouring atoms increases, electrons in the
analyzed orbital are more tightly bound and the binding energy increases. This is seen as
a shift in the peak position in the XPS spectra. In most materials, where various chemical
states of the same element exist, a series of peaks are obtained for an orbital. Figure 2.16
shows the peak for the carbon ls orbital (Cls) for polycarbonate. Polycarbonate has
carbon atoms in three different bonding states with carbon-carbon bonds (C-C, CzC),
carbon bonded to a single oxygen atom (C-O), and the carbon atom in the carbonate
linkage with four bonds with oxygen atoms (O-CO-O). As the number of electrons shared
with oxygen increases, the binding energy of the Cls orbital increases by ca. 1.5 eV per
bond [15]. Figure 2.16 shows the overall Cls peak envelope and the deconvoluted

Gaussian-Lorentzian statistical fit showing the constituent peaks. The areas of these

66

peaks are proportional to the number of atoms in that particular bonding state. This
allows indirect determination of the chemical structure of the material surface.

A Perkin Elmer Phi 5400 ESCA system was used for experimentation with a Magnesium
Ka x-ray source. Samples were analyzed at pressures between 10'9 and 10'8 torr with a

pass energy of 29.35 eV and 45° take-off angle unless stated otherwise.

2.7 Nanoindentation Tests

Nanoindentation tests were developed for the purpose of probing the mechanical
properties of very small volumes of materials. It is an ideal technique for the
characterization of thin films, coatings, and surface layers. The advantage of
nanoindentation tests is that material properties in the top 1-2 pm of the substrate can be
measured as a function of depth. A MTS Nanoindenter (MTS Systems Corp.) was used to
probe the elastic modulus (E) and hardness (H) of UV treated surfaces.

A nanoindentation test consists of three main steps. An indenter is pushed into the
material surface causing elastic and plastic deformation in the material up to a pre-
determined contact depth, hc. The indenter is held at the indentation depth for a period of
time with a constant indenter load. The indenter is subsequently withdrawn and the
elastic deformation in the material is recovered. It is the elastic recovery which allows the
determination of the elastic properties of the surface layers.

Figure 2.17 shows a typical loading and unloading cycle in the nanoindenter. The
important quantities are the peak load and displacement (Pmax and hum), the residual

depth after unloading (bf), and the slope of the initial portion of the unloading curve (S)

67

Pt

Pt

Ft

also known as the elastic stiffness of contact. The hardness of the test surface (H) is

determined by equation 2.16 [16].
P
H = — 2.16
A ( )

where P is the applied load and A is the projected contact area at that load. The elastic

modulus of the surface is determined from the reduced modulus (E,):

M (2.17)

Er = zflJX

where b is a constant depending on the geometry of the indenter. The elastic modulus of
the material surface (E5) is calculated using the expression:

(Hz) (13.2)
1 s l

E E E.
r s z

 

where E, and V, are the elastic modulus and Poisson’s ratio of the indenter and v5 is the
Poisson’s ratio of the test material. The indenter used for testing was a diamond
Berkovich pyramidal tip with B = 1.034, E, = 1141 GPa, and Poisson’s ratio = 0.07.
Poisson’s ratios for most polymers range between 0.25 and 0.35.

For testing of polymers, which can exhibit large plastic deformations, a series of 36
indents were made in a 6 x 6 matrix with horizontal and vertical spacing of 50 um
between indents. Indents were made to a depth of 2 mm. For UV treated polycarbonate,

significant changes in modulus were observed in the first 500 nm of the surface.

68

 

 

Load

Residual
impression

 

 

 

 

 

Load

Loading Slope =
Elastic stiffness

Unloading

 

 

- +
Displacement

Figure 2.17 Typical nanoindentation load displacement
curve showing loading and unloading behavior.

69

.569; :'- .

./.‘>r‘vv‘_nr- A-Al-
- '(kf"\i~‘l‘ >1 ;: ”31" '5 ’S‘fl‘lr/q‘bh‘?

”3:5: germs...

-. ya: 24";

fl.
«a;

.A I
33,3 r

E
.

fiftfeé'(¢:¢?e>2 012’me ‘

,- mew

(Ah/Av
I, _ .
at)? 'v (

~' 222-irate:

lf‘t

2.8 Summary

The UV lamps used for surface treatment of polymers in this study and the measurement
of their output has been discussed in this chapter. Insufficient characterization of the lamp
outputs in many published works in the area of UV surface treatment has limited a
thorough characterization of the process in the past. The biggest advantage of UV surface
treatment is that the process can be run under ambient conditions. However, ambient
conditions present severe problems for measurement of lamp output, especially in the
UV-C region, because of the absorption of radiation by air. The use of vacuum or noble
gas purges during radiation measurement is also not ideal because the measured
irradiances can be very different from the irradiance received by samples under treatment
conditions where oxygen and ozone are both present in relatively high local
concentrations. The difference in irradiance under oxygen and ozone purge in the sample
chamber for identical lamp configurations is significant. Using the output at a single
wavelength (254 nm) to draw conclusions about the entire spectral output of the lamp is a
simplification, but in the absence of a full spectrum N IST calibration standard UV lamp,
it is a reasonable strategy which can be used to make relative comparisons between
different process conditions.

In addition to the experimental methods described in this chapter, UV treated
polycarbonate was also characterized by attenuated total reflection Fourier transform
infrared spectroscopy (ATR-FTIR), UV spectroscopy, and electron paramagnetic
resonance spectroscopy (EPR) to examine the changes occurring on the surface due to

UV oxidation. These techniques will be discussed briefly in later chapters.

70

10.

11.

l2.

13.

14.

15.

16.

References

Rabek, J .F. Experimental Methods in Photochemistry and Photophysics Part I ,
Wiley-Interscience, New York, 1982.

Walzak, M.J.; Flynn, S.; Foerch, R.; Hill, J .M.; Karbashewski, B; Lin, A.; Strobel,
M. Polymer Surface Modification; Relevance to Adhesion, K.L. Mittal (Ed.), pp. 253-
272, 1995.

Lothian G., Absorption Spectrophotometry, 3rd Ed., pp. 101, Hilger, London, 1969.

Jaffe H. and Orchin M., Theory and Applications of Ultraviolet Spectroscopy, pp.8-
10, Wiley, New York, 1962.

ASTM D5110-94.

Watanabe, F.; Shimomura, 0.; Journal of Hygienic Chemistry, Vol. 37, No. 5, pp.
323-331, 1991.

Adamson, A.W. Physical Chemistry of Surfaces, 5'h ed., Wiley Interscience, New
York, 1990.

Kaelble, D.H.; Uy, K.C. J. Adhesion, Vol.2, p.50, 1970.

Fowkes, F.M.; McCarthy, D.C.; Mostafa, M.A. J. Colloid Interface Sci., Vol. 78, p.
200, 1980.

VanOss, C.J.; Good, R.J.; Chaudhury, M.K. J. Colloid Interface Sci., Vol. 111, p.
378, 1986.

Good, R.J. J. Adhesion Sci. Technol., Vol. 6, No. 12, pp. 1269-1302, 1992.
Vickerman, J .C. Surface Analysis — The Principal Techniques, Wiley, UK, 2000.
Scofield, J.H. J. Electron Spect..°Relat. Phenom, Vol. 8, p. 129, 1979.

Seah, M.P.; Dench, W.A. Surf. Interface Anal, Vol. 1, p. 2, 1979.

Briggs D.; Seah, M.P. Practical Surface Analysis, Wiley, New York, 1983.

MTS Systems Corp, Nanoindenter XP user guide.

71

CHAPTER THREE

SURFACE MODIFICATION OF POLYCARBONATE

3.1 Description of Polycarbonate

The polymer chosen for this study was a commercial grade bisphenol-A based
polycarbonate (GE Plastics, GE8040). The material is available as an extruded film of
175 um thickness and is packaged with a protective polymer (polyolefinic) film on both
sides. One of the protective films adheres to the polycarbonate (PC) film by electrostatic
attraction while the film on the other side of the polycarbonate adheres with a pressure
sensitive adhesive. Although XPS analysis of the side protected by pressure sensitive
adhesive film did not show any evidence of transfer of the adhesive to the polycarbonate
film, only the side protected by the electrostatically attached film was used in all
experiments to avoid the possibility of artifacts. Figure 3.1 shows the chemical structure
of bisphenol-A polycarbonate [1]. The molecular weight of the polycarbonate used is ca.
20,000 Dalton. This chapter describes the changes occurring on the polycarbonate surface

as a result of UV induced oxidation.

(EH3 <3 1
Woo-—
a.
— — n

 

 

Figure 3.1 Chemical structure of polycarbonate

72

3.2 Process Parameters

UV/Ozone (UVO) surface treatment depends on a variety of process variables
including: irradiance, exposure time, ozone concentration, temperature, and humidity. In
addition, the changes occurring on the surface are strongly dependent on external mass
transfer, nature and surface chemistry of the substrate, and surface chemical reactions on
the surface. Understanding the process involves a systematic study of these process
parameters. In this section, results of studies on the effects of mass transfer, ozone
concentration, irradiance, and exposure time will be presented. The effect of various
process variables was determined by contact angle measurements which are known to be
sensitive to monolayer level changes in surface chemistry and topography and are

relatable to adhesion.

3.2.1 Flow Rate

The reaction of ozone with the UV irradiated surfaces is a gas-solid reaction and
the overall rate of reaction depends on the relative rates of the surface chemical reaction,
and the mass transfer of reactive species to the surface. The slowest step in the process is
the rate determining step and for a given chemical reaction, the rate constants cannot be
changed easily. The external mass transfer in a gas-solid surface reaction, a measure of
the ability to transport reactants to the reaction sites on the surface, is a more easily
modified parameter. If the system is mass transfer limited then the overall rate of the
reaction is dependent on the mass transfer coefficient and an increase in mass transfer
results in an increase in the total reaction rate to the limit allowed by the reaction kinetics.

To determine if mass transfer limitations existed in the experimental setup, contact angles

73

were measured on samples treated at different flow rates. Flow rates of 10, 20 and 30 scfh
(standard cubic feet per hour) were used while keeping the ozone concentration and UV
treatment time constant. When the flow rate is decreased, the gas velocity decreases and
the mass transfer coefficient decreases. Any mass transfer limitation in this case would be
seen as a reduction in the treatment efficiency and higher than normal contact angles for a
set of given treatment conditions as the transport of ozone to the surface is hindered by
the slow flow rate.

Figure 3.2 shows the variation of contact angles as the flow rate is changed by a
factor of three for treatment times ranging from 0 to 90 seconds UV exposure. The
contact angles for untreated polycarbonate were found to be approximately 90°. After
UVO oxidation, the contact angles decrease sharply for treatments as short as 10 seconds.
As the treatment time is increased, the rate of change of contact angles decreases and
angles of less than 20° are obtained after treatments of 90 to 120 seconds. Further
changes in contact angles are difficult to measure accurately. Contact angles measured
for treatments at all three flow rates were found to be identical within the limits of
experimental error and leads to the conclusion that no mass transfer limitation is present
when flow rates on the order of 10-30 scfh are used. All further experiments in this
chapter were performed at a flow rate of 30 scfh to ensure no external mass transfer

limitations were present.

74

 

 

Contact Angle (deg)

 

 

 

 

 

Control 10 sec 20 sec 30 sec
UV Treatment

Figure 3.2 Contact angles of UV treated PC for ozone flow rates of 10, 20, and 30

 

 

I30 SCfl'l

IZOscfh

EllOscih

 

 

 

scfh indicating no mass transfer limitations

75

 

 

 

3.2.2 Ozone Concentration

The Ozone concentration is a potentially important process variable. Depending
on the kinetics of the surface reaction between ozone and the UV treated surface, the
reaction can be strongly dependent on the concentration of ozone on the surface. To
determine the influence of ozone concentration on the UV treatment process,
polycarbonate samples were treated at various ozone concentrations at identical operating
conditions of irradiance and exposure time. Sessile drop equilibrium contact angles were
measured at various locations on multiple samples as shown in Figure 3.3.

It was found that ozone was necessary for UV treatment, but the concentration of
ozone did not have any effect on the quality of treatment achieved. Without additional
ozone, contact angles between 70-80° were measured for most samples after a 30 second
UV exposure. With 400 to 800 ppm ozone in the treatment chamber, contact angles of
30° were obtained for the same exposure. With decreasing ozone concentration, the data
shows a large amount of scatter in the individual measurements. However, the lowest
value of contact angles measured (shown by horizontal line) was ca. 30° for all samples
irrespective of the ozone concentration. This was observed to be true even at the lower
limit (ca. 10 ppm) of our capability to measure ozone concentration accurately. In terms
of process parameters, UV treatment of PC can be considered to be independent of ozone
concentration because, for a surface reaction, ideally only a small amount of ozone is
needed to form a monolayer and obtain full surface coverage. The scatter in data at low
ozone concentrations is suggested to be a result of poor sample coverage, starvation of
ozone near some parts of the sample due to inadequate mixing of ozone in the treatment

chamber, or non-uniform air flow patterns and channeling in the UV treatment chamber.

76

At high concentrations, the mixing and diffusion rate of ozone inside the UV treatment
chamber is increased and the scatter decreases. This is likely an artifact of the treatment
chamber design and not the process itself. To avoid data scatter, ozone concentrations in
the range of 700-800 ppm were used for all experimentation and generally very low

errors in measured contact angles were observed at this concentration.

90
80
70 i

50 7

40 T.
C

30 ,n_._' 4

20 e

10 e

O l l l 1
O 200 400 600 800

Ozone conc (ppm)

 

Contact Angle (deg)

 

 

Figure 3.3 Contact angles of PC after 30 sec UVO treatment as a function of ozone
concentration. Horizontal line indicates the lowest angles measured at the lower
limit of measured ozone concentrations.

77

3.2.3 Irradiance

Irradiance is the most important variable in the UV treatment process. The
irradiance received by the sample determines the extent of modification possible on the
surface. As described in the previous chapter, irradiance levels were measured and varied
by controlling the distance between the lamp and sample (however, the path length of
ozone the radiation travels through remains constant). Polycarbonate film was exposed to
UV radiation from the RC-747 lamp at distances ranging from 1 to 5 inches in the
presence of approximately 700 to 800 ppm supplemental ozone concentration and a flow
rate of 30 scfh. Contact angles of the UV modified surfaces with deionized water are
shown in Figure 3.4.

As the treatment time is increased, a reduction in contact angles is observed till
the contact angles reach 20° and further reduction in contact angles is difficult to measure
accurately as the liquid spreads on the surface. At treatment distances over 3 inches
(irradiance < 2.6 mJ/mz) the rate of change of contact angles is almost linear. As the
treatment distance is decreased, the increase in irradiance causes the rate of change to be
increasingly non-linear. It is of particular interest to relate the rate of change of contact
angles with an absolute process parameter such as UV irradiance. From observations of
the trends in contact angles, it was found that it is possible to super-position the work of
adhesion curves for different combinations of irradiance and exposure time. This is
similar to the concept of time-temperature super-positioning in polymer creep properties
[2]. The time-irradiance superpositioning of the thermodynamic work of adhesion
(proportional to the cosine of the contact angle) can be expressed as a function of the total

irradiation incident on the surface. Figure 3.4 shows the work of adhesion for deionized

78

water on UV treated polycarbonate treated at various combinations of distances and UV
exposure times. The irradiance was varied by a factor of five and exposure times of 20 to
150 seconds were used to generate work of adhesion data. All of the data in Figure 3.4
can be reduced to a universal curve shown in Figure 3.5. To test the robustness of this
assumption, work of adhesion data was also collected using the low power RC-SOO lamp
to create irradiance levels differing by almost an order of magnitude compared to the RC-
747 lamp. The data from both lamps follow the universal wettability curve for
polycarbonate shown in Figure 3.5. This trend was seen to exist in other polymers as well
and will be discussed in Chapter 5.

The construction of a universal curve for the change in wettability as a function of
irradiation is significant in terms of developing a process model. One of the implications
is that time ceases to be a controlling variable in the UV treatment process and a specific
targetted level of wettability can be achieved by a variety of combinations of irradiance
and time. Short treatments at high irradiances, and long treatments at low irradiances can
yield the same work of adhesion. However, there can be qualitative differences between
treatments at high and low irradiances. Surface modification by UV depends on the
ability of chromophores in the material to absorb radiation, which is necessary to start a
chain of complex reactions. Absorption in the material follows Beer’s law [3] and the
intensity of radiation at any point in the thickness of the material depends on the intensity
of the incident radiation. Consequently, as irrandiance levels are increased to reduce
exposure times, the depth at which chemical changes occur in the material can potentially
increase and vice-versa. The penetration depth of radiation can thus become an important

consideration in choosing combinations of irradiance and time to suit a given application.

79

Contact Angle (deg)

 

 

 

 

 

100

+ w -- 1 inch (7.458 mW/sq.cm.)
90 -Ei--2inch (3.508 mW/sq.cm.)
80 4 1‘ "war-- 3 inch (2.662 mW/sq.cm)
704 g Xx. --C--4inch(1.899 mW/sq.cm.)
60 d «‘1‘ \I‘g...‘ ~. --->l(--5mch(1.618 mW/sq.cm)

n“ ‘.\:~ X .....

50 ‘ i‘. .2! . ' "x...
40— . .. ~.
3Ofi ~.-.‘~ Q . .~‘~..~ ~~~§
20“ fl! :3:::::,a .-‘-:E g
10 —
O l l l l

0 30 60 90 120 150

Time(sec)

Figure 3.4 Equilibrium contact angles of deionized water on UV treated
polycarbonate as a function of time at 1 to 5 inches from the UV source.

80

 

3.329;.le h: v-Lqug

 

150

 

 

 

 

 

 

14o — my? ......... 39"”? ....................................... «9
I30 ~ i #§ é.
a 120 ~ ¥ .
2 ,.. a
.5 E, 110 ~ §
< 3 g.
f; E 100 ~
‘5 " if
3 90 _. .
: O RC747 lamp (1500 W)
80 4 ORCSOO lamp (300 W)
70
60 W T f f T f T,
0 100 200 300 400 500 600 700

Irradiation (mJ/sq.cm)

Figure 3.5 Work of adhesion of UV treated polycarbonate with deionized water as
a function of irradiation for data obtained at various combinations of irradiant
powers and exposure times for multiple lamps.

81

3.3 Wettability and Surface Energy

Contact angles were measured with five liquids of varying acid-base and polar-
dispersive character to calculate the surface energy of modified polycarbonate using acid-
base and polar—dispersive models [4]. As polar oxygen containing functional groups are
incorporated on the polymer surface, the wettability and surface energy increases. The
previous section showed the dependence of the change in wettability on the irradiation
received by the polymer. The surface energy after UV treatment also exhibits an identical
trend. Samples exposed to the same levels of irradiation had the same surface energy and
polarity regardless of the level of irradiance and exposure time.

The change in the polar-dispersive components of surface energy for UVO treated
polycarbonate is shown in Figure 3.6. The untreated polycarbonate has a total surface
energy of 38.6 mJ/mz. Most of the surface energy is due to dispersive interactions and
only a small polar component of 0.2 mJ/m2 was measured. After UVO treatment the
dispersive van der Waals interactions remain almost constant while a sharp, linear
increase in the polar component is observed as a function of irradiation. The total surface
energy increases to over 60 mJ/m2 after irradiation of 300 mJ/cm2 in ozone.

Similar trends are seen from results of the acid-base analysis shown in Figure 3.7.
The surface energy increases from approximately 38 mJ/m2 to approximately 55 mJ/m2
after UVO treatment. The Lifshitz-van der Waals (39-41 mJ/mz) and acid (0.5-1 mJ/mz)
contributions remain constant and most of the increase in surface energy is from a sharp
increase in the basic component from 2 mJ/m2 to 60 mJ/mz. The increase in the basic
character is due to the addition of hydroxyl, carbonyl, and carboxylate functional groups

on the surface.

82

70

 

 

 

 

 

 

 

 

 

60 """" .
,9, ‘
,ar‘
50~ e a ,A’ _ f i
~/
,/
2'6
Egmafé-p #0 o +- e
‘5 a. V/ 00 ————— o_-__<> o 00
’ o 0 ---------- Q_
8 & -~~a ......
a 3 ~-—)‘-.
t v 30 4 a 7 , , % ’/’ ~—-0*
: /
U) x/"
/,4
20 ~ - * ~ ‘3)" ~
9" . .
)’/- <>Dispers1ve
10 Cf.’ . OPolar -2
,/
a“. 0 Total Surface Energy
/" I
O * A T l
0 50 100 150 200 250 300
Irradiation (mJ/sq.cm.)

Figure 3.6 Polar-Dispersive surface energy of UV treated polycarbonate for various

combinations of irradiance and exposure time.

83

 

 

 

 

 

 

 

 

 

 

6O 70
"‘r 60
55‘ * ” " *' is“ f‘
' g,fi"’" 9 ””‘7'1\
. ””__~ l’ ‘.W 50
50 2° — ~ .»<—--- H ——
>. _____ __ __.°9__o_..<>——-e——-—°”-9’ ------- '0
g f? 000% 1;:3 1‘ w 40
5545-2 g -
g E '/ . )~ 7’7 30
a: V ’J . ’1’
a": 40 ~43; VA ’ .4
V’ ’3‘. OTotal Surface Energy 20
. J.” OBase
35 /, . OACid _.s
33’ OLW ” '0
0‘
A
30 ’ - "0 --°".-90-"7 """ - o
O 50 100 150 200 250 300
Irradiation (ml/sq.cm)

(mJ/sq.m.)

Acid, Base Components

Figure 3.7 Acid-Base surface energy of UV treated polycarbonate for various
combinations of irradiance and exposure time.

84

3.4 Molecular Spectroscopy

Chemical changes occurring in polycarbonate because of UV oxidation were
investigated by ATR-FTIR and UV spectroscopy. The two techniques are capable of
providing molecular information [5,6], but suffer from a lack of surface selectivity. ATR-
FTIR is more surface selective than transmission FT IR, but the sampling depth is on the
order of several hundred nanometers and resembles a bulk spectroscopic method in the
context of probing changes in surface properties which may be limited to the top few
hundred Angstroms of the material. However, changes in chemistry were observed and

are presented here.

3.4.1 ATR-FTIR Spectroscopy

Polycarbonate films were UV treated and analyzed by ATR-FTIR. Thin film
samples of PC spin-coated on KBr pellets were also evaluated for use in transmission
FTIR, but it was observed that UVO treatment of the films caused changes in the KBr
baseline, invalidating any further analysis. ATR-FI‘IR spectrum of PC is characterized by
characteristic absorption peaks at 1770 cm'l (C=O, carbonate stretch), 1620 cm'1 and
1500 cm'1 (C=C, benzene ring stretch), 1250 cm"1 (O—C-O carbonate asymmetrical
stretch), and 830 cm"1 (C-H, benzene out of plane bending) [7]. Figure 3.8 shows the
ATR-FTIR spectra for untreated, 90 sec UVO, and 180 sec UVO treated PC. The spectra
could not be normalized with respect to each other because of the lack of a stable
reference peak. The structural changes caused by UV oxidation are broad and it is
suggested that both bisphenol—A and carbonate linkages in PC are subject to UV

oxidative attack. The major features observed after surface modification include the

85

appearance of a broad hydroxyl peak between 3500 cm’1 and 3100 cm'l after UV
exposure and the broadening of the carbonate peak at 1700 cm". The broadening of the
carbonate peak and the appearance of a shoulder at 1690 cm'l are indicative of scission of
the carbonate group and the addition of substituents to the aromatic structures, including
rearrangement to form phenyl salicylates according to the mechanisms proposed in
literature for photo-Fries rearrangement of PC and the associated oxidative pathways
previously described in Chapter 1 [8]. Evidence is obtained from the relative intensities
of the carbonate peak at 1770 cm-1 and the aromatic peak at 1500 cm'1 shown in Table
3.1 where the intensity of the carbonate peak decreases and the intensity of the broad

hydroxyl peak increases compared to the aromatic phenyl peak.

Table 3.1 ATR-FI‘IR peak ratios of carbonate, hydroxyl, and
aromatic structures in PC after UVO treatment.

 

 

 

 

Ratio of peak intensities at Ratio of peak intensities at
Treatment 1770 cm'1 and 1500 cm'1 3460 cm'1 and 1500 cm'1
(Carbonate/Aromatic) (Hydroxyl/Aromatic)
Untreated 1 .64 0.04
90 sec UVO 1.40 0.13
180 sec UVO 1.36 0.15

 

 

 

 

 

86

 

 

 

 

 

w ##JKVUW

 

 

 

 

 

 

90 sec uvo LU ' l
‘ "W" #LWJ

Absorbance
1
g % “é.

 

 

 

 

 

 

at: 21.0, A 51.1.1

I 1 I I

4000 3700 3400 3100 2800 2500 2200 1900 1600 1300 1000 700
Wavenumber (cm -1)

Figure 3.8 ATR-FI‘IR spectra of PC after UV treatment
at 2.662 mJ/m2 irradiance and 700 ppm ozone

87

3.4.2 UV Spectroscopy
Polycarbonate samples were spincoated on quartz slides using tetrahydrofuran
(THF) as solvent and analyzed by transmission UV spectroscopy. Figure 3.9 shows the
UV spectra for untreated, UVA and UVO modified PC. The untreated samples have a flat
absorption profile till 280 nm and peaks characteristic of phenyl groups in the bisphenol-
A structure are seen in the 260-280 nm range. After exposure to UV in air and ozone,
broad peaks are formed in the 280-380 nm region and peaks in the 260-280 nm region
increase in intensity. The spectra of UV treated PC are similar to those reported in
literature for the degradation of polycarbonate by photo-Fries rearrangement and chain
scission [9-12]. The increase in intensity at 245 nm and 315 nm is attributed to the
formation of phenyl salicylate while the set of peaks between 260 - 280 nm indicate the
formation of phenols as the carbonate linkages are cleaved by UV radiation. The higher
amount of ozone present in UVO treatment compared to UVA treatment results in a
Slight reduction in the intensity of peaks of the degradation products indicating further
Oxidation of the phenolic end-products of UV degradation. It must, however, be
remembered that the spectra obtained are essentially bulk spectra of spincoated films
Several microns thick and are not surface selective. Subtle changes in the surface
Chemistry can easily be masked by stronger changes occurring in the bulk polymer.
Figure 3.10 shows the evolution of the UV absorbance profiles as a function of UV
CXposure. The spectra obtained at 30 second UV exposure intervals show the rate of
formation of products of photo-Fries rearrangement and degradation is maximum at the
start of the UV exposure cycle and gradually decreases up to 120 seconds exposure, after

which no significant changes are observed.

88

rr. C
SUCSCLCIC<

 

 

0.5

 

 

0-45 ‘ “x — Untreated
()4 — w -— l2OSec UV/Air
------ 1203ec UV/O3

 

 

 

0.35 —

 

 

.9
U.)
1

0.25 -

Absorbance

.9
N
1

0.15 e
0.1 e
0.05 4

0 r
220 240 260 280 300 320 340 360 380 400
Wavelength (nm)

 

 

 

 

Figure 3.9 UV spectra of UVO, UVA, and untreated PC showing UV induced
chain scission and the formation of phenolic groups (260-278 nm) and phenyl
salicylates (245 nm and 315 nm) in the polymer because of photofries
rearrangement.

89

Eh

 

Absorbance (A.U.)

 

 

 

 

 

 

 

 

 

320

Wavelength (nm)

Figure 3.10 UV spectra of PC spincoated on quartz after UV Treatment
at 2.662 mJ/m2 irradiance and 700 ppm ozone.

90

3.5 EPR Spectroscopy

Electron Pararnagnetic Resonance (EPR) spectroscopy is used to detect the
presence of free radicals or any electron with an unpaired spin. The UV activation and
subsequent reaction with oxygen (itself a paramagnetic molecule) leads to a high
concentration of unpaired spins in the polycarbonate. These unpaired electrons may serve
as further reaction sites. The presence of unpaired spins was detected and quantified
using room temperature EPR spectroscopy at a magnetic field of 3364 Gauss and a
microwave frequency of 9.46 GHz. Torikai et a1. [13-14] have reported similar spectra in
PC irradiated in vacuum and air and assigned the signal to phenoxy and phenyl radicals;
both products of chain scission in polycarbonate. EPR is not a very robust quantitative
technique for direct measurement of spin concentrations. Quantitative EPR requires the
use of a constant amount of sample with identical surface areas which is difficult to
achieve with solid samples. Although the amount of sample used for each spectra was
kept as uniform as possible, slight changes in sample location, alignment, and weight can
cause changes in the results obtained. All samples were analyzed at the same receiver
gain in order to make some quantitative comparisons between samples possible. The data
provided here is only intended to compare changes in spin concentrations differing by
large values.

Figure 3.11 shows the spectra of polycarbonate film exposed to ozone, 120 sec
UVA, and 120 sec UVO treatments. The y-axis shows the amplitude of the integrated
signal in the material. Spin concentrations are usually calculated by measuring the area
under the absorbance peak or by the product of the amplitude and FW HM (full width at

half maximum) of the peak, but here only the amplitudes will be compared directly as a

91

 

measure of the approximate spin concentrations because all analyzed samples had similar
peak widths.

PC exposed to ozone does not show any peaks and only a low background is
observed with a peak amplitude of less than 4000. After UV treatment the amplitude
increases to ca. 70,000 for UVA and 110,000 for UVO samples indicating either the
presence of free radicals or paramagnetic oxygen in the material bulk or surface.

In addition to increasing the surface energy, UV treatment also creates free
radicals on the surface that can react with applied paints and adhesives, forming chemical
bonds and improving adhesion. Thus it is of interest to determine if the free radicals
formed during UV treatment, and seen in the EPR spectra, are on the surface or in the
bulk. Free radicals in the bulk can neither react with the adhesive nor contribute to
surface energy and may be detrimental to the bulk properties of the material in much the
same way as long term weathering. An indication of the amount of free radicals on the
UV treated polymer surface which can react further was obtained by analyzing samples
of UV treated PC immersed in an epoxy resin. The resin used was a DGEBA based
epoxy (Shell, Epon 828) cured with triethylene tetramine (Jeffamine T403). UV treated
samples were immersed in the epoxy resin and allowed to cure for 24 hours before EPR
spectra was obtained. Baseline spectra of the epoxy and the treated polymer were also
obtained.

Figure 3.12 shows the EPR spectra of the cured epoxy, UVO treated PC and UVO
treated PC cured in epoxy resin. The neat epoxy does not show any paramagnetic
response. The background is very low and on the amplitude is approximately 2000. The

UVO treated polycarbonate film has an amplitude of ca. 110,000 and the treated film

92

cured in epoxy shows a much reduced amplitude of ca. 20,000. These changes are
significant despite the approximate nature of the comparison, and indicates that the
radicals which do not react with the ozone during treatment, or indeed the radicals or ions
which may be formed due to the reaction with ozone during treatment, are largely located
on the surface and are capable of reacting with an adhesive or paint applied to the

surface.

93

 

800

110]

400 e A

0 I
L .. .
-\ l ”U ‘ A W
V/\ \ 0 rd “3
-400 _ Va ~JLJ

 

 

 

 

 

 

 

 

 

 

 

 

— V
'800 1 l l l l l 1 1
3330 3350 3370 3390
[G]
mo 100 '-
h‘ /7
50 ~ ,/ \\
// \
,/' \
0 ‘—-‘”’ *—‘ ~‘.
-50 _
'100 —
3340 3360 3380 3400 3420
[G]
13 100 ~ /// \\\
50 ~ \
\
\‘H‘
0-**‘”“—’*”‘#—"_ *‘w‘ ~¥.—*_z
-50 _
*100 —
3330 3350 3370 3390

[G]

Figure 3.11 EPR spectra of PC exposed to ozone (top), 120 sec UVA treatment
(middle), and 120 sec UVO treatment (bottom).

94

[10]

[1031

[102]

400

200

-200

-400

100

50

 

if!

1

 

 

 

 

 

 

 

 

300

200*-

100

-100

-200 -

 

-300+-

 

I l l

 

3340

3360

3380

3400
[6]

Figure 3.12 EPR spectra of epoxy (top), 120 sec UVO treated PC (middle), and 120

95

sec UVO treated PC cured in epoxy (bottom).

3.6 Nanoindentation

Nanoindentation tests on UVO treated polycarbonate films were performed using
a Berkovich pyramidal tip to make 36 indents in the sample surface to a depth of 2000
mm. The spacing between indents was 50 um, and Poisson’s ratio of 0.35 was used for the
polycarbonate. Three samples were tested for each treatment condition and the data for
the each sample (36 indents) was averaged separately. Nanoindentation uses the load
displacement data obtained by pushing the diamond tip into the sample surface to
calculate surface modulus and hardness. On soft polymer samples this analysis is
complicated by the difficulty experienced by the tip in determining the topmost layer of
the surface. The surface is found by the instrument by increasing displacement till a pre-
determined load is experienced. Due to the arbitrary nature of the surface find segment,
properties measured very near the surface have large standard errors. Thus data obtained
in the top 20—25 nm of the surface is not reliable and should be considered as machine
noise. Figure 3.13 shows the surface modulus of UVO treated PC as a function of depth
up to 500 nm. In spite of the large standard deviations in the top 100 nm of the surface,
definite trends in modulus can be observed as a function of UVO treatment. The
untreated material has a modulus between 2.7-2.8 GPa in the t0p 100 nm which increases
to 2.9-3.0 GPa after UV treatment. The trend is clearer and becomes statistically
significant as the probe depth increases beyond 100 nm. The increase in modulus of the
polymer can be due to cross-linking of the surface layers under UV oxidation conditions.
Similar increases in modulus and hardness of physically aged polycarbonate have been

recently reported and attributed to changes in free volume and chain scission induced

crosslinking [15].

96

 

 

 

3.8

3.6 00schVOPC

3.4 012080cUVOPC
' . 305ccuvopc

 

 

9°
N

 

 

Modulus (GPa)
U)

 

 

 

 

 

 

1111mm:

2.6..“ !!§!!!!!
2.4,

2.2

0 100 2(1) 3(1) 400 500

Displaoermnt Into Surface (nm)

Figure 3.13 Modulus vs depth profiles of UVO treated polycarbonate measured by
nanoindentation tests. Higher modulus in the surface layers is observed with
increasing UVO exposure.

97

3.7 Adhesion

The adhesive performance of UV treated polycarbonate was measured using two
adhesion tests: tensile stub-pull and stub shear. Tensile tests used a pneumatic piston to
pull off an aluminum stub bonded to the PC film and the failure load was recorded. Shear

tests were performed on similar samples and the shear load-displacement curves were

recorded.

3.7.1 Tensile Stub-Pull Tests

UVO and UVA treated polycarbonate was bonded to 0.5 inch diameter aluminum
stubs with a two part epoxy adhesive (Vantico, Araldite 2011). Figures 3.14 and 3.15
show the peak adhesive strength for UVO and UVA samples respectively. The untreated
polycarbonate has a bond strength of ca. 300 psi. After UV treatment no statistically
significant change in the adhesive strength was observed for treatments up to 120
seconds. Increased adhesive bond strengths were expected from UV treated samples
because of the increase in work of adhesion and surface energy. The locus of failure
appeared to be at the polycarbonate-epoxy interface from visual observations, but
detailed fracture surface analysis, described in the following chapter, proved that the
failure occurs within the polycarbonate substrate and the adhesive bond strength values
reported here are not representative of the strength of the interface between the PC and
adhesive. Samples were also prepared using a two part polyurethane adhesive (Vantico,
Araldite 2040) in an effort to see if the failure mode changed for a different adhesive
chemistry. Figure 3.16 shows the adhesive strength of UV treated PC with polyurethane

and similar trends to the epoxy adhesive are observed.

98

“a"

 

é

(A)

O

o
r

 

Pull Off Tensile Strength (psi) .
N
8

 

 

 

' 0 15 30 45 60 90 120
UVO Treatment Time (sec)

Figure 3.14 Adhesive strength of UVO treated PC with epoxy adhesive.

§

 

§§§

 

N
1

3

Pull Off Tensile Strength (psi)

 

 

 

 

 

 

0 5 15 30 60 90 120
UVA Treatment Time (see)

Figure 3.15 Adhesive strength of UVA treated PC with epoxy adhesive.

99

 

    

 

 

 

800

W i 7
m m m m m m
6 5 4 3 2 1

$8 355 case to :3

120

20 30

UVO Treatment Time (sec)

10

Figure 3.16 Adhesive strength of UVO treated PC with polyurethane adhesive.

 

 

 

 

  

 

f ,. a , w.“
m m m m
4 3 2 1

3.3 5308 255a to :5

., . . EOE + O>D 08 03
an... EOE + O>D 08 co
. . :05 + O>D com Om
M. M., moam + O>D 0% cm

Scam + O>D com S
O>D 8m o2

O>D 09. co

O>D 8w Om

O>D 08 cm

O>D com S

EOE + 280 0% GS
E05 + onoNO 08 OM
58>? EOE

.. \. .. _ OCONO 00m ONm
I 285 02 cm

.. 3:8

Figure 3.17 Adhesive strength of ozonated and UVO treated polycarbonate before

and after ethanol wash to remove low molecular weight material.

100

It has been reported in literature that UV treatment of polymers can lead to the
fo:rmation of low molecular weight (LMW) fragments on the surface due to chain
sc ission [16,17]. LMW fragments are typically loosely bonded to the surface and can be
washed away with weak solvents such as water or ethanol. The presence of weak LMW
material at the interface can interfere with adhesion by creating a weak boundary layer
that fails under low loads. Adhesion test samples were prepared using UV treated PC
which was rinsed in ethanol prior to bonding. The effect of ozonation (no UV irradiation)
was also studied for unwashed and ethanol washed PC. Figure 3.17 shows the results for
ozonated and UVO treated samples before and after washing. As in the case of previous

results, no significant changes in adhesion were observed.

3.7.2 Stub-Shear Tests

One of the possible drawbacks of the tensile stub pull test for thin film samples is
the possibility of debonding between the untreated film surface and the rigid backing to
which the PC is attached. This can lead to the development of peeling forces which can
reduce bond strengths. Shear tests were performed to address this potential flaw. Shear
tests also allowed recording of full load-displacement curves for calculation of the
adhesive bond stiffness as well as peak failure loads. Figures 3.18 shows the bond
stiffness for UVO and UVA treated samples bonded with epoxy (Vantico, Araldite 201 1).
As in the case of tensile tests, no statistically significant changes in adhesion bond
stiffness and peak failure loads were observed. This is an unexpected result and an
explanation will be proposed in the following chapter based on XPS failure analysis of

adhesion test fracture surfaces.

101

T—

 

Shear Adhesive Bond Stiffness (kpsi)

140

 

120 -

100-

80~

40_,

20~

 

           

Y‘ Y"
493 03

N 0

0°
0° "5° 6° 09°” '5

Figure 3.18 Slopes of load displacement curves for stub shear adhesion tests.

102

 

 

3.8 Summary

The surface treatment of polycarbonate by UV oxidation was characterized by
wettability and surface energy measurements. UV exposure in ozone and air was found to
impart strong hydrophilic nature to the normally hydrophobic PC surface. UV process
variables such as the ozone flow rate were carefully chosen to avoid mass transfer
limitations and under these conditions, it was found that the UV modification was
independent of the ozone concentration in the treatment environment. The most
important process parameter was found to be the UV irradiance. The changes occurring
in the polymer are strongly dependent on the total irradiant energy, or irradiation,
incident on the surface, which makes the UV treatment process independent of the
exposure time. Short exposures at high irradiances and long exposures at low irradiances
were found to yield identical surface properties (wettability, work of adhesion, and
surface energy) when irradiances were varied from 1.6 to 7.5 mW/cmz.

The molecular changes on the surface were probed by ATR-FT IR and UV
spectroscopic techniques. The results show that UV exposure leads to the photo-Fries
rearrangement of the bisphenol A carbonate to form phenyl salicylate and di-hydroxy
benzophenones. The products of UV degradation on the surface can undergo oxidation by
ozone to form hydroxyl, carbonyl, and carboxylate functional groups. The net result of
this oxidation is the formation of conjugate bases on the surface which is seen as a strong
increase in the base component of the surface energy. This increase in surface polarity is
responsible for the enhanced wettability of UV modified surfaces.

EPR spectroscopy of UV exposed PC showed a strong paramagnetic response from

the material after UVA and UVO treatments. This is believed to arise from unpaired

103

3 pins on oxygen containing functional groups, trapped free radicals, or radical ions in the
polymer. The majority of paramagnetic sites were found to be near the surface of the PC
as they can be effectively quenched or undergo reactions with an amine cured epoxy.
The mechanical properties of the modified surface were probed by nanoindentation
tests. UVO treated PC showed an increase in the modulus in the top 500 nm with
increasing UV irradiation. This is expected to be a result of UV irradiation induced
crosslinking in the near-surface material. Despite the increase in surface modulus,
adhesion tests showed no change in the adhesive bond strength with epoxy and
polyurethane adhesives. This is an unexpected result and an explanation will be proposed

in the following chapter based on XPS failure analysis of adhesion test fracture surfaces.

104

10.

ll.

12.

13.

14.

15.

16.

References

. Brydson, J .A.; Plastics Materials, 5‘h ed., London, 1989.

Crawford, R.J. Plastics Engineering, 2nd ed., Oxford, 1987.

Rabek, J .F. Mechanisms of Photophysical Processes and Photochemical Reactions in
Polymers, John Wiley, New York, 1987.

Good, R.J. J. Adhesion Sci. Technol., Vol. 6, No. 12, pp. 1269-1302, 1992.

White, R. Chromatography/Fourier Transform Infrared Spectroscopy and its
Applications, Marcek Dekker, New York, 1990.

Jaffe H. and Orchin M., Theory and Applications of Ultraviolet Spectroscopy, Wiley,
New York, 1962.

Ming-Shin, L.; Chem-Chi M. M.; Jyh-Luen C.; Miaw-Ling L.; Feng-Chih C.
Macromolecules, Vol. 29, No.2, pp. 499-506, 1996.

Rabek, J .F. Mechanisms of Photophysical Processes and Photochemical Reactions in
Polymers, John Wiley, New York, 1987.

Humphrey, .l.S.; Shultz, A.R.; Jaquiss, D.B.G. Macromolecules, Vol. 6, No. 3, pp.
305-314, 1973.

Pankasem, S.; Kuczynski, J .; Thomas J .K. Macromolecules, vol. 27, pp. 3773-3781,
1994.

Gupta, A.; Rembaum, A.; Moacanin, J. Macromolecules, Vol. 11, No. 6, pp. 1285-
1288, 1978.

Gupta, A.; Liang, R.; Moacanin, J .; Goldbeck, R.; Kliger, D. Macromolecules, Vol.
13. pp. 262-267, 1980.

Torikai, A.; Murata, T.; Fueki, K. Polymer Photochemistry, vol. 4, p. 255, 1984.
Torikai, A.; Suzuki, K.; Fueki, K. Polymer Photochemistry, vol. 3, p. 379, 1984.

Soloukhin, V.A.; Brokken-Zijp, J .C.; VanAsselen, 0.; DeWith, G. Macromolecules,
vol. 36, pp. 7585-7597, 2003.

Strobel, M.; Lyons, C.S. , J. Adhesion Sci. Technol., Vol. 17, No. 1, pp. 15-23, 2003.

105

l7. Walzak, M.J.; Flynn, S.; Foerch, R.; Hill, J.M.; Karbashewski, E.; Lin, A.; Strobel,
M. Polymer Surface Modification; Relevance to Adhesion, K.L. Mittal (Ed.), pp. 253-
272, 1995.

106

CHAPTER FOUR

SURFACE ANALYSIS OF UV TREATED POLYCARBONATE

UV oxidation induced changes in the atomic composition and chemical bonding
state of polycarbonate generally occur in the top few atomic layers of the surface.
Qualitative and semi-quantitative characterization of these chemical changes was
accomplished using X-ray Photoelectron Spectroscopy (XPS). UV treated polycarbonate
samples were analyzed with MgKor x-rays in a Perkin Elmer Phi 5400 ESCA system at
pressures between 10'9 and 10'8 torr, pass energy of 29.35 eV, and a 45° take-off angle

unless stated otherwise.

4.1 Surface Chemistry of Polycarbonate

The major change occurring in polycarbonate because of UV oxidation is an
increase in the surface oxygen content. Figure 4.1 shows an example of the Cls peak and
the deconvoluted fits for polycarbonate samples before and after UV treatment. As
explained in Chapter 2, deconvolution of the carbon peak gives indirect information
about the chemical environment of the carbon atoms. The untreated polycarbonate shows
three distinct peaks in the Cls envelope. The peak at 284.7 eV is assigned to aliphatic and
aromatic hydrocarbon bonds. Carbon atoms bonded to more electronegative atoms like
oxygen experience an increase in the binding energy and Figure 4.1(a) shows two
additional peaks around 286.2 eV and 290.9 eV corresponding to carbon atoms in
hydroxyl/ether groups and carbonate groups respectively. The stoichiometric ratio of

carbon and oxygen in the polycarbonate repeat unit is 16:3 (C16H14O3) corresponding to a

107

stoichiometric O/C ratio of 0.187. The atomic compositions of untreated polycarbonate
films determined by XPS were generally in good agreement with the stoichiometric
composition, however, the material used is a commercial grade engineering polymer and
the surface cannot be considered a pure polycarbonate surface. Included and adventitious
impurities are expected to be found on the film surface and the state of the material more
closely resembles materials likely to be used in industrial applications. Polycarbonate
from two different batches was used in this study and small variations in the overall O/C
ratio between batches or sheets within batches were observed as shown in Table 4.1. This
variation in the baseline material was not considered significant compared to the changes
in O/C ratios after UV treatment. From repeatability studies, the standard error in O/C
ratios was found to be less than 5% and the error in deconvoluted curve fit data was
approximately 10%.

The polycarbonate repeat unit has two ether carbon atoms and one carbonate
carbon atom. Table 4.2 compares the stoichiometric and experimentally observed
distribution of carbon. The polymer has ca. 11.5% ether groups and 5% carbonate which
is in good agreement with the expected distribution. Figure 4.1(b) shows the Cls peak for
a typical UV oxidized sample. After UV oxidation of the polycarbonate in air or ozone,
the oxygen content increases sharply. Deconvolution of the Cls peak shows an increase
in the hydroxyl/ether peak (286.2 eV) and the appearance of two new peaks between the
hydroxyl and carbonate (290.9 eV) peaks corresponding to carbonyl (287.5 eV) and
carboxylate/ester (289.8 eV) functional groups. A more detailed description of the
changes in chemistry of UV/Ozone (UVO), UV/Air (UVA), and UV/Vacuum (UVV)

treated polycarbonate is provided in the following sections.

108

 

 

 

290 285 290 285
Binding Energy (eV) Binding Energy (eV)

Figure 4.1 Cls peak for polycarbonate (a) before and (b) after UV treatment.

109

Table 4.1 O/C ratios in untreated polycarbonate.

 

 

 

 

 

 

 

 

 

 

Polycarbonate C(%) O(%) O/C
Sample 1 85.1 14.9 0.175
Sample 2 86.2 13.5 0.156
Sample 3 84.4 15.6 0.184
Sample 4 87.8 12.2 0.139
Sample 5 84.4 15.6 0.185
Sample 6 84.6 15.4 0.182

 

 

 

Table 4.2 Carbon distribution in untreated polycarbonate.

 

 

 

 

 

 

 

 

 

Carbon distribution Stoichiometric (%) Experimental (%)
Hydrocarbon 81.3 83.4
Ether / Hydroxyl 12.5 1 1.4
Carbonyl 0.0 0.0
Carboxylate / Ester 0.0 0.0
Carbonate 6.3 5.1

 

*Values expressed as percentage of total carbon.

110

 

4.2 Surface Chemistry of UV/Ozone Treated Polycarbonate
There is significant oxygen uptake in the surface after UV treatment and the O/C
ratio reaches an asymptotic value of 0.5 after prolonged exposure. Samples were UVO
treated at irradiance levels ranging from 1.6 to 3.0 mW/cm2 for times ranging from 10 to
1 20 seconds. Figure 4.2 shows the O/C ratio for UVO treated samples. It is apparent that
the increase in O/C ratio depends solely on the total irradiation and all data points follow
a universal trend for polycarbonate.

A clear correlation can be drawn between the increase in surface energy, work of
adhesion, and O/C ratio for UVO treatment. The deconvoluted Cls curve fit data for
various treatment conditions is shown in Table 4.3. The detailed Cls spectra provides
some information regarding the mechanism of oxidative degradation. The hydrocarbon
and carbonate peaks decrease with increasing irradiation suggesting that the UVO
treatment attacks multiple sights on the PC chain. The carbonyl and carboxylate peaks
increase following a trend similar to the overall O/C ratio. The ether/hydroxyl ratio
increases by a small amount initially and remains constant even after extended
tr eatments. The decrease in the carbonate peak is indicative of a chain scission
r1lechanism involving cleavage of almost 50% of the carbonate linkages in the polymer.
The number of carbon atoms with C-C and C-H (hydrocarbon) bonds also decreases
gradually indicating breaking of bonds in the bisphenol-A structure including the Opening
of aromatic rings and possible cleavage of the primary carbons (methyl groups). The
re(inaction of carbonate and hydrocarbon peaks by 50% after extended exposures points to
e"'iliensive chain scission on the surface, possibly followed by further reactions such as

0x idation or crosslinking of the chain fragments.

lll

OIC Ratio

 

0.6

0.5~ - - ~- --- u A-.. _. was s

 

 

 

 

,,§-——0 """"""" 3
04+- g ,«i _ _ s s s _
K".

o3---+ , _—.ufi -smus-s-z. s+_+
02 “2;?! W i —— — — — — —

6
0.1 -_.. —- — ‘- e s - s, , or e
0.0 1 l l l l 1

o 50 100 150 200 250 300

UV Irradiation (mJ/cmz)

Figure 4.2 O/C ratio of UV treated PC as a function of irradiation.

112

350

Table 4.3 XPS surface chemical composition of UV treated polycarbonate
showing the O/C ratio and Cls curve fit analysis.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Irradiant Total Hydro- Ether/
Energy Carbon carbon* Hydroxyl* Carbonyl* Carboxyl“ Carbonate“
(mJ/cmz) (%) (”7Q (%) (%) (%) (%)
0 87.6 73.1 10.0 0.0 0.0 4.5
27 82.2 63.9 13.9 0.9 0.9 2.6
53 76.8 55.0 11.6 3.9 3.3 3.1
80 75.5 50.6 12.8 4.6 4.8 2.7
85 76.0 52.2 13.7 3.6 4.6 1.9
114 74.0 44.8 15.3 5.7 5.7 2.5
120 72.6 44.5 12.1 6.3 6.9 2.8
160 71.0 43.4 13.4 5.9 6.4 1.9
194 68.0 38.6 13.1 6.9 8.5 0.9
200 68.8 38.9 13.8 7.1 7.9 1.1
228 67.7 35.9 14.3 6.8 9.1 1.7
240 66.4 36.8 12.5 5.5 10.0 1.7
319 67.9 37.8 12.4 6.9 8.5 2.4

 

 

*Values expressed as percentage concentration of total surface.

113

 

4.3 UV Treatment in Vacuum

The samples described in previous sections were treated in an external aluminum
treatment chamber and subsequently transferred to the XPS for analysis. To avoid sample
exposure to ambient conditions between treatment and analysis, a UV treatment pre-
chamber was designed to interface with the XPS as shown in Figure 4.3. The XPS pre-
chamber has a UV transparent sapphire window and provides the ability to irradiate
surfaces in a clean environment, including exposure in vacuum (pressure < 10'6 torr) as
well as other gases. The UV lamp was mounted over the sapphire window and after
treatment, samples were directly introduced into the XPS for analysis. Due to size
restrictions, only the RC500 lamp was used with the prechamber. The more restrictive
opening for the lamp on the prechamber causes a reduction in irradiance compared to the
external chamber, hence irradiances measured for the lamp in Figure 2.4 do not apply to
UV exposures in the prechamber. However, irradiance in the prechamber is constant due
to the inability to vary the distance between the lamp and sample, and the total irradiation
in the prechamber is in direct proportion to the exposure time. The irradiance in the pre-
chamber could not be measured directly because of restricted space and large size of the
radiometer detector assembly. The dependence of most properties on the irradiation has
already been demonstrated, so trends in time in the pro-chamber experiments can be
considered to be linearly equivalent to trends in irradiation in the external chamber.

The ability to treat samples in vacuum in the pre-chamber also allows the
separation of UV exposure and gas exposure reactions. Experiments were performed to
measure the oxygen uptake in samples irradiated in vacuum by introducing ozone and

other gases into the pre-chamber after the UV exposure. This will be termed ‘post-

114

treatment’ in this text and indicates a consecutive, but separate, operation in the treatment
process. Table 4.4 shows the chemical composition of polycarbonate samples treated in
vacuum, air, and ozone.

Polycarbonate treated in UV/Vacuum (UVV) did not show any change in the
surface chemical composition (O/C), although for very long UVV treatment on the order
of 120-180 seconds, a small amount of carboxylate was observed, most likely due to
cleavage of some carbonate bonds. The possibility of creating free radicals in the surface
with UVV exposure and the reaction of such an activated surface with a gas was
investigated. In samples exposed to air (ca. 20% oxygen) for 10 minutes following UVV
exposure, there was no noticeable change in the chemical composition from the untreated
polymer. However, when activated samples were exposed to ozone for 5 minutes, the
surface O/C ratio changed to values expected from direct UVO treatment.

Untreated PC exposed to ozone for such short times does not account for the
change in O/C ratios seen in these samples. Figure 4.4 shows the increase in O/C ratio of
PC exposed to ozone. The O/C ratio increases initially but reaches a maximum value of
ca. 0.23 to 0.24 after 30 seconds of ozone exposure and remains relatively constant
thereafter. O/C ratios of UVV treated PC post-treated with ozone ranged between 0.27 to
0.30. This suggests a mechanism which may involve the formation of activated radicals
or unsaturated bonds on the surface which are susceptible to selective oxidative attack by
ozone and not oxygen. This result also provides confirmation of the necessity of low
wavelength (below 200nm) UV light which is capable of producing trace amounts of
ozone in-situ during UV exposure. The independence of UV activation of the surface and

the actual surface modification also opens avenues for the use of different gases to

115

 

achieve specific functional groups on the treated surfaces. Activated surfaces can be
grafted with different groups once the specific reaction sites are identified.

Activated surfaces also proved to be stable in air and can be post-oxidized with
ozone after aging in ambient conditions for long periods. Table 4.5 shows the O/C ratio
for samples aged up to 6 days in normal laboratory environment after UVV exposure.
Samples were exposed to ozone for 5 minutes immediately before XPS analysis to react
any active sites not quenched by storage in ambient conditions. The O/C ratio of aged
activated samples remained stable between 0.27 and 0.30 over the period of six days and
only a slight decrease was seen up to 139 hours of aging. No increase in oxygen
concentration was found in samples stored in air if they were not exposed to ozone prior
to analysis. This strongly suggests that the active radical or ionic sites within the XPS
sampling depth do not get quenched by exposure to oxygen on time scales spanning tens
of hours. This may be a combined effect of the formation of stable ions or trapped
radicals under the surface which do not recombine easily.

The amount and rate of uptake of ozone in UVV treated samples is studied in

more detail in the following section.

116

Table 4.4 Surface chemical composition of polycarbonate
UV treated in vacuum, air and ozone.

 

 

 

 

 

 

 

 

 

 

Treatment Conditions C (%) O (%) O/C
Untreated 86.0 14.0 0.16
120 sec UV/Vacuum 86.4 13.6 0.16
120 sec UVNacuum + 10 min Air 86.6 13.5 0.16
120 sec UV/Air 81.4 18.6 0.23
120 sec UV/Air + 5 min ozone 78.4 21.6 0.28
120 sec UV/Ozone 77.7 22.3 0.29
120 sec UV/Vacuum + 1 min Ozone 78.8 21.2 0.27
120 sec UV/Vacuum + 5 min Ozone 77.0 23.0 0.30
12503:: [3:13;]:cuum + 10 min Air 74.6 22.3 0.30

 

 

 

 

 

 

Post-treatment with gas indicated by ‘+’.

117

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

52:
H E V Q Ozone outlet Kl trap
Ozone , 1
generator I 5
Ozone J _
inlet ~
Oxygen , ,
apphire
WifldOW A
XPS
(3 (X) 1 chamber
Sample V
XPS sample G
loading fork ate
valve

 

 

To vacuum pump

/)®/\

 

 

 

Sample load port

 

Figure 4.3 Schematic of the XPS pre-chamber for UV treatment

118

0.30

 

0.28 _ . z
0.26 -* Q

0.24 - 3 ° 0

OIC Ratio

0.22 _

 

020 1, o Untreated
9 180 sec UVV

 

 

 

 

0.18

 

 

0.16 1 T 1 l 1 l l I I 1 I
0 30 60 90 120 150 180 210 240 270 300 330
Ozone Exposure Time (see)

Figure 4.4 O/C ratio of untreated and U V V treated PC exposed to 700 ppm ozone.

Table 4.5 Surface chemical composition of UVV treated PC aged in air.
Aged samples oxidized in ozone for 5 minutes prior to analysis.

 

 

 

 

 

 

 

 

 

Aging Conditions O/C Ratio

0 hrs 0.18

95 hrs 0.18

0 hrs aged + 5 min. Ozone 0.30

22 hrs aged + 5 min. Ozone 0.28
46 hrs aged + 5 min. Ozone 0.29
90 hrs aged + 5 min. Ozone 0.27
95 hrs aged + 5 min. Ozone 0.28
139 hrs aged + 5 min. Ozone 0.27

 

 

 

 

119

4.4 Oxygen Uptake in UV/Vacuum Treated PC

As mentioned in the previous section, UVV treated samples post treated with
ozone exhibit uptake of oxygen in the surface which is greater than that expected by
ozonation of untreated polycarbonate. A detailed study of the kinetics of oxygen uptake
was performed by ozonation of samples UVV treated for different times. The UVV
activated samples were exposed to ozone for times ranging from 1 to 300 seconds in an
external ozonation chamber with 700-800 ppm ozone in oxygen carrier gas flowing at 30
scflr. The high concentration of ozone guarantees availability of enough ozone to the
surface. Figure 4.5 shows the rate of ozone uptake in samples UVV treated for 30, 60,
and 180 seconds. The data in Figure 4.5 is reported in terms of percent oxygen on the
surface instead of the O/C ratio in order to compare the concentration of oxygen with an
approximate concentration profile of a first order surface reaction between an active
surface site and an ozone molecule. This is valid because only two elements are present
on the surface and the percent composition can be converted to O/C ratio. Thus 19% 0
corresponds to a O/C ratio of 0.23.

The rate of oxygen uptake in Figure 4.5 clearly shows the uptake is proportional
to the level of UVV irradiation at long ozone exposure times. However, for short ozone
exposures between 1 and 30 seconds, the rate is similar for all three samples. After an
exposure of approximately 30 seconds, the curves begin to diverge. A hypothetical first
order kinetic model, shown by the dashed line, was used to examine the possibility of the
initial part of the curves being indicative of the reaction between ozone and an active site
on the surface. Since the precise reaction mechanisms for a complex process such as UV

oxidation are as yet undetermined, the elementary reactions and rates are unavailable.

120

The first order kinetic model is overly simplified and presented for purely illustrative
purposes. The assumptions made for this simple model are that a surface active site reacts
with one ozone molecule, the reaction adds a single oxygen atom at that site, excess
ozone is available for the surface reaction and the reaction is of the first order with a rate

expression as shown in Equation 4.1.

s*+o3—>so+... r=ks* (4.1)

* o o o o
where, S is the active surface site, S0 is the site after oxrdatron, and k rs the reaction

constant. The first order kinetic solution for the reaction is of the form
>1: :1:
[s ]= [s 10 exp(— kt) (4.2)

lll 11!
where, [S ]0 and [S ] are respectively the initial and final concentrations of active
surface sites. For an initial concentration of surface sites bonded to oxygen, [SO]0, and

* fl
the increase in surface oxygen due to oxidation reactions given by [S ]0-[S ], the final

amount of oxygen on the surface can be expressed as
at:
[50145010 +[s ]0{l—exp(— 1a)} (4.3)

An arbitrary reaction rate constant of 0.15 and a 3.5% surface concentration of active
sites after UVV treatment were assumed to generate an illustrative concentration profile
of oxygen for a first order surface reaction in Figure 4.5.

The UVV treated samples exhibit oxygen uptake beyond the initial part of the
curves that is dependent on the amount of irradiation received. Since absorption of
radiation by the polymer follows Beer’s law, it is reasonable to assume that irrespective

of the irradiation levels, the surface of the sample receives the maximum dosage and

121

increasing the dosage progressively activates sub-surface material. In such a scenario, the
rate of oxidation of the polymer under the surface must be limited by diffusion of gaseous
ozone into the material and the maximum oxygen uptake at any level of irradiation is
indicative of the number of active sites created beyond the first few monolayers.

The separation of the oxygen uptake into a surface reaction limited to the top few
monolayers and a diffusion limited oxidation of the sub-surface layers was further

investigated using surface derivatization experiments described in the following section.

 

 

 

 

 

 

 

23
22 — A I
21

A A‘

.\° 20 — I 1 1*

a 19 _ ‘ i O? L Sub-surface Diffusion

t .u -------------------------------
18 ~ -

O 0 Surface Reaction O 308 UVNac
17 —¢ o 603 UVNac

- A 1803 UVNac

16 - - - 1st Order Rxn
1 5 T 1 1 l l T 1 r f r

 

O

30 60 90 120 150 180 210 240 270 300 330
Ozone exposure time (sec)

Figure 4.5 UV/Vacuum treated polycarbonate post oxidized with 700 ppm ozone.

122

4.5 Silver Nitrate Derivatization

XPS is a fairly surface sensitive analytical technique, yet the sampling depth
averages over multiple monolayers of the material, making it difficult to attribute any
properties to the top monolayer. Angle resolved XPS (AR-XPS) can provide some depth
information by using shallow take-off angles to reduce the effective sampling depth. All
XPS data shown here was obtained at a 45° take-off angle. AR-XPS was performed at 22°
and 78° on UV treated samples, but did not show any significant differences in
composition as a function of sampling depth. This may be likely if the chemical changes
occur in the material to depths greater than 100 A.

The C Is spectra of UV treated PC shows an increase in the carboxylate peak
which follows the same trend as the overall oxygen ratio. To differentiate between the
surface reaction occurring in the top few monolayers and the reactions occurring in the
sub-surface monolayers to the XPS sampling depth of 75 to 100 A, UV treated surfaces
were derivitized with silver nitrate. Derivatization techniques have been used in XPS
analysis to differentiate between different functional groups [1-3]. Silver nitrate was used
to react with the carboxylate groups on the surface [1] to give silver carboxylate as shown
in Figure 4.6. The silver on the surface can provide a high cross—section tag for XPS
analysis. The assumption in this derivatization experiment is that the probability of
formation of carboxylate groups in the first 100 A due to UV oxidation is uniform and
that the silver nitrate in liquid phase does not penetrate into the bulk of the sample and
limits the reaction to the top monolayer. The silver atom is large compared to oxygen
and has much lower mobility in the liquid phase compared to an aggressive gas phase

oxidant such as ozone. Silver nitrate solutions were made in ethanol, a poor solvent for

123

polycarbonate. UV treated samples were dipped in 2% w/w ethanolic AgNO3 solution
for times ranging from 1 to 30 minutes. Silver concentrations on the surface were found
to be invariant after 3-4 minutes of reaction. All derivatization data reported in this
section was obtained from samples dipped in ethanolic silver nitrate for 5 minutes. Figure
4.7 shows the results of the silver nitrate derivatization experiments. The absolute
concentration of silver and carboxylate concentration obtained from Cls curve fit are
plotted for a 60 sec UVV treated polycarbonate exposed to ozone up to 300 seconds. The
carboxylate concentration follows the general trend seen in the total oxygen
concentration and statistically significant increases are seen till 120 seconds ozone
exposure. The silver concentration increases rapidly to ca. 0.5% for a 30 second ozone
exposure and remains constant thereafter even with increasing carboxylate functionality.
The difference in the maximum measured absolute concentrations for silver (0.5%) and
carboxylate (2.5%) validates the assumption that the silver ions do not diffuse into the
sub-surface to react with all available carboxylate groups and selectively provides
information about the top monolayers of the surface. The other important observation is
that all samples with more than 30 second ozone exposure do not show increases in the
silver concentration indicating that the surface layers are oxidized by ozone in the first 30
seconds and all subsequent increases in oxygen content are the result of ozone diffusion
and reaction into the material. It is also important to note that the surface reaction limit of
30 seconds is identical to the exposure for which the oxygen uptake in Figure 4.5
diverges for samples with different levels of UVV irradiation and the surface saturation

indicated by the simple kinetic model used to fit the initial oxygen uptake profile.

124

 

 

 

 

 

 

 

 

0
ll
_C'OH + AgNO3

O

_C—R

_OH

 

\\\\\\\\\\\\\\\\

Q
XPS
Sampling
ept
v
1’
—’ _C-O-Ag+

_C—R

—OH

 

\\\\\\\\\\\\\\\\

Figure 4.6 Reaction mechanism of silver nitrate with carboxylate groups
on UV treated polycarbonate surface. Silver, unlike ozone, is limited to
reactions on the top few monolayers of the surface.

125

swerve-2x >

9°
or

 

 

<> Carboxylate
0 Silver

 

 

 

 

 

 

Composition (%)

 

r "3 “r 1‘ r * r r r 5*“

.0 7‘ 7‘ 1° 1° 9°
or o or o 01 o
l 1 1 l l l

 

 

0.0 . I l 1 1 l l l I
0 30 60 90 120 150 180 210 240 270 300 330
03 exposure time (see)

Figure 4.7 Silver and carboxylate concentration in silver nitrate derivatized
polycarbonate with 60 sec U V V treatment and post-oxidation with 700 ppm ozone.

126

4.6 UV Treatment in Inert Atmosphere

The proceeding section described experiments performed in vacuum to separate the
effects of irradiation and oxidation. The disadvantage of treating samples in the XPS pre-
chamber is that irradiance levels and ozone concentrations are difficult to measure and
since the prechamber requires evacuation after any process gas is introduced, exposure
times to gases cannot be controlled precisely.

Results similar to UVV treatment can be obtained by irradiating polymers in an inert
atmosphere to prevent oxidation. A nitrogen purge was used in the UV treatment
chamber to prevent in-situ oxidation of samples during irradiation. This was followed by
ozone exposure to oxidize the activated surfaces. Experiments in inert atmosphere were
designed to compare the properties of in-situ and ex-situ oxidized polycarbonate.

Figure 4.8 shows the O/C ratio of polycarbonate treated in ozone and nitrogen using
the RC747 lamp. UV/Nitrogen treated samples were exposed to 700 ppm ozone for 120
seconds after irradiation. The increase in O/C ratio for UVO treated samples is rapid and
reaches a maximum value around 0.5. The UV/Nitrogen treated post-oxidized samples do
not show a comparable increase despite the fact that irradiation in nitrogen is typically
higher than the irradiation through the UV absorbing ozone atmosphere. The O/C ratio
does increase in nitrogen, but only to a value of 0.3, the extent seen in UVV treatment.

The same trend was observed in the physical properties of samples treated in
nitrogen and ozone. Figure 4.9 shows the work of adhesion with deionized water for
UVO and UV/Nitrogen post-oxidized samples. With UVO treatment the work of
adhesion increases sharply while only a slight positive change is observed for samples

exposed in nitrogen. The different results of the two types of treatments can be explained

127

in terms of recombination of radicals. During UVO treatment, any radicals created on the
surface can be immediately attacked by ozone adsorbed on the surface, whereas in inert
or vacuum environments, the radicals can live long enough to undergo recombination and

only a fraction of the original radicals remain on the surface in a state susceptible to

ozone attack.

 

 

 

 

 

 

 

 

 

 

 

0.6
0.5 + f g ”was? - __. -_o._ 2 2 s ___ s- __
<> 0
0.4—»- e g s Wflg‘ # f s a 2 ._ _
<>
.3
E 0.3-_- ice 9.- f- s -3 v- s w. a -3
Q . 0
° 6
02.. Q g ‘e- s- es s- ”s s, -_-_‘__
<>
OUV/OS
0" OUV/N2+1203 03
o . . . .
0 30 60 90 120 150

UV Exposure Time (sec)

Figure 4.8 O/C Ratio of UVO and WIN; + 120 sec 03 treated PC

128

é

 

 

 

ll'_lll
I I!

EI130

a

2120—

Erro—

63

E

:100“

.2

.r°:’ 90

”G

<1

“5 80~

J

I-I

E 70—
60

 

 

 

o uvo
r: uvmz + 1203 03

 

 

 

 

60 90
UVExposureTimfisec)

120

Figure 4.9 Work of adhesion of UVO and WIN; + 120 sec 03 treated PC

129

150

4.7 Washed Films

The formation of low molecular weight (LMW) polymer fragments on the surface
after UV oxidation is expected in a polymer such as polycarbonate where both carbonate
linkages and hydrocarbon bonds undergo chain scission. The formation of LMW
fragments for various polymers has also been reported in literature [4,5]. The LMW
fragments are often soluble in water and can be removed by mild rinsing. UVO treated
PC was rinsed in deionized water and ethanol to test the presence of LMW fragments and
determine the effect of their removal on surface properties.

LMW fragments can increase the wettability in polymers by dissolving in the
probe liquid and reducing the overall energy. The presence of labile LMW can also lead
to undesirable properties such as the development of a weak boundary layer in the
material. Figures 4.10 and 4.11 show the O/C ratio and work of adhesion with deionized
water for UVO treated PC (RC747 lamp, 2.6 mW/cm2 irradiance, 800 ppm ozone)
washed with water and ethanol prior to analysis. The removal of LMW fragments from
the surface is apparent by the drop in O/C ratio beyond 20 to 30 seconds of treatment.
The work of adhesion is reduced after rinsing and ethanol was observed to be slightly
more efficient at removing LMW material compared to water.

Table 4.6 shows a comparison between washed UVV and UVO treated samples.
The UVV samples were oxidized by post-treatment exposure to ozone. The O/C ratios
and Cls curve fit distribution for oxidized UVV samples before and after washing in
water are identical. UVO samples treated for the same time show a significant change in
the O/C ratio and the Cls curve fit shows a decrease in the carbonyl and carboxylate

concentrations that is consistent with the removal of oxidized low molecular weight

130

material from the treated surface. Similar trends are seen in samples treated for very long
times (600 sec UVO). Comparison of the UVV and UVO samples after washing indicates
that UVV (or UV/Nitrogen) treatment creates surface sites capable of reacting with
ozone, but recombination of radicals in the inert environment lowers the concentration of
such sites and the resultant surface has a lower amount of LMW material that can be
removed easily. In UVO treated materials the simultaneous presence of ozone and UV
decreases the probability of recombination of radicals and leads to extensive formation of
oxidation products, which, when exposed to irradiation, can undergo further reactions,
ultimately leading to extensive chain scission on the surface and formation of a higher

number LMW fragments as indicated in Table 4.6.

 

 

 

 

 

 

 

 

 

 

 

 

0.6
0.5 - - - 4 s, f s .# ._- MI .2 .2- z s s
I
I
0.4... a 7 s s_ ls s- , __‘__ss
I
.2
"‘ I
‘3 0.3 — if. h- 0+“- -+ , - —<>- -- -~-
g 0
° 0
02*. 22+. -222 -,s s
o O/C (washed)
0.1
I O/C (no wash)
0.0 . . , T
0 30 60 90 120 150

UVO exposure (see)

Figure 4.10 OIC ratio for UVO treated PC before and
after rinsing with water followed by ethanol.

131

140

 

 

 

 

 

 

 

E.
g 120
.,
i
E. 110 - i
s a 5
'g 100 ~ 5 g i
3
. 9. —. r r
E 1 oUVO
3 80 AUVO+water rinse
3 70 oUVO+ethanol rinse
60 r 1 T ‘ l 1 ‘ T T

 

0 15 30 45 60 75 90 105 120 135
UVO Exposure Time (sec)

Figure 4.11 Work of adhesion for UVO treated PC
after rinsing with water and ethanol.

132

Table 4.6 Comparison of UVV and UVO treatments after water rinse.

 

 

 

 

 

 

 

 

g "3}
a 0 ">1 3'. g
a A IE‘A C A K A C A
Treatment 35% z 5: §§ §§ _8 §
i E u u 5
c o O/C :1: g
(%) (%) Ratio
6°SCCUW 80.7 19.3 0.24 60.2 13.9 1.7 1.7 3.2
+Ozone
60 sec UVV
+Ozone 79.8 20.2 0.25 59.4 14.3 1.9 1.6 2.6
+Washed
60860 UVO 71.0 29.0 0.41 43.4 13.4 5.9 6.4 1.9
60 sec UVO
+Washed 78.5 21.5 0.27 57.0 13.1 3.6 3.1 1.8
600schVO 64.5 32.6 0.50 28.8 13.8 7.5 13.3 1.0
6005‘” UVO 83.0 16.5 0.20 62.7 12.2 3.7 3.5 0.9
+Washed

 

 

 

 

 

 

 

 

 

133

 

4.8 Analysis of Adhesion Test Fracture Surfaces

Tensile and shear adhesion tests performed on UV treated polycarbonate, (see
Chapter 3) did not show statistically significant changes in adhesive bond strength or
modulus after UV treatment in spite of the higher surface energy and work of adhesion.
The formation of an adhesive bond between two materials is a complicated process that
depends on many factors including the mechanical properties of the surfaces,
morphology, microstructure, and thermodynamic favorability. A strong thermodynamic
driving force alone is not sufficient to guarantee strong adhesive bonds.

The failure of adhesive bonds in UV treated polycarbonate was investigated by
fracture surface analysis using XPS. Tensile and shear adhesion samples were prepared
using epoxy adhesive (Araldite® 2011) to bond an aluminum stub to the UV treated
films. During testing, failure of the bond consistently occurs near the polycarbonate—
epoxy interface based on visual examination. The two fracture surfaces thus generated
are henceforth referred to as ‘polycarbonate-side’ and the ‘adhesive-side’ fracture
surface. Determination of the locus of failure is an important step in analyzing the failure
of adhesive bonds and both fracture surfaces were analyzed by XPS. Analysis of the
polycarbonate-side fracture surface showed a surface resembling untreated polycarbonate
even in UV treated samples. The analysis of the adhesive-side fracture surface provided
more information. Figure 4.12 shows the Cls spectra for the cured epoxy adhesive and
fracture surfaces for various UV treated samples.

The adhesive used was an amine cured epoxy that provides a unique nitrogen
peak, which is absent in polycarbonate. Table 4.7 shows the atomic composition of cured

epoxy adhesive and the adhesive-side fracture surfaces for untreated and UV treated PC

134

samples. Three samples were analyzed for each condition to ensure accuracy. The cured
epoxy has 8.5% nitrogen and 6.1% oxygen on the surface. The adhesive-side fracture
surfaces have a much lower nitrogen concentration ranging from 3-4% and a high oxygen
concentration between 12—15%. The composition of the fracture surface was identical
within experimental error for untreated PC, 30 and 120 sec UVO treated PC, and 120 sec
UVO treated PC washed with water to remove low molecular weight material prior to
bonding. The composition of the adhesive-side fracture surface is consistent with the
presence of a thin layer of polycarbonate on top of the epoxy adhesive. The unique
nitrogen tag atom on the epoxy allows further analysis by using the C/N ratio of the pure
epoxy to apportion part of the carbon signal from the fracture surfaces to carbon atoms in
the epoxy molecule. Using this methodology, the amount of Cls signal attributable to
epoxy was calculated and is shown in Table 4.7. The remainder of the carbon signal on
the fracture surfaces was attributed to polycarbonate.

The results show that the adhesive-side fracture surface, within the XPS sampling
depth, is composed of 50-60% polycarbonate for all samples irrespective of surface
treatment or removal of LMW material by washing. The percentage of polycarbonate
may be slightly overestimated if the polycarbonate is present in the form of a continuous
film on the adhesive surface because of a higher bias towards the top surface layers. Of
more importance is the fact that UV treatment does not change the amount of
polycarbonate removed. To determine if the material removed is indeed polycarbonate,
weighted spectral subtraction of the epoxy Cls spectra from the fracture surface Cls

spectra was performed.

135

 

   
  
  
  
   

120 sec UVO
+ water rinse

120 sec UVO

30 sec UVO

0 sec UVO

Epoxy adhesive

 

 

 

l i i l i l | i i

298 296 294 292 290 288 286 284 282 280 278
Binding Energy (eV)

Figure 4.12 Cls spectra of epoxy adhesive and fracture surfaces (adhesive side) for
0 sec UVO, 30 sec UVO, 120 sec UVO, and 120 sec UVO + water rinsed PC.

136

Table 4.7 Atomic composition and analysis of adhesion test fracture surfaces
(adhesive side) for UVO treated polycarbonate - epoxy bonds.

 

 

 

 

 

 

 

C N O C/N Epoxy PC
(%) (%) (%) Ratio (%) (%)
Epoxy 85.4 t 0.3 8.5 :l: 0.2 6.1 i 0.3 10.1 :t 0.2 100 0
0 sec UVO
PC — Epoxy 82.0 i 3.1 3.2 10.8 14.8 t 2.4 27.2 :t 8.1 39 :11 61 :11
Failure Surface
30 sec UVO
PC — Epoxy 83.1 :1: 1.1 3.6 $0.9 13.3 i 0.6 23.8 :1: 5.3 44 :11 56 :11
Failure Surface
120 sec UVO
PC-Epoxy 81.8 12.7 4.0: 1.0 14.3 $1.9 21.5 15.6 49 :14 51 i 14
Failure Surface
120 sec UVO +
water “used 84.9 i 2 3.0 i 0.7 12.1:13 29.4 :3: 8.1 36 i 8 64 a; 8
PC — Epoxy

Failure Surface

 

 

 

 

 

 

 

137

 

The results of spectral subtraction are shown in Figure 4.13. The resultant spectra
are very similar to polycarbonate with a pronounced shoulder indicating the
ether/hydroxyl peak and the presence of the characteristic carbonate peak at 291 eV.
Deconvolution of the Cls spectra showed a carbon distribution similar to that of
polycarbonate.

The results of the fracture surface analysis are indicative of failure occurring
approximately within the top 30—60 A of the polycarbonate surface irrespective of
whether the low molecular weight polymer formed by UVO oxidation is removed prior to
bonding. Failures such as these are indicative of a weak boundary layer within the
substrate or very high interfacial adhesion which propagates the crack into the relatively
weaker substrate. This type of cohesive failure in the substrate also explains the
invariance of adhesive bond strength after surface treatment. The results from
nanoindentation tests show that the modulus of the polycarbonate in the top 500 nm of
the surface increases after UVO treatment, but this increase in modulus may either be too
small to effect a change in the macroscopic adhesive bond failure or of no consequence if

bond failure is caused by a weak boundary layer near the surface.

138

 

   
  
  
   

120 sec UVO +
water rinse

120 sec UVO

30 sec UVO

0 sec UVO

 

 

 

I I I I 1 l l l l i

298 296 294 292 290 288 286 284 282 280 278
Binding Energy (eV)

Figure 4.13 Cls spectral profile of adhesive fracture surfaces after subtraction of
epoxy Cls signal for 0 sec UVO, 30 sec UVO, 120 sec UVO, and 120 sec UVO +
water rinsed PC.

139

4.9 Summary

Polycarbonate surfaces were analyzed using XPS to characterize the chemical
changes occurring on the surface during UV oxidation. The O/C ratios in UVO treated
PC were found to depend on the irradiation received by the sample and, like the surface
energy, showed no dependence on the irradiance levels or exposure time. Study of the
functional groups created on the surface during UVO treatment show that the oxidation
process occurs by reactions of ozone, and possibly the by-products of ozone dissociation,
with the carbonate groups as well as the hydrocarbon bonds in the bisphenol-A structure.

When in-situ oxidation by ozone is prevented by irradiating samples in vacuum or
inert gas atmosphere, active sites are created on the polycarbonate surface. The active
sites thus generated are unreactive in air and do not show any uptake of oxygen even after
many days of aging in ambient conditions. However, the surfaces retain their activity and
readily undergo oxidation when exposed to ozone. The increase in oxygen content of
these samples is beyond that expected from unirradiated, passive polymer samples
exposed to ozone.

The rate of oxidation of activated surfaces, determined by measuring the increase in
oxygen content as a function of ozone exposure, showed two distinct processes: 1) a fast,
initial incorporation of oxygen into the surface that is independent of the amount of
irradiation received by the surface, and ii) a slower region of growth which shows direct
dependence on the level of irradiation provided to the sample during treatment in a non-
oxidative environment. The initial rapid uptake is attributed to the reaction of ozone with
active sites in the top few monolayers of the material. Once the surface is depleted of

active sites, further oxidation involves the slow diffusion of ozone into the sub-surface

140

layers to oxidize active sites in the polymer. The validity of this model was tested by
derivatization experiments with silver nitrate which allowed differentiation between the
oxygen signal contribution from the surface and sub-surface layers of the polymer.

The formation of water soluble low molecular weight oxidized material on the
surface during UV treatment was observed. Differences in the amount of LMW material
formed on in-situ and ex-situ oxidized samples provided indication of the different
behavior of the polymer when irradiated in inert and oxidizing environments. Irradiation
in non-oxidizing environments increases the possibility of radical recombination, or
rearrangement, to form a smaller number of stable surface active sites. Irradiation in
oxidizing environments enhances the total extent of oxidation because of aggressive
reactions between the oxidizer and radicals created on the surface, but the higher extent
of reaction can lead to extensive chain scission and formation of LMW fragments for
longer treatment times.

Fracture surface analysis was used to determine the locus of failure in adhesion test
samples. Fracture surfaces of the epoxy adhesive and polycarbonate substrates were
analyzed and results obtained using spectral subtraction techniques showed the removal
of approximately 30-60 A of polycarbonate material by the epoxy adhesive. The cohesive
failure in the substrate may be indicative of a weak boundary layer within the
polycarbonate surface or a very high level of interfacial adhesion between the epoxy and

polycarbonate which forces the locus of failure into the relatively weaker polycarbonate.

141

References

. Briggs, D.; Seah, M.P. Practical Surface Analysis Volume 1, 2nd ed., pp. 450—454,
Wiley, New York, 1983.

. Dickie, R.A.; Hammond, J .S.; deVries, J .E.; Holubka, J .W. Anal. Chem, Vol. 54, pp.
2045-2049. 1982.

. Everhart, D.S.; Reilley, C.N. Anal. Chem, Vol. 53, pp. 665-676, 1981.

. Walzak, M.J.; Flynn, S.; Foerch, R.; Hill, J.M.; Karbashewski, E.; Lin, A.; Strobel,

M. Polymer Sutface Modification; Relevance to Adhesion, K.L. Mittal (Ed.), pp. 253-
272, 1995.

Strobel, M.; Lyons, C.S. , J. Adhesion Sci. Technol., Vol. 17, No. 1, pp. 15-23, 2003.

142

CHAPTER FIVE

UV TREATMEN T OF POLYMERS

One of the advantages of the UV treatment process is its applicability to a wide
range of materials. Surface modification of polymers such as polyethylene (PE) and
polypropylene (PP) has always generated interest in the plastics industry because of the
widespread use and poor adhesive performance of these materials. In recent years, the
growing use of composites has drawn attention to surface treatments for thermosetting
composite matrices such as epoxy. Rubbers and thermoplastic polyolefins (TPOs), which
typically use hazardous chlorination treatments to enhance wettability, can also benefit
greatly from UV surface treatment. The previous chapters described the results of UV
modification of polycarbonate. In this chapter, the results of UV modification of several

other polymers will be discussed in brief.

5.1 Epoxy

A diglycidal ether of bisphenol A (DGEBA) epoxy resin (Shell, Epon 828) cured
with a polyoxypropylene triamine (Huntsman, Jeffamine T403) was used in this study.
The structures of the two components are shown in Figure 5.1. The epoxide equivalent
weight of the resin is 188 and the amine hydrogen equivalent weight of the curing agent
is 81. The cured epoxy has some structural similarity with polycarbonate because of the
bisphenol-A unit in the repeat unit. The resin and curing agent were mixed at 100:45
weight ratio and cured at 85°C for 2 hours in an open Teflon® mold followed by

postcuring at 150°C for 2 hours. The cured epoxy was cut with a diamond blade saw and

143

washed with detergent and deionized water to remove traces of cutting lubricants from
the surface. The epoxy surface exposed to air during cure was used as the test surface.
The polymer was UVO treated using the RC 747 lamp at various irradiance values and
contact angles were measured with five liquids, as described in Chapter 2, to determine

the surface energy.

0 O
/ \ 8‘3 ‘1’” CH“ / \
CHz—CH—OHCIIO—CH—C Hz—O-H—O—CHz—CH—C H2
C H3 CH;

(“3H2[OCH2CH(CH3)],-,- NH2
CH3CH2(|3—CH2[OCH2CH(CH3)]y— N112
CH2[OCH2CH(CH3)]z—NH2

x+y+z=53

Figure 5.1 Chemical structures of Epon 828 (top) and Jeffamine T403 (bottom).

Figure 5.2 shows the work of adhesion of the UVO treated surfaces with water as
a function of total irradiation. The work of adhesion increases after treatment to an
asymptotic value of I30-140 mJ/mz. Samples were treated at various combinations of
irradiance and time, but, as in the case of polycarbonate, the work of adhesion does not
show any independent influence of the irradiance or time and only depends on the total
irradiation. The change in the work of adhesion is caused by the incorporation of oxygen
on the surface. XPS analysis shows an increase in the surface O/C ratio from 0.18 to 0.53

after 320 mJ/cm2 UVO irradiation. The detailed XPS results are given in Table 5.1.

144

Analysis of the spectra shows significant increases in the hydroxyl, carbonyl and
carboxylate functionality after UVO treatment while the concentration of carbon in
hydrocarbon and ether bonds decreases suggesting chain scission and oxidation of the
ether linkages and aromatic/aliphatic structures in the epoxy [1]. Although ethers and
hydroxyls cannot be distinguished from the C 1s spectra alone, the 018 peak clearly
shows an increase in the hydroxyl concentration on the surface implying the decrease in
the ether/hydroxyl peak in the Cls spectra must be because of a decrease in the ether
concentration.

The addition of oxygen functional groups on the surface leads to an increase in
the surface energy as shown in Figures 5.3 and 5.4. Increases in surface polarity (polar-
dispersive model) and the Lewis base component (acid-base model) are appreciable and

are the main contributors to the total increase in surface energy.

Table 5.1 Surface chemical composition of UVO treated epoxy.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

”a $3 A 9
SA :A gx 5A g:
C N O 8'25 $726 :81 3% :9.
O/C .o a e g, >. a 2 a >. .
:3 "O :1 >5 Nfi
(%) (%) (%) 82 >2 3!: 3:2 8
58 §8 89 DB Ea
>4 _g D 5 I: 2
Untreated 82.4 2.4 15.2 0.18 49.9 27.9 4.6 0.0 1.4
608 UVO 70.6 2.9 26.5 0.38 39.2 17.8 7.2 6.3 11.3
1208 UVO 63.4 3.3 33.4 0.53 23.7 19.8 9.6 10.3 11.2

 

Values expressed as percentages of total surface.

145

 

Work of Adhesion (mJ/sq.m.)

 

150

140 - 9

1.11"":5‘*""¥"""“f1 T

I

120 4 i
110 d I, g 0 3.5 mW/sq.cm

 

 

 

 

 

 

 

100 - O 2.6 mW/sq.cm
, I 1.9 mW/sq.cm
90 ¥ A 1.6 mW/sq.cm
80 1 1 T 1 l 1 T
0 50 100 150 200 250 300
Irradiation (mJ/sq.cm.)

Figure 5.2 Work of adhesion of UVO treated epoxy with deionized water
as a function of irradiation. Data obtained at various combinations of
irradiance and exposure time.

146

350

 

 

 

 

 

 

 

 

 

 

50 — --
S;40Kuo$u<i...o....6.<>....6....‘.<.> ........... 9
m 0 __.___.‘_..

"\é A/At/A \A
E13 30 ~ ‘A ..
8 20 H /.- " A OLW
.g . " 9 Acid
3 10 3' ABase
. 4 0 Total Surface Energy
0 b—————4¢ """" 1"i""""¢.""”1"‘¢"", """ l
0 50 100 150 200 250 300
Irradiation (mJ/sq.cm.)

Figure 5.3 Acid-Base surface energy of UVO treated epoxy
at various combinations of irradiance and exposure time.

147

Surface Energy (mJ/sq.m.)

60

 

 

 

 

 

 

 

 

..... Q -_-.,,,
Q 2"”..‘3". . ...... "
50 a t’ "”.
,V
40
1 ......... ..... ..... . ..... . .............. 9
3O - i A
l ---- *' ”"r ~~
..... “a
20 ~ y A ‘
‘5 " 'A o Dispersive ]
10 a .0"; APolar
K... 0 Total Surface Energy 1
O 1
0 50 100 150 200 250 300
Irradiation (ml/sq.cm)

Figure 5.4 Polar-Dispersive surface energy of UVO treated epoxy

at various combinations of irradiance and exposure time.

148

5.2 Crosslinked Tire Rubber Compound

A polydiene rubber compound was obtained in the form of sheets from Goodyear
Tire Co. The exact composition of the rubber, used in tire manufacture, is unknown but
typically tire rubber compounds are blends of natural and butadiene based synthetic
rubbers with high carbon black loadings, UV stabilizers, and anti-oxidants to retard
photodegradation and oxidative aging of the material [2]. The untreated elastomer is
highly hydrophobic with a water contact angle over 120°. The surface of the rubber
sheets has a rough texture which may cause hydrophobic contact angles to be higher, and
hydrophilic angles to be lower, than that measured on a chemically equivalent smooth
surface [3]. UVO treatment of the material yields a highly hydrophilic surface with a
contact angle of 20° and a corresponding increase in the work of adhesion with water
from approximately 30 mJ/m2 to 140 mJ/mz.

Rubber samples were treated using the RC 747 lamp at different irradiances and
times, but the work of adhesion was found to be independent of both and only varies with
total irradiation as shown in Figure 5.5. The rate of change in the work of adhesion for
rubber is very high because of the high susceptibility of the unsaturated butadiene
segments in the elastomer backbone to photooxidation in the presence of ozone.

The presence of carbon black in rubber leads to heating of the surface as the
material is exposed to radiation, especially the infrared radiation emanating from the UV
lamp. The effect of surface temperature on the observed properties was studied using the
RC 500 lamp. Surface temperatures were measured by an infrared pyrometer using an
emissivity value of 0.95 for the polymer. Table 5.2 shows the temperature rise in the

surface of a rubber sample placed 20 mm away from the UV lamp.

149

 

160

 

 

 

A 140 -- _ ________,____,-_.t_.
S ,,,,,,,,,,,, f
g 120 e xi" I
“a“ 100 - x I;
‘8’ ,3
“a 80 2 E’
8’
<1 60 ~ f
8 1
fi 40 7i
O E’
3 20 -

O 1 1 1

0 100 200 300 400 500

Irradiation (ml/sq.cm)

Figure 5.5 Work of adhesion of UVO treated crosslinked tire rubber compound
with deionized water as a function of irradiation. Data obtained at various
combinations of irradiance and exposure time.

Table 5.2 Temperature of crosslinked tire rubber compound surface after UV

 

 

 

 

 

 

exposure in RC 500 lamp.
Exposure Time
(RC 500 lamp, 20 mm) Temperature (°C)

0 sec 24 i 1

30 sec 79 :1: 2

60 sec 107 :1: 3

120 sec 139 i 12

240 sec 158 i 5

 

 

 

 

150

The surface temperature can be a very important variable in the treatment of low
Tg polymers such as elastomers because the high energy polar functional groups
introduced on the surface by oxidation experience a thermodynamic driving force to
reorient below the surface in an effort to reduce the overall surface energy of the material.
This reorientation can occur more easily in polymers above the Tg due to enhanced chain
mobility caused by an increase in the free volume. The effect of reorientation was studied
in rubber by comparison of contact angles in samples allowed to reach high temperatures
and samples whose surface temperature was maintained below 75°C by intermittent UV
exposure to allow cooling of the surface. The intermittent exposure cycle used for
lowering the rubber surface temperature when exposed to the RC 500 lamp was: 30
seconds on, 30 seconds off. Only the ‘on’ segments are used to calculate the total UV
exposure time. Table 5.3 shows the strong effect of temperature on the rubber contact
angle. A similar effect of temperature was observed in the surface composition and

wettability of UVO treated polyethylene (section 5 .5) and Bexloy® (section 5.6).

Table 5.3 Effect of temperature on UVO treated crosslinked tire rubber compound.

 

 

 

 

 

Maximum
Treatment Exposure Cycle Temp. (°C) Contact Angle (deg)
Untreated -- 24 i 1 120 i 5
120 sec Ozone Continuous Ozonation 24 :1: 1 120 1 6
120 sec UVO Continuous UVO (120 Hz) 139 i 12 64 :1: 29
Intermittent UVO (120 Hz)
120 sec UVO 30 sec On + 30 sec Off 69 :1: 5 20 :1: 4

 

 

 

 

 

 

151

5.3 Poly(ethylene terephthalate)
The polyethylene terephthalate (PET) polyester (Melinex®, Dupont Teijin Films)
was obtained in the form of biaxially oriented sheets of 125 micron thickness. The

chemical structure of PET is shown in Figure 5 .6.

_ —

O 0

II II
‘r—O—(CHflZ—O—C‘C—t‘

— ‘11

 

 

Figure 5.6 Chemical structure of PET.

Samples were UVO treated using the RC 747 lamp at different irradiances and
times. Figure 5.7 shows the work of adhesion with water for PET at various values of
irradiation. The work of adhesion was found to be independent of both irradiance and
time, and only varies with the total irradiation.

Table 5.4 describes the chemical changes on the PET surface before and after
UVO treatment. The untreated PET has a O/C ratio of ca. 0.33 which increases to ca.
0.44 after 120 seconds of UVO exposure. Details of the Cls spectrum for the untreated
PET shows approximately 48% hydrocarbon, 15% ether/hydroxyl, and 10% carboxylate
carbon atoms on the surface. The errors associated with the above concentrations are
approximately 10%. Following UVO treatment, there is a 11-12% decrease in carbon
atoms with hydrocarbon bonds and a slight increase in the carboxylate carbon content

indicating oxidation of the polymer.

152

 

 

 

 

 

 

 

 

”a“ 130 ~
2‘ _ ..... {0141“} --------- “@-
E 1 411' ‘ 11
t:
o ,1
g 110 — ,I
:33 § A 1.618 mJ/sq.crn.
“a 100 “ :' 1:11.899 ml/sq.cm.
E 90 fl 0 2.662 ml/sq.cm
80 1 1 1
0 100 200 300 400
Irradiation (ml/sqcm.)

Figure 5.7 Work of adhesion of UVO treated PET with deionized
water as a function of irradiation. Data obtained at various
combinations of irradiance and exposure time.

Table 5.4 Surface chemistry and atomic composition of UVO treated PET surface

 

 

 

 

 

 

 

 

 

 

 

 

Ether/
C O O/C C-C Hydroxyl Carbonyl Carboyxlate
(%) (%) Ratio (%) (%) (%) (%)
Untreated 74.3 24.5 0.33 48.4 14.9 0.0 10.0
60 sec UVO 69.6 28.8 0.41 39.6 15.7 0.0 13.2
120 sec UVO 68.1 29.9 0.44 36.6 15.7 1.1 13.4

 

153

 

5.4 Poly(tetrafluoro ethylene)

Polytetrafluoroethylene (PTFE, Teflon®) is an inert polymer with very high
resistance to thermal and photodegradation, oxidation, and chemical attack because of the
stable bonds between carbon and the highly electronegative fluorine atom. Figure 5.8
shows the chemical structure of PT FE. The UV radiation available from xenon lamps is
of insufficient energy to cause 0'——>0'* transitions in CF bonds. Surface modification of
fluorinated polymers is very difficult and often involves changing the topography of the
surface to enhance adhesion by means of increasing contributions from mechanical
interlocking processes. Aggressive treatments such as plasma etching and oxidation are

more often needed to modify PTFE.

1 1
113-11
F F n
Figure 5.8 Chemical structure of PTFE.

The strong carbon-fluorine bond has very low absorption in the wavelength range
available from xenon lamps and no change in surface energy is expected after UVO
treatment. PTFE samples were UVO treated using the RC747 lamp and approximately
700 ppm ozone. The contact angles before treatment were approximately 110° and did

not change even after extended UV exposure as shown in Figure 5.9.

154

 

00
O

 

 

’7 70 ~
5
8‘
E 60 2
5051 ——————— inf-n1 ------ 1 --------------------------
, 1 1
'8
95 4O -
i
o
3 30 -
20 1 1 1 1 1 1

 

0 50 100 150 200 250 300 350
Irradiation (mJ/sq.cm)

Figure 5.9 Work of adhesion of UVO treated PTFE with deionized water as a
function of irradiation.

155

5.5 Polyolefins

Polyolefins represent one of the major groups of polymers used in the consumer
goods industry. Polyethylene (PE), polypropylene (PP), and polypropylene based
thermoplastic polyolefin elastomers (TPO) are the commercially important polymers in
this class. Polyolefins have simple structures comprising of OH bonds that are very
stable under UV irradiation. The stability of OH bonds is a result of the high energy
required for o—>0'* transitions. Typically this energy is available only in wavelengths
below 160 nm. UV irradiation with a xenon lamp should theoretically not have any effect
on saturated polyolefins. However, saturated polyolefins have been the subject of many
studies [4-13] which show that it is possible to oxidize and modify these materials. The
consensus is that oxidation in these materials is sensitized by internal and processing
impurities. Internal impurities include main chain unsaturation, catalyst residues from the
manufacturing process, and carbonyl groups formed by adventitious oxidation.
Processing impurities can include hydroperoxides, carbonyl groups, and transition metal
ions from polymer processing equipment [1]. UV irradiation forms free radicals at these
impurity sites and further reactions of the polymer with the free radicals leads to chain
scission, crosslinking, and oxidation. Three types of polymers, an ultra high molecular
weight polyethylene (UHMWPE), PP, and TPO were studied.

UHMWPE samples were abraded with 360 and 600 grit wet sandpaper for 5 mins
to remove any contaminated surface layers and washed with water prior to irradiation
using the RC500 lamp. Initial experiments showed the UHMWPE surfaces to be very
difficult to treat. Table 5.5 shows the results of contact angle measurements on

UHMWPE after UVO treatment. The contact angles of untreated UHMWPE did not

156

 

show any appreciable statistically significant change. The surface temperature of the PE
was measured to be approximately 100°C after 120 seconds of UV exposure. The Tg of
polyethylene is —78°C, which, as mentioned in the previous section, can lead to surface
reorientation at high temperatures. Samples were treated at low temperatures by
intermittent irradiation for 15 seconds followed by a dark cooling period of 60 seconds.
The maximum temperatures recorded for the intermittently irradiated samples was
approximately 40°C. The low temperature treatment enhances the wettability of the
polymer. The adhesive bond strength of the treated surfaces with epoxy adhesive
(Araldite 2011) is shown in Figure 5.10. Increases in the adhesive bond strength up to
300% were measured for UVO treated UHMWPE.

Polypropylene was obtained in the form of powder (KE020, Montell) and
injection molded into plaques. A reactor grade TPO based on polypropylene and
ethylene-propylene rubber (Equistar) was also studied. The two materials behave very
similarly, and contact angles of the untreated material range from 90°-100° and UVO
treatment does not change the contact angles significantly. However, adhesion tests show
a large increase in the adhesive bond strength of these materials. Table 5.6 shows the
XPS analysis of UVO treated PP and TPO. The O/C ratio of both PP and TPO increases
following UVO treatment and the oxygen incorporated into the surface is mainly in the
form of ethers which are relatively non-polar and hence do not contribute to an increase
in surface energy. The irradiation of saturated polyolefins can lead to the formation of
extensive crosslinks in the polymer [5,6,9]. The increase in the adhesive bond strength of
UVO treated materials is thus most likely a result of surface crosslinking which increases

the modulus of the surface. In the case of TPO, some silicon is measured and this is likely

157

from a process additive blooming to the surface from the bulk. After UVO treatment, the
silicon content increases suggesting an increase in the bloom caused by heating of the

substrate during treatment. In spite of the silicon bloom, adhesive strength increases by

400-500% in the case of TPO as shown in Figure 5.11.

Table 5.5. Water contact angles of Ultra-High Molecular Weight Polyethylene after
UVO treatments with and without thermal management.

 

 

 

 

 

Treatment Contact Angle Contact Angle
(Temp z 100°C) (Temp z 40°C)
Untreated 100 i 8 100 i 8
30 sec UVO 97 i 4 _-
60 sec UVO 94 i 3 89 i 2
1208UVO 88i2 67i6

 

 

 

 

 

Table 5.6 Surface chemistry of PP and TPO after UVO treatment.

 

 

 

 

 

Material C (%) O(%) Si(%) O/C Si/C
Untreated PP 97.3 2.7 -- 0.03 --
120 sec UVO PP 87.0 13.0 -- 0.15 --

Untreated TPO 91.9 6.3 1.9 0.06 0.02

120 sec UVO TPO 82.2 14.0 3.8 0.17 0.05

 

 

 

 

 

 

 

 

158

700

 

 

5005
400:
300-
200:
100-

Adhesive Strength (psi)

 

 

 

0 sec 30 sec 60 sec 120 sec
UVO Treatment Time

Figure 5.10 Stub pull adhesive bond strength of UVO treated UHMWPE.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

600 145

500 : ‘1': 1 ~ 125
:5} l. l {3 E9
5 400 ‘1 i‘ wi+_..-1— _ 105 V 3
°0 ‘~~ __..-—i.-"-" 1'~~ 8 4)
§ "7" ,I “i .3 '50
a 300 ‘1 ”/ _. 85 a 2
Q) ’ H
.2. F E 8
3 200 « ,I 1 65 12 a
‘6 I E 8
<3 ,’

100 - I I — 45

I
It
0 1 1 1 1 25
0 sec 30 sec 60 sec 90 sec 120 sec
UVO Treatment

 

 

 

1: Adhesion - -I- - Contact Angle — + - Temperature

 

Figure 5.11 Adhesive strength, contact angles and surface temperature of UVO
treated TPO showing no significant change in contact angles despite the increase
in adhesive bond strength.

159

5.6 Ionomers (Bexloy®W)

Bexloy®W (DuPont) is a glass filled composite based on Surlyn®, an ionomer of
ethylene methacrylic acid copolymers. The ionomer is made by neutralizing part of the
methacrylic acid with metal ions such as sodium. The polymer structure has three
regions: amorphous polymer, crystalline polymer, and ionic clusters which act as
crosslinks in the material. The polymer has a low glass transition temperature (Tg) and
increasing the temperature of the polymer causes the domains to lose structure and
undergo reorientation. The Tg of the Bexloy®W used was 50°C.

UVO treated Bexloy®W sheets were analyzed by XPS and contact angle
measurements'. Samples were treated under continuous and intermittent UV exposure.
Continuous exposure caused the polymer surface temperature to exceed the Tg whereas
intermittent exposure cycles with 30 second exposure and 30 second dark cooling period
was sufficient to limit the surface temperature to below 50°C.. The results are shown in
Figure 5.12. The O/C ratio and wettability increases with UVO exposure as long as the
temperature was maintained below the Tg. When the temperature was allowed to exceed
the Tg, and immediate decrease in the O/C ratio and wettability was observed. The time
of irradiation was found to have no effect on this reorientation as samples treated for

longer times at temperatures below 50°C did not show any reorientation.

 

' Experiments performed by Dr. Per A. Askeland, Composite Materials & Structures Center, Michigan
State University.

160

 

 

 

12:1 O/C Ratio - 9- -Temp - 9- - Contact angle

 

 

ADV 8:223th

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m 0 0 0 0 0 0 0 0 0
1 9 8 7 6 5 4 3 2 l 0
mooneceoo
O, _ T. oom
/ \ GS
/
/ x 1
x a
/ V , 30:52.8
/ \ m _ 0% oo
\ \ Al
\ \ / 1
\ \ « “=65“?an
fl _ 4 11.1.1 oom ova
/ _
z _ -
z _ 8036:85
on]. 4
1 1. i 08 cm“
/ g
I y r
/
Ii , Ems—Euge—
6 It . .
08 cm
4 / 1
I I
r.
TI. _obcou
5. A 3. 2 .J O.
0 0 0 0 0 0

8 88 2w=< 83:00

23% 05

UVO Treatment

Figure 5.12 OIC ratio, contact angles and surface temperature of UVO treated

Bexloy®W showing surface reorientation effect above the Tg (50°C).

161

5.7 Summary
The surface modification of some common polymers with UVO treatment was
described in this chapter. For most materials, sharp increases in the surface oxygen,
wettability, and adhesive strength are observed. The properties of epoxy, PET, and
rubber were studied in detail and the wettability was found to depend entirely on the
amount of irradiation given to the sample. In the case of rubber, a strong effect of
temperature on the surface properties was also observed. The rise in surface temperature
caused by exposure to UV and IR radiation from the lamp was observed to cause
reorientation of the oxidized polar groups. The effect of this regeneration is readily
observed by the decrease in wettability. This regeneration of the low energy surface was
also observed very clearly in thermoplastics like polyethylene and Bexloy®W polymers
above their glass transition temperatures.
The adhesive strength of surfaces depends on many factors other than surface
energetic favorability. In polymers like PP and TPO, the wettability of the surface does
not change after UVO treatment. However, the adhesive strength increases by 400% to
500% of the untreated material. This is believed to be an effect of the crosslinking of the
surface and near-surface polymer leading to the formation of high modulus material at
the interphase which can enhance adhesion.
Almost all materials treated with UVO show beneficial changes in properties
except fluorocarbon based polymers. The fluorocarbon bond is very strong and UV
radiation in the range provided by xenon lamps is not energetic to cause excitations in the
C-F bond. Teflon did not show any changes in the surface properties even after long

exposures to UVO. With the help of the phenomenological data obtained for various

162

 

types of polymers, the following chapter will attempt to identify trends in various UV
treatment process and model the changes in a way that can be easily applied in industry to

understand and predict the properties of UVO treated materials.

163

10.

11.

12.

13.

References

Rabek, J .F. Mechanisms of Photophysical Processes and Photochemical Reactions in
Polymers, John Wiley, New York, 1987.

Brydson, J .A.; Plastics Materials, 5‘h ed., London, 1989.

Adamson, A.W. Physical Chemistry of Surfaces, 5‘h ed., Wiley Interscience, New
York, 1990.

Dasgupta, S. Journal of Applied Polymer Science, Vol. 41, pp. 233-248, 1990.
Qu, B.; Ranby, B. Journal of Applied Polymer Science, Vol. 49, pp. 1799-1807, 1993.
Ranby, B. Polymer Engineering and Science, Vol. 38, N0. 8, pp. 1229 —1243, 1998.

Wilken, R.; Hollander, A.; Behnisch, J. Macromolecules, Vol. 31, pp. 7613-7617,
1998.

Strobel, M.; Jones, V.; Lyons, C.S.; Ulsh, M.; Kushner, M.J.; Dorai, R.; Branch, M.C.
Plasmas and Polymers, V01. 8, No. 1, pp. 61-95, 2003.

Yu, J.R.; Chen, Z.L.; Zhu, J .; Liu, Z.F.; Wang, QR. Intern. Polymer Processing,
XIV, 4, pp. 331—335, 1999.

Pasternak, M. Journal of Applied Polymer Science, Vol. 57, pp. 1211-1216, 1995.

Macmanus, L.F.; Walzak, M.J.; Mcintyre, N.S. Journal of Polymer Science: Part A:
Polymer Chemistry, Vol. 37, pp. 2489-2501, 1999.

Gijsman, P.; Sampers, J. Polymer Degradation and Stability, Vol. 58, pp. 55-59,
1997.

Gijsman, P.; Meijers, G.; Vitarelli, G. Polymer Degradation and Stability, Vol. 65,
pp. 433-441, 1999.

164

CHAPTER SIX

DISCUSSIONS AND PROCESS MODEL

In the preceeding chapters, the effects of UV modification on the surface
properties of various polymers have been discussed. This chapter will provide an
overview of the process and discussions about the effects of different process variables,
followed by a proposed process model that can be used as a predictive tool for estimating

surface properties of UV modified polymers.

6.1 Process Variables

The main variables which can be controlled in the UV treatment process are:
spectral output of the UV lamp, lamp pulse frequency, irradiance, exposure time,
temperature, humidity, and ozone concentration. The effect and importance of these

variables are discussed below.

6.1.1 Spectral Output

Pulsed xenon flashlamps were used in this study because of the strong spectral
output of these lamps in the UV-C region, which provides radiation at 185 nm and above.
Analysis of polymers UVO treated through a Pyrex® window acting as a 280 mm high
pass filter showed no appreciable change in the surface properties. This is because of two
reasons: i) the ozone absorption maximum (71m) is at 254 nm and in the absence of this
wavelength, photodissociation of ozone does not occur, and ii) most engineering and

commodity polymers are designed to withstand ground level exposure to solar irradiation

165

 

and consequently do not have chromophores with very high absorption coefficients above
300 nm, or are modified to enhance stability at those wavelengths [1]. The presence of
continuous radiation in the 185 nm — 280 nm is thus essential for UV treatment.
Considering the ultimate goal of developing the UV treatment process to modify a vast
range of polymers, which may have widely different UV absorption bands, the use of
lamps with continuous spectral output, as opposed to line output, is also essential. In this
respect, xenon flashlamps are preferred over mercury lamps because of their broader
output especially at wavelengths below 254 nm.

Excimer lamps and lasers also suffer the handicap of narrow spectral outputs
which may be able to excite a specific chromophore in a polymer but may have no effect
on a slightly different chromophore. The absorption profiles of most organic bonds
become virtually continuous in the VUV region (below 185 nm) because of the close
spacing of the excited energy levels [2], and conceptually, surfaces can be UV oxidized
more efficiently under VUV irradiation, but the drawback of using VUV radiation is the
strong absorption by oxygen which necessitates exposure in vacuum. In a commercial
process, the use of vacuum is a severely limiting factor, not only in terms of complicated
equipment, but also in terms of low throughputs and the necessity of batch operations as
opposed to continuous operations. The use of xenon flashlamps with strong output
between 185 and 280 nm makes it possible to treat the widest range of polymers in
ambient conditions and makes the process more adaptable to continuous processing in

manufacturing environments.

166

6.1.2 Irradiance

The effect of irradiance on the changes observed in surface properties for a wide
range of polymers has been previously described in detail in Chapters 3, 4, and 5. For all
the polymers evaluated in this study, it was possible to achieve identical surface
properties even when the irradiance (power) was varied by a factor of five as long as the
total energy incident on the surface (irradiation) was equal. In terms of irradiation of the
surface, this implies a trade-off between the irradiance and time. Samples exposed to long
exposures at low power, and short exposures at high power, were found to have
equivalent surface properties. Thus it can be concluded that irradiance itself is not a
controlling process variable in the UV treatment of surfaces.

However, such a broad conclusion can only be made for the surface of the
material and does not describe the changes occuring within the material. For a polymer
following Beer’s law of absorption [1], which states that the intensity of radiation 1, at a
depth of b from the surface in a material with absorptivity 8, concentration c, and incident
intensity 1,), is given by I = Io 10‘°°°, the irradiance at any depth is dependent on the
incident irradiance. If the modification is a function of irradiation, then a change in
irradiance will affect the depth of penetration of the radiation, and possibly the depth of
modification. Equivalent irradiation at low irradiance (low power) should then be
expected to have lesser penetration into the material compared to irradiation at high
irradiance (high power). Thus, the trade-off between power and time could affect the
depth of treatment and possibly the properties of the bulk material. It is important to
reflect on the use of the term ‘treatment’ in the context of changes occuring below the

polymer surface. The modification of a polymer by UV and ozone encompasses a variety

167

of processes including photolysis, photorearrangement, photooxidation, and crosslinking.
Moreover, the concentrations of quenching species like oxygen and ozone in the bulk
polymer are very different from that on the surface and can affect the rates of free radical
recombination in the bulk polymer. The balance between these multiple parallel
processes can change with depth and affect what is conveniently termed the ‘treatment’
of the bulk polymer. Although the effective depth of penetration of the photons in the
polymer is known, the effect of that irradiation on the polymer can be very different from
what is observed on the surface. Due to the difficulties involved in measuring properties
at spatial resolution of monolayers, the effect on the depth of treatment has not yet been
measured. The evolution of properties in the bulk polymer as a function of depth in the
near surface polymer is one of the areas which needs to be investigated in the future.
Another assumption in this study concerns the measurement of irradiance. UV
irradiance at 254 nm was measured with a 10 nm bandpass filter. The spectral irradiance
over the entire UV-C spectrum ranging from 185 — 280 nm was not measured due to the
unavailability of a NIST calibrated standard lamp. The irradiance at 254 nm is a narrow
band measurement and characterization of the UV treatment process based on irradiance
involves the implicit assumption that the complete spectral profile of the lamp is similar
at all irradiances within the measured range. Measurement of irradiance in such a manner
also makes it imperative to state that the results obtained in this study are only verified
for pulsed xenon flashlamps operating at 120 Hz frequency and whether other types of

UV sources yield the same results must be verified experimentally.

168

6.1.3 Exposure Time

As in the case of irradiance, the overall UV exposure time is not a controlling
variable and the exposure time is important only in the determination of the total
irradiation incident on the surface for a chosen value of irradiance. Other time factors
which may be important are the time required for the process gas (ozone) to reach a
surface newly introduced in the treatment environment. The amount of ozone required in
the surface oxidation process is very small and a monolayer of ozone gas adsorbed on the
surface is adequate to oxidize the top surface layer of the material. No dependence on the
exposure time was observed at the irradiance levels used in this study (i.e. for exposures
on the order of seconds.) This is likely the result of the rate of formation of a
physisorbed monolayer of a gas at standard pressure and temperature which typically

occurs within a fraction of a second.

6.1.4 Ozone Concentration

The actual ozone concentration used in UVO treatment did not have a pronounced
effect on the surface properties, although at very low concentrations (less than 200 ppm),
the variability in data was observed to increase. This was studied in detail for
polycarbonate, but the effect was observed in elastomers as well, and is believed to be
due to the inefficiencies in the treatment chamber used in this study. For many samples
treated at low ozone concentrations, some areas of the sample showed high wettability
while other areas had properties similar to the untreated material. This type of non-
uniform surface can be caused by channeling of the gas flow where there is incomplete

mixing between the incoming stream of process gas and the existing gas in the chamber,

169

causing local areas of starvation where no ozone is present. At higher concentrations,
these inefficiencies are alleviated to some degree by the excess amount of ozone as well
as increased diffusion throughout the chamber. Thus, we can conclude that ozone is
necessary for UV oxidation of polymers, but the concentration of ozone near the sample
surface is far in excess of the ozone necessary for the surface reaction, hence the
concentration does not have any effect on the degree to which the surface can be treated
when ozone concentrations on the order of a few hundred ppm are used and would not be
a rate limiting factor.

The ozone concentration can, however, affect the UVO treatment in other ways. It
has been demonstrated earlier that the extent of surface modification is dependent on the
total irradiation and the wavelengths necessary for UV oxidation range from 185 nm to
300 nm. A concentration dependence can arise indirectly in the process at very high
ozone concentrations because ozone has a strong absorption maximum centered at 254
nm and ranging in width from 200 to 300 nm as shown in Figure 2.6. This wide
absorption in the UVC region can cause attenuation of irradiance which has already been
shown to be the main variable in the UV treatment process. In this study the attenuation
by ozone was accounted for by measuring the irradiance under actual processing
conditions, including passage of the UV radiation through a layer of ozone at the typical

concentrations used in UVO treatment.

170

6.1.5 Temperature

The surface temperature during UV treatment can have a pronounced effect on the
extent of surface modification for polymers above their glass transition temperature (Tg).
Studies on rubber, Bexloy®, and polyethylene suggest that extensive surface
restructuring, including thermodynamically driven reorientation of the polar functional
groups away from the surface in order to reduce the total surface energy [12], can occur
in the polymer as the surface temperature increases during treatment. This effect is not
seen in polymers treated below their Tg because of the reduced chain mobility.

The increase in the surface temperature during UV treatment can occur because of
various reasons. UV lamps can have significant output in the infrared (IR) region. IR
radiation is absorbed by the polymers resulting in vibrational excitation to higher levels.
The relaxation from the excited vibrational states to the ground state is accompanied by
release of thermal energy. Similarly, UV radiation causes electronic excitations in
molecules and the relaxation processes from the excited state often include intermediate
vibrational relaxations (leading to the phenomena of fluorescence and phosphorescence)
which ultimately leads to the release of thermal energy[3]. In addition heat generation by
radiation, the exothermic oxidation reactions occurring on the surface can also contribute
to the increase in temperature. Thermal management strategies have been shown to work
well for such polymers. The use of water filters, or dichoric reflectors on UV lamps can
reduce the amount of IR radiation reaching the surface. Similarly, lamp design can be
modified to reduce the fraction of IR emission from the lamps. A more practical approach
is to remove the generated heat from the polymer surface during treatment. One of the

approaches adopted in this study is the use of intermittent exposure. Samples were

171

irradiated for short times and allowed a dark cooling period to maintain low surface
temperatures and the process was repeated till the desired irradiation was achieved.

The most striking results were obtained in Bexloy® which has a Tg of ca. 50°C.
With temperatures maintained below the Tg by the use of intermittent exposures, both the
OIC ratio measured by XPS, and the wettability measured by contact angles, increases
with UVO treatment. When no thermal management is present, and the surface
temperature is allowed to exceed the Tg, the polymer loses its ionic crosslinks and
undergoes extensive reorientation, resulting in a decrease in the OIC ratio and wettability.
The influence of temperature on the measured O/C ratio, which is an average of the
composition over a sampling depth of 75-100 A, illustrates that restructuring can occur

well below the top monolayer of the material.

6.1.6 Pulse Frequency

The experimental work described so far was performed on xenon flashlamps
operating at a pulse frequency of 120 Hz. The UV lamps used also have the capability to
be operated at different frequencies. The RC740 lamp operates at 10 Hz while the RC747
lamp has a variable frequency module allowing operation at 3H2 and 120 Hz. The
advantage of using a pulsed lamp is the dark period available for cooling of the surface
between two UV pulses. Table 6.1 shows the average cooling time, pulse widths, and
pulse energy for different operating frequencies. It can be readily observed that as the
pulse frequency increases, the cooling time available between two pulses, as well as the
average energy output per pulse, decreases as the reciprocal of the frequency. The

considerations of low sample surface temperatures have been discussed in the previous

section. To keep the temperatures low, and yet have high enough irradiance to treat
surfaces in as short a time as possible, a balance must be reached between the energy

given to the surface per pulse and the cooling time available between two such pulses.

Table 6.1 Cooling time and energy per pulse for a 2.6 mW pulsed UV output.

 

Pulse Frequency Pulse Width Cooling Time/Pulse Energy/Pulse

 

 

 

3 Hz 150 us 333.1 ms 887 uJ/pulse
10 Hz 150 us 99.9 ms 266 uJ/pulse
120 Hz 150 us 8.2 ms 22 uJ/pulse

 

 

 

 

 

 

lt was found from observation that increasing the cooling time between pulses
was not as effective a thermal management strategy as reduction of the energy per pulse.
The reason for this is that the radiation is absorbed by the polymer to some depth and this
absorption depth increases with increasing intensity of the pulses. At low frequencies, the
high intensity pulses can cause liberation of more thermal energy, and at greater depths,
compared to high frequency, low intensity pulses. In poor thermal conductors such as
polymers, the high intensity pulses give rise to thermal gradients and the low thermal
conductivity can cause higher transient temperatures on the surface than low intensity
pulses. The greater cooling time between low frequency pulses was found to be
insufficient compensation for the amount of heat generated in polymers. Surface
temperatures of polymers treated at 3Hz pulse frequency were invariably found to be

higher than the temperatures recorded for equivalent irradiation at 120 Hz. In fact, at 3

173

Hz, the heating effect of the UV lamp is so strong that even thermally stable, high
temperature polymers such as silicone rubber were found to char and combust after 4 -5
minutes of UVO exposure. Thus, for polymers, higher frequencies with low intensity

pulses are more suitable than low frequency high intensity pulses.

6.1.7 Humidity

The effect of humidity was not determined in this study, although some
experiments were performed to successfully demonstrate the possibility of treating
polymers through a film of water. The interaction of UV with water has been well
reported in literature [4,5]. UV radiation leads to the formation of hydroxyl and
hydroperoxy radicals in water which have the ability to react with chemical waste
compounds as well as polymer substrates. This chemistry is widely used in water
purification and remediation systems and similar reactions can occur in water in the gas
phase. The influence of humidity is one of the UV treatment process variables that needs

to be studied in greater detail in the future.

6.2 Process Model

The UV oxidation process involves far too many different sub-processes and
pathways to describe the changes occurring in a wide range of materials with a unique
universal model. The oxidation of each polymer follows different pathways and unless all
the intermediate reaction mechanisms are known and characterized, analytical
descriptions are bound to fall short. In addition, the polymers studied here are

commercial polymers, many of which are compounded with stabilizers, anti-oxidants,

174

and processing additives. These low molecular weight additives often migrate to the
surface and create a chemistry which is almost always different from that expected from a
pure material. The surfaces of polymers processed at high temperatures can also be
oxidized to a certain extent. Processing conditions can also affect the surface morphology
and a polymer processed at two different conditions can have very different surface
properties, be it due to changes in crystalline structure, molecular weight, or composition.
Unfortunately, such ‘real world’ materials are the rule rather than the exception in most
industrial applications and the description of a process to modify these materials must
take into account the inherent variability of the surface. It is fair to say that for a
commercial polymer we often only have a general idea of the composition, structure, and
processes occurring on the surface. To describe the effect of a surface treatment process
on such materials, an empirical approach is the most direct route.

In that respect, UVO treatment has many similarities to long term
photodegradation and weathering in polymers. Some investigators in the field of polymer
degradation have developed the concept of polymer sensitivity curves to describe the
susceptibility of a polymer to long term photodegradation. Allan et al. proposed an
approach to model the UV degradation of the bulk properties of a polymer with different
loadings of photostabilizers [6]. The model is based on the determination of a
“wavelength sensitivity spectrum” over a range of solar wavelengths to model the
degradation of polystyrene with benzotriazole, benzophenone, and triazine UV stabilizing
additives. Their approach was to model the change in any property (P) of the bulk
polymer in terms of the incident irradiance and an empirically determined sensitivity (S).

The Allan model is described by Equations 6.1 to 6.4.

175

AP(/i.,x) = I(
t

4.105(2) (6.1)
where I(X,x) is the irradiance in the sample at wavelength )1. and depth x, and SO.) is the
wavelength sensitivity spectrum. The depth of penetration of the radiation is modeled for

a solid obeying Beer’s law such that I = IO 10°“.

APOt, x)

1 =1010‘ 80" 5(a) (6.2)

 

Integrating Equation 6.2 over the over depth x = 0 to (1, gives

 

 

_ -8Cd
APOt) :10 l 10 SO.) (6.3)
t 8C

The total degradation rate in the polymer is determined by numerical integration over all
wavelengths using experimentally determined source spectra, absorbance spectra, and

sensitivity spectra.

 

__ —8cd
g: [101 1° s(i)di (6.4)
1 8C

The wavelength sensitivity spectrum was determined by measuring the degradation of the
property using a series of Schott glass cut-off filters. Differences in property degradation
between samples exposed behind different pairs of cut-off filters were ascribed to the
radiation in that band of wavelengths by assuming that the sensitivity is constant over the
measured irradiance band between the two filters. The irradiance bandpass between two
filters spanned almost 100 nm at 10% transmittance in their study. A similar approach

has been followed to model the UV surface treatment process.

176

6.2.1 UVO Surface Treatment Model for Xenon Flashlamps

Investigation of the various process parameters described in section 6.1 show that
for UV treatment, the only dependence of the evolution of surface properties in the
presence of excess ozone is on the irradiation, H, received by the sample. The spectral
band mainly responsible for the photooxidation reactions has also been determined to
range from 185 - 280 nm. Given the ca. 100 nm bandwidth of the radiation of interest,
irradiance measurement at 254 nm was used as a one—point measurement for the entire
spectral band. The use of irradiance measurement at 254 nm can be used in a manner
similar to the irradiance bandpass measured by Allan for a pair of Schott glass filters. The
most general relationship for the rate of change in an arbitrary property of the polymer

surface can be expressed as

_Apfif’“ *1 = 10.18811) (65)

where P is the change in the measured property, t is the exposure time, x is the distance
from the surface, S is a depth dependent spectral sensitivity function, I is the irradiance,
and H is the total irradiation given by the product of irradiance and time. Due to the
limitation of using a single spectral bandpass, an assumption can be made that the
empirical sensitivity of the polymer is constant over the entire range from 185 — 280 nm
provided future comparisons with this model are made with lamps having a similar
spectral output (i.e. xenon arc flashlamps). Equation 6.5 can then be simplified to

AP(H, x) = H(x)S(H) (6.6)
To describe the surface treatment of polymers, the above equation only needs to be
evaluated for depth x = 0

AP(H) = AH - S(H) (6.7)

177

For a differential change in a surface property caused by a differential increase in
irradiation, the sensitivity function, S(H), can be expressed as

d<AP(H))

SH:
() dH

(6.8)

The change in the property can be obtained by integration of Equation 6.8 over the
irradiation range to give

H2
H1

The sensitivity function can be empirically determined for a particular polymer, and the
integral of the sensitivity function gives the net change in the measured property, P, from

the reference value.

6.2.2 Determination of Sensitivity Functions

Work of adhesion sensitivity functions for five ambient temperature UVO treated
polymers (PC, PET, Rubber, PT FE, and Epoxy) were calculated using data obtained from
contact angle measurements with water. The change in the work of adhesion (Wa) for all
polymers, except PTFE, increases rapidly with irradiation and then reaches an asymptotic
value at high values of irradiation. The rate of change of Wa decays exponentially as the
susceptible surface sites on the polymer are oxidized. Wa was plotted as a function of the
irradiation, H, and the resulting plots were analyzed using a curve fitting software
(CurveExpert 1.34 for Windows). Details of the curve fit are given in Appendix B. From
observation of the trends in the surface properties (work of adhesion and O/C ratio) of the
polymers, as well as the derivatives of the work of adhesion curves, the use of an

exponential decay function to describe the trends is strongly indicated. The Wa vs H

178

curves for the polymers were fit to a three parameter exponential association function of

the form y=a(b-exp(-cx)) as shown in Equation 6.10.
wa = (b — e‘CH) (6.10)

where a, b, and c are empirical correlation constants, the work of adhesion has units of
(mJ/mz), and irradiation has units of (mJ/cmz). This function should only be applied
within the range of irradiances measured and does not extrapolate beyond the measured
values. This type of function was fit to the data for four polymers which showed a change
in the work of adhesion. The work of adhesion sensitivity function S(H) for the polymers

is calculated by differentiating Equation 6.8 to give
dW _. H
SH =——a—=ac(e C 1 6.11
( ) dH ( )

For polymers like PTFE that do not absorb radiation in the specified range, the
exponential association function does not apply and the data is fit with a linear function
of the form Wa = aH + b, which gives a sensitivity function of S(H) = a. Other polymers
may not necessarily fit this shape in which case numerical differentiation can be
performed. Wa sensitivity curves were generated using the empirical constants a and c
over the measured irradiation range. Table 6.2 lists the empirical correlation constants for

the five polymers. The corresponding sensitivity functions are illustrated in Figure 6.1.

179

Table 6.2 Correlation constants for the work of adhesion sensitivity
functions for PC, PET, Epoxy, Rubber, and PTFE.

 

 

 

Pol mer Work of Adhesion a b C
3’ S(H) =
PC (a*c) exp(—cH) 72.6 1.97 0.0107
PET (a*c) exp(-cH) 37.7 3.31 0.0292

 

Epoxy (a*c) exp(-cH) 45.2 3.00 0.0155

 

Rubber (a*c) exp(-cH) 107.2 1.27 0.0107

 

PTFE” a 0 -- --

 

 

 

 

 

 

 

# PTFE does not absorb in the tested UV range, hence no modification is possible.

180

 

 

 

 

 

.0

dWa/dH (sq.cm/sq.m)
_o

 

 

 

0 50 100 150 200 250 300 350
Irradition (H) (mJ/sq.cm.)

Figure 6.1 Work of adhesion sensitivity curves for UVO treated polycarbonate
(PC), epoxy, polytetrafluoroethylene (PTFE), crosslinked tire rubber compound,
and polyethylene terephthalate (PET) showing the rate of property change as a
function of irradiation.

181

 

 

6.2.3 Relation to Physical Properties

Examination of the empirical correlation constants a, b, and c shows that b is
simply a scaling parameter related to the reference value of the work of adhesion of the
untreated polymer, thus a higher value of b indicates a higher starting value of the work
of adhesion of the polymer. The parameter c defines the rate of exponential decay of the
sensitivity function. The parameter a is related to the extent to which changes in the work
of adhesion occur, but has no effect on the rate of decay.

Assuming the changes in the work of adhesion are a combined result of
photooxidation and ozonation of the photolysis products on the polymer surface, the rate
of decay of the sensitivity, c, must be related to the rate of depletion and, by inference,
the susceptibility, of chromophores in the polymer to UV oxidation (absorptivity), as well
as susceptibility to oxidative attack by ozone. The extent of modification, affected
primarily by the parameter a, can be related to the relative population of susceptible sites
on the surface.

The parameter a describes the populations of susceptible sites on the polymer
which can be oxidized by UVO exposure. A higher value of a can indicate a chemical
structure with high frequency of susceptible sites, or a polymer microstructure which
causes a large number of susceptible bonds to be exposed on the surface. An estimate of
the relative populations of reactive sites on the surface of different polymers for the
purpose of comparison can be obtained by determining the fractional number of
susceptible bonds in the polymer. A first approximation for the relative density of
susceptible bonds can be obtained by normalizing the number of bonds or sites expected

to react during UV oxidation with respect to the molecular weight of the repeat unit of the

182

polymer. This can provide a very rough estimate for rapid qualitative comparisons. The
utility of this approach is limited because the number of sites are normalized in terms of
the molecular weight and the effects of surface microstructure and configuration are not
considered. A more rigorous calculation should take into account the fractional area or
volume of the surface covered by susceptible bonds, the physical configuration of the
polymer chains, steric shielding effects, and any other factors which may affect the
density of surface sites capable of reacting in UVO exposure conditions. The fractional
volume of susceptible sites can be calculated using the additive group contributions
approach of Bondi and Van Krevelen [7] to determine the Van der Waals volume of the
polymer. This is a more reliable method than comparison based on mass. The two

methods are compared below.

Approximation of ‘a’ by number of sites per unit mass.

Assuming the diene rubber to be pure polybutadiene, there is one unsaturated
bond in the repeat unit of molecular weight 54 Daltons, giving a reactive site p0pulation
density of 1/54 = 1.85 sites/ 100 Daltons. Similarly, the polycarbonate can be assumed to
have 4 sites sensitive to direct UVO attack — the two carbonate linkages and the two
methyl groups in the Bisphenol A unit. The reactive site population density for PC is
4/254 = 1.57 sites/100 Daltons. PET has a repeat unit of molecular weight of 192 and has
two aromatic ester chromophores which are expected to react, giving a site density of
1.04 sites/100 Daltons. The epoxy is a two part system of DGEBA (MW = 582) and
Jeffamine T 403 (MW = 470) in 100:45 weight ratio. The average molecular weight of

the epoxy “repeat unit” is (582*1)+(470*0.45) = 793.5 Daltons. The epoxy resin has 10

183

potential sites in the backbone capable of reacting if the ON crosslinks are considered
reactive, and the curing agent has three ether groups. The three amine groups in the
curing agent are not counted to avoid double counting. The average number of sites for
epoxy are (10*1)+(3*0.45) = 11.3 sites/unit and the site density is 11.3/793 = 1.43
sites/ 100 Daltons. These values are provided only to aid qualitative comparisons and

should not be used for any quantitative purposes for the reasons outlined previously.

Approximation of ‘a ’ by Van der Waals volume from group contribution methods.

The concept of additive group contributions is based on the assumption that a
physical property of a molecule is in some way determined by a sum of contributions
made by the structural and functional groups in the molecule. In a polymer, the
contributions from various sub-groups in the repeat unit can be added to approximate the
properties of the polymer. The Van der Waals volume of a molecule is defined as the
space occupied by the molecule, which is impenetrable to other molecules having normal
thermal energies. The Van der Waals volume of a molecule is assumed to be the volume
of several interpenetrating spheres. The radii of the spheres corresponds to the atomic
radii of the elements involved and the distance between their centers correspond to bond
lengths. Van Krevelen has tabulated the group increments to Van der Waals volumes
determined by Bondi, Slonimskii, and Askadskii, along with group increments for
amorphous, crystalline and semi-crystalline polymers [7]. The molar volumes of
polymers in the rubbery state (Vr), glassy state (Vg), crystalline state (Vc), and semi-
crystalline state (Vsc) have all been related to the Van der Waals volume as shown in

Table 6.3.

184

Table 6.3 Ratio of amorphous, glassy, crystalline, and semi-crystalline molar
volumes of polymers to Van der Waals volumes by group contribution [REF].

 

 

 

 

 

Ratio at 298 K Mean Values
V1/V w 1.6
Vg/Vw 1.6
VC/Vw 1.435
Vsc/VW 1.6 - 0.165xc

 

 

 

 

* xc = crystalline mole fraction.

Since the ratios of the molar volumes of polymers in different physical states to
the Van der Waals volume are constant, calculation of the volume fraction of the polymer
repeat unit susceptible to UVO oxidation is independent of the physical state of the
polymer and scales directly with the Van der Waals volume except in certain extreme
cases such as a highly crystalline polymer subjected to a big temperature change
spanning a strong transition. Moreover, many polymers are semi-crystalline and the
crystalline fraction of the surface is not always known, so the Van der Waals volume is
the simplest way to calculate volume fractions of susceptible groups without requiring
extensive a priori characterization of the polymer.

Van der Waals molar volumes for the entire repeat unit and the molar volumes of
the susceptible groups in the repeat unit were calculated for PC, Polybutadiene, PET, and
Epoxy (Epon 828 + Jeffamine T403). Susceptible groups were chosen on the basis of
absorption of UV in the 185 - 280 nm range as well as possibility of oxidative and free

radical attack based on photodegradation mechanisms from literature. The details of the

185

calculations are provided in Appendix C. Since the Van der Waals volume is calculated
assuming a series of spheres, an approximation of the projected surface area of the

susceptible groups can be given by
2
1
A = (%VJA 112A (6.12)

where V is the total or fractional volume, and A is the total or fractional area of the
molecule or repeat unit.

The results of the Van der Waals fractional volume and fractional area of
susceptible sites on the surface are compared with the experimentally determined values
of the parameter a for the four polymers in Figure 6.2. The results of the approximate
estimation of a by mass normalization are also shown for comparison. It can be readily
observed that the Van der Waals fractional volume and fractional area are excellent
predictors of the value of the parameter a provided the susceptible groups are chosen
properly.

Parameter c describes the susceptibility of the above mentioned surface sites to
photo-oxidation and photolysis. PC has the highest stability to UV radiation because the
Photo-Fries rearrangement products (hydroxybenzophenones) of the polymer are very
good UV stabilizers, resulting in shielding of the rest of the polymer from further UV
degradation [8]. The epoxy resin has aromatic ethers and hydroxyl groups in the structure
which are not very strong absorbers at higher wavelengths and the curing agent has a
high percentage of hydrocarbon content which is also not expected to be very sensitive to
UV degradation, hence the susceptibility of epoxy is considered to be moderate. PET is

reported to be extremely sensitive to photodegradation because of the high absorption

186

coefficient of the aromatic ether group. Studies on PET have shown that even short UV
exposures can cause significant yellowing and cracking in PET [9], hence the PET should
be rated as having very high susceptibility. Virgin rubber (polybutadiene) is also highly
susceptible to ozone and UV attack because of the strong absorption coefficient of the 1:
bond in the polymer backbone and should ideally have a value of c that is comparable
with PET.

It can be seen from Table 6.2 that the value of c agrees well with this qualitative
comparison for all polymers, except rubber, which shows a much reduced susceptibility
to UV oxidation than expected. This apparently anomalous behavior of rubber may be
due to two possible reasons: i) surface restructuring of the rubber due to entropic and
thermal effects, and ii) the effect of UV stabilisers and anti-oxidant additives. The Tg of
polybutadiene is approximately —98°C and at room temperature the increase in free
volume of the polymer allows a high degree of chain mobility. Even though bonds in the
polymer may absorb strongly and get oxidized, the polar groups experience a
thermodynamic driving force to rotate away from the surface and present a fresh low
energy surface, which, from a macroscopic view of the surface, is equivalent to a bond
having low absorptivity. Evidence of this mechanism was provided in Table 5.3 where
the rubber shows an increased rate of overall surface oxidation when temperatures are
lowered by intermittent treatment. The second reason for this behavior is the high loading
of carbon black, which acts as a physical screen for photons, and anti-oxidant additives in
the compound which are designed to retard the oxidation process by quenching and
scavenging free radicals. The low value of c, compared to the expected behavior, can be

explained in terms of an apparent stabilizing effect due to surface reorientation and

187

restructuring, and a real stabilizing effect due to additives. However, the net change in the
work of adhesion of rubber is the highest of all four polymers indicating that even though
the rate of oxidation may be retarded by restructuring and stabilization, the ultimate
extent of oxidation described by parameter a remains unaffected and can be reached at
higher levels of irradiation.

A modelistic determination of c is difficult because several processes govern the
decay of the sensitivity function. The molar absorptivities of various chromophores are
widely tabulated in literature, but the precise mechanisms of oxidation and the
corresponding rate constants are often not known. In addition, the influence of stabilizing
additives in the polymer needs to be studied in detail. Lastly, the effect of free volume
also needs to be accounted for polymers above the Tg. Empirical group contribution
models for temperature dependent molar volume changes may be of use in such cases.
But the biggest stumbling block is the insufficient quantitaive characterization of the
individual oxidative processes which are responsible for the majority of the changes
occurring on the surface.

The three parameters discussed in this section are summarized below in terms of
the physical properties of the polymer which can possibly affect them:
a = f (population of sites, microstructure, chain configuration, steric effects)
c = f (absorptivity, oxidation rates, free volume, chain flexibility, stabilizers & additives)
b = f (initial properties of the surface)

The trends in the three empirical parameters and their correlation to the physical

properties of the polymer are given in Table 6.4.

188

0.65 2.5

 

 

 

 

0 Van der Waals volume fraction 2 3

0-6 - OVan der Waals area fraction ’ ,_ '
:‘é 5'; ASites/Mass I,’ -_ 2.1
g g 0.55 . ,z’
0. 0. ,’ -_ 19
5:3 3 ," ’A
g a 05 H /’ ’/’ U)
-: .g /’ / “'- 1.7 c
8 g /’ A’I/ %
i— i—o
t“; 0.45~ e" ,/ o ~~15 O
E 2 A /’ l’ 8
.2 < / / ,_,
° 9 x’ /’ ~— 13 R
2 73 04 I,’ I,’ g
3 3 I, + '
33 '3 0.35 ~ A ,’ .
'1: a: ,’ __
r: > I 0.9
5“ 0’

3 z
03 ‘ 4- 0.7
0.25 1 1 1 1 0.5
20 40 60 80 100 120
a(mJ/sq.m)

Figure 6.2 Comparison of the Van der Waals fractional volume and fractional area
of susceptible sites in PC, PET, Crosslinked Tire Rubber Compound, and Epoxy
with the parameter a calculated from curve fit to the work of adhesion data. Data

for rapid estimation of a by counting number of sites per mass is also shown for
comparison.

189

Table 6.4 Correlation between empirical parameters and physical properties of
UVO treated PET, Epoxy, PC, and crosslinked tire rubber compound.

 

 

 

 

 

 

 

 

 

 

 

 

 

Dimensionless Maxglzlgffimble Rate of change of
Scallng Factor Work of Adhesion Work of Adhe31on
Polymer . Fractional
Baseline . . .
Wa b volume of a 2 SUSCCptlblllty 2c
2 susceptible (mJ/m ) of sites (cm lmJ)
(mJ/m ) . *
sues
PET 91 3.31 0.30 37.7 High 0.0292
Epoxy 91 3.00 0.31 45.2 Moderate 0.0155
PC 69 1.97 0.36 72.6 Low 0.0107
Rubber 28 1.27 0.45 107.2 High“ 0.0107

 

* from group contributions to Van der Waals volume [7].

# expected for rubber with no UV stabilizers or anti-oxidants.

190

 

6.2.4 Extention to Bulk Polymer

In the previous section a model was presented to describe the oxidation of
polymers by UVO treatment. The model is applicable only to the evolution of surface
properties because the effect of UV irradiation on the bulk polymer is ignored. However,
photolytic reactions can occur in the polymer up to the depth of penetration of the
radiation. The penetration depth of deep UV radiation in many polymers is estimated by
Rabek to be around 300-400 nm [1]. To describe the changes occurring in the near
surface polymer, hereafter referred to as ‘bulk polymer’ for convenience, a general form

of Equation 6.5 can be written using Beer’s law
AP(H, x) = Hoe_ 8‘5" -S(H, x) (6.12)

where H0 is the product of exposure time (t) and surface irradiance (Io), and S(H,x)
describes an overall sensitivity function which varies with depth. It was assumed in
modelling the surface reaction that excess oxygen or ozone is available at the surface. In
the polymer bulk, this does not hold true and the chemical changes in the bulk will be
. severely limited by the concentration of the oxidizing species. The variation of the
sensitivity (S) with depth (x) can be considered to be a function of the availability of
oxygen or ozone at that depth by defining a term R(co) to describe the reaction rate
dependence on the concentration of the oxidizing species. This dependence can be built
into the model as

S(H, x) = S(H)R(c0 )00 (x) (6.13)

where co(x) is the concentration of the oxidizing species at depth x in the polymer.
In this study the ozone concentration was reduced to approximately 10 ppm, the

lower limit of detection of the ozone measurement experimental setup, and no change in

191

the rate of surface modification was observed. However, when no ozone or oxygen is
present, no changes in the surface properties are evident. Thus the reaction rate
dependence on the ozone concentration must be rate determining at some value of ozone
concentration between 0 and 10 ppm. In the future, a detailed study of the rate of surface
modification at very low ozone concentrations can provide information about R(co).

The concentration gradient of the oxidizing species in the polymer, co(x), is more
easily accessible. The concentration profile of ozone can be modeled analytically using
well established gas diffusion models. The diffusion of gases in polymeric solids is well
described by Fick’s second law [10, 11]. The concentration depth profiles for one

dimensional diffusion can be obtained from solutions to

2
992:1)[3 CO] (6.14)

 

at 8x 2

where D is the diffusion coefficient. A more pertinent approach The overall modification
of the depth dependent material property in the bulk polymer can be expressed as

AP(H, x)
t

= 106‘ 30" -S(H)-R(co)-c0(x) (6.15)

The biggest hurdle in this approach is the determination of property changes in
the top layers of the surface with good spatial resolution. Analytical techniques like XPS
can characterize surfaces over average depths of 7-10 nm, monolayers can be

characterized by techniques like SIMS, but depth profiling of bulk polymer properties

remains a challenge.

192

 

6.3 Application to Unknown Polymers

One of the motivating factors for developing a process model for UVO surface
treatment is the ability to predict a priori the properties of UV modified polymers. From
the results described in the previous sections, it can be seen that for a polymer that
absorbs radiation in the spectral output available from pulsed xenon lamps, and
undergoes oxidation, the change in the work of adhesion can be expressed in terms of the
empirical parameters, a and c. This is not limited to the work of adhesion and can be used
for any property that evolves due to UV oxidation, although the values of the empirical
parameters for different properties are likely to vary. A knowledge of these two constants
can fully describe the evolution of the property under investigation as a function of
irradiaton.

The parameter a has been shown to vary with the Van der Waals fractional
volume for four polymers belonging to different classes: elastomers, thermosets,
amorphous thermoplastic, and semi-crystalline thermoplastics. This relationship between
a and the molar volume fraction must be verified over a larger range for more polymers.
It is unlikely for the trend to be linear over large ranges of the fractional volume of
susceptible sites. It is also possible that various types of polymers may follow diverging
patterns. If a verified relationship between the fractional volume of susceptible sites and a
is available, an estimation of a for an untested polymer can be made using group
contribution models. The parameter c refers to the susceptibility of the material to UV
oxidation and hence must be experimentally determined, but it can be reasonably

estimated for homologous polymers if they have similar chromophores and structures.

193

An alternate, empirical approach can also be used to measure a if the c for a
homologous polymer is known and assumed to be equal to that of the untested polymer.
Two measurements of the work of adhesion can be obtained; one for the untreated
surface and one for the surface treated to saturation. The difference between Wasaturawd) -
Wmummwd) is the experimentally determined a. The value of b can be calculated as b =
1+(Wa/a) and the profile for the evolution of the work of adhesion as a function of
irradiance can be generated for all intermediate values. The value of c can be adjusted by
measurements at two or more intermediate points and successive iterations can converge

on the true value of c.

6.4 Summary

The variables in the UV treatment process have been systematically studied. It
was found that radiation in the 185 - 280 nm band is necessary for the application of the
process to a broad range of polymers. Radiation in this band is capable of in-situ
production of ozone. Although ozone is necessary for the process to work efficiently,
within the range of concentrations measurable by our equipment (>10 ppm) , no effect of
the ozone concentration was observed. The variable with the largest effect on the process
is the irradiance and the surface properties were found to have a direct relationship with
the irradiant energy incident on the surface. Temperature of the substrate can affect
properties, especially above the glass transition temperature in polymers, by providing a
driving force for surface reorientation. Temperature was also affected by the pulse
frequency of the lamps. At lower frequencies, the pulses are more intense and lead to

higher surface temperatures than those reached by high frequency, low intensity pulses.

194

This is most likely a limitation posed by the poor thermal conductivity of polymers.
However, lower frequencies also provide more dark cooling time for the surface. The
lamps used in this study could be operated at 3 Hz, 10 Hz, and 120 Hz. Among these
lamps 120 Hz was the most effective frequency for treatment of polymers. UV treatment
with lamps operating at intermediate frequencies may provide interesting results. The
effect of humidity is also an avenue which needs to be explored. Preliminary work shows
that it is possible to treat polymers effectively through thin layers of water. The presence
of hydroxyl and hydroperoxy radicals created in water by UV irradiation can yield very

different products compared to direct attack by ozone and oxygen radicals.

 

 

 

 

 

0.25
=a(b-exp(-cx))
0 2 _. a = 21.1
' b = 1.59
A c = 0.0102
E 0.15 5
U
2‘
E
> 0.1 ~
6“
9
"a
0.05 -
O I 7 I
0 100 200 300 400 500

Irradiation (H) (mJ/sq.cm)

Figure 6.3 Sensitivity function for surface oxygen content in UVO treated PC.

195

 

An empirical process model has been proposed to describe the UVO treatment
process. The model is analogous to the wavelength sensitivity spectrum approach used to
model long term bulk photodegradation of polymers. The model describes the evolution
of surface properties after UV irradiation in ozone for pulsed xenon lamps. Sensitivity
functions for work of adhesion with water were determined for five polymers: PC, PET,
Epoxy, Rubber, and PT FE. The functions were fit with a three parameter exponential
association model. The parameters correlate with physical properties of the polymer: the
parameter 0 provides information about the density of reactive sites on the polymer, [7 is a
parameter that describes the initial properties of the polymer, and c provides information
about the general reactivity of the polymer sites to photo-oxidation and other
photoinduced reactions. Sensitivity functions can be determined for other surface
properties as well. Figure 6.3 shows the oxygen content sensitivity function for
polycarbonate. It is observed that the c value for the OIC ratio of UVO treated
polycarbonate (c = 0.0102 O%lmJ) has very good agreement with that for the Work of
Adhesion (c = 0.0107 cmz/mJ) although the two properties are unique and measured over
different sampling depths. Sensitivity functions for other polymers can be developed in
the future to create a database of parameters which can be applied to predict the
properties of new UVO modified polymers, blends, copolymers, and composites.

To use this model in a predictive capacity for previously untested polymers, at
this stage, prediction of extrapolated parametric values may not be reasonable except
over a very narrow range for specific homologous polymers. The initial verification for
untested polymers can be performed by measurements at four to five points ranging from

H=0 to H = 00 (saturation of the surface property) and the value of a can be determined

196

from this data. The value of c can be determined by iteratively fitting the profile at
intermediate measurements. With the generation of more data for a wider variety of
polymers and homologues, predictive accessment of the properties can be accomplished

with increasing reliability.

197

 

10.

ll.

12.

References

Rabek, J .F. Mechanisms of Photophysical Processes and Photochemical Reactions in
Polymers, John Wiley, New York, 1987.

Hollander, A.; Klemberg-Sapieha, J .; Wertheimer, M.; J. Polymer Science Part A:
Polymer Chemistry, Vol. 23, pp. 2013-2025, 1995.

Atkins, P.W.; Friedman, R.S. Molecular Quantum Mechanics, Third Ed., Oxford
University Press, Oxford, 1997.

Gonzalez, M.; Martire, D.O. Water Sci. Tech. Vol. 35, No.4, pp. 49-55, 1997.

Hashem, T.M.; Zirlewagen, M.; Braun, A.M. Wat. Sci. Tech., Vol. 35, No. 4, pp. 41-
48, 1997.

Allan D.S.; Maecker, N.L.; Priddy, D.B.; Schrock, N .J . Macromolecules, Vol. 27, pp.
7621-7629, 1994.

VanKrevelen, D.W. Properties of Polymers, 3rd ed., Elsevier, Amsterdam, 1990.
Pickett, J .E. Permanence of UV Absorbers in Plastics and Coatings, Handbook of
Polymer Degradation, Second Edition, S. H. Hamid ed., Marcel Dekker, New York,
1998.

Praschak, D.; Bahners, T.; Schollmeyer, E. Appl. Phys. A, Vol. 66, pp. 69-75, 1998.

Bird, R.B.; Stewart, W.E.; Lightfoot, E.N. Transport Phenomena, Wiley, New York,
1960.

Poulsen, L.; Zebger, 1.; Klinger, M.; Eldrup, M.; Sommer-Larsen, P.; Ogilby P.R.
Macromolecules, vol. 36, pp. 7189-7198, 2003.

Naito, T.; Horie K.; Mita, I. Macromolecules, vol. 24, pp. 2907-2911, 1991.

198

 

 

CHAPTER SEVEN

MARKET AND ECONOMIC ANALYSIS

Adapted from a report submitted by Alekh Bhurke] and Matthew Jensonz towards
completion of an independent project in Management of Technology & Innovation
(MT&I Program, ML890) administered by the Case Center for Computer-Aided

Engineering and Manufacturing at Michigan State University, December I 997.

7.1 Abstract
With the automotive industry using more polymers in the manufacture of
components, there is a need for proven surface treatment technologies that will provide
enhanced performance at reasonable costs. Growing concerns over the environmental
impact of existing processes and the advent of ISO 14000 standards have also contributed
to the need for development of environmentally friendly processes. The manufacturing
industry is looking to incorporate new surface treatment technologies for several reasons:
0 Commercial applications involving painting and adhesive bonding of low energy
polymers are increasing.
0 Existing surface pretreatment techniques are labor intensive, time intensive,
hazardous, or present environmental problems.
0 There is a push to develop and transition to environmentally benign manufacturing

processes due to the increasing environmental awareness in society.

 

' Department of Chemical Engineering & Materials Science, Michigan State University, Ml 48824, USA
2 School of Supply Chain & Management, Michigan State University, MI 48824, USA

199

Based on our research, the UV surface treatment developed by Michigan State
University shows great promise — provided an actual, pilot scale process is developed to
mimic real manufacturing environments. Not only is there a great need in the automotive
industry for this process, but there are also markets in the biomedical industry which
require modern surface treatment technologies. A basic cost analysis indicates the UV
surface treatment is a viable alternative to existing processes and information obtained
from communications with leaders in the automotive industry (General Motors and Ford
Motor Co.) shows a willingness to consider this process if certain conditions and

qualifications are met.

7.2 Introduction

Polymers usually have low surface energies, making the material difficult to wet
and presents problems with adhesion of paints and other materials. Various surface
treatments are available in industry to improve the wettability and adhesion of such
materials. The more common surface treatments used for this purpose are flame
treatment, plasma treatment, corona discharge, acid etching, solvent washes,
halogenation, etc. The disadvantages of these treatments are that they are typically
expensive, labor intensive and present several environmental and safety concerns. The
release of volatile organic compounds (VOCs) into the atmosphere via these processes is
also a growing concern in today’s world. An alternative process for surface treatment
using ultraviolet (UV) light has thus been studied. This process is economical, easy to
incorporate into a manufacturing environment, presents minimal safety concerns, and is

environmentally benign.

200

Ultraviolet light in the 180 - 280 nm wavelength (UVC region) is known to
interact with atmospheric oxygen to produce ozone and nascent oxygen, both of which
are strong oxidizing agents. Exposure of a material to UVC radiation for short times in
the presence of oxygen or ozone leads to oxidation of the surface. Surfaces thus oxidized
have higher surface energies leading to lower contact angles and higher wettability. The
by-products of such a process are usually small molecules like water and oxides of
carbon.

With growing concern over environmental impact and the advent of ISO 14000
standards, there is a need for development of an environmentally friendly process. In the
UV treatment process, no volatile or particulate emissions are produced and no wet
chemicals are used, making the process viable for use in future manufacturing
environments. A brief comparison of the various surface treatment processes used in

industry with the UV treatment process is made in the following section.

7 .3 Surface Treatment Processes
7.3.1 Chemical/Solvent Wash

In many current painting and bonding processes, a chemical wash is used in
which the surface to be painted or bonded is cleaned with solvents, coated with primers
and adhesion promoters, and finally coated with multiple layers of paints. This method is
neither economical nor environmentally sound because it uses primers which are
typically organic solvent based, and the solvents used are known to cause volatile organic
compound (VOC) emissions. Water based paints and primers are typically difficult to use

in polymer applications because of the difficulties involved in spreading them on low

201

energy polymer surfaces. This process is suited to continuous flow manufacturing and
generally inexpensive. Within the auto industry, chemical, detergent and solvent washes
are the norm. Solvent washes are primarily used to clean surfaces not to modify the
surface chemistry. With the danger of VOC emissions, the chemical wash process must

take place in a controlled environment.

7.3.2 Flame Treatment

Flame treatment is performed by passing a hydrogen flame (temperatures higher
than 3000°F) rapidly over a surface. The adhesive characteristics of the surface are
enhanced by a combination of two factors: an increase in surface energy by oxidation and
the degradation of weak boundary layers or contaminants. The flame creates oxygen
radicals in the air which attack the polymer and form polar groups on the surface. The
substantial oxidation and increase in polarity enhances the attraction between the
substrate and the paint, adhesive, or ink being applied. The hydrogen flame head is
typically placed 0.25 inch from the product’s surface which allows throughput speeds of
up to twenty inches per second. The treated surface may remain stable for several weeks.
Flame treatment is not particularly suitable for three dimensional objects and the use of
combustion of fuel as the primary source of energy has obvious environmental

CODSCQUCDCCS.

7 .3.3 Plasma Treatment

Plasma treatment exposes a material to gaseous oxygen, nitrogen, argon, or any

other gas that is energized or excited to a plasma state in a vacuum chamber. The plasma

202

can be high temperature (energized by electrical discharge) or cold (energized by radio
frequency). The components of the gas mixture can be chosen to seed certain functional
groups on the surface, and during the 0.5 to 10 minute exposure to the gas plasma, the
surface undergoes chemical, morphological and topographical changes that can improve
adhesion. As with the flame treatment, the surface is also cleaned in the process and
organic compounds or contaminants are removed from the surface. The plasma treatment
can be time limited or permanent depending on the material. An advantage over flame
treatment is that the entire surface of three-dimensional objects can be treated, but the

requirements of a controlled atmosphere and vacuum are disadvantages.

7.3.4 Corona Treatment

Corona discharge, also known as ‘non-vacuum plasma’, is commonly used to
pretreat polymers in web type applications. Treatment occurs when high voltage (10,000
volts) is applied to an electrode positioned a short distance from the substrate. The air
gap is ionized and this corona of ionized air oxidizes the surface, resulting in a polymer
surface which is better suited for printing or laminating and has better adhesive
properties. Corona treatments are less likely to mar or damage treated surfaces than other
treatment processes. The corona treatment is environmentally friendly because the use of
volatile solvents and potentially hazardous process gases is eliminated. Corona is well
suited to continuous manufacturing environments but not suited to treat complex surface
geometries as well as other methods. Typical cycle-times for treating three dimensional
objects like car bumpers with corona is 30 seconds. A concern with corona treatment is

the build up of static charges that could cause explosions in a paint booth if the static

203

electricity is not removed. The high-voltage of the corona discharge is another safety
concern. The only environmental concern with corona treatments is the production of

large amounts of ozone.

7.3.5 Ultraviolet Treatment

UV treatment consists of exposing the substrate to UV radiation. The UV energy
converts oxygen in the air to ozone and nascent oxygen, and these reactive species have
the ability to oxidize organic material on the surface. The ozone also activates the
surface material — making it more reactive with the ozone. Materials are generally placed
within a few centimeters from the UV source. UV treatment combines the benefits of
low temperature with the convenience of operating at ambient pressures and
temperatures. Basic precautions must be taken when using UV surface treatments to not
expose the skin or eyes to the UV light. The ozone produced in UV treatment is typically
several orders of magnitude lower than that produced in corona treatment and the only
environmental issue is the removal of this low level of ozone. In summary, the UV
process has the following favorable characteristics:
0 Environmentally benign (no harmful by products)
0 Inexpensive (low capital and operating costs)
0 Fast (treatment times on order of 1-2 minutes)
0 Non-contact process
0 Ability to treat complex geometries

o Easily integrated into the manufacturing environment

204

The process under investigation and analysis is the modification of an existing
process. Surfaces can be cleaned and oxidized by exposing them to UV radiation of
appropriate wavelength in ambient atmospheres. The resulting oxidized surface has
beneficial chemistry for adhesive bonding, painting, printing, or metallization. The UV
process has the potential to be less expensive than chemical and solvent washes
modification, it is able to treat three dimensional parts, large surface areas, and it is
environmentally friendly. The initial research study has identified UV oxidation and
surface activation as a viable surface treatment process that can be used in a
manufacturing environment. The goal of the technical research at Michigan State
University is the development of a model to reduce UV treatment times and increase the
reliability of the process, making it an attractive process for use in manufacturing

environments.

7.3.6 Surface Treatment Process Comparison Matrix3

Using information obtained from literature searches and position papers, a matrix
(Table 7.1) was constructed to compare the available surface treatment techniques. The
ratings are subjective opinions of the authors and based on a scale of 1 (best) to 5 (worst).
With the UV surface treatment under consideration, treatment times and cost are both
rated “2” but with further development, it is believed these two areas will improve.4

Corona and UV seem to be the two most viable surface treatment technologies for
the future. As an improvement over baseline treatments, plasma has gained popularity

and market share in European markets but not in the United States because of the

 

3 All opinions presented reflect the state of the industry in 1997.
4 State of the UV treatment process as evaluated by the authors in 1997.

205

unsuitability to continuous manufacturing. The United States markets are now looking at
the other surface treatment technologies — bypassing the original alternative surface
treatments such as plasma. Solvent washes, even though popular today, are subject to
future competition from environmentally friendly techniques. With further research and
experience, UV treatment times and costs will be reduced — making UV one of the more

attractive surface treatments for the future.

Table 7 .1 Surface treatment process comparison matrix.*

 

 

 

 

 

 

o g E 3 a =1
Technology a *9 E g g 5 z: 3 E ... 5% 'g _. 00
1.. 2) o 0 O E u: z: E o o .D 8 '3 on c:
O :4 O :3 E: H q) ”—1 H H .2: CU N 'U H .S
‘1: -— c: > “’ e D “‘ m > a. r: v .. "g
a) .D Q o x: 8 ..... S 8 o c: E :1 8 > ca
9.. < m U m E— t— m [- U m H m :1. O a: a:
Chemical
Wash 3 1 4 2 l 1 5 1 18 4
Flame 2 5 2 1 1 2 2 1 l6 3
Corona l 2 l 1 1 2 1 1 10 1
Plasma 1 1 5 3 1 3 2 4 20 5
UV 1 1 2 2" 1 2” 1 1 11 2

 

 

 

 

 

 

 

 

 

 

 

 

 

*Subjective scale from 1—5 based on 1997 data. Lower is better.
# Properties targeted for improvement.

206

 

7 .4 Surface Treatment Cost Comparisons

The qualitative cost comparison made in the previous section was based on
information received from surface modification equipment manufacturers. Since
comparisons are made between a wide variety of existing surface treatments, all on a
different scale, it is important to find a good basis for detailed cost-benefit analysis. To
avoid any inconsistencies in the analysis, a simple approach was adopted. The costs
associated with each method of treatment were normalized on the basis of costs incurred
to treat a square inch of a surface. Only the cost of operating the equipment for a unit of
time was considered and the amount of surface area treated in this unit time was
calculated. This serves the advantage of providing a base value for cost comparison.
These values, calculated for various processes, can then be used to estimate the
approximate costs that would be incurred given a certain performance requirement or
application for a given treatment method. The downstream and labor costs were ignored
on the assumption that the UV surface treatment technique will parallel other existing
technologies in terms of these associated parameters. Another reason to adopt this
approach is the fact that these associated costs differ widely depending on the type of
process used. For example, the costs associated with using robotic arms or any automated
process would be very different from a manual batch operation and thus cannot be taken
into account when comparing the cost of the technology itself. It was observed from
some examples obtained during our communications with equipment manufacturers and
users, that the associated costs could be several orders of magnitude higher than the cost
of the treatment. The essential information would thus be lost if associated costs were

included.

207

The cost comparison was done for a hypothetical production-line application of
treating car bumpers at the rate of 200 bumpers per hour. The area of the bumpers
needing surface treatment was assumed to be 5 feet x 1 feet. Five surface treatment
systems: plasma, corona, UV lab scale, UV pilot scale, and UV production scale were
evaluated. The cost data for plasma, corona and UV laboratory scale equipment was
obtained through communication with existing commercial product vendors. The cost
data for UV pilot scale and UV production scale equipment was taken from the best
engineering estimates obtained from manufacturers and users of the technologies
involved.

The results of the cost analysis for the five surface treatment processes are given
in Tables 7.2 to 7.6. The capital and operating costs were used to define a normalized
cost for treatment of a unit area of the surface. These base costs were extended to a
production line treating car bumpers at the rate of 200 bumpers per hour with 5 sq. ft.
area per bumper. The normalized annual costs for each of the processes for this
application are given in Table 7.7.

From the above analysis it can be seen that, as expected, the UV treatment
process, at the laboratory scale, cannot be very competitive with other production scale
processes. However, the forecasts for the UV-Pilot and UV-Production scale cost
estimates, based on information gathered from UV equipment suppliers, show that the
normalized operating and capital costs decrease sharply at larger scales The low capital
and working costs of a production scale UV treatment process makes it a particularly
attractive choice and is cost-competitive with the established treatment processes such as

corona and plasma treatment.

208

1c. .

 

 

Table 7.2 Cost analysis of plasma treatment process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Capital Costs

Equipment Life (hours) Total Cost Cost/hour
Plasma Unit Cost 87600 $800,000 $9.13
Total Capital Investment $800,000 $9.13
Operating Costs

Description Hours Cost Cost/hour
Maintainance (5% Annually) 1 $4.57 $4.57
Total Operating Costs $4.57 $4.57
Total Costs

Total cost per hour of operation $13.70
Surface area treated per hour (sq. inch) 144,000
Cost per 10 square feet of treated area $0.14
Annual Cost $120,000
Area Treated Annually (sq. inch) 876,000

 

 

 

209

 

Table 7.3 Cost analysis of corona treatment process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Capital Costs

Equipment Life (hours) Total Cost Cost/hour
Corona Unit Cost 35040 $150,000 $4.28
Total Capital Investment $150,000 $4.28
Operating Costs

Description Hours Cost Cost/hour
Electricity (10 kWh) 1 $1.00 $1 .00
Maintainance (10% Annually) 1 $1.71 $1.71
Total Operating Costs $1.00 $2.71
Total Costs

Total cost per hour of operation $6.99
Surface Area Treated per hour. (sq. inch) 86,400
Cost per 10 square feet of Treated Area $0.12
Annual Cost $61,260
Area Treated Annually (sq. inch) 525,600

 

 

 

210

 

Table 7 .4 Cost analysis of UV laboratory scale treatment process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Capital Costs

Equipment Life (hours) Total Cost Cost/hour
UV Lamp 1000 $500 $0.50
Power Supply 35040 $9,500 $0.27
Ozone Generator 35040 $500 $0.01
Air Supply Unit 35040 $600 $0.02
Total Capital Investment $11,100 $0.80
Operating Costs

Description Hours Cost Cost/hour
Electricity (1 kWh) 1 $0.10 $0.10
Total Operating Costs $0. 10 $0. 10
Total Costs

Total cost per hour of operation $0.90
Surface Area Treated per hour. (sq. inch) 300
Cost per 10 square feet of Treated Area $4.33
Annual Cost $7,906
Area Treated Annually (sq. inch) 1,825

 

Lamp area = 5 in. x 2 in., treatment time = 120 seconds.

211

 

 

 

 

Table 7.5 Cost analysis of UV pilot scale treatment process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Capital Costs

Equipment Life (hours) Total Cost Cost/hour
UV Lamp 1000 $500 $0.50
Power Supply 35040 $26,500 $0.76
Ozone Generator 35040 $500 $0.01
Air Supply Unit 35040 $600 $0.02
Total Capital Investment $28,100 $1.29
Operating Costs

Description Hours Cost Cost/hour
Electricity (2 kWh) 1 $0.20 $0.20
Total Operating Costs $0.20 $0.20
Total Costs

Total cost per hour of operation $1.49
Surface Area Treated per hour. (sq. inch) 1,200
Cost per 10 square feet of Treated Area $1.79
Annual Cost $13,032
Area Treated Annually (sq. inch) 7,300

 

Lamp area = 20 in. x 2 in., treatment time = 120 seconds.

212

 

 

Table 7.6 Cost analysis of UV production scale treatment process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Capital Costs

Equipment Life (hours) Total Cost Cost/hour
UV Lamp 1000 $400 $0.40
Power Supply 35040 $50,000 $1.43
Ozone Generator 35040 $500 $0.01
Air Supply Unit 35040 $600 $0.02
Total Capital Investment $51,500 $1.86
Operating Costs

Description Hours Cost Cost/hour
Electricity (10 kWh) 1 $1.00 $1.00
Total Operating Costs $1.00 $1.00
Total Costs

Total cost per hour of operation $2.86
Surface Area Treated per hour. (sq. inch) 43,200
Cost per 10 square feet of Treated Area $0.10
Annual Cost $25,039
Area Treated Annually (sq. inch) 262,800

 

 

Lamp area = 60 in. x 12 in., treatment time = 60 seconds.

213

 

 

Table 7 .7 Summary of cost analysis based on a production facility
treating 200 bumpers (5 ft. x 1 ft.) per hour.

 

 

 

 

 

 

Area treated Cost of treatment Annual cost for treating
Treatment per hour ( er 10 3 ft) 200 bum ers/hour
(sq. in/hour) p q. ' p
Plasma 144,000 $ 0.14 $ 120,000
Corona 86,400 $ 0.12 $ 102,100
UV (Lab Scale) 300 $ 4.33 $ 3,794,880
UV (Pilot Scale) 1,200 $ 1.79 $ 1,563,840
UV (Production) 43,200 $ 0.10 $ 83,463

 

 

 

 

 

 

7 .5 Potential Markets

Based on our market surveys, in addition to the known market in the automobile
industry, the biomedical industry represents the most developed opportunity for UV
surface treatment. To be successful in the biomedical industry a process or technology
must be efficient and reliable. In this industry reliability is a matter of life and death;
customers depend on a product that performs flawlessly every time. The technology must
deliver uniform results each time. The demands on product performance are paramount in
the biomedical industry. With increased pressure on margins, a cost-effective solution is
needed to allow the products to perform at the level required. In general the biomedical
industry is more concerned about environmental consequences than other industries. This
industry requires technologies and processes with the following characteristics: no use of

CFC’s or solvents, clean operation, and low energy process. With several different types

214

 

of products and seemingly endless applications, this industry requires an innovative

process that can treat many different types of materials. The industry requires perfect

quality and an innovative approach —- all characteristics of a UV surface treatment.
Several companies have developed surface treatment applications for the medical

device industry. All types of surface treatments are used: corona, gas plasma, flame, and

UV for curing applications. Our research showed over 40 surface treatment companies

are competing in the biomedical industry. Devices treated include catheters, angioplasty

balloons, intraocular lenses, sensors, pacemakers, filter membranes, assay trays, filter

housings, and syringe hubs. Surface treatments utilized in the biomedical industry are

typically used to:

0 increase adhesive bond strengths

0 improve lubricity and biocompatibility

0 increase plastic to metal adhesion

0 ultra-cleaning and surface modification for better wetting characteristics

0 activate surfaces of sensors

0 enhance other surface characteristics such as microbial resistance and cell adhesion.

The biomedical industry is known as an industry that embraces innovation

provided the innovation delivers some sort of perceived benefit; often measured in terms

of the number of new biomedical products and growth. There is a high level of product

awareness and customer acceptance. While there are several firms competing in this

industry, innovative products and processes providing an improvement over current

technology have the best chances of success.

215

 

7.5.1 Industry Comparison Matrix
Based on the information obtained through surveys, an industry comparison
matrix was constructed. Table 7.8 shows the comparison between the automobile and

biomedical industry.

Table 7 .8 Comparison between the automobile and biomedical industry.

 

 

 

 

 

 

 

 

 

 

 

Automobile Biomedical
Current Process Chemical Plasma, Corona, UV
Awareness Low High
Process Flow Continuous Batch & Continuous
Investment in New Technology Low High
Relationship Power Manufacturer Shared
Life Cycle 1 Long Short
Environmental Concerns Moderate Important
Primary Concern Cost Performance
Growth Potential Low High
Industry Traditional Innovative

 

 

 

 

 

216

 

7.5.2 Marketing Considerations

Before the UV surface treatment technology can be marketed or licensed, several
issues must be considered. At this point, the process is only a lab scale process and no
real product exists. For this process to create any sort of return for Michigan State
University, a product must first be developed. According to the information received
through the surveys of engineers and managers at Ford Motor Company and General
Motors, the process must meet certain threshold requirements before it can be considered
a viable product. The following information must be obtained before this process is

considered a ‘real’ alternative:

0 Demonstrate consistent performance on varying substrates.

0 The process must be reliable, durable and duplicable.

- A reasonable estimate of costs in a production environment must be developed.
0 The process must work in a continuous flow environment.

0 Some sort of validation and quality control must exist to verify results.

0 The affect on throughput must be minor.

- Suppliers must be motivated to use such processes and technology.

Both automotive manufacturers contacted indicated that the supplier base is an
integral part of any solution. In most situations the components arrive at a manufacturing
facility already processed or treated. The supply base would need to adopt and develop
the UV treatment process to their current manufacturing environments. Granting

companies like Ford and General Motors exclusive rights, will not ensure the adoption of

217

 

this process. The intellectual property agreements must be drafted to allow MSU and the
supply base to develop an actual working product that incorporates the UV surface
treatment process. With such arrangements, MSU could partner with a supplier or an
alliance of suppliers to develop this new technology into an actual product that could be

marketed to automobile manufacturers.

7 .6 Recommendations and Implementation Plan

The UV surface treatment developed by Professor Lawrence Drzal at Michigan
State University and preliminary laboratory results indicate that the process will generate
interest in both the automotive industry and the biomedical industry. Based on the
information received from industry, the process would not yet be considered a viable
technology. In its current state, the process has little value to the interested parties. The
process must be proven on a larger scale and an actual product must be developed. In an

effort to push this technology into the marketplace, the following steps are recommended:

0 Patent the process ($40,000).

0 Develop pilot facility focused on refining the process (outside funding).
0 Receive waiver of exclusive rights.

0 Market process at industry trade fairs ($5000).

0 Seek out a partner to develop actual product — perhaps an alliance of suppliers as
defined by GM and Ford.

0 Reach out to the biomedical industry.

218

 

Since the process has the potential to be less costly than current processes,
delivers improved performance, and is environmentally friendly process, the UV surface
treatment process has value in the marketplace if it can be developed into a product. With
the increasing use of polymers, the process has the potential to be used on every
automobile. Furthermore, the process could be used by the biomedical industry,
representing another sizable market. With over 6,000,000 passenger vehicles produced in
the United States last year and potentially five areas of the vehicle subject to a UV
surface treatment, there was the potential of 30,000,000 UV surface treatments in 1996 -
not including biomedical applications. The real value would be determined by the
licensing arrangement and the fee structure agreed between the parties. For the process to
be valuable, the automotive suppliers would be required to pay an up-front or per use
licensing fee.

With the size of the markets under consideration, the UV surface treatment
process does have potential as a revenue generating technology for Michigan State
University, but the exact value is difficult to determine. At this point, an actual product

needs to be developed which can either be sold or licensed.

7.7 Summary

Research is under way at Michigan State University to determine if a few seconds of
UV light treatment in air can clean and oxidize polymer, plastic, and polymer composite
surfaces with no environmentally hazardous by products. It has been shown that there are
beneficial chemical and physical effects on adhesion of paint and adhesives to polymer

and plastic surfaces resulting from short, high intensity exposure to ultraviolet (UV) light

219

 

in air. This study has shown that UV oxidation and surface activation is a viable surface

treatment process which can be used in a manufacturing environment. At this juncture in

the research project, it has been demonstrated that the UV treatment process is:

A new, alternative, environmentally benign, surface pretreatment process for treating
the majority of polymer, plastic and polymer composite surfaces using ultraviolet
light.

The non-contact process is adaptable to treat flat or convoluted external surfaces and
has the potential to require exposures on the order of a few seconds to produce
significant increases in wettability and adhesion of paints and adhesives to treated
surfaces.

The UV process is environmentally benign since it does not create VOCs or
suspended airborne particulates.

A process comparison and cost analysis of the UV process indicates that this new
technology is strongly competitive with existing processes (flame, corona and
plasma). It is at least equal to current technologies in terms of cost and effectiveness
of cleaning and pretreating surface for painting and adhesive bonding.

The UV process is a low cost (ea. $0.01 per square foot) and high speed process.
MSU has been granted two US patents based on the results from this research and

several other applications are pending.

220

 

References

l. Bhurke, A.S.; Jensen, M.; UV Treatment - A Low Cost, Environmentally Friendly
Surface Treatment Technique, ML890 Report, Case Center for Computer-Aided
Engineering and Manufacturing, Michigan State University, USA. (1997)

221

 

CHAPTER EIGHT

CONCLUSIONS AND RECOMMENDATIONS

8.1 Conclusions

A wide range of polymers including thermosets, elastomers, amorphous, semi-
crystalline, and crystalline thermoplastics were successfully modified by UV treatment.
One of the many advantages of UV surface treatment is the ability to use it to modify
surfaces under ambient conditions. UV irradiation consisting of wavelengths in the 185
nm — 280 nm (UVC) band can oxidize polymers in air by producing ozone. The
combined effect of UV activation of the surface and oxidative attack by ozone creates
high energy polar functional groups on polymer surfaces, increasing the surface energy
and thereby providing the thermodynamic driving force required for good wettability and
adhesion.

The investigation into the mechanisms of surface treatment of polymers by UV
radiation was characterized by wettability, surface energy, and surface chemistry
measurements. UV exposure in ozone and air was found to impart a strongly hydrophilic
nature to most polymer surfaces that have absorption in the range of wavelengths
provided by xenon arc lamps. With the exception of high purity saturated polyolefins and
fluorocarbon polymers which do not have absorption in the UVC band, this process can
be used for oxidative treatment of a wide variety of materials. Even in the case of
polyolefins which do not have significant absorptivity above 185 nm, the presence of
internal and process impurities can lead to the formation of radicals which can be

oxidized or crosslinked, thereby improving the adhesive properties of the polymers.

222

 

 

polyrr
indep
proce
polyr
and t
be va
founc
advar
[ream
T
surfac
expos
pheny
Photoc
PC an

P0111111

 

The use of supplemental ozone to maintain high local ozone concentrations on the
polymer surface during treatment was helpful in ensuring a steady state process
independent of mass transfer limitations. Under these conditions, the most important
process parameter is the UV irradiance (power). The changes in surface energy of the
polymer are strongly dependent on the total irradiation (energy) incident on the surface
and the process is virtually independent of the exposure time provided the irradiance can
be varied. Short exposures at high irradiances and long exposures at low irradiances were
found to yield identical surface properties. This gives UV treatment a significant
advantage over competitive processes in manufacturing environments where short
treatment times are necessary.

The molecular properties of UV treated bisphenol-A based polycarbonate (PC)
surfaces studied with ATR-FI‘IR, and UV spectroscopic techniques show that UV
exposure leads to the photo-Fries rearrangement of the bisphenol-A carbonate to form
phenyl salicylate and di-hydroxy benzophenone. This is consistent with mechanisms of
photodegradation of PC reported in literature. The photo-Fries rearrangement products of
PC are strong UV stabilizers and in the normal course of weathering, stabilize the
polymer from further degradation. However, in the case of UVO treatment, the
intermediate and final products of UV degradation on the surface can undergo oxidation
by ozone to form hydroxyl, carbonyl, and carboxylate functional groups. The net result of
this oxidation is the formation of conjugate bases on the surface which is seen as a strong
increase in the basic component of the surface energy. This increase in surface polarity is

responsible for the enhanced wettability of UV modified surfaces.

223

 

EPR spectroscopy of UV exposed PC showed a strong paramagnetic response from
the material after UVA and UVO treatments. This is believed to arise from trapped free
radicals, or radical ions in the polymer. The majority of the paramagnetic sites were
found to be near the surface of the polymer and they can be effectively quenched or
undergo reactions with an amine cured epoxy adhesive.

The mechanical properties of the UVO modified PC probed by nanoindentation
tests showed an increase in the modulus in the top 500 nm, which can be attributed to
crosslinking of the irradiated polymer by processes similar to those seen in thermally
oxidized and aged PC. Despite the increase in surface modulus, adhesion tests showed no
change in the adhesive bond strength with epoxy and polyurethane adhesives. It was
determined from fracture surface analysis that the PC-epoxy interphase failure occurs
within the PC substrate and material equivalent to approximately 30-60 A of PC is
removed by the epoxy adhesive. The cohesive failure in the PC may be indicative of a
weak boundary layer or a very high level of interfacial adhesion between the epoxy and
PC which forces the locus of failure into the relatively weaker PC.

The surface chemistry, probed by XPS, shows that the correlation of measured O/C
ratios with the surface energy of the polymer is very good and both properties vary
predictably with the total irradiation received by the sample and have no dependence on
the irradiance or exposure time. The functional groups created on the surface during
UVO treatment show that the oxidation process occurs by reactions of ozone, and
possibly the by-products of ozone dissociation, with the carbonate groups as well as

hydrocarbon bonds in the bisphenol-A structure.

224

 

When in-situ oxidation by ozone is prevented by irradiating samples in vacuum or
inert atmospheres, active sites are created on the PC surface, but these active sites are
unreactive in air and the polymer do not exhibit any uptake of oxygen even after many
days of aging in ambient conditions following irradiation. However, the surfaces retain
their activity and readily undergo oxidation when exposed to ozone. The increase in
oxygen content of these samples is beyond that expected from unirradiated, passive
polymer samples exposed to ozone alone.

The rate of oxidation of activated surfaces, determined by measuring the increase in
oxygen content as a function of ozone exposure, showed two distinct processes: 1) a fast,
initial incorporation of oxygen into the surface that is independent of the amount of
irradiation received by the surface, and ii) a slower region of growth which shows
dependence on the level of irradiation provided to the sample during treatment in a non-
oxidative environment. The initial rapid uptake is attributed to the reaction of ozone with
active sites in the top few monolayers of the material. Once the surface is depleted of
active sites, further oxidation involves the slow diffusion of ozone into the sub-surface
layers to oxidize active sites in the polymer. This was demonstrated by derivatization
experiments with silver nitrate which allowed differentiation between the contributions to
the total oxygen signal from the surface and sub-surface layers of the polymer.

Formation of water soluble low molecular weight oxidized material on the surface
after UV treatment was also observed. Differences in the amount of LMW material
formed on in-situ and ex-situ oxidized samples provided indication of the different
behavior of the polymer when irradiated in inert and oxidizing environments. Irradiation

in non-oxidizing environments increases the possibility of radical recombination, or

225

 

rearrangement, to form a smaller number of stable surface active sites. Irradiation in
oxidizing environments enhances the total extent of oxidation because of aggressive
reactions between the oxidizer and radicals created on the surface, but the higher extent
of reaction can lead to extensive chain scission and formation of LMW fragments.

For most materials, sharp increases in the surface oxygen, wettability, and adhesive
strength are observed following UVO treatment. The properties of epoxy, polyethylene
terephthalate (PET), and polydiene rubber were studied in detail and trends similar to
polycarbonate were observed. The wettability was found to depend entirely on the
amount of irradiation given to the samples. In the case of rubber, a strong effect of
temperature on the surface properties was also observed. The rise in surface temperature
caused by exposure to UV and IR radiation from the lamp is believed to cause
reorientation of the oxidized polar groups. The regeneration of low energy surface was
also observed in thermoplastics like polyethylene and Bexloy®W polymers when treated
above their glass transition temperatures (Tg). This is believed to be an effect of the
higher free volume and increased chain mobility in polymers above the Tg. In polymers
like PP and TPO, the wettability of the surface does not change after UVO treatment
because the inherent stability of the hydrocarbon bonds to radiation above 185 nm
prevents extensive chain scission and oxidation. However, impurities such as carbonyl
groups, main chain unsaturation, and metal atoms present in commercial polyolefins
sensitize the polymer by creating free radicals which can attack other polymer chains and
ultimately cause cross-linking in the polymer. The adhesive bond strength of polyolefinic
materials (PP, TPO) with epoxy adhesives increases by 400% to 500% compared to the

untreated polymers because of photo-induced crosslinking which increases the modulus

226

 

 

of the polymer. Almost all materials treated with UVO show beneficial changes in
properties except fluorocarbon based polymers such as Teflon®.

The variables in the UV treatment process were systematically studied. It was found
that radiation in the 185 nm to 280 nm band is necessary for the application of the
process to a broad range of polymers. Although ozone is necessary for the process to
work efficiently, within the range of concentrations measurable by our equipment (>10
ppm), no effect of the ozone concentration was observed. For all polymers investigated,
the surface properties were found to have a direct relationship with the irradiant energy
incident on the surface. Temperature of the substrate can affect properties, especially
above the glass transition temperature in polymers, by providing a driving force for
surface reorientation. Surface temperature of the material can be affected by the pulse
frequency of the lamps for a given power output. At lower frequencies, the high intensity
pulses lead to higher surface temperatures than those reached by high frequency, low
intensity pulses. However, lower frequencies also provide more dark cooling time for the
surface. The lamps used in this study could be operated at 3 Hz, 10 Hz, and 120 Hz.
Among these lamps 120 Hz was the most effective frequency for treatment of polymers.
UV treatment with lamps operating at intermediate frequencies may provide interesting
results.

A process model has been proposed to describe the UVO treatment process. The
model describes the evolution of surface properties after UV irradiation in ozone for
pulsed xenon lamps. A sensitivity function approach was employed to describe the
evolution of any property of the surface as a function of the irradiation. Sensitivity

functions for work of adhesion with water were determined for PC, PET, epoxy, and

227

 

polydiene rubber. The functions were fit with a three parameter exponential association
model. The parameters correlate with physical properties of the polymer: the parameter a
provides information about the density of reactive sites on the polymer, b is a
dimensionless scaling parameter that describes the initial properties of the polymer, and c
provides information about the general susceptibility of the polymer sites to photo-
oxidation and other photoinduced reactions. A good correlation between a for the four
tested polymers and the Van der Waals volume fraction of reactive sites on the surface
was obtained. Sensitivity functions can be determined for other surface properties as well.
Sensitivity functions for other polymers can be developed in the future to create a
database of parameters which can be applied to predict the properties of new UVO
modified polymers, blends, copolymers, and composites.

To use this model in a predictive capacity for previously untested polymers, at this
stage, prediction of extrapolated parametric values may not be reasonable except over a
very narrow range for specific homologous polymers. However, rapid initial verification
for untested polymers can be performed by measurements at three to four irradiation
values ranging from H = 0 to H = 00 (saturation of the surface property) and the value of a
can determined from this data. The value of c can be determined by fitting the generated
profile at intermediate measurements. With the generation of more data for a wider
variety of polymers and homologues, predictive assessment of the properties can be

accomplished with increasing reliability.

228

 

8.2 Recommendations for Future Work

The UV oxidation surface treatment process in polymers was studied with the
objective of formulating a process model that would be applicable to a wide range of
polymers. The effect of irradiation on the surface properties is measured and well
understood, but the effect of irradiation on the polymer below the surface, more
specifically the near surface polymer, needs to be investigated because adhesion and
other interfacial phenomena are rarely limited to a two dimensional interface, and are
more often controlled by the formation of three dimensional interphases. The degradation
of properties below the first few monolayers can negate the enhancement of properties in
the top monolayers if, for example, the degradation leads to the formation of a weak
boundary layer. Bulk polymer degradation phenomena are reported in literature, but these
are usually applicable to long term weathering for UVA and UVB (solar) wavelengths.
Specifically, it would be interesting to model the behavior of polymers in the top 500 nm
of the surface. Assuming the same mechanisms observed on the surface also occur in the
sub-surface polymer, the information required for this would be the depth profiles of
ozone concentration in the polymer and the rate dependence of UV induced oxidation on
the ozone concentration. The former can be based on models used to determine the
diffusion of oxygen in polymers. The experimental verification can be accomplished by
measurements on well characterized thin films. The determination of reaction rate
dependence on ozone concentration is more challenging because of difficulties in
accurately measuring low values of ozone concentration and the variety of parallel
reactions occurring in the polymer. In this study no reaction limitation due to ozone

concentration was observed at ozone concentrations as low as 10 ppm. However, a more

229

 

sensitive ozone sampling setup or treatment at low pressures may be able to yield this
information. The diffusion of ozone in the polymer on the short time-scales used in UV
treatment may not be significant, but the influence of pre-diffused oxygen must also be
considered. Molecular oxygen plays two important roles in this process. Oxygen can
quench the polymer from the excited state, and it can also prevent the recombination of
free radicals formed by photolysis of the polymer by reacting with them to form peroxy
radicals which can undergo further reactions. On the polymer surface, the presence of
ozone and atomic oxygen, both strong oxidizers, can mask the contribution of oxygen in
these processes. In the bulk polymer, these mechanisms may be quite different.

The proposed process model for surface oxidation is described in terms of three
parameters and the physical properties expected to contribute to the two material

dependent parameters a and c are given in Table 8.1.

Table 8.1 Process model parameters and relation to physical properties.

 

Parameter Physical Properties

 

Population of reactive sites
Microstructure

 

a Chain configuration
Steric effects
Absorption coefficient
Oxidation rates

C Free volume (temperature)

Chain flexibility
Stabilization (internal and external)
Anti-oxidant additives

 

 

 

 

230

 

The effect of the population of reactive sites has been demonstrated, but the
effects of differing rnicrostructures and chain configurations of the polymer on the
population of sites exposed to UV oxidation can be studied further. Specifically, the
effect of crystallinity on the rate of oxidation for semi-crystalline polymers could be
interesting. The parameter c, related to the reactivity of the surface can be easily
characterized in terms of the UV absorptivity, but it is also strongly dependent on the rate
of oxidative attack by ozone and the reorientation of reacted surfaces at temperatures
above the Tg of the polymer. The free volume of the polymer determines the rate at
which reorientation or isomerism can occur within the polymer, and it is perhaps the key
variable to characterize the constant c for polymers above the Tg at treatment
temperatures. The effect of internal (photo-rearrangement products) and external
(additive) UV stabilizers and anti-oxidants can be modeled as functions decreasing the
absorptivity of the polymer.

The verification and extension of the process model for other polymers also needs
to be performed. A detailed study of polymers belonging to a homologous series, or
binary random copolymers with different ratios of monomers, can be used to measure the
effect of changes in populations and reactivity of surface sites on the rate and extent of
UV surface oxidation. The effect of free volume or temperature (Tmmm - Tg) on surface
reorientation can also be studied on copolymers which exhibit variation in the Tg as a
function of composition. Alternatively UV treatment can be performed at different
temperatures. The latter approach to measuring the effect of temperature presents
challenges such as the accurate measurement of transient surface temperatures on the

polymers as they undergo UV oxidation.

231

 

APPENDIX A

SURFACE ENERGY CALCULATIONS

Results of contact angle measurements with five liquids were used to determine
the surface energy of the solid using polar-dispersive (Equation A.l) and acid-base

(Equation A.2) models.

7L(l+cosl9)=2(‘/}gyz +‘/y§y£] (AJ)
7L()1+0086=)Z[\/7LWLW+\/757L+1/757L] (A.2)

Polar-dispersive model analysis was performed by finding the linear regression fit

 

”2

to five data points obtained by plotting yL(1+cost))/2(YLd) against (pr NS)“2 for each

liquid. The slope of the linear fit is given by (759)”2 and the intercept is (75 d).”2

The analysis for acid-base model was performed using the sum of least squares to
reduce the five three-dimensional equations to three three-dimensional equations. The
following procedure was used for data reduction. Equation A.3 shows the general form
Equation A2 for each liquid (Li, i=1 to 5) and the three unknown variables ysl‘w, 75+, and

75'. The liquid values of surface tension are known and the five linear simultaneous

equations will be henceforth expressed as shown in Equation A4.

71.- (W 1+711(JZ]+71,1\lgj-[L'(l—E3fl]=o (A3)

a,.x + by + ci.z - Pi = 0, i: 1 to 11 (A4)
The sum of least squares approach is used to find the values of variables x, y, and z to

minimize the RHS of equation A.4. This is accomplished by taking partial derivatives of

232

 

the summation of the square of equation A.4 (Equation A5) with respect to variables x,

y, and 2 (Equations A6, A7, A.8).
2(ai.x + by + ci.z - P, )2: +5 i= 1 to n
2 a, (ai.x. + bi.xz + Ci.X3 - Pi) = 0 i: l to n
2 b, (ai.x. + bx; + ci.x3 - Pi) = 0 i = l to n

2 Ci (ai.x1 + bi.X2 + Ci.X3 - Pi) = 0 1: l 10 n

(A.5)
(A.6)
(A.7)

(A.8)

Equations A.6 - A.8 can be solved simultaneously to obtain the values of the three

variables corresponding to the Lifshitz-van der Waals, acid, and base components of the

solid surface energy.

233

 

APPENDIX B

SENSITIVITY CURVE FITS

l. Polycarbonate

Exponential Association (3): y=a(b-exp(-cx))
Coefficient Data:

a = 72.607165

b 1.9687052

c = 0.010709325

Standard Error: 4.1877087

Correlation Coefficient: 0.9752584

Comments: The fit converged to a tolerance of le-006 in 8 iterations. No weighting used.

 

 

 

  

 

 

 

8:4.18770865
r=0.97525845
140: F
A .1
E120-
o a:
F
X
V
€100-
E
'u
so-
60'ITtllrlllllurrllllfiltlllltlirtwttltwrr
o 100 200 300 400 500 600 700 800

Irradiation (H) (mJ/sq.cm.)
Figure B.1 Curve fit for work of adhesion of PC.

234

 

2. Epoxy

Exponential Association (3): y=a(b-exp(-cx))
Coefficient Data:

a = 45.179984

b = 2.9992781

c= 0.015529191

Standard Error: 4.5793659

Correlation Coefficient: 0.9295959

Comments: The fit converged to a tolerance of 1e-006 in 8 iterations. No weighting used.

S = 4.57936589
r = 0.92959587

 

 

 

 

 

140- o O
. g. .
‘5" : s l: f I
'3 J . 1
v; 120‘ r i
v .. 1*.
g ‘ .5.
E ‘ 9 ' .
% 100-
80 I I I I l I I I I l I I I r l I I I 1
o 100 200 300 400

Irradiation, H (mJ/sq.cm)

Figure B.2 Curve fit for work of adhesion of epoxy.

235

 

 

3. PET

Exponential Association (3): y=a(b—exp(-cx))
Coefficient Data:

a = 37.703992

b = 3.3117358

c = 0.029163071

Standard Error: 4.8290926

Correlation Coefficient: 0.9248849

Comments: The fit converged to a tolerance of 1e-006 in 9 iterations. No weighting used.

S = 432909258

 

  
  

 

 

 

 

r=0.92488494
14o
' 9. 9 i
A .
E 120‘
o
1- 1
x -
v
:5
E
g 100-
'o
80 I I I I I I I I I I I j I I I I r I I I
o 100 200 300 400

Irradiation, H (mJ/sq.cm)

Figure B.3 Curve fit for work of adhesion of PET.

236

 

4. Crosslinked Rubber Compound
Exponential Association (3): y=a(b—exp(-cx))
Coefficient Data:

a = 107. 18317

b = 1.2668587

c = 0.010698666

Standard Error: 6.0723686

Correlation Coefficient: 0.9884397

Comments: The fit converged to a tolerance of le-006 in 11 iterations. No weighting

used.

8 = 6.07236864
r = 0.98843969

 

150

130

—l
.5
O

CO
O

70 7‘

dWa/dH (x1OE4)

50

30

 

10' I I I T I I I I I I I I I I l I I I I I I I I I I I I I I

0 1 00 200

Irradiation H (mJ/sq.cm)

300

400

500

 

600

Figure B.4 Curve fit for work of adhesion of crosslinked rubber compound.

237

 

APPENDIX C

CALCULATION OF POLYMER VOLUMES

1. Polycarbonate

Table C.1 Reactive volume fraction of polycarbonate

from group contribution theory.

 

 

 

 

 

 

 

 

 

 

Groups per Volume Reactive Reactive Net
Structural Group Repeat per Group Site Volume Volume
Unit (cc/mole) (Y es/No) (cc/mole) (cc/mole)
2 43.3 N -- 86.6
1H3
_CI3— 1 30.7 Y 30.7 30.7
CH3
0
ll 1 18.9 Y 18.9 18.9
—O—C—O——
Total Volume 49.6 136.2

 

 

Reactive Volume Fraction = 0.36

 

238

 

2. Polybutadiene

CHz—CH=CH—CH2

Table C.2 Reactive volume fraction of polybutadiene

from group contribution theory.

 

 

 

 

 

 

 

 

 

Groups per Volume Reactive Reactive Net
Structural Group Repeat per Group Site Volume Volume
Unit (cc/mole) (Y es/No) (cc/mole) (cc/mole)
—CH2— 2 10.23 N -- 20.46
CH=CH 1 16.9 Y 16.9 16.9
Total Volume 16.9 37.36

 

 

Reactive Volume Fraction = 0.45

 

239

 

3. Poly(ethylene terephthalate)

it it
—C‘C—O—CH2-CH2-O—

Table C.3 Reactive volume fraction of PET

from group contribution theory.

 

 

 

 

 

 

 

 

 

 

Groups per Volume Reactive Reactive Net
Structural Group Repeat per Group Site Volume Volume
Unit (cc/mole) (Y es/No) (cc/mole) (cc/mole)
1 43.3 N -- 43.3
—CH2— 2 14.03 N -- 28.06
O
ll 2 15.2 Y 30.4 30.4
—C—O—
Total Volume 30.4 101.76

 

 

Reactive Volume Fraction = 0.30

 

4. Epoxy (Epon 828 + J effamine T403)

0 O
/\ <1" 1": /\
CHg—CH—O@é@O—CH—CHz-O@C@O—CHZ-CH—CHz

A“, In,

(IIHZ[OCH2CH(CH3)]-x-— NH2

CH3CH2(II—CH2[OCH2CH(CH3)]y—NH2

CH2[OCH2CH(CH3)]-z—NH2

240

x+y+z=5.3

 

 

Table C.4 Reactive volume fraction of epoxy (Epon 828 + Jeffamine T403)

from group contribution theory.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Groups per Volume Reactive Reactive Net
Structural Group Repeat per Group Site Volume Volume
Unit (cc/mole) (Yes/No) (cc/mole) (cc/mole)
Epon 828 Resin
4 43.3 N -- 173.2
1H3
_CI:_ 2 30.7 Y 61.4 61.4
CH3
—CH2— 4 10.23 N -- 40.92
OH
1 3 14.8 Y 44.4 44.4
_CH—
—O— 4 5 Y 20 20
Total Volume (Epon 828 Resin) 125.8 339.92
Jeffamine T403
CH3— 1 13.67 N -- 13.67
_CH2— 9.3 10.23 N -- 95.14
—T— l 3.3 N -- 3.3
—O— 5.3 5 26.5 26.5
CH3
1 5.3 20.45 N -- 108.38
N/ 3 4 3 Y 12 9 12 9
Total Volume (Jeffamine T403) 39.4 259.89
Mixture (100:45 wt. ratio)
1 mol Epon 828 (M01 wt. 598) 125.8 339.92
0.62 mol Jeffamine T403 (Mol wt. 432) 24.42 161.13
Total Volume of Mixture 150.22 501.05

 

 

 

Reactive Volume Fraction = 0.30

 

241

 

   

1|11111111111111111111111111111111111
3 1293 02504 9721