Professional Documents
Culture Documents
Martin Goosey (Auth.), Martin Goosey (Eds.) - Plastics For Electronics-Springer Netherlands (1999)
Martin Goosey (Auth.), Martin Goosey (Eds.) - Plastics For Electronics-Springer Netherlands (1999)
Edited by
Martin Goosey
Chief Scientist and Technology Fellow
Shipley Europe Ltd
Coventry, UK
List of Contributors ix
Preface Xl
John A. Chilton
Bosworth College, Nazareth House, Barrack Road, Northampton, UK
David Dalman
The Dow Chemical Company, Midland, Michigan, MI 48674, USA
John Everett
Dow Deutschland Inc., Industriestrasse I, D-77836 Rheinmunster, Germany
Klaus Feldmann
F APS, Lehrstuhl fUr Fertigungsautomatisierung und Produktionssyste-
matik, University of Erlangen-Nurnberg, Germany
Philip Garrou
The Dow Chemical Company, c/-MCNC Research Triangle Park, NC
27909, USA
Martin Goosey
Shipley Europe Ltd, Herald Way, Coventry, UK
Nigel Hackett
Dexter Electronic Materials, Langley, UK
Thomas Kainmiiller
Ciba Specialty Chemicals Inc., Basel, Switzerland
David Merricks
Shipley Company Inc., Marlborough, MA, USA
Jean-Paul Mollie
Dow Corning SA, Seneffe, Belgium
x I IL-_______________L_IS_T_O__F_C_O_NT__RI__B_U_T_O_R_S______________~
Mike Plant
Kemtron International, South Woodham Ferrers, UK
Frank Pohlau
F APS, Lehrstuhl fUr Fertigungsautomatisierung und Produktionssyste-
matik, University of Erlangen-Numberg, Germany
Martin Roth
Ciba Specialty Chemicals Inc., Basel, Switzerland
Wolfgang Seiz
Ciba Specialty Chemicals Inc., Basel, Switzerland
Preface
Martin Goosey
Shipley Europe
Coventry
1998
Introduction to Polymers and 1
their Important Properties for
Electronic Applications
MARTIN GOOSEY
Shipley Europe Ltd, Coventry, England
The word plastic has its origins in the Greek language where 'plastikos'
meant to form or shape. Today plastic has a number of meanings but the
one which best fits the description of materials detailed in this work is that
referring to any of various substances which harden and retain their shape
after being moulded or formed by heat and pressure. This description is still
rather broad, covering a large number of materials including many such as
tar, clay and even mud which do not really agree with our current
perception of a plastic. In order to qualify the word plastic, a more accurate
description would be synthetic plastic, since this eliminates the naturally
occurring materials that have plastic properties and only includes those
derived by chemical means from relatively simple raw materials. Plastics are
thus man-made polymers where the word polymer is an abbreviation of the
two words poly and monomer and literally means many monomers. It
should also be noted, however, that the word plastic is generally taken to
refer to thermoplastic materials and therefore does not include thermoset-
ting materials. Consequently, in this revision of the original chapter the
word polymer will generally be used unless the reference is specifically to a
thermoplastic material.
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
2 I I INTRODUCTION TO POLYMERS FOR ELECTRONIC APPLICATIONS
During the nineteenth century much interesting work was performed on
polymeric materials beginning with the vulcanization of rubber in 1839
and also including the development of nitrated cellulose (celluloid) in 1868.
It was not until after the tum of the century that the first commercially
successful fully synthetic polymer was produced. This was Bakelite made
from phenol and formaldehyde by Leo Baekeland in 1909. Before the end
of the 1920s a large number of other synthetic polymers had been created,
notable examples being the development of polyvinyl chloride in 1927 and
urea formaldehyde in 1929. Today there are literally hundreds of synthetic
polymers commercially available with ranges of properties making them
suitable for applications in many industries including the electrical and
electronics industries. For every material commercially available, many
more have been synthesized, examined and discarded as of no present use
for various technical or economic reasons. The year of introduction
and typical electronics applications for various polymers are shown in
Table 1.1.
Much of the driving force behind the development of many modem
polymeric materials came from the electrical and electronics industries,
largely because of the numerous areas for application in electrical and
Table 1.2 Typical applications for polymers in the electronics and electrical
industries
Integrated circuits, Transfer moulding compounds, injection
transistors, diodes and moulding compounds, casting, potting,
other discrete devices dip and powder coating compounds, die
attach adhesives, photoresists, intermetal
dielectrics, junction coatings.
Wires and cables Sleevings, coatings, varnishes.
Connectors Transfer moulding, injection moulding,
compression moulding compounds.
Hybrids Conductive and non-conductive
adhesives, sealants, conformal
coatings.
Transformers, coils, Transfer moulding compounds, coatings,
bushings potting compounds, coil impregnants,
wire insulation.
Printed circuit boards Laminates, conformal coatings, solder
masks, masking tapes, component
attachment adhesives, vibration
dampers, photoresists and direct plate processes.
Moulded interconnect devices.
Polymers are different from the majority of simple inorganic and organic
molecules commonly encountered in that they typically have very large
molecules with molecular weights sometimes exceeding one million. Even
though these molecules are big, they can be defined by the basic repeat unit
that forms the polymer chain. Many of the overall properties of a particular
type of polymer will be a direct result of the individual atoms and the way
they are arranged in the repeat unit. One of the most important atoms in
~_____________C_H
__EM
__IS_T_R_Y_O_F __L_Y_M_E_R_S______________~I
__PO I 5
polymer chemistry and indeed in all organic chemistry is the carbon atom.
In polymer chemistry carbon is the atom that forms the basis of the main
structural chains found in most polymer molecules. Carbon has a valency of
four, which simply means that it has four sites at which it can combine with
other elements. If a single atom of carbon combines with four atoms of
hydrogen the gas methane is formed:
H
I
H-C-H
I
H
Carbon can combine with many other elements besides hydrogen and
importantly can also form bonds with itself. Examples of carbon atoms
combined with other carbon atoms and hydrogen atoms are the alkanes,
such as ethane and propane, where two and three carbon atoms are joined
together, respectively, and the rest of the unoccupied sites are used to
combine with hydrogen atoms:
H H H H H
I I I I I
H-C-C-H H-C-C-C-H
I I I I I
H H H H H
Propane (C3HS)
One of the key features of carbon which can ultimately lead to the
formation of polymers is its ability to form multiple bonds with itself. For
example, in the molecule ethylene (C2H 4) there are not enough hydrogen
atoms to occupy all of the available bonding sites and so the carbon atoms
form a double bond between themselves
6 I I INTRODUCTION TO POLYMERS FOR ELECTRONIC APPLICATIONS
These double bonds are very reactive and under the right conditions
ethylene molecules can polymerize in a chain reaction to form a long linear
molecule of ethylene:
-
H H
I I
n C-C
I I
H H n
Ethylene Polyethylene
Carbon also has the capability of forming ring structures containing
various numbers of carbon atoms; the simplest molecule containing six
carbon atoms in a ring structure is benzene:
H H
I I
C-C
©
I
H n
Styrene Polystyrene
These are known as polymers derived from vinyl monomers and the word
vinyl refers to an ethylene molecule with one of its hydrogen atoms replaced
with another atom or molecule. There is no reason why only one of the four
hydrogen atoms can be replaced and it is quite normal to replace more than
one of these atoms. When two atoms are replaced the resultant polymers are
said to be derived from vinylidene monomers, where vinylidene refers to an
ethylene molecule with two of its hydrogen atoms replaced with other atoms
or molecules. When two hydrogen atoms are replaced they are both from
the same carbon atom of the ethylene molecule. Some examples of linear
polymers based on vinylidene monomers are shown below with their
substituents.
F F
I I
C-C
I I
F CI
n
The above examples have been based upon ethylene and its derivative
monomers, all of which have led to the formation of polymer backbones
containing chains of linked carbon atoms. Carbon may join with other
atoms and molecules such that they can also be incorporated into the
polymer backbone. The previous examples have all been formed by the
addition polymerization process in which carbon-carbon double bonds have
been changed to single bonds in order that further carbon atoms could join
together. In many of the polymers containing different atoms in their
backbones polymerization may have occurred by another means such as
condensation polymerization. Typical examples of linear polymers contain-
ing other atoms in addition to carbon in their backbones are shown below:
Structural unit Polymer
toc-{5\-c-0-~_1-of.
.II~II
0
H1'rH
n
Polyethylene terephthalate
Polycarbonate
Although carbon has the ability to form very large long chain molecules it
is not completely unique in this characteristic. Nitrogen and oxygen can also
form long chain molecules with carbon atoms. Sulphur can be incorporated
either as an individual sulphur atom as in polyphenylene sulphide or it may
form short chains of sulphur atoms. Perhaps the main examples of long
chain polymers not requiring carbon atoms to form their backbones are the
silicones. In these materials the main chain is composed of alternating
silicon and oxygen atoms;
R R R
I I I
-Si-O-Si-O-Si-O-
I I I
R R R
These materials are covered in detail in Chapter 2.
One of the main reasons why polymers have become so widely used in
electronics and electrical applications is that they are good dielectric
materials with readily controllable electrical properties. However, the fact
that polymers are good insulators does not mean that they are inert in an
electrical field and the intrinsic properties of a material can usually be related
to performance under specific test conditions. The electrical properties that
have to be considered when selecting a material include dielectric constant,
10 I I INTRODUCTION TO POLYMERS FOR ELECTRONIC APPLICATIONS I
volume resistivity, surface resistivity, dissipation, power and loss factors, arc
resistance and dielectric strength. When selecting a polymer for a specific
application the variation of these properties with changes in the likely service
environment has to be taken into account since many of them are
temperature, frequency, voltage and environmentally variable.
Most polymers serve as electrical insulators and may be required to
operate in conditions of varying voltage with both direct current and
alternating current at frequencies up to and above the GHz region. The
voltage a polymer is exposed to may be as little as a few microvolts in
communications equipment or up to several million volts in power
distribution systems. Similarly, currents can vary from a few picoamps to
many thousands of amps.
All of these variations may be encountered over a wide range of
temperatures, humidities and operating environments. The selection of a
material capable of operating under such variable conditions is obviously
important and thus an understanding of the electrical properties of polymers
and how they vary under these conditions is essential. Some of the more
important electrical properties are now discussed.
1018 Polytetratluoroethylene
1017 Polystyrene, polypropylene
1016 Polyethylene, silicone rubber
10 15 Amine-cured epoxides
10 14 Diamond
1013 Polyamides Insulators
1012
1011
1010 Glass
109
108 Silver bromide
107
106
105
104 Silicon Semiconductors
103
102 Germanium
101
100
10-1
10-2 Quinolinium salts
10-3
10-4 Bismuth, mercury, graphite Conductors
10-5 Polypyrrole and polythiophene
10-6 Silver, copper
10-7
electronics applications such as the direct plate processes that form part of
printed circuit board manufacturing.
Air 1.00053
Fluoropolymers 1.9--2.6
Polypropylene 2.2-2.3
Polystyrene 2.45-2.65
Benzocyclobutenes 2.6-2.8
Epoxides 2.8-5.5
Polyetherimide (Ultem) 3.15
Polyimide (Pyralin) 3.50
Epoxy-glass laminates 4.4-5.2
Phenolics 5.0-6.5
Ceramic filled PTFE 6.0--10.2
Nitrobenzene 34.9
Water 79.5
Lead titanate 200
Lead zirconate titanate (PZT) 3400
14 I I INTRODUCTION TO POLYMERS FOR ELECI'RONIC APPLICATIONS
Relative permittivities of polymers vary with frequency and can either
increase or decrease depending upon molecular structure, and the presence
of any additives or fillers. In order to be of use in electrical and electronic
applications a polymer should have reasonably constant dielectric properties
over a broad range of temperatures, frequencies and humidities. It is for
these reasons that materials such as polyethylene, polystyrene and the
fluorocarbons find such extensive usage as electrical insulators.
Strain
Table 1.6 Glass transition temperatures (Tg) and melting points (Tm) of some
common polymers
Polymer
1.5 SUMMARY
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
26 1 1-1__S_I_L_IC_O_NE_M_A_T_E_R_IA_L_S_F_O_R_E_L_E_C_TR_O--,--NI_C_CO_M_PO_N_E_NT_S_------'
R R R R
1 1 1 1
-O--Si-O--Si-O--Si-O--Si-
1
R
1
I.
°I.
1
R
1
-Sl-O--Sl-O--
°
1 1
R
1
Fig. 2.1. Model of silicone fluid molecule showing helical arrangement of polymer
chain with silicone atoms bonded to oxygen atoms and to outer organic groups.
28 II SILICONE MATERIALS FOR ELECTRONIC COMPONENTS
~----------------------------------------------------~
mixing is apparent when a uniform dark grey colour is obtained. Figure 2.3
shows the use of this encapsulant to insulate and protect high-voltage
cascade units for visual display terminals, with processing in a fully
38 1 ,--I__S_I_L_IC_O_N_E_M_A_T_E_R_IA_L_S_F_O_R_E_L_E_CT_R_O_NI_C_C_O_M_PO_N_E_NT_S_-----'
automated vacuum, mixing and transfer system. The silicone encapsulant is
highly resistant to fire, passing the stringent UL 94-YO standard specified by
the computer manufacturer end user, and it also protects against corona
discharge at high voltage. (It should be noted, however, that in order to
achieve optimum performance a primer is sometimes needed with this
encapsulant.)
Product (2) in Table 2.3 is an example of a product that has been
developed with an integral primer, though a heat cure is necessary for its
activation. This type of self-priming silicone encapsulant is very useful in
cases where priming may be difficult because of small gaps. It is also a cost-
saving process since separate primers need a drying time before encapsu-
lation.
The next product example (3) is a transparent encapsulant, again
processed by mixing two fluid components as supplied. One version will
cure either at room temperature or by heating, while a variant has much
longer pot life at room temperature and is normally cured by heating in the
temperature range 70-150°C. The cured product is optically clear, permit-
ting visual inspection of an encapsulated circuit or components. One specific
application, outside the definition of electronic encapsulation, is the
protective coating of optical fibres. 9 In a continuous process the silicone is
cured in a few seconds as the fibre passes through a radiant curing zone at
over 3OO°C. The absence of any solvent or by-products prevents blistering
and the silicone provides a chemically pure coating which is also soft enough
to prevent mechanical damage to the glass fibre.
Example (4) is a silicone gel. Silicone dielectric gels are polymers with a
low level of crosslinking and they exhibit good pressure-sensitive adhesion
to most substrates. Therefore, they are extremely good in protecting
electronic circuits against corrosion and metal migration. Silicone gels form
a cushioning gel-like mass which combines the stress relief and self-healing
properties of a liquid, thereby preventing stress damage to delicate
components, even when subjected to heat ageing. The unique silicone
dielectric gels retain their softness, adhesion and electrical properties after
more than 1000 cycles from 50°C to 200°C (Fig. 2.3). Silicone gels also have
outstanding dielectric properties. Silicone gels have proven their unique
characteristics in the protection of automotive electronic modules (Fig. 2.4)
where the exposure to environmental extremes including high and low
temperatures, mechanical shock and vibration, water and salt from road
surfaces require excellent and reliable protection. 1o
Typical applications for silicone gels are:
• hybrid encapsulation
• connection sealant
• clean room filter sealing
• potting of high voltage units
L -_ _ _ _ S_IL_I_C_O_N_E_A_P_P_Ll_C_A_T_IO_N_S_IN_E_L_E_CT_R_O_N_I_C_S_ _ _ _ ---'I I 39
Fig. 2.4. Silicone dielectric gels function as durable dielectric insulation, as barriers
against environmental contaminants, and as stress-relieving shock and vibration
absorbers. Here it is applied in an Engine Management Unit from Bosch Germany.
Special gels have been developed for extra low temperature applications
(-80°C) and for the protection of units and circuits in fuel or solvent-rich
environments.
In microelectronic packages high purity silicone gels ll are used to provide
premium moisture protection, due to their good adhesion, and to keep the
packaging material cost low (Fig. 2.5). See also the further information in
section 2.4.4 on high purity electronic coatings and adhesives.
Fig. 2.5. Delco Electronics Corporation's Sensing and Diagnostic Module (SDM) is
an active part of a car air bag system. A one-part, solventless silicone conformal
coating protects the unit's electronic components without stressing delicate soldered
joints.
reinforced circuit board, the silicone coating did not improve the assembly
to the same degree of fire resistance as the silicone itself, but the lower value
for the uncoated board was not impaired. Conversely, other types of coating
reduced the fire resistance of the uncoated epoxy-glass fibre circuit board.
The third coating (8) belongs to a new family of one-part solventless
silicone elastomers that cure with heat to a give a flexible and transparent
material. Heat cured conformal coatings offer a significant range of
processing benefits. Their most common application in electronics is to
protect printed circuit boards, and they have been used successfully in this
way for many automotive applications.
One useful attribute of all these silicone coatings is their repairability.
Solder joints can be re-made, often even without removal of the coating. No
carbon is formed and fresh silicone may then be applied to the area which
has been repaired. They also comply with the MIL-I-460S8C specification
for PCB coatings. One limitation, however, is their poor resistance to strong
solvents, such as fluorinated hydrocarbons.
this family of products, the user can choose from a wide variety of
properties to select the material that is best suited for each specific
application. 14 Properties to consider in product selection are hardness,
elongation, green strength, and tensile strength. Very good adhesion can be
achieved with most metal and reactive plastic surfaces. To meet various
process requirements, the materials are made in a wide viscosity range, from
flowable to thixotropic or non-slump. In addition to the manufacturing time
and space savings offered by a relatively fast cure, some systems are
formulated with a UV dye to offer the capability of optical detection and
inspection. These silicone adhesives also help to accommodate the thermal
expansion coefficient mismatch generated stresses created when two
materials with different coefficients of thermal expansion (CTE) are bonded
together.
Another advantage silicone adhesives bring is moisture resistance.
Moisture absorption is very low and has little effect on these adhesives or
their electrical properties. This quality, together with adhesion and
corrosion protection, is important in harsh environments, such as in
under-the-hood automotive-type applications.
In the same family of primerless silicone adhesives there are materials
available with the additional capability of dissipating thermal energy away
from delicate components. IS A thermally conductive silicone adhesive
basically has four components:
L -_ _ _ _ --'I I
S_1_L_IC_O_N_E_A_P_P_LI_C_A_T_I_O_N_S_I_N_E_L_E_CT_R_O_NI_C_S_ _ _ _ 43
Fig. 2.7. Thermally conductive adhesive to quickly dissipate thermal energy away
from delicate components and absorb stresses created when two materials with
different CTE are bonded together.
only for the semiconductor surface but also over the connecting bond wires
(Fig. 2.8). In practice it has been found best to cover the bond wire near to
the chip surface but not along its whole length, in order to avoid failures due
to wire breakage. Similar types of coating are also used to prevent
corrosion,17 current leakage or failure under inverse voltage at thyristor
junctions.
The rheology desired for many of these applications includes high
thixotropy, so that shearing forces during dispensing allow the coating to
flow closely round the chip surface and bond wires. The original viscosity
must then be rapidly restored so that the silicone coating remains in situ
without further flow during the curing stage, whether at room temperature
or by heating at, say, 120-1 50°C. The material is cured by heating and,
typical of silicone polymers, shows a much smaller fall in viscosity as
temperature rises prior to cure than an organic glob top coating such as an
epoxy. This ensures that the glob does not flow away from the chip during
the curing process.
Outside the field of discrete component protection, high-purity silicone
coatings contribute to the growing use of microcircuits combined with
newer forms of interconnection and packaging. In wafer fabrication,
flowable oxide or hydrogen silsesquioxane (HSQ) provides an interlevel
dielectric coating with excellent planarity and gap filling in a simple spin-on
L -_ _ _ _ ---ll I
S_1_L_IC_O_N_E_A_P_P_L_IC_A_T_I_O_N_S_I_N_E_L_E_CT_R_O_N_I_C_S_ _ _ _ 45
Fig. 2.8. High performance protective silicone materials are integral to product
advancement for the semiconductor industry.
Fig. 2.9. Microcomputer keypad made from a single moulding of a liquid silicone
rubber, with its underlying membrane switching circuit (Sinclair Research, Ha-
ffenden and NFl Electronics).
L -_ _ _ _ _ _ _ F_UT_V_R_E_D_EV_E_L_O_P_M_E_NT_S_ _ _ _ _ _ _ ---'I I 47
elastic properties over a long life without the adverse changes due to ageing
characteristics of other elastomers. When the keys are depressed, they
actuate a membrane switching sandwich below and then spring back into
place. The resilience of the keypad shown was tested over five million
operational cycles before entering volume production.
2.6 REFERENCES
1. Kipping, F. S. and Lloyd, L. L., Organic derivatives of silicon, J. Chem. Soc., 79,
449 (1901). .
2. Alperowicz, N., Dow Corning poised for growth in Europe, Chemical Week,
April 3, 1996.
3. Eabom, C., Organosilicon Compounds, Butterworth Scientific Publications,
London, 1960.
4. Noll, W., Chemie und Technologie der Silicone, Verlag Chemie, Weinheim, 1960.
5. Von Hippe!, A., Dielectric Materials and Applications, Technology Press of MIT
and John Wiley & Sons Inc., New York, 1954.
6. Joint Assessment of Commodity Chemicals No. 26: Linear Polydimethylsilox-
anes, ECETOC Brussels, September 1994.
7. Hilado, C. J., Casey, C. J., Christensen, D. F. and Lipowitz, J., Toxicity of
pyrolysis gases from silicone polymers, J. Combustion Toxicology,S, 130-140,
May 1978.
8. Lynch, W., Handbook of Silicone Rubber Fabrication, Van Nostrand Rheinhold
Co., New York, 1978.
9. France, P. W., Dunn, P. L. and Reeve, M. H., Plastic coating of glass fibres and
its influence on strength, Fibre and Integrated Optics, 2, (1979).
10. Mollie, J.-P., and Paquet, R., Performance of new silicone adhesives and
encapsulants at high and low temperature, Proc. lEE Automotive Eighth
International Conference, London, 1988.
11. Kookootsedes, G., and Antonen, R., Selecting protective materials for coating
hybrid circuits, Hybrid Circuit Technology, March 1986.
12. Day, A. G., and Weller, M.G., Flammability testing of printed wiring board
material and the effect of the conformal coating, Proc. Internepcon Conference,
Brighton, 1980.
13. Mollie, J.-P., Klebstoffe aus Silicon, Electronik Produktion Pruftechnik, March
1995.
14. Vanwert, B., One-part thermal cure silicone adhesives, Proc. NEPCON WEST
'96, California, 1996.
15. Wilson, S., Thermally conductive adhesives for high thermally stressed
assembly, Proc. NEPCON WEST '96, California, 1996.
16. Verhelst, V., Capillary zone electrophoresis determination of ionic impurities in
silicone product used for electronic applications, J. Chromatography, Special
Issue, International Symposium, Reading, 1996.
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _R_E_F_ER_E_N_C_E_S______________ ~I I 49
17. Otsuka, K. et al., The mechanisms that provide corrosion protection for silicone
gel encapsulated chips. IEEE Transactions on Components, Hybrids, CUMT-12
(4), December 1987.
18. Ahlburn, B. T. et al., Hydrogen silsesquioxane-based flowable oxide as an
element in the inter-level dielectric for sub 0.5 micron ULSI circuits, Multilevel
Interconnection Conference, Santa Clara, 1995.
19. Ballance, D. S., Scheibert, K. A. and Tietz, J. V., Low temperature reflow
planarisation using a novel spin-on inter-level dielectric, Multilevel Interconnec-
tion Conference (VMIC), 1995.
20. Waeterloos, J. and Meynen, H., Integrating a hydrogen-silsesquioxane spin-on
dielectric in a quarter-micron technology, DUMIC, California, February 1997.
21. Cush, R.J., LSR: the versatile alternative, British Plastics and Rubber, 29-31
(October 1982).
22. Lipowitz, J., Combustion, Flammability and Fire Hazard Properties of Silicones,
Publication NMAB-342, National Academy of Sciences, Washington, DC,
1978.
23. Weyenberg, D. R., and Lane, T. H., Future Directions for Silicon-based
Polymers, American Chemical Society, 1990.
Engineering Thermoplastics 3
HARRY VAN BAAL
General Electric Plastics bv, Bergen op Zoom, The Netherlands
3.1 INTRODUCTION
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
52 I ~I____________E_N__G_IN_E_E_R_IN__G__T_H_E_RM__O_P_L_A_S_T_IC_S____________~
thermoplastic materials in electrotechnical applications is also given later in
this chapter.
3.2.2 Acrylics
Acrylics, specifically those from the polymethylmethacrylate (PMMA)
family, are widely used in the electrotechnical industry. Typical applications
include lighting diffusers, such as those used in street lighting where impact
is not critical. These amorphous materials are characterized by sparkling
ENGINEERING THERMOPLASTICS II
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~ 53
H0f-<Q)-rQ-0- LOtQ-rQ-OH
CH) n CH)
Polycarbonate
Flame retardency can normally be achieved through the incorporation of
brominated components such as tetrabromobisphenol A.
Amorphous polycarbonate (PC) resins are distinguished by their unique
balance of toughness, heat and flame resistance, dimensional stability and
natural transparency. In the electrotechnical industry they are used in a
broad range of end-use applications including connectors, wiring devices,
terminal blocks and indoor and outdoor lighting diffusers. The insulating
and related electrical properties of polycarbonate are excellent and remain
almost unaffected by changes in temperature. Certain combinations of
chemical environments, temperature and stress can adversely affect moulded
polycarbonate parts. However, grades with improved chemical resistance
are available and special coating systems can be applied to provide
t°-@-o-@-~~. PEEK
t°-@-~-@-~~. PEKK
{
II ~~
~:J§lo-@-T-@-o-©r! 0
0
" 0II .
Me
°II 0
:~-@-CH2-@-~-~ Polyamide-imide
II H
o
Polyphenylene sulphide
Polyethersulfone
This section introduces some of the principal standard tests that are used to
characterize thermoplastic materials used in electrotechnical applications.
These can be grouped into flammability, electrical, thermal, mechanical and
weathering tests.
3.3.1 Flammability
In nearly all electrical and electronic applications of polymers, due attention
must be paid to the flammability of the materials used in the construction of
a particular piece of equipment. Polymeric materials must usually pass strict
flammability rating tests and must ideally also be known not to produce
toxic decomposition products if burning or thermally induced decomposi-
tion should occur.
Burning behaviour is not only a material chemistry and formulation
characteristic but it is also dependent on physical dimensions including the
shape and wall thickness of the application, and on the processing
conditions employed. As there are possibly more flammability test methods
than polymer types, this section gives details of only the most commonly
recognized tests.
In the USA, testing is carried out according to Underwriters' Labora-
tories UL 94. The main tests for this classification fall into three categories:
• UL 94HB
• UL 94V
66 I ~1____________E_N_G_I_N_E_E_ru_N_G__TH__E_R_M_O_P_L_A_S_TI_C_S__________~
• UL945V
Of the three tests, UL 94HB is the least stringent and UL 94 5V is the most
stringent with two classifications, 5VA and 5VB. UL 94V classifications
range from V-O, V-I to V-2, with a V-O rating giving the best flammability
behaviour.
In Europe, the flammability of polymeric materials is commonly tested
using two different methods. The first of these is the glow wire test (GWT),
according to lEC 695-2-1, and the second is the needle flame test according
to lEC 695-2-2.
The glow wire test simulates thermal stresses which may be produced by
sources of heat or ignition, such as might be encountered, for example, in
electrical applications by overloaded resistors or glowing elements. A
sample of the insulating material is held vertically for 30 seconds with a
force of I Newton against the tip of an electrically heated glowing wire. The
travel of the glow wire tip through the sample is limited. After withdrawing
the sample, the time for the flames to extinguish plus the presence of any
burning drops is noted. The sample is deemed to have passed the glow wire
test if one of the following occurs:
• if there is no flame and no glowing;
• if flames or glowing of the specimen, of the surroundings and of the layer
below, extinguish within 30 seconds after removal of the glow wire and
the surrounding parts and the layer below have not burned away
completely.
In the glow wire test, actual test temperatures can vary considerably from
650°C for non-current carrying parts, to 850°C and even 960°C depending
on end-use requirements. Most engineering thermoplastics pass this test at
650°C with a I mm sample thickness. To pass at the higher temperatures,
materials are generally at least UL 94V-2 rated, and examples include flame-
retardant grades of modified PPE and polycarbonate.
The Limited Oxygen Index (LOI) according to ISO 4589 is defined as the
minimum concentration of oxygen, expressed as a percentage by volume, in
a mixture of oxygen and nitrogen, which will just support the flaming
combustion of a material. It is a test which measures the relative
flammability of materials by burning them in a controlled environment.
The test atmosphere is an externally controlled mixture of nitrogen and
oxygen and the supported specimen is ignited with a pilot flame which is
then removed. In successive test runs the oxygen concentration is reduced to
a point where the sample can no longer support combustion. The higher the
value indicated, the less the likelihood of combustion. In general, a material
with an LOI of > 21 % will not bum under atmospheric conditions. (Several
French standards, however, require an LOI of > 25%, and even> 28% in
some cases.)
PRINCIPAL STANDARD TESTS
~----------------------------------------------------~
II 67
3.3.2 Electrical
The Comparative Tracking Index (CTI) is one of the key electrical tests
which are carried out to determine the safety of components carrying live
parts. According to IEC 112 and UL 746, the CTI test indicates the voltage
at which a material continues to resist tracking when the surface is exposed
under stress to contaminated water. The CTI test is carried out using two
platinum electrodes of given dimensions placed on the sample surface. A
potential difference is applied between these and the voltage is gradually
increased in steps of 25 volts up to a maximum of 600 V. The surface of the
sample is wetted with 50 drops of a 0.1 % ammonium chloride solution by
allowing the drops to fall centrally between the two electrodes. (The size and
frequency of the drops is specified.) When no current is flowing at the
applied voltage the test is repeated with increasing voltages until there is a
flow of current. This voltage, decreased by one step of 25 volts is called the
CTI. The test is then repeated at this voltage but this time using 100 drops of
the solution. For direct support of live parts, a CTI greater than 175 V is
required. The CTI 'M' test is a similar procedure which creates more severe
conditions for the testing of industrial parts.
Engineering thermoplastics in general have CTIs of > 175 volts. It should
be noted, however, that the CTI can be affected by the addition of flame-
retardant additives, colours/pigments and fillers. Carbon black, flame
retardants and glass fibres for example, have a distinct negative effect on the
CTI whilst minerals such as Ti02 tend to raise the values. Materials meeting
the CTI requirements at 600 volts are known as 'high tracking resins'.
3.3.3 Thermal
All thermoplastics will soften at elevated temperature. According to ISO 75,
the heat deflection temperature (HDT) or the deflection temperature under
load (DTUL) is a relative measure of a material's ability to perform at
elevated temperatures while supporting a load. The test measures the effect
of temperature on stiffness and a standard test specimen is given a defined
surface stress and the temperature is· raised at a uniform rate. In both the
ISO and ASTM standards a loaded test bar is placed in a silicone oil-filled
heating bath. The force is allowed to act for 5 minutes and then the original
bath temperature of 23°C is raised at the rate of 2°Cfminute. The deflection
of the test bar is continuously observed and the temperature at which the
deflection reaches 0.32 mm (ISO) or 0.25 mm (ASTM) is reported as
deflection temperature under load or heat deflection temperature.
In amorphous materials the HDT is almost the same as the glass
transition temperature (Tg ). However, it is worth noting that this measure-
ment does not necessarily represent the upper use temperature of a material.
Many crystalline materials can be used at temperatures well above their
68 I I~___________E_N_G_I_N_E_E_R_IN_G__T_H_E_R_M_O_P_L_A_S_TI_C_S__________~
HDTs, depending on factors such as the load, the time and whether the
material is filled or unfilled. .
The ball pressure test, according to IEC 335-1, is the most common
softening temperature test and is similar to the Vicat method. During this
test a steel ball is pressed against the sample in a heating cabinet for one
hour. After one hour the ball is removed, the sample is cooled in water for
10 seconds and the remaining impression caused by the ball is measured. In
general, the ambient temperature is 125°C for current carrying parts and
75°C for non-current carrying parts. The diameter of the impression caused
by the ball should not exceed 2 mm.
In accordance with UL 746B, the Relative Thermal Index (RTI) of a
material is an indication of the material's ability to retain a particular
property when exposed to elevated temperatures for an extended period of
time. There can be up to three independent RTI ratings assigned to a
material: electrical, mechanical with impact and mechanical without impact.
3.3.4 Mechanical
To withstand stresses imposed during installation and use, components are
tested to ensure that they can resist mechanical damage. The range of
mechanical properties which are tested includes impact strength, E-
modulus, tensile elongation and hardness. Equipment will often be required
to withstand heat ageing and exposure to humidity while retaining its
mechanical properties.
Impact testing gives the opportunity to evaluate a material under
conditions typical of what might be encountered in real life if a component
is dropped or made to absorb rapidly an impact in some other way. Two
tests used to evaluate the effects of impact on a sample are the Izod and
Charpy methods and these are detailed in ISO 180 (ASTM D256) and
ISO 179 (also ASTM D256) respectively. These methods are used to
investigate the behaviour of suitably defined specimens under specified
impact stresses and to estimate the brittleness or toughness of specimens.
Both tests are performed using a pendulum impact machine in which a
clamped notched sample is struck with a hardened steel hammer released
from a pre-determined height. The impact of the hammer causes the sample
to shear from the sudden load and any residual energy carries the pendulum
upwards. The difference in the drop height and the return height can be used
to determine the energy required to break the test bar. The test can be
carried out at various temperatures and also with a number of test bar
designs.
The modulus of a sample can be determined by monitoring the amount of
deformation (strain) introduced by a given load (stress) and stress/stress
strain relationships under tension are the most widely reported property for
comparing materials or designing for a specific application. Tensile stress/
F_U_R_TH
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ AP_P_L_I_C_A_T_IO_N_S______________~I
__E_R__ I 69
3.3.5 Weathering
Weathering is the deterioration of a material under the influence of
temperature, oxygen, relative humidity and UV radiation. These can result
in complete degradation of a polymer with the loss of many important
properties and also its colour. The weatherability performance of a material
is therefore vitally important, particularly where the sample is used in critical
exterior applications. Weathering tests are divided into 'real time' exterior
testing and laboratory-accelerated testing and, according to ISO 4892-2,
include both wet and dry cycles for applications intended for external use.
It should be noted that colour, an integral part of a material's
formulation, can have a significant influence on visual appearance in UV
ageing performance as well as affecting other properties.
3.4.2 Enclosures
The characteristic properties required of materials for enclosures include,
most importantly, -surface finish, impact strength and low flammability. In
accordance with the standards EN 60670 and lEC 439, most high
performance engineering thermoplastics meet the diverse requirements of
enclosures for accessories for household and similar fixed electrical instal-
lations, including both current carrying and non-current carrying parts.
Ongoing development in both materials and process technology ensures
that material suppliers keep in step with market trends. Predominantly,
these trends are for an optimum balance of performance and cost, with a
drive towards improved low temperature performance, chemical and UV
resistance and halogen-free flame retardancy. Colourability is also an
important feature as is the suitability for laser printing.
Each engineering thermoplastic offers a different balance of key properties
to meet individual application needs. Power tool housings, for example,
require the chemical resistance and impact performance of a material like
PBT. Housings for fuse boxes and cable distribution frames also require
high impact resistance combined with good tracking resistance and constant
electrical properties in humid conditions. PBT, modified PPE and PC resins
are ideal for these applications. Uninterrupted power supply (UPS) batteries
require flame-retardancy, chemical resistance, dimensional stability and ease
of moulding for which flame retardant grades of modified PPE resin or ABS
are a popular choice.
Structural foam moulding materials are a sound choice for larger
enclosures as they offer cost-effective parts with good aesthetics, rigidity
and strength.
3.4.3 Installations
PVC is the established material for applications such as cable channels,
conduits and industrial trunking. However, where specifications require
L -_ _ _ _ _ _ _ _ FV_R_T_H_E_R_A_P_P_L_I_C_A_T_IO_N_S_ _ _ _ _ _ _ ---->I I 71
3.4.7 Connectors
In the connector industry, no specific IEC material criteria exist. However,
different segments require an individual balance of material properties. The
predominant trends in connector technology are high heat soldering,
miniaturization and MIDs (see Chapter 11). Materials must be easily
moulded in intricate moulding tools with thin sections and very close
tolerances.
Manufacturers are offered a wide range of high performance flame
retardant materials which can be broken down into amorphous resins,
offering high dimensional stability, and glass-reinforced semi-crystalline
resins, which feature high chemical and temperature resistance. Materials
which are generally used in this sector are thermoplastic polyesters,
polycarbonate, nylon and modified PPE. Polysulphones offer high heat
distortion temperatures but at a high cost which limits their use to specific
applications.
3.4.12 Lighting
With an extensive portfolio of engineering thermoplastics satisfying low,
medium and high heat requirements, lighting manufacturers have wide
design freedom to replace metal and thermoset materials in a broad range of
applications. The leading European standard for lighting applications is EN
60598: Part 1 covers general requirements on flammability, GWT and cn
and Part 2 covers more specific requirements by application.
3.4.12.2 Diffusers
Polycarbonate is an ideal material for diffusers, produced by injection
moulding, extrusion or blow moulding. This high impact, transparent
material provides an ideal balance of mechanical, optical and thermal
76 I I~___________E_N_G_I_N_E_E_R_IN_G__T_H_E_R_M_O_P_L_A_S_T_IC_S__________~
The choice of material for the lens is often dependent on the height at
which it is mounted. Where risk of damage to the lens through vandalism is
minimal, then PMMA is widely used because of its excellent UV stability
and cost advantages. However, where superior impact resistance is required,
PC is generally specified. PC is also the material of choice for the structural
components, providing excellent creep performance and mechanical prop-
erties. PP or PC are commonly used for the covers, depending on the
required cost/performance balance.
3.4.12.6 Spotlights
The broad range of engineering thermoplastics provides the design latitude
to meet the varying levels of service temperature, with high performance
polyphenylenesulphide providing a relative thermal index (RTI) of up to
220°C.
Excellent dimensional stability, high temperature resistance and the
ability to achieve tight tolerances make polyetherimide a good choice for
critical components.
3.4.12.9 Reflectors
Engineering thermoplastics offer reflector manufacturers a brighter future in
terms of greater design freedom and cost efficiency. The excellent combi-
nation of heat resistance and good surface finish make thermoplastic
polyester, polyetherimide and modified PPE/PA alloys ideal for the
FURTHER APPLICATIONS
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ II
~
79
3.5.1 Metallization
The metallization of engineering thermoplastics is normally undertaken for
either decorative or functional reasons. A very large and growing quantity
of plastic parts is metallized for reasons of electromagnetic compatibility
(EMC) or to shield electronics from electromagnetic interference (EMI), as
plastics are inherently transparent to EM!. EMI is a recognized form of
pollution and since European legislation was introduced in 1992, there have
been strict limitations on the levels of interference that can be emitted from
any computing device. There are various ways to metallize engineering
thermoplastics using both chemical and physical deposition of metal. A
typical route would involve the deposition of metals such as copper, nickel
and chrome depending on the application. For example, for electromagnetic
shielding, the part could be metallized using electroless copper and nickel
chemistry. A layer of around 5 l1IIl of copper would be deposited onto the
textured surface of the part and a thin top coating of nickel would provide
the outer protective layer. The key to the successful metallization of
polymers is the adhesion of the deposited metal and this is largely
determined by the surface morphology of the polymer sample prior to
metallization. Typically, the surface of the polymer may be treated with
strong oxidizing agents such as chromic acid and alkali permanganate
solutions to achieve the desired result.
3.5.2 Painting
Plastic parts today are primarily painted to enhance aesthetics and provide
colour uniformity. The paint layer can also have an additional beneficial
effect on the surface gloss and offer improved scratch and UV resistance.
However, a key advantage of high performance engineering thermoplastics
is that they offer self-coloured grades with superior surface quality which
often avoids the necessity to apply additional paint.
initial investment has been recovered, unit cost savings are achieved through
simplified, faster production processes, reduced inventories and the recy-
clability of in-plant scrap. Furthermore, set-up and change-over times in
production are minimized, allowing greater manufacturing flexibility and
reduced down-time.
A broad range of laser markable materials has been developed using
proprietary combinations of pigments and additives. Furthermore, expand-
ing on the choice of 'light mark on dark background' or 'dark mark on light
background', the technology has recently expanded to allow laser marking
in many contrasting colour combinations, including both transparent and
translucent colours.
3.7 CONCLUSION
4.1 INTRODUCTION
The epoxide resins are a class of materials that possess all the properties
required to make them useful for a wide range of applications throughout
the electronics industry. When fully cured they have good electrical
properties, low shrinkage, good adhesion and resistance to thermal and
mechanical shock, whilst also possessing resistance to moisture, solvents and
general chemical attack.
Epoxide resins are all based on the epoxide group, a strained three-
membered carbon, carbon, oxygen ring structure (also known as the oxirane
group);
o
/\
-CH-CH2
The first synthesis of an epoxide group was made early in the nineteenth
century, but it was not until 100 years later that the possible uses of epoxides
began to be realized. They were brought to the attention of industry in 1939,
with the publication of a German patent by I. G. Farben l describing liquid
epoxides. P. Castan of De Trey Freres2 in Switzerland (patents licensed to
CIBA) and S. O. Greenlee of Devoe Raynolds3 first identified the
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
84 I LI____E_P_O_XI_D_E_R_E_S_IN_S_A_ND
__T_H_E_I_R_F_O_R_M_U_LA_T_I_O_N_ _ _ _ .....l
/\ /\
diglycidyl ether of bisphenol A (DGEBA):
o CH 0
-o-I~
H2C-CH-CH2"0 ~ h 9~O-CH2CH-CH2
CH3
Diglycidyletber of bisphenol A
In order to form the familiar crosslinked networks with the desired
physical properties, epoxide resins are cured by the addition of a suitable
chemical known as a hardener or curing agent. The term 'curing' is used to
describe the process in which an epoxide and a curing agent react to give a
highly crosslinked network. The curing agents for epoxides are numerous
and come from well over 100 different classes of chemical compounds. The
curing of epoxides occurs by an addition type reaction, with the interaction
of resin and hardener or by a catalytically induced homopolymerization to
form a crosslinked structure. These types of curing reaction give epoxides a
particular advantage over some other polymers in which a condensation-
type cure leads to the evolution of a small unwanted molecule such as water
or an alcohol. Condensation molecules are an obvious potential source of
problems when polymerization occurs in the proximity of an electronic
circuit, since they can lead to device malfunctioning or corrosion of
metallization. Another advantage offered by the addition type curing
reaction is reduced shrinkage and epoxides are well known to have
unusually low shrinkage values (usually less than 2%).
In order to be useful, a cured epoxide compound should possess as many
of the following properties as possible:
• good adhesion
• environmental and chemical stability
• thermal stability
INT
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ____R_O_D_U_C_T_IO
__ N__________________~I I 85
The epoxide resins, whilst being characterized by the epoxy group, are
available with greatly varying physical and chemical properties. They
exhibit widely ranging molecular structures, molecular weights and viscos-
ities. This wide range of epoxide resins means that they find broad usage in
all types of application, particularly in the electronics industry. By careful
selection of the correct resins and other components of the formulation, the
epoxy resins can be made to yield a variety of forms, ranging from soft and
flexible materials to hard, tough, highly crosslinked, chemically resistant
products.
One of the most commonly used groups of epoxide resins is that
comprised of the diglycidyl ethers, and these are made by reacting
epichlorohydrin with a bisphenol such as bisphenol A. Bisphenol A is
easily and inexpensively formed from the reaction between acetone and
phenol, whilst epichlorohydrin is derived from propylene via allylchloride.
88 II
L ____ E_P_O_X_I_D_E_R_E_S_I_N_S_A_ND
__T_H_E_I_R_F_O_R_M_U_L_A_T_IO_N
_ _ _ _-----I
HO-o-to-OH + 2
o
~CI
Bisphenol A Epichlorohydrin
For liquid epoxy resins n is generally less than 1, while for the solid types n
is generally 2 or larger, although it should be remembered that a product
will always be a mixture of various molecular weight resins. Higher
molecular weight resins with a value of 1 < n < 4 can be obtained by reducing
the amount of epichlorohydrin relative to bisphenol A and increasing the
reaction temperature gradually to about 95°C. This so-called 'Taffy' process
is used for solid resins up to a molecular weight of about 1500. An
alternative is the 'advancement' process, where difunctional glycidylethers
and bisphenols are submitted to a polyaddition reaction to give higher
molecular weight, solid resins with n as high as 20 or 30. The relatively high
reaction temperatures of 160--200°C necessitate a careful selection of
suitable catalysts in order to avoid side reactions such as branching or even
gelation.
As their average molecular weight increases, the resins become more and
more viscous until they eventually turn into solids with increasing melting
points (Table 4.2).
Besides the diglycidyl ether of bisphenol A, many other glycidyl ether-
type epoxides are of considerable importance. Closely related is the
diglycidyl ether of bisphenol F (dihydroxydiphenylmethane), offering lower
viscosities and being used as an additive to prevent crystallization of
bisphenol-A-based liquid resins with high purity.
L-_________________E_PO
__X_ID
__E_R_E_S_IN_S__________________ ~I I 89
Table 4.2 Melting points of typical epoxide resins
Diglycidyletber of bisphenol F
The preparation of an epoxide novolac type resin is shown below:
6 -
OH OH
0
+ HJl H CH 2
n
Novolac resin
o~
0
0
Novolac resin + ~CI- CH 2
n
Epoxide novolac
The epoxide novo lacs are a very important class of compounds, since they
are the basis of the materials used in the formulation of many of the epoxide
moulding compounds. They are ideal in these applications because very high
crosslink densities are achieved in the cured polymer network. This is a
direct result of the large number of epoxy groups available in the resin for
reaction and these also give the added advantage of high temperature
stability. A typical epoxy novolac might have a molecular weight of around
650, with a weight per epoxide of 175-182 and a very high viscosity. In order
90 1 ,-I____E_PO_XI_D_E_R_E_SIN_S_A_ND
__T_H_E_I_R_F_O_RM_U_L_A_TI_O_N_ _ _ -----l
o _~ ~ 0
~O~W-CH=CH~O~
o
Dihydroxychalconediglycidy lether
In addition to the diglycidyl ether-based resins, other epoxide resins are
known, that are also synthesized via epichlorohydrin. Glycidyl esters based
on tetra or hexahydrophthalic acid have low viscosities and after cure, e.g.
with cyclic anhydrides, form materials with improved light and arc
resistance.
o
II
c
o~
CC o~
c
I
o
Hexahydrophthalic acid-diglycidylester
Especially suitable for the manufacture of carbon fibre reinforced
materials are triglycidyl-p-aminophenol and N,N,N',N'-tetraglycidyl-4,4'-
diaminodiphenylmethane. 1,3,5-triglycidylisocyanurate is another very im-
portant N-glycidyl-derivative, finding use mainly in powder coatings.
o o 0
o ~~ ~~~~
~O~N'---D ~N~CH~N'---D
o o 0
Triglycidyl-p-aminophenol N,N,N',N'-Tetraglycidyl-4,4'-diaminodiphenylmethane
92 I LI________E_P_O_XI_D_E__R_E_SI_N_S_A_ND
___T_H_E_I_R_F_O_R_M_U_L_A_T_I_O_N________~
1,3,5-Triglycidylisocyanurate
Further materials can usually be prepared by carrying out an epoxidi-
zation reaction on an unsaturated compound using hydrogen peroxide or
peracetic acid:
\ /
C=C
/ \
Products synthesized in this way contain no hydrolysable chlorine and have
very low ash content, two advantages for their use in electronic applications.
They are also available in a variety of forms, but can essentially be divided
into those which contain a ring structure as well as an epoxide group
(cycloaliphatic) and those which have a linear structure attached to the
epoxide group (acyclic aliphatic).
The cycloaliphatic resins are of paler colour than the standard diglycidyl-
ether resins, they show better light stability and often have lower viscosities.
Typical examples are resins based on two cyclohexeneoxide rings connected
by ester or acetal bridges.
3,4-EpoxycycJohexylmethyl- 2-(3,4-EpoxycycJohexyl)-5,5-spiro-
3' ,4'-epoxycycJohexane-carboxylate (3' ,4'-epoxy)-cyclohexane-m-dioxane
These resins are preferably combined with anhydride curing agents or, if a
very quick cure is asked for, with cationic curing catalysts. The more
compact structure leads to a greater crosslinking density than with the
standard diglycidyl-esters. The cured cycloaliphatic resins thus show very
C_U_R_IN
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ __S________________~I
G_A_G_E_NT I 93
high glass transition temperatures, lower dissipation factors, good arc and
arc tracking resistance as well as high heat stability. They are also well suited
for the manufacture of electrical insulators. On the other hand they are
somewhat brittle. Proper formulation with flexibiIizing or toughening agents
leads to high performance cycloaliphatic epoxide systems for electrical
apparatus which provide a good balance of high heat distortion temper-
ature, thermal shock resistance and arc resistance and high temperature
electrical properties. In both filled and unfilled forms they are also suited to
the encapsulation of coils, transformers and other electrical components.
The acyclic aliphatic resins differ from the cyclic aliphatic epoxy resins in
that they do not possess ring structures. They can be subdivided into two
groups: the epoxidized diene polymers and the epoxidized oils. The
epoxidized diene resins are generally more viscous than the diglycidyl
ethers, except when volatile diluents are used. They are also less reactive
with amines than diglycidyl ethers because of the absence of the ether
linkage, although they have similar reactivity with anhydride hardeners. The
epoxidized oils have the following general skeletal structure:
0 0 o
t: "
II I \ I \
0 0
I \
0 0 o
0 II I \ I \
The number of epoxide groups per molecule can vary, although an average
four to six groups would be usual. As with the epoxidized dienes, their
reactivity with amine curing agents is also low because of steric reasons and
the absence of an ether linkage. The epoxidized oils are usually modified
before use by the introduction of hydroxyl groups into the molecule in order
to reduce the spacing between potential crosslinking sites. By introducing
these hydroxyl groups, the modified material can be made to exhibit
increased toughness and rigidity.
The special properties that have allowed such widespread usage of epoxide
resins are only obtained when the resins are properly cured. The final cured
characteristics of an epoxide resin are largely dependent upon the curing
system that is used to convert the resin into a thermoset product. This
reaction is brought about by a compound generally known as a hardener,
which causes the epoxide resin to undergo either an addition or a catalytic
94 II
L-_ _ _ _ E_PO_X_ID_E_R_E_SIN_S_A_ND
__TH_E_IR_F_O_R_M_V_L_A_T_I_O_N_ _ _------'
~NH2
H2N
Ethylenediamine (EDA) Diethylenetriamine (DETA)
lN~NH
~ 2
Table 4.3 Properties of commonly used polyamines and optimized amounts required
to cure epoxides
(CH~NH....,/""-NH-......,.,NH2
o
~ (CH 2);CH 3
Polyetber-polyamine Polyaminoamide
Another group of curing agents based upon the polyamine concept is the
cycloaliphatic amines. These have properties somewhere between those of
the aliphatic amines and the aromatic amines. Typical examples are
isophorone-diamine, 3,3'-dimethyl-4,4'-diaminodicyclohexylmethane, 1,2-
diaminocyclohexane, aminoethyl piperazine and 1,8-diamino-p-menthane.
__S__________________~I
C_U_R_I_N_G_A_G_E_NT
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ I 97
Xylylene diamine is closely related in reactivity and properties of the cured
materials.
£: I NH2
m-Xylylenediamine
HN
1\
~
N~
NH2
Aminoetbyl-piprazine
~H'
NH2
1,8-Diamino-p-mentbane
These materials generally give higher glass transition temperatures than the
aliphatic amines but have poorer thermal stabilities than the aromatic
amines.
2,2'-Dietbyl-4,4'-diaminodiphenylmethane
DADM, for example, is a solid, melting at around 90o e. It contains four
active hydrogens and is usually used in concentrations of 25-30 parts per
hundred of resin. Mixing is achieved by heating part of the resin to just
above 90 0 e and adding the molten amine. After thorough mixing, the batch
is cooled by the addition of more resin and mixing is continued. DADM
gives relatively high temperature strength and enhanced electrical proper-
ties, such as dielectric constant and dissipation factors.
DADS has an advantage over other aromatic amines in that it gives very
high glass transition temperatures and its strength properties are maintained
to higher temperatures. It is a solid with a melting point around 175°e and
is generally used in the range of 40 phr with DGEBA type epoxides
(although 20-30 phr may be adequate if an accelerator is utilized). In order
to mix the resin and hardener, the resin is heated to above nooe and the
DADS is added slowly with stirring and maintenance of temperature until
all the hardener is dissolved. If an accelerator is to be used, the temperature
should be lowered to 1000e and one part by weight of boron trifluoride-
monoethylamine blended in. This type of system would be suitable for
electrical casting requiring extreme service durability and elevated temper-
ature strength.
As well as the principal aromatic amines reported above, various other
aromatic amines find use in certain applications. Typical examples of these
compounds are:
NH2
2,4-Diaminotoluene 2,6-Diaminopyridine 2-Aminobenzylamine
__S__________________~I
C_U_R_I_N_G_A_G_E_NT
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ I 99
4.3.5 Anhydrides
Acid anhydrides are the second most commonly used curing agents, and
they are characterized by the anhydride group, which is derived from a
diacid by dehydration.
aGo
o
OH
OH
o o
They represent a low cost group of hardeners which provide long pot lives,
have excellent heat stability and are especially suited to electrical applica-
tions. Anhydrides also offer advantages in handling, since, unlike the
amines, they are not skin-sensitizing agents. Also, unlike the aliphatic amine
cured epoxide systems, the acid anhydride cured epoxides usually require
long, high temperature cure schedules before optimum properties can be
achieved. This is probably the major limitation to their use, although certain
accelerators are useful in enhancing the reactivity. Acid anhydrides can,
however, provide more rapid cures than amines when used as curing agents
for the cycloaliphatic and epoxidized olefin based resins.
100 II'- -____ E_P_O_X_I_D_E_R_E_S_I_N_S_A_ND
__T_H_E_I_R_F_O_R_M_V_L_A_T_IO_N
_ _ _ _--'
The curing reactions of anhydrides with epoxide resins are quite complex
and may be dependent upon a number offactors, such as the temperature, the
type of epoxide, the amount of anhydride used and the presence and type of
accelerator. The first stage of the curing reaction is the opening of the
anhydride ring by a hydroxyl group, such as may be found in the epoxide resin.
o
~O-R
~o + R-OH
~OH
o o
There are then five further possible types of reaction that can take place but
the two most important ones are (l) the reaction of a carboxylic acid group
«
with an epoxide group:
~O-R +
o I
OO_R
OH
~OH ~ ~ O~
o o
and (2) the etherification of a hydroxyl group by an epoxide group:
I o OH I
H-C-OH + ~ H-C-OA
I I
Anhydrides react relatively slowly and, especially with cycloaliphatic resins,
it is quite often found that approximately equal amounts of the ester and
ether linkages are formed, thus leading to brittle products. Therefore, the
reaction is usually modified in commercial systems by the use of accelerators
such as tertiary amines, quaternary ammonium salts or, in the case of
cycloaliphatic resins, sodium alcoholates.
The tertiary amine accelerators are commonly used with standard resin/
anhydride systems to speed up their curing reactions. They are used in
concentrations of 0.5 to 5 phr and typical examples are:
'NyO~
I I I h
N/ fvNH
N/
I
Benzyl-dimethyl- Dimethyl- Tris- 2-Ethyl-4-methyl-
amine aminomethyl- (dimethylaminomethyl)- imidazole
phenol (DMP 10) phenol (DMP 30)
L -_ _ _ _ _ _ _ C_U_R_IN_G_A_G_E_NT_S_ _ _ _ _ _ _ -'I I 101
There are many types of anhydride available, but one of the most
important groups is that based upon anhydrides with cyclic structures.
Typical examples of anhydrides in common use are:
o 0 0
~o ~o ~o
o 0 0
Phthalic anhydride Tetrahydrophthalic Hexahydrophthalic
m.p.131-133°C anhydride anhydride
m.p. 65-68°C m.p. 35-36°C
CI
o 0
HO~
CI
o~o ~
I 0
CI
0
o 0 0
CI
Pyromellitic Trimellitic anhydride Chlorendic anhydride
dianhydride mp.168°C mp.239°C
m.p. 283-286°C
similar cure times to phthalic anhydride, with slightly higher glass transition
temperatures. Mixtures of isomeric methyl-tetrahydrophthalic-anhydrides,
as well as methyl-hexahydrophthalic anhydride, have the advantage of being
liquid at room temperature and easily mixable with the resin.
Where higher glass transition temperatures are required with diglycidyl-
ether of bisphenol A based resins, anhydrides with higher functionality have
been employed, such as pyromellitic dianhydride, which contains two
anhydride groups, and trimellitic anhydride which has an additional
carboxylic acid group. Chlorendic anhydride is a member of the family
that, because of their halogen content, imparts flame retardancy upon a
cured resin. Systems cured with chlorendic anhydride have excellent
electrical and mechanical properties although it does have a somewhat
limited thermal stability and a tendency to be hygroscopic.
Similarly to the aromatic amines, eutectic mixtures of two anhydrides are
sometimes used to overcome difficult mixing procedures such as the
requirement for high temperature mixing. One such example is a 70/30
mixture of chlorendic anhydride and hexahydrophthalic anhydride which
imparts nearly the same properties on the cured system as pure chlorendic
anhydride yet allows much simpler mixing to be achieved.
4.3.6 Novolacs
Novolacs mentioned above as starting materials for multifunctional epoxy
resins are also very useful curing agents. Curing needs temperatures in the
range of 160 to 200°C even when accelerators such as tertiary amines or
imidazole-derivatives are added. The resulting crosslinked products com-
bine high Tg with excellent long-term thermal and chemical resistance as
well as relatively good mechanical properties. Novolacs find use in epoxy
moulding compounds as well as in coatings.
[~ 0 ~R [E-{O~~~ RJ
E] ~
1+ +)
o
.. X- _ X-
~R
~
E+ X - =Lewis acid
X =non-nucleophilic anion
+HX
(V)
104 I LI________E_PO__XI_D_E__R_E_S_IN_S_A_ND
___T_H_E_I_R_F_O_RM~U_L_A_T_I_O_N________~
Boron trifluoride is one of the more important Lewis acids currently used.
However, as it is a corrosive gas which is extremely reactive at room
temperature, it is usually complexed with an amine such as monoethylamine
or benzylamine:
FI H
I
FI H
I
F-S:N~
I I \ F-S:Nb
I I
F H F H -
~ h
BF3-ethylamine complex BF3-benzylamine complex
Qx-
R'VS'Q R"'
Triarylsulfonlum salt Diaryliodonium salt lron-arene complex
(metallocene)
4.4 EXTENDERS
The term extender is very broad and covers a wide range of materials but it
generally applies to inorganic materials that are added to a polymer in
relatively high concentrations. One of the major reasons for using extenders
in an epoxide resin system is to reduce the cost of the final product and this
has become particularly important since the onset of crude oil price
increases in the early 1970s. By using extenders in a resin system, not only
can the price be controlled but their use also allows the properties of the
finished product to be tailored for a particular application. In the electronics
"--_ _ _ _ _ _ _ _
EX_T_E_ND_ER_S_ _ _ _ _ _ _ _ --'I I 107
industry, the use of extenders with epoxides is usually more important from
the point of view of tailoring properties than of cost reduction.
Extenders can be used to control curing exotherms and moisture
permeability, to impart increased strength or to tailor properties such as
hardness, heat distortion, thermal expansion coefficients and thermal
conductivity. More novel fillers can be used to impart electrical conductivity
and magnetic properties or to reduce the compound's density by the
incorporation of small hollow glass spheres. The selection of the correct
filler is of the utmost importance and all the critical desired properties of the
cured compound must be taken into account before a particular filler can be
decided upon.
One of the first problems that is encountered is the selection of an
extender with the correct particle size, shape and particle size distribution.
Since the properties a filler imparts upon a formulated epoxide resin system
may well be very dependent upon surface area, the particle size distribution
of a given type of filler may be critical. For a moulding compound
formulation, particle size will control the flash and bleed characteristics and,
indeed, the mouldability. Another important factor that has to be
acknowledged is the possibility of a reaction occurring between the extender
and the resin system, and since reactivity will be influenced by surface area,
this should be borne in mind when selecting a particular particle size
material.
Perhaps some of the most important considerations of extenders for
epoxides to be used in electronic applications are the changes that they
impart to the electrical properties of the formulated product. If good
insulative properties are required of the compound, the condition and type
of extender selected will be important since ions, salts, moisture or water
soluble impurities may be adsorbed on the particle's surfaces, all of which
can serve to increase the conductivity of the compound. Conversely, high
conductivity may be required, as in the case of conductive die attach
materials where the extender may be a finely divided metal such as silver or
gold. Care also has to be taken when selecting organic extenders, since high
voltages or areas of high current density can cause tracking or carbon
bridges to form. The use of extenders can thus be seen to play an important
part in determining the electrical properties of a finished product and
extreme care should be exercised in their selection.
Extenders are often used to alter the thermal expansion coefficient of a
material and this is particularly true with electronic applications, where the
principal substrate is usually silicon. If a silicon die is encapsulated in an
epoxide based moulding compound, reliability problems can occur because
of the thermal expansion coefficient mismatch between the silicon and the
encapsulant. During any temperature changes, the differential expansion of
the two materials can cause stresses to build up in the package which lead to
wire bond fatigue and possibly even cracking of the chip. The same applies
108 II~________E_P_O_X_I_D_E_R_E_S_I_N_S_A_ND
___T_H_E_I_R_F_O_R_M
__ ________~
U_L_A_T_IO_N
Advantages Disadvantages
with die attach materials, particularly when the adhesive used is a fairly
rigid, highly crosslinked type.
It is generally true that nearly all commercial epoxide resin formulations
have extenders incorporated in their formulation, the final selection being
determined by the properties to be altered, the application and the other
components of the system. Some of the properties affected by the addition of
extenders are listed in Table 4.4. (For some specialized applications, certain
of the listed disadvantages may in fact be advantages.)
As can be seen, for the majority of applications the advantages of using
extenders far outweigh the disadvantages and it is hardly surprising that the
inorganic extenders find such widespread usage. Several of the more
important extenders will now be discussed in more detail.
product than the same weight of a denser material. The densities of various
extenders are shown in Table 4.5.
Another commonly used form of silica is quartz, which has advantages
over fused silica for certain applications due to its higher thermal
conductivity. Thus, where enhanced thermal conductivity is required, as
in the encapsulation of power devices or integrated circuits (lCs) with high
circuit density, quartz would be a preferred filler.
Another popular form of silica that is particularly useful with liquid
systems is fumed silica and this is an amorphous form that has a very large
surface area. It can be used in liquid epoxide systems to provide thixotropic
and anti-settling characteristics and in moulding compounds to control flash
and bleed. The thickening effect of fumed silica results from the interaction
of silanol (SiOH) groups on adjoining particles, as well as between the
silanol groups and the reactive groups of the hardener (such as amine
groups). Fumed silica is also useful in elastomeric materials, such as the
silicone rubbers, as a reinforcing agent to improve mechanical properties. It
has the distinct advantage of allowing the formulation of silicones that are
translucent or even transparent.
4.4.2.1 Talc
Talc is the name given to hydrated magnesium silicate, which has the
theoretical formula 3MgO.4Si02 ·H20 and it occurs in several ores from
which talc products can be made. The form of most use in plastic
formulations is finely ground talc, which consists of thin platelets. It is
considered to be a reinforcing filler in many compound applications,
although it can also be described as an extender because of its low cost.
Talc possesses a good degree of chemical resistance, being inert to most
acids and it is also a good insulator. This latter property makes it
particularly useful in the manufacture of high frequency electrical insulators.
It has good ~hermal stability up to 900°C, which is considerably higher than
any temperature likely to be encountered in an epoxide resin application.
110 I ~I________E_PO__X_ID_E__R_E_SI_N_S_A_ND __F_O_R_M_V_L_A_T_I_O_N______~
___T_H_E_IR
The use of talc in epoxide formulations leads to higher stiffness and
improved creep resistance at both room and elevated temperatures. In order
to obtain the full benefits from using talc, it is very important that the
correct compounding techniques are used to incorporate high loading into
the resin system.
The machinability of epoxides varies considerably from filler to filler, and,
if ease of machinability is required, the use of talc may provide this property.
Talc is also of use in epoxide resin formulations when the thermal expansion
coefficient has to be lowered by a large amount, since loadings up to 80 parts
per hundred of resin can be tolerated. Loading epoxides with talc has also
been shown to increase their lap shear strengths when they are used as
adhesives.
4.4.2.2 Wollastonite
Wollastonite is a naturally occurring calcium-silicate that is available in
different particle size distributions with or without surface treatment. Due to
its lower coefficient of thermal expansion (CTE ca. 7 ppm/K) compared to
ca. 14 for (crystalline) silica the CTEs of metal substrates can be better
matched by formulations containing Wollastonite. In terms of lowering
CTE, fused silica, with a CTE near to zero is still more efficient. Beside the
effect or reducing stresses from CTE mismatches Wollastonite can also
improve the fracture mechanical properties of epoxy formulations due to its
particular needle shaped form. This property profile leads to materials that
are very resistant to thermal shocks. Wollastonite shows other useful
properties, e.g. a low abrasivity. Disadvantages with this filler are its lower
thermal conductivity compared to crystalline silica and the fact that
additional technical efforts are necessary to manufacture really homoge-
neous dispersions.
4.4.2.3 Clay
Another extender occasionally used in epoxide formulations is clay. Clay is
a hydrated alumino silicate mineral, also known as kaolin, which is available
in a calcined, anhydrous form. The hydrated form is non-abrasive,
chemically inert and has a large surface area. It is naturally acidic and
this must be taken into account when formulating for applications where
corrosion of metallic conductors may be encountered. Hydrated clay
disperses readily in most formulations, particularly when dispersants or
surfactants are used, and is utilized in epoxide transfer moulding com-
pounds to allow control of Bow.
The calcined clays are considerably harder than their hydrated counter-
parts but they impart better electrical properties to both thermosetting
compounds and thermoplastics. Clay is also of particular use in fibre-filled
L -___________________ __E_R_S__________________~I
E_X_T_E_N_D I 111
Br Br
HO OH
Br Br
Tetrabromobisphenol A
Bromine 26--30
Chlorine 13-15
Tetrachlorobisphenol A 39
Tetrabromobisphenol A 59
Even though there are literally hundreds of different epoxide resins and
hardeners commercially available, giving countless formulations with a
diversity of cured properties, further modifications of a material are
L -_ _ _ _ _ _O_T_H_E_R_U_S_E_F_U_L_C_O_N_STIT
__ U_ENT
__S_ _ _ _ _ _ --'I I 117
Q 6 «:==: o
m-Xylene Toluene Dibutyl-phthalate
When used in the range of 15-20 phI' with DGEBA-based resins, four-
fold reductions in viscosity can be achieved, although, by nature of its
chemical structure, it may also bring about a reduction in crosslinking
density. Other materials that are useful under certain circumstances as non-
reactive diluents are the low molecular weight styrenes and the bisphenol-
118 I LI________E_PD__X_ID_E__R_E_SUN ___T_H_E_I_R_F_O_R_M_V_L_A_T_I_O_N______~
__S_A_ND
~o~
o
o
~o~o~
o
Butanediol diglycidylether Neopentylglycol diglycidylether
n
The value of n would be in the range 2-7.
The glycidyl derivatives of fatty acids are also good flexibilizing resins and
such materials generally have viscosities of between 400 and 600 cP at room
120 I ~I________E_PO__XI_D__E_R_E_S_UN_S_A_ND
___T_H_E_I_R_F_O_R_M_U_L_A_T_I_O_N______~
temperature. They can be used with most epoxide curing agents, but their
incorporation into a system generally leads to long pot lives and the need for
longer cure times.
Flexibilization of a formulation can also be achieved by using flexibilizing
curing agents or other additives. Examples are diester-dicarboxylic acids or
low molecular weight polyesters with acidic end groups to be used together
with anhydride curing agents. These materials again have reactive end
groups separated by flexible long chain molecular segments. Another typical
example is dodecenyl succinic anhydride which has the structure:
o
Dodecenyl succinic-anhydride
In certain applications this material can impart outstanding electrical
properties along with increased flexibility. When it is used in a 50/50 adduct
with hexahydrophthalic anhydride it gives a useful eutectic mixture with a
melting point of rvO°C. One other group of flexibilizing hardeners also
worthy of mention is the polysulphides. They have a polymeric structure
which is terminated at each end with a thiol group (-SH). Although the thiol
group can react slowly with the epoxide resins, they are generally used in
conjunction with catalysts such as DMP 10 and DMP 30. The polysulphides
have found considerable usage in the formulation of epoxide systems for the
encapsulation and potting of delicate components.
The third possible way of improving toughness of epoxy formulations,
besides adding plasticizers or flexibilizing agents, consists of dispersing soft
rubbery inclusions in the brittle epoxy matrix. A system family that has
received much attention are the carboxyl-terminated butadiene acrylonitrile
(CTBN) copolymers, that react via end groups with the epoxide and form a
second phase during curing. This second phase acts as a crack stopper. In
contrast to the other two mentioned methods of increasing toughness, the
heat distortion temperature of such two-phase systems is practically not
reduced. The morphology and fracture/mechanical properties of such two-
phase systems formed during curing depends, of course, very much on the
curing conditions. The toughness of other two and multiphase systems that
already contain pre-formed particles before curing is relatively independent
of the curing conditions. The size of such particles can be less than 1 pm.
Core shell type elastomers are typical examples of materials used to provide
such pre-formed particles.
L -_ _ _ _G_EN_E_RA_L_F_O_R_M_U_L_A_T_IO_N_C_O_N_S_I_D_E_R_A_T_IO_N_S_ _ _ ---'I I 121
Having been presented with the large number of materials that are
commonly used in the formulation of epoxide encapsulants, the reader
new to this area may be bewildered at the vast range of possibilities for the
composition of a given compound and may be wondering how one begins to
select the individual components to impart the desired properties in the
cured product. It is relatively easy to define the properties required of an
epoxide for a specific application, but it is more difficult to select ingredients
that will allow all of these desired properties to be accommodated and, quite
often, some degree of compromise is necessary. The formulator will have to
make a choice of components from a range ofliterally hundreds of available
resins, hardeners, catalysts and other ingredients, each of which is often
available in a variety of forms, in an attempt to balance their contrasting
properties. The properties of a given formulation can be conveniently
divided into those required before, during and after cure.
The main characteristics of the material required before cure that will
influence ingredient selection are:
1. The overall cost limitations for the product.
2. The required physical form of the compound. Is a liquid or a solid
required?
3. Chemical stability. What storage life is required and does the
compound have to be one or two parts?
Having taken these characteristics of the compound into account, the
properties of the compound during the curing reaction now have to be
considered. The major factors here are:
1. What cure schedule is possible? Does the compound have to cure
within a specific time? Is there a maximum permissible curing
temperature?
2. Are the components of the formulation chemically compatible with the
substrate being encapsulated and coated?
3. How much cure shrinkage can be tolerated?
4. What viscosity is required for both liquid systems and moulding
compounds?
5. What curing exotherm is tolerable under the determined curing
conditions?
Finally, we come to the properties of the cured product. Many of these
properties will obviously be influenced by or influence the properties
required before and during curing of the compound. The more important
characteristics commonly considered when formulating epoxides for elec-
tronic applications are:
122 I LI________E_P_O_X_I_D_E_R__ES_I_N_S_A_ND
___T_H_E_I_R_F_O_R_M ________~
__U_L_A_T_IO_N
1. Colour and optical properties; an epoxide may need to be a specified
colour or it may need good transmissive properties at a given
wavelength for electro-optic applications.
2. Density; there may be limitations on the weight of a product,
particularly if it is intended for aerospace use.
3. Thermal expansion coefficient; the thermal expansion coefficient mis-
match between the epoxide and a substrate may need to be minimized.
4. Adhesion; good adhesive properties may be required, as in die attach
applications.
5. Thermal conductivity; good thermal conductivity may be required for
heat dissipation from power devices and integrated circuits such as the
latest microprocessors.
6. Chemical and solvent resistance; good chemical and solvent resistance
may be required for epoxides exposed to harsh chemical environments.
7. High purity; this is required for epoxides in contact with delicate
microcircuits where impurities may cause corrosion of conductor
metallization.
8. Low gas, moisture and vapour permeability; again, this is required for
epoxides in contact with delicate microcircuits, where corrosion or
parametric shifts may occur.
9. Internal stress; low internal stress may be important with epoxides in
contact with large microcircuits, microcircuits with multi-layer metal-
lization and large I/O count packages.
10. Low alpha particle emission; ingredients may need to be pure with
respect to actinide elements, to avoid alpha particle emission that can
cause soft errors in dynamic random access memories.
11. Thermal stability; stability at high temperatures may be required for an
epoxide operating in a harsh environment where elevated temperatures
are experienced or with high power devices.
12. Tensile and flexural strength, hardness and glass transition tempera-
tures (TJ are all important properties where good mechanical perfor-
mance and physical integrity is required.
13. Flammability; an epoxide may need to be flame retarded to meet certain
customer requirements. Examples would be to achieve Underwriters,
Laboratory listings (UL), such as UL94 1/16 in VO.
14. Electrical properties such as volume resistivity, dielectric constant,
dielectric strength, arc resistance, dissipation factor and loss factor are
all important where the epoxide is in contact with high voltages or
where there is intimate contact with a microcircuit surface. Certain of
these electrical properties are particularly important with high-speed
devices operating in the frequency range above 100 MHz, where low
values are generally desirable.
L -_ _ _R_E_C_y_C_L_IN_G_O_F_CU_R_E_D_E_P_O_X_y_-B_A_S_E_D_M_A_T_E_R_IA_L_S
_ _ _--.JI I 123
Since the oil crisis in the 1970s, the question of recycling has gained
importance with the experience of increasing waste streams and with space
for landfills becoming expensive and more and more scarce.
Epoxies in structural applications with their long usable life and their
excellent chemical resistance offer ecological advantages and play, with less
than about 0.5%, only a minor role in the generation of the plastic waste
streams; nevertheless they are also included in the emerging regulations and
methods for recycling them have to be found.
Particle recycling, e.g. the use of milled cured epoxy-based materials as
additional filler in new formulated epoxy products, is already used in the
processing of epoxy moulding compounds. Up to about 15% of milled
runners can be added to new material. With a proper choice of particle size
distribution (low fines), up to about 30% of the mineral filler in quartz-filled
anhydride cured epoxy-casting formulations can be replaced by recycled
epoxy-based insulating materials etc. without affecting the properties of the
cured materials too much. The use of milled recycled epoxy-based printed
circuit boards as reinforcing filler in thermoplastic materials as well as in
concrete or road building materials is under evaluation.
Feedstock recycling as a way to split plastic waste back to basic chemicals
is also applicable to cured epoxy-based products. Pyrolysis results in a gas
fraction with an energy content similar to that of natural gas and a liquid
fraction consisting mainly of aromatic compounds. Hydrolysis with alkali at
elevated temperature is very easily achieved with anhydride-cured epoxies
and leads to phenolic ethers with OH end-groups and acid-salts. Hydrolysis
of dicyanodiamide-cured epoxies, e.g. from printed circuit boards, needs
more severe conditions. In the case of bromine-containing resins, the
bromine is completely split off and aromatic compounds similar to those
found after pyrolysis are formed. Catalytic hydrogenation of e.g. anhydride-
cured epoxy resin at elevated temperatures under pressure results in a
fraction of phenolic and other aromatic compounds, neopentylglycol-
derivatives and lactones.
Energetic recycling (incineration in power plants) is no problem with
epoxy-based cured materials that contain no critical by-products. Inciner-
ation of chlorine- or bromine-containing materials requires modem equip-
ment with sufficient corrosion resistance and perfect cleaning of the smoke
gases.
124 I LI________E_P_O_X_ID_E__R_E_SfIN
__S_A_ND __F_O_R_M_U_L_A_T_I_O_N________~
___T_H_E_IR
Use of waste plastics instead of coal in blast-furnaces or similar
equipment is a method lying between pyrolysis and incineration. Use is
already made of this technique in the case of printed circuit boards. The
concept behind this is to recover the copper and, if present, the noble metals
contained in the electronic scrap.
Surveys show that the methods used for the recycling of the commodity
plastics are, with proper choice of the reaction conditions, also applicable to
cured epoxies.
Epoxy resins and curing agents are reactive chemicals and have to be
handled with care. Toxicological properties of the numerous products used
in epoxy resin formulations differ very much. Bisphenol-based conventional
epoxy resins, for example, have a low order of acute toxicity and are mild to
moderate irritants; low molecular weight resins are skin sensitizers. Skin
irritation and sensitizing effects are stronger, when certain reactive diluents
are added.
Amongst curing agents, the aliphatic amines are strong bases and thus
corrosive and volatile materials which can cause tissue damage (bums) to
skin, eyes and mucous membranes. Several aromatic amines show evidence
of carcinogenicity in animals and exposure should be strictly avoided.
Anhydrides are eye, skin and respiratory irritants. Solvents may cause fire
hazards and even mineral or reinforcing fillers may be dangerous to health.
It is therefore very important in every case to consult the safety data sheet
for the individual product to be used. It is also important to make a person-
and workplace-oriented risk analysis followed by intensive instruction and
training before work with epoxy-based products starts. To minimize
exposure, the following protective equipment should be provided and used:
rubber or plastic gloves and sleeves, overalls and aprons (to be washed and
cleaned regularly), goggles or face shields, special protective skin cream to
be used before work starts.
After work, and in the case of skin contamination, but also before eating
or using the we etc., operatives should wash with warm water and mild
alkali-free soap and dry themselves with disposable paper towels. Never use
solvents to cleanse contaminated skin.
The working area should be separated from other areas in the factory and
be provided with general and local exhaust systems to remove vapours,
mists etc. where they are generated. Rooms for eating, taking a break or
changing clothes should be separated from the working area.
________________~I
F_U_R_T_H_E_R_R_E_A_D_I_N_G
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ I 125
4.10 SUMMARY
This chapter has described the principal ingredients more commonly used in
epoxide formulations and illustrated the vast selection of choices available
to the prospective compound formulator. The important factors that may
need to be taken into account for a given application have also been
reviewed. For further information and assistance with formulation, the
reader is referred to the list of further reading at the end of this chapter.
4.11 REFERENCES
Lee, H. and Neville, K., Handbook of Epoxy Resins, McGraw-Hill Inc. New York,
1967, repro 1982.
Lohse F. and Seiz, W., Houhen Weyl Methoden der Org. Chemie, Vol. E 20, Part 3,
Georg Thieme Verlag, Stuttgart, New York, 1987.
May, C. A. and Tanaka, Y., Epoxy Resin Technology, 2nd edn., Marcel Decker, New
York,1988.
McAdams, W. and Gannon, J. A., Encyclopedia of Polymer Sci., Wiley Interscience
New York, 6, 322-82, 1986.
Potter, W. G., Epoxide Resins, Butterworth and Co. (Publishers) Ltd., London,
1970.
Potter, W. G., Uses of Epoxy Resins, Newness Butterworth, Sevenoaks, 1975.
4.12.2 Reviews
Kishore, K., Joseph, M. and Vijayshree, M. N., Epoxide curing, a literature survey.
J. Sci. Ind. Res. 47(11), 662-74, 1988.
Lohse, F. and Zweifel, H., Photocrosslinking of epoxy resins. Adv. Polym. Sci., 78,
62-81, 1986.
Scola, D. A., Novel curing agents for epoxy resins. Dev. Reinf Plast., 4, 165, 1984.
Schultz, N., Potting and encapsulating electronic components. Adhesives Age, 19,
(July 1983).
Snyder, P. M., Developments in semiconductor device encapsulation, Microelec-
tronic Manufacturing and Testing, 41, 1983.
Spitsbergen, c., Epoxy resins and electronics reliability. Int. SAMPE Electron.
Conference, 7, 517-25, 1994.
Stadtmiiller, R., Use of adhesives in electronic industry. Adhiision, 28 (6), 22-5, 1984.
Suzuki, H., Sato, M., Muroi, T. and Watanabe, Y., New moulding compounds for
electronic devices. Soc. Plast. Eng. Tech. Paper 19, 6, 1973.
Wong, C. P., Polymers in encapsulation. Adv. Polym. Sci., 84, 63-83, 1988.
Recent Developments in the 5
Encapsulation of
Semiconductors by Transfer
Moulding
MARTIN GOOSEY· and MIKE PLANTt
·Shipley Europe Ltd Coventry, England and tKemtron
International South Woodham Ferrers, England
5.1 INTRODUCTION
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
132 IENCAPSULATION OF SEMICONDUCTORS BY TRANSFER MOULDING I
pressure, is transferred into the cavities where it solidifies giving the familiar
plastic encapsulated device. Typical plastic encapsulated parts are shown in
Fig. 5.1.
Whilst the early materials and machinery were fairly basic, considerable
progress has been made in both equipment design and compound
formulation. The main areas where improvements were required were in
device reliability, production rates and material wastage. The major
achievements of the equipment manufacturers in both mould and press
design have allowed the number of parts moulded per cycle to increase
dramatically, so that in some cases 400-500 parts can be finished per shot.
Much larger active devices with higher pin count packages can now be
moulded, while the amount of material wasted in the runner systems has
been reduced. A further breakthrough has been the introduction of
automatic moulding equipment, offering advantages in terms of labour
costs and quality (see section 5.2.7).
The most commonly used materials for transfer moulding both active and
passive components are epoxide based. The longer flowing epoxide
moulding compounds for semiconductors are now largely unrivalled, the
only other materials occasionally used in these applications being silicones,
silicone-epoxides and more rarely, polyimides and polyphenylene sulphide.
(In 1995 about 97% of the total consumption of semiconductor encapsulants
was epoxide resin based.) The main reason for the popularity of epoxides in
this field is that they possess advantages such as:
I. Excellent compatibility with other components of the formulation, e.g.
good bonding to fillers, particularly when coupling agents are used.
2. Dimensional stability in the cured product.
3. Good chemical and solvent resistances due to high crosslinking density.
Transfer moulding was the term chosen by L. E. Shaw, the originator of the
process, 'to describe the method and apparatus for moulding thermosetting
materials, whereby the material is subjected to heat and pressure and then
forced into a closed mould cavity by this same pressure and held there under
additional heat and pressure until cure is complete'. Transfer moulding
equipment for electronic components consists of a hydraulic press equipped
with platens, one of which contains a chamber known as a pot, in which the
moulding powder is placed and liquefied by a combination of pressure and
heat. A piston or ram transfers the molten material into cavities of the
mould via a series of channels known as runners. Figure S.2a shows a
modem FICO (Holland) fully automatic moulding machine (with close-up
detail in Fig. S.2b), making an interesting comparison with the conventional
press shown in the first edition of this book (many of which continue to give
excellent and reliable service around the world).
The moulding compound can be introduced into the pot as a ground
powder or more usually as a preheated preform. In this latter case the
powder is compressed to form a solid pellet of material which is
dielectrically heated to around 90°C using a high-frequency preheater.
The preheated preforms are loaded into the pot and transferred by the ram
at pressures typically around 1000 psi, although as little as 50 psi may be
sufficient for some specially formulated epoxides. For an epoxide compound
used for moulding many tens of semiconductor devices in a single shot, a
transfer time of between IS and 20 s is usually typical, although for some
compounds and moulds, operation outside these limits may be necessary.
The mould temperature would normally be around 17SoC and the parts
could be removed after a mould cycle time of I to 2 min. The mould is
usually designed so that, when it opens, ejector pins in the cavities and
runners push the moulded devices free. The moulded strips of components
will then be removed from the mould and subjected to a number of finishing
processes such as defiashing, cropping and forming ofthe lead-frames, post-
curing and marking before the final product is obtained.
Transfer moulding can also be used for the packaging of resistors, resistor
networks, capacitors, coils, reed relays and other discrete components. The
process only becomes economically feasible, however, for mass production
L -_ _T_RA_N_S_FE_R_M_O_V_L_D_I_N_G_EQ_V_IP_M_E_NT
__ C_O_N_S_1D_E_RA_T_I_O_N_S_ _ - 11 I 135
Fig. S.2b. Detail of dual strip mould for matrix frames (courtesy FICO).
operations since the initial cost of the press, moulds, preheater and
preformer can only be amortized over the production of many hundreds of
thousands of parts.
136 IENCAPSULATION OF SEMICONDUCTORS BY TRANSFER MOULDING I
In order to design a mould for the transfer moulding of semiconductors, it
is essential to define the basic parameters and processing variables that are
likely to be important in order to gain a thorough understanding of what is
actually required before the mould can be manufactured. A mould will
generally be developed for a specific device and encapsulant combination.
Some important basic factors which need to be taken into account before a
mould design can be initiated are:
1. Shrinkage of the moulding compound during cure.
2. Mould release behaviour of the compound and draft angles needed to
enhance component release.
3. Siting of gates to optimize filling of cavities and to give good cosmetic
appearance.
4. Degree of venting required to prevent air and volatile entrapment in the
moulding. .
5. Whether or not ejector pins will be required to release the moulded parts
from the mould and, if so, where they must be located, taking into
account the rigidity and hot hardness of the cured epoxide and the
sensitivity of the encapsulated devices to mechanical damage during
ejection.
6. Whether the surface of the moulded components needs to be glossy or
matte and whether conventional marking or laser marking is to be used.
5.2.2 Venting
All transfer moulds have to be vented in order that air and volatiles in the
mould can escape, preventing the build-up of back pressure and allowing
complete void-free filling. The size and location of the vents will depend
upon the size and design of the moulding and the position of any pins or
inserts. A vent is a small groove machined from one face of the mould that
allows the expulsion of unwanted gaseous products from the cavity, but
L -_ _ T_RA_N_S_F_E_R_M_O_V_L_D_I_N_G_E_Q_V_I_P_M_E_NT
__ C_O_N_S_ID_E_RA
__ T_IO_N_S_ _-----'I I 137
which is too narrow to allow the passage of any of the moulding compound.
For the moulding of dual in-line packages, a single vent is usually machined
at the opposite end of the package to the gate, as this is normally the area
which fills last with compound. Dimensions for a vent of this type would be
between 0.0005 and 0.0015 in deep and 0.125 to 0.250 in wide, although it is
normal to begin with a minimum gate dimension and enlarge as necessary.
Fig. 5.3. A 320 cavity MTI encapsulation mould for 14/16 pin dual in-line ICs.
Fig. 5.4. A close-up view of the 320 cavity mould illustrating the position of gates,
knock-out pins, runners and runner turn-offs.
at the opposite side ready moulded, trimmed and formed. The lead-frames
are placed in the mould cavities automatically by a robot loading arm fed
from a cassette or magazine. The moulds are of a multiplunger design, with
140 IENCAPSULATION OF SEMICONDUCTORS BY TRANSFER MOULDING I
each plunger having runners to typically only four devices. Each plunger is
automatically fed with mini preforms or tablets or, with some equipment, a
specially ground powder. After loading the preforms and transfer moulding,
the devices are automatically removed from the mould and transferred by a
conveyor system to an automated trim and form operation, which is part of
the equipment, or to storage cassettes ready for trimming and forming on a
separate machine. After each mould cycle a cleaning operation can be
specified and this is performed by revolving brushes.
The principal advantage claimed for these automated presses, apart from
the fact that they can be largely unsupervised when operational, is the
increase in throughput that is achievable. Conventional moulding is limited
not only by the cure time of the moulding compound but also by the speed
at which the operator can load and unload the machine. The main
automated moulding machine manufacturers are Towa, Yamada and
Daichi Seiko of Japan and FICO of Holland. A number of American and
other European manufacturers are also pursuing the manufacture of
automated moulding systems.
t
....e-
S
.re
;>
20 30 40
Spiral F10wrmches
Fig. 5.5. Variation of spiral flow with viscosity and filler content.
flow requirement thus has a direct bearing upon the filler content of most
compounds, giving a finite range. Another factor influenced by the filler
content of a compound is the cure shrinkage, with high filler contents giving
lower values. Other aspects of the filler composition can also add further
limitations to the development of a compound. For example, the filler
particle size and particle size distribution affect the flash and resin bleed
characteristics of the finished compound. Oversized filler particles can cause
gate blockage, resulting in incomplete filling of cavities. This is particularly
important when moulding small-sized packages such as chip carriers and
SOT23 devices, where gate depths may be thinner than for conventional
transistor and dual in-line packaged integrated circuit moulds.
45
30
15
60 70 80
Filler ContentJ%
Fig. 5.6. Variation of thennal expansion coefficient with filler content.
30
10
60 70 80
Filler Contentl%
Fig. 5.7. Variation of thermal conductivity with filler content.
20 20
10
fi
~
1
o L -____________________________~
o
o 0.5 1.0
Release Concentrationl%
Fig. 5.S. The relationship between mould release concentration, time to stain and
release force.
....w
~
J
6
4
120 150 180
Glass Transition TemperaturerC
Fig. 5.9. Variation of internal stress with T g •
170
oU
~:I
~
100
<II
S'
~ 150
1:1
Q
;e
~
E-4
~
6
130
0 12 24
Post Cure/hours
Fig. 5.10. Variation of Tg with cure time and temperature.
148 IENCAPSULATION OF SEMICONDUCTORS BY TRANSFER MOULDING I
time and temperature to different maximum levels depending upon the post-
cure temperature.
The choice of post-cure conditions has a direct bearing upon the degree of
crosslink density achieved and, in effect, on the overall properties of the
cured compound. The zero point on Fig. 5.10 represents the as-moulded Tg
before post-curing and thus variations in the permissible moulding
conditions play a large part in determining the catalyst type and level to
be selected for a given formulation. Whilst it would seem obvious to use
additional catalyst where a curing schedule is limited, it should be borne in
mind that device reliability usually decreases somewhat with increasing
catalyst concentration. The use of a more reactive catalyst may also lead to
problems with reduced shelf stability.
l00rr=====================~-'
~
1:1
:8
j SO 25"C
&!
~
3SOC
0
0 30 60
Storage Time/days
Fig. S.11. Loss of spiral flow with time at different temperatures.
TRA
L-..._ _ ___N_SF_E_R_M_O_U_L_D_IN_G_CO_M_P_O_U_ND_F_O_R_M_U_L_A_T_IO_N
___ -'I I 149
not supposed to, the size of the gap along which flow occurs being too small
to allow the filler particles and pigment to migrate. Resin bleed adheres
tenaciously to the surfaces it comes into contact with leading to problems
with lead-framed devices in which the lead-frame has to be plated or
soldered in a later operation.
Flash is generally above 0.001 in thick and the colour of the moulding
compound whereas bleed is translucent, difficult to see and below 0.001 in
thick. There are a number of causes of flash and bleed, the principal ones being:
1. High transfer pressures. Excessively high transfer pressure may result in
flash and bleed along the parting line area. The transfer pressures
recommended by the moulding compound manufacturer should be used
and it is essential that the clamping pressure of the mould is sufficient
(e.g. five times the transfer pressure).
2. Worn moulds. As with most mechanical equipment, continued use leads
to wear and, in the case of a mould, the parting line area becomes worn
allowing flashing and bleeding to occur. As the wear increases it will be
necessary to replace or recondition the mould. It may be possible, as a
palliative measure, to prolong the use of worn moulds by using moulding
compounds with shorter spiral flows and/or increased melt viscosity.
This can be accepted for use with rigid, non-fragile inserts such as TAB
and bump chip type assemblies, but with standard wire-bonded lead-
frames, wire sweep will eventually become a problem giving reduced
yields (see section 5.4.4).
COMMON MOULDING PROBLEMS AND HOW TO AVOID THEM II 153
here, but links up to it in tenns of the design of the insert. For example, in
order to balance flow between the two halves of a cavity, a bent 'flap' is
sometimes introduced immediately inside the gate of the cavity to act as a
baffle and modify the flow characteristics. In the case of uneven filling, the
purpose of a baffle is to equalize the flow and to reduce turbulence
immediately inside the gate. It can also playa part in restricting the velocity
of the material, in effect steering it away from those wires nearest to the gate
which are often most likely to be damaged by the in-rush of material.
Finally, some transfer presses have been sold in the last 10 years which
attempt to address this problem in a rather sophisticated manner, including
the technique of altering the available pressure on the ram at different points
in its travel. At this time, however, it is not clear that this sort of
sophistication is necessary. The working of the restrictor valve in the
transfer speed control is a perfect example of a simple idea that works well
and covers every eventuality, but the important thing is that the transfer
speed control is often misunderstood or taken for granted, and yet it
remains, closely behind actual mould tool temperature settings, one of the
most important controls on a hydraulic press.
Fig. 5.15. Typical cull and spiral from a spiral flow test.
but the principal three are the ram follower/spiral flow test, the Koka flow
test, and the torque rheometer (Brabender plastograph) test.
O~ ________________________- - J
o 1 2
Time/mins
Fig. 5.16. Typical torque versus time curves for a moulding compound at different
temperatures.
the torque reading will be high, but as the compound heats up and melts its
viscosity drops and the torque reading falls. However, at the elevated
temperature of the chamber the curing reaction occurs and so the viscosity,
and thus the torque reading, begins to increase until the compound gels.
From this type of experiment the minimum melt viscosity, the reactivity and
the gel time of a compound can all be evaluated. Typical results for this type
of test performed at various temperatures are shown in Fig. 5.16.
With these three tests it should be remembered that the behaviour of the
compound will be different to that occurring in a moulding operation and
some allowances will need to be made, particularly since the flow of a
compound into mould cavities will be mould design dependent.
basis a system having some major advantages but producing a far from ideal
pH can have the pH modified or buffered by the incorporation of an acidic
or alkaline additive that will have little or no other influence on the
formulation.
5.7 REFERENCE
Goosey, M. (ed.) Plastics for Electronics, Elsevier Applied Science, London, 1985.
Manzione, L. T., Plastic Packaging of Microelectronic Devices, Van Nostrand
Reinhold, New York, 1990.
Seraphim, D. P., Lasky, R. C., and Li, C. Y., Principles of Electronic Packaging,
McGraw-Hill, New York, 1989.
Sinnadurai, N., Handbook of Microelectronics Packaging and Interconnect Technol-
ogy, Electrochemical Publications, Ayr, Scotland, 1985.
F_U_R_TH
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __E_R_R_E_A_D_I_N_G________________ ~I I 169
Sinnadurai, N., Plastic packing is highly reliable, IEEE Trans. on Reliability, June
1996.
Tummala, R. R., and Rymaszewski, E. J., Microelectronics Packaging Handbook,
Van Nostrand Reinhold, New York, 1989.
Materials for Advanced 6
Encapsulation
NIGEL HACKETT
Dexter Electronic Materials, Langley, England
6.1 INTRODUCTION
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
172 I 1~_______M_A_T_E_R_lA__L_S_F_O_R_AD
___V_A_N_C_E_D_E_N
__CA
__PS_U_L_A_T_I_O_N______~
Fig. 6.1. Ball grid array substrate with encapsulated wire bonded and flip chip
attached devices.
Tall Glob Top lor COB Potting V.",lon lor COB or Modul. .
(No potting dam)
Fig. 6.2. Various methods used for the encapsulation of wire bonded and flip chip
attached chip on board (COB) devices.
hole component mounting techniques, has meant that many new challenges
have been encountered.
Whilst the traditional packaging techniques still continue to evolve, the
emphasis in more recent years has been towards the use of new
alternatives. In particular, the natural progression towards new device
packaging methods such as chip scale packaging (CSP), where the package
174 I ~I________M_A_T_E_R_I_A_L_S_F_O_R_A_D__V_A_N_C_E_D_E_N__CAP
__S_U_L_A_T_I_O_N______~
Table 6.1 The market for encapsulants and die attach materials
area is only fractionally larger than the silicon it encases (not more than 1.2
times), has demanded the design and development of new packaging and
assembly techniques. This in tum has changed both the application methods
and performance demands upon the polymeric encapsulants being used.
Interestingly, there has been a significant readoption of the so-called
'glob-top' approach (as once used in early plastic encapsulated transistor
production), albeit with much improved polymers on vastly different device
types. Currently, no one package type has emerged as the future industry
standard; indeed most forecasts still show the significant continuing use of
leaded packages, with newer packaging styles being adopted in addition.
What is very clear, however, is that the continuing evolution of package
and device types will require newer generations of encapsulants and
packaging materials to be produced. The current market size and forecast
for these encapsulants (and die) attach materials is shown in Table 6.1.
This chapter concentrates on the liquid materials that are finding wider use
as flip chip underfills and glob-top encapsulants in new electronics
packaging applications. These applications are shown schematically in
Figs 6.1 and 6.2.
Aliphatic
Glob-top type coatings have the tendency to exhibit large thermal shrinkage
stresses due to their high thermal expansion coefficients and high moduli.
With high loadings of fillers the thermal expansion coefficients can be
significantly reduced to values typically around 40 ppm;oC. For this reason
the aliphatic resins generally used in these applications are those with low
viscosities that enable high filler contents to be used. They are also normally
required to be flexible so that they can be used at low temperatures. This
type of material often tends to have relatively high levels of chlorides due to
the manufacturing techniques used in their production and care must
therefore be taken when selecting a resin, particularly if it is likely to come
into contact with the aluminium metallization on a semiconductor surface.
Overall, the biggest weakness of these resins, however, when used in
microelectronics applications is their poor water resistance.
176 II MATERIALS FOR ADVANCED ENCAPSULATION
~------------------------------------------------------~
Cycloaliphatic
The cycloaliphatic epoxides used in glob-top applications are also generally
low viscosity materials, offering much lower viscosities than are typical with
the standard diglycidyl ether resins. They thus again find use where high
filler loadings are required. Their compact structures allow much higher
crosslink densities to be achieved but they lack flexibility due to the absence
of the ether linkage found in other systems. They do, however, give
inherently high Tgs when cured, allowing use at elevated temperatures. Due
to the use of an alternative manufacturing process, a much lower level of
chlorides is present than in the aliphatic epoxides, although they still have
relatively poor water resistance. Examples of cYcloaliphatic epoxides are
dicyclopentadiene dioxide and vinyl cYclohexane dioxide which are solid
and liquid at room temperature respectively. When fully cured the
cycloaliphatic resins give very high glass transition temperatures, lower
dissipation factors and good arc resistance.
Aromatic Glycidyl Ethers
These materials tend to exhibit moderately high viscosities and this limits
the amount of filler that can be used in conjunction with this type of resin.
Despite this limitation, the inherent high strength and good water resistant
properties make these epoxide resins a good choice, especially when used in
conjunction with others.
Epoxide-novolacs
Although widely used in transfer moulding encapsulant formulations, their
very high viscosities limit the use of these resin types to specialized
applications. They do, however, offer low moisture absorbance coupled with
high strength and high glass transition temperatures. They are used, for
example, in liquid encapsulants in small quantities to help raise the glass
transition temperature.
Speciality Structures
More and more speciality resins are becoming available based both on
epoxide chemistry and, for example, cyanate esters. By allowing effectively
tailored properties, these types of materials are best used in conjunction with
more standard resins. As with most new technology, the disadvantages
include high cost and possible compromises in more basic properties.
6.2.2.2 Hardeners
Anhydrides
Anhydride-based curing systems are the main choice of hardener for liquid
encapsulants used in microelectronic applications. Their low viscosities
coupled with high purities and high functionalities make them ideal for
many applications. In addition, the latent reactivity at low temperatures
'--_ _ _ _M_A_T_E_R_I_A_L_C_H_O_I_C_E_S_F_O_R_E_N_C_A_PS_U_L_A_NT_S_ _ _ _ -----'I I 177
ensures extended pot life. Anhydrides are also relatively low-cost materials
and, unlike the aliphatic amines, they are not skin-sensitizing agents.
Amines
Amines have for many years been one of the mainstays of epoxy curing
technology and are still used in less critical areas, although they have largely
been replaced by anhydride technology for newer microelectronics applica-
tions. They are available in a wide range of forms, both aliphatic and
aromatic. The aromatic amines are less reactive than the aliphatic amines
and generally offer longer pot lives and higher glass transition temperatures.
Speciality
Many other materials are available, again with properties tailored to meet
specific application requirements. They are mainly employed in specific
formulations such as those used in the 'snap cure' products for smart cards
or flip chip underfills.
6.2.2.3 Catalysts
In the formulation of glob-top type encapsulants the catalysts used are often
defined as follows:
• Latent reaction. Latent catalysts are particularly important because they
allow longer pot life and longer flow times. They only induce significant
reaction above a certain temperature.
• Blocked reactivity. By using a catalyst that is chemically blocked it is
possible to achieve a much faster curing system while retaining a long pot
life and flow time.
• Chemically stable. Chemical stability is vital if an even cure is to be
achieved over the pot life and shelf life of the product.
For more details of the formulation of epoxide systems the reader should
refer to Chapter 4.
6.2.2.4 Fillers
A large number of fillers are used in the formulation of encapsulants for
electronics applications and they are typically used to lower thermal
expansion coefficients, improve thermal conductivities and reduce costs.
They also tend to improve the mechanical properties and modulus. Some
advantages and disadvantages of using fillers in encapsulant formulations
are shown in Table 6.2.
When formulating an encapsulant it is therefore vitally important that
careful consideration be given both to selection of the appropriate fillers for
178 1 1-1____M_A_T_ER_IA_L_S_F_O_R_A_D_V_A_N_C_E_D_E_N_C_AP_S_U_L_A_T_IO_N
_ _ _- - '
Advantages Disadvantages
a specific application and to the level to be used. Perhaps one of the most
important considerations when selecting fillers are the changes that they
could make on the electrical properties of the encapsulant. If good
insulating properties are required, as is typically the case with electronic
encapsulants, the condition and type of filler employed will be important
since ions, salts and moisture-soluble impurities may be adsorbed onto the
filler particles' surfaces which can degrade the final electrical properties.
Some of the key fillers used in encapsulant applications are as follows.
Alumina (A 12°3)
Conventional alumina is also used in encapsulant formulations particularly
where high thermal conductivity is required, although it is very abrasive and
imparts poor mechanical strength. It also has a relatively high coefficient of
thermal expansion. Alumina is typically higher in cost than calcium
carbonate and still has the disadvantages of relatively high ionic impurity
levels and a high thermal expansion.
Silica (Si0 2 )
Silica is usually thought of as being simply Si02 but there are in fact over 20
distinct phase of silica, each one still chemically being Si02 . Several types of
silica find use as fillers in electronic encapsulant applications.
One of these phases is known as fused silica and it is the most important
type of silica used in glob-top type encapsulant formulations for microelec-
tronics applications. Fused silica is produced by subjecting high purity
crystalline quartz to very high temperatures such as those found in an arc
furnace. It has a relatively low thermal expansion coefficient of approx-
imately 9.54 ppmrC over the temperature range O-lOOO°C and a compar-
atively low density (2.15 g/cm3). Fused silica is also readily available in high
purity grades. It has the advantage of being available in many distinct
physical forms, including a spherical type. The disadvantages of fused silica
can include abrasiveness (although it is typically less abrasive than alumina)
and a lower thermal conductivity than some other inorganic fillers.
Another type of silica that finds use in some encapsulant applications,
although not usually in glob-top type formulations, is quartz. This has the
advantage of a higher thermal conductivity than fused silica and could be
considered for use in applications where an encapsulant was to be used with
power devices.
A third type of silica which finds widespread use in the formulation of
liquid encapsulant systems is the so-called fumed silica. This is an
amorphous form of silica that has a very large surface area. It can be
used in, for example, liquid epoxy encapsulant systems to provide
thixotropic and anti settling properties. The thickening effect of fumed silica
is brought about by the interaction of silanol (SiOH) groups on adjoining
particles, as well as between the silanol groups and the reactive groups of the
hardener system (such as amine groups). Fumed silica is also used as a
reinforcing agent to enhance mechanical properties in elastomeric materials,
such as the silicone-based encapsulants.
Talc
Talc is the name given to hydrated magnesium silicate which has the
theoretical formula 3Mg0.4Si02.H20 and it occurs in several ores from
which talc products are made. The form most used in encapsulant
formulations is finely ground talc which consists of thin platelets. It
possesses a good degree of chemical resistance, being inert to most acids and
it is also a good insulator. The use of talc in epoxide encapsulant
applications leads to higher stiffness and improved creep resistance at both
room and elevated temperatures. Encapsulants loaded with talc have also
been shown to have increased lap shear strengths at their adhesive
interfaces; a property that is important in encapsulant applications because
the possibility of moisture ingress along interfaces must be avoided. In order
to obtain the full benefits from using talc it must be incorporated into an
encapsulant using the appropriate compounding techniques.
Carbon Black
Although not directly comparable to the other materials listed above as
fillers, the carbon blacks are sometimes used in encapsulant formulations at
relatively low levels in order to give the material a black appearance. They
are produced by the thermal decomposition of oil or gas and exist in a
number of forms such as lampblack, channel black and thermal black.
Carbon blacks are also good at imparting stability against ultra-violet light
exposure, as well as increasing modulus and tensile strength.
6.2.2.5 Elastomers
Low Viscosity
Low viscosity elastomeric polymers are used in encapsulant formulations to
enable higher filler loadings to be achieved where it is critical to reduce the
thermal expansion coefficient of the cured encapsulant. These are often
based on silicone chemistry.
Heterophase
Heterophase additives in a formulation form discrete isolated areas within
the encapsulant during curing. It is possible to incorporate, for example,
rubber-like molecules into an encapsulant which can be clearly seen as
discrete structures using a microscope. By being effectively a separate part of
the material structure it is possible to maintain overall properties of the
system (such as high TJ despite using an elastomer.
Hydrophobic
Hydrophobic elastomers have chemical structures that enable the encapsul-
ants in which they are used to minimize water absorption. This is
particularly critical in surface mount applications, where resistance to the
moisture-induced 'popcorn effect' may be a requisite.
Pigments
Pigments are often used to impart a particular colour to an encapsulant and
they are particularly important in encapsulants used with light-sensitive
devices where minimal optical transmission is required (see also carbon
blacks above). They also playa role in ensuring security and have even been
used to produce encapsulants with a characteristic company colour. Typical
pigments include titanium dioxide, cadmium sulphide, iron oxide and
barium sulphate, although it should be noted that not all of these are used in
electronic encapsulant applications.
Surfactants
A surfactant (or, more correctly, surface active agent) is a substance that,
when present at a relatively low concentration in a particular system has the
ability to adsorb onto the surfaces or interfaces of the system, thereby
altering to a significant extent the surface or interfacial free energies of those
surfaces or interfaces. Although there are many types of surfactants, they all
exhibit a characteristic molecular structure consisting of two different
structural groups known as lyophilic and lyophobic groups. This is known
as an amphipathic structure and it is because of this structure that
surfactants are able to have such a big influence on the nature of interfacial
interactions. Surfactants are therefore used in encapsulant applications to
modify the interfacial interactions and also help to reduce the possibility of
void formation.
Rheological Modifiers
Rheology modifiers are used to modify and tailor the flow and thixotropy
characteristics of an encapsulant. These could be used, for example, where it
is necessary to prevent flow through small holes in a package.
Ion Scavengers
As mentioned above, ionic impurities in an encapsulant can comprise
semiconductor reliability, and ion-scavenging additives are sometimes added
L -_ _ _ _ _ _ _ _M
__A_T_E_R_IA
__ L_C_H_O_I_C_E_S_F_O_R__
EN__
C_A_P_SU
__LA_NT
___ S ________ ~I I 183
6.2.3 Combinations
As already mentioned the possible combinations of ingredients used for a
specific encapsulant formulation are virtually limitless. In order to give an
idea of the possible effect of just one of these combinations, Table 6.3 shows
the effect of using different resins with one version of the anhydride curing
system. The effect on coefficient of thermal expansion is dramatic, especially
when it is considered that this is in an unfilled system. Addition of a low
thermal expansion coefficient filler would further reduce the thermal
expansion, with similar differences.
Most systems used today are formulated with a combination of different
resins in order to obtain the necessary compromise in properties. Newer
resins are also now becoming available that are more tailor-made for a given
specific application. Whilst this can assist in the basic formulation, there is
usually a down side to be considered, not the least of which is the relatively
high cost and limited availability of these newer resins.
Similarly, multiple options of other ingredients are also often used. In a
case where improved reliability performance is required through the use of
materials such as ion scavengers, it may be necessary to combine several
types to compensate for the use of less pure resins with each scavenger being
specifically targeted at a single particular ionic impurity.
6.2.4 Limitations
There are of course specific limitations to what can be achieved by
combining the various formulation ingredients and whilst these limits are
constantly being extended, there are a number of areas where specific trade-
Table 6.3 The thermal expansion coefficients (CTE) of basic anhydride cured
encapsulants formulated with different epoxy resins
Aliphatic 80
Cyc10aliphatic 75
Aromatic glycidyl ether 60
Aromatic glycidyl novolac 55
Speciality resins 45-60
184 , ''--_ _ _M_A_T_E_R_IA_L_S_F_O_R_AD_V_A_N_C_E_D_E_N_C_AP_S_U_L_A_T_IO_N
_ _ _----'
often require more intensive (largely high shear) mixing, usually under
vacuum and at an elevated temperature. For this reason most microelec-
tronic encapsulants are now typically premixed by the manufacturer.
Depending upon the specific formulation, the material will then be
packaged in some form of bulk container or in a syringe or cartridge. The
former are generally referred to as one-component systems and the
challenge for the formulator is to provide them with a reasonable room
temperature shelf life. The latter are normally referred to as premixed or
frozen products and usually require low temperature storage (e.g. -40°C) to
minimize pre-use reaction and thus to give good shelf lives e.g. in excess of
four months.
6.3.2 Packaging
The correct choice of packaging can also be crucial to ensuring quality and
productivity. Factors that should be taken into account include shelf life,
pot life, shipment, storage and handling. The type of packaging should be
optimized for both shipment and handling. The size will largely be
determined by the chemistry of the material and the dispensing equipment.
Too large a size will result in material being wasted, either at the end of a
production run, or as a result of increased viscosity as a material nears its
pot life. Too small a size and frequent replacement will be needed, thereby
adding unnecessary work for the user and affecting overall productivity.
In addition to the inner packaging, thought should be given to the outer
packaging. As an example, if material is to be shipped by air it is essential
that the outer packaging meets lATA regulations. In addition, the
packaging should be sufficiently robust to withstand the shipment and be
able to hold dry ice (in the case of frozen products) sufficient to maintain a
low temperature throughout the journey.
6.3.3 Storage
Specific storage requirements for each product should be provided by the
material supplier. Generally, frozen products should be kept at -40°C or
below since storage at a higher than recommended temperature will
adversely affect the material's ultimate performance. Too high a temperature
will generally result in a reduced pot life, increased viscosity or reduced flow.
Although it may still be possible to use material that has been incorrectly
stored, there is always a risk of incurring quality problems by doing so.
6.3.4 Thawing
Once removed from cold storage it is important that the material is allowed
to thaw properly prior to use. The time taken to thaw will depend upon the
188 11L _ _ _ _M_A_T_E_RIA_L_S_F_O_R_A_D_V_A_N_C_E_D_E_N_C_A_P_S_U_L_A_T_I_O_N_ _ _-...l
size of the packaging, the ambient temperature and the storage temperature.
Even thawing is best accomplished by leaving the material in a well-
ventilated area at room temperature (20--25°C). Although higher temper-
atures may accelerate the thawing process they may also create an uneven
temperature, especially for larger syringes and cartridges. A typical thaw
time for syringes in the 5-50 em3 range would be 30--40 minutes (from
-40°C), whilst a larger syringe (200--400 em3) might require somewhat more
than one hour.
Although it may be acceptable (dependent upon the material and
application) to start dispensing with the material a few degrees below the
ambient, it is critical to ensure that, at the very least, the material has
exceeded the dew point to avoid moisture condensation when dispensing.
For similar reasons the material should not be left exposed to the atmosphere
for any length of time, including after dispensing and prior to curing.
6.3.5 Preheating
In most applications preheating is a key factor in achieving good material
flow and wetting, whilst minimizing void formation. Generally, the substrate
to which the encapsulant is to be applied is heated although it is also
possible to use a needle heater if deemed essential. Syringe heating is not
recommended as it will in most cases lead to a reduction in the pot life of the
encapsulant.
For most applications, substrate temperatures in the region of 60--1 10°C
should be used. Reference to the supplier's literature showing viscosity and
onset of cure properties may be helpful in establishing initial operating limits
but for a specific application it is normally preferable to conduct a series of
trials to determine the optimum temperature. Dispensing onto a warmed
substrate will often aid bubble release although excessive warming will also
impair release if the epoxide gels too quickly.
6.3.6 Dispensing
Many different techniques can be used to dispense liquid encapsulants.
Historically, simple time/pressure dispense techniques have mostly been
employed. Whilst this is still an acceptable way to dispense materials for a
wide range of applications, newer developments in equipment have resulted
in better consistency. Use of a positive displacement valve ensures a
constant amount of material is delivered, regardless of variations in viscosity
that might occur due to changing temperature or pot life of the material.
For dispensing very small volumes, where the positive displacement valve
may not be sufficiently accurate, it is also possible to program a time/
pressure dispense system with the viscosity/temperature/time relationship of
a particular material thus ensuring consistent dispensing.
L -_ _ _ _ _ _ _ AP_P_L_I_C_A_Tl_O_N_T_E_C_H_NI_Q_V_E_S_ _ _ _ _ _-----l1 I 189
6.3.7 Cure
Data supplied by the material manufacturer should also be used as a
starting point to establish the optimum cure schedule. Due to the wide range
of possible applications, these data should only be used as a guide. Factors
such as temperature limitations may mean an alternative cure schedule is
preferable. Whilst a longer time at a lower temperature can often be
substituted for a short time at a high temperature, it should be remembered
that most materials have an absolute minimum temperature that must be
reached to enable full curing to be achieved. This is especially true when
using a material with a latent or blocked catalyst. Generally, a lower
temperature cure will result in lower stress, whilst a higher temperature will
result in higher T g • Often a two (or more) step cure is used to help minimize
stress, with the final high temperature step being designed to ensure full
curing.
To ensure full curing has been achieved, the Tg and degree of cure should
be measured using thermo-mechanical analysis (TMA) and differential
scanning calorimetry (DSC) techniques. The variation in degree of cure with
time and temperature for a typical encapsulant is shown in Fig. 6.4.
~ 80
~
~ 60
'0
~ 40
g» 20
c
o
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Time/Hours
Fig. 6.4. Variation of cure with time and temperature for a typical encapsulant
(Hysol FP4450).
,--
.
.. • • .. • • .
I
. • • •• .. •.. • .. •• •. .. .,. • .
.. . • . • ~-- • • .... . .. ..
• • .. . ..
.. ..
.,. .. .. • ~~-:~ ffp"
.~
• •,. • •. .... ..
. e • .
. • ... • • . ~'.. ~....."';~~
,r,': _ .,..<~
.. • • ..
.. . • .. • • •
6
•••••.
• • • • • ..
....,, ~
.. • • . .
•
•.. .. • •• ,.
<:0'
• .. •.. .. • • • . .. •
..
.. ..• ... ..• .. •
~~
-1.:1{~: ~ ..5.f;
using indexing holes in the tape carrier, into the correct dispensing
position.
The traditional method of dispensing involves use of a medium viscosity
encapsulant. The dispense needle follows a circular or rectangular outline
just beyond the outermost wire bond. The dispensing then continues in a
circular pattern towards the centre of the die. In order to maintain the
correct profile the substrate is then subjected to a post-heat stage. Control of
both the pre-heat and post-heat stages is critical to maintain the correct
thickness of encapsulant.
More recently alternative systems have been introduced using either a
shower head principle, or utilizing hot air levelling to 'flatten' the
encapsulant surface. Some early systems also used a grinding system to
remove the top of the glob but this technique has largely disappeared as it
relied upon the use of encapsulants with soft fillers which were poorly suited
to large die encapsulation.
L-_____________ A_P_P_L_IC_A_T_I_O_N_T_E_C_H_N_I_Q~U_E_S______________~I I 193
Fig. 6.6. Smart Card modules showing outside contact surface arrangement and
inside surface with glob top encapsulant covering the semiconductor device.
Organic or c:enmk ~
Fig. 6.7. Schematic representation of flip chip assembly with polymeric underfill/
encapsulant.
The use of an underfill with flip chips reduces the thermal stress on
connections and adds environmental protection. The underfill plays a vital
role in lowering the thermal expansion mismatch between the flipped chip
and the substrate. This is essential for ensuring that the solder joint
interconnects are reliable. The use of an epoxide-based underfill with a
controlled thermal expansion coefficient has been shown to improve the
temperature resistance of flip chip devices by 5 to 10 times. Also, with the
use of an epoxy underfill, encapsulated flip chips on flexible circuitry have
been able to pass severe automotive temperature cycling requirements. By
Property 1 2 3
Fig. 6.S. Vnderfilling of flip chips (a) flip chip on foil, (b) flip chip on a multichip
module.
196 II MATERIALS FOR ADVANCED ENCAPSULATION
~----------------------------------------------------~
necessary to repeat the dispensing along the edge of the die to completely
flow under the chip. Curing is then normally carried out in-line.
6.6 SUMMARY
the requirement for more and more I/Os within a diminishing packaging
area. Ultimately, the most likely scenario is that for advanced applications
the use of the flip chip approach i.e. the assembly of inverted unpackaged
devices directly onto a substrate will become widely adopted. This in turn is
driving the need for the new types of encapsulants and encapsulation
methods detailed above. There will also continue to be significant challenges
for encapsulant formulators such as the need for further increased thermal
conductivity and reduced dielectric constant materials. The future for
advanced encapsulants is both interesting and challenging but there is no
doubt it is also full of opportunity.
6.7 ACKNOWLEDGEMENTS
The author would like to extend his thanks to Maurice Edwards, Director of
Research and Development at Dexter Electronic Materials for many helpful
technical discussions and detailed background information.
Recent Developments in the 7
Chemistry of Lithography for
Electronics Production
DAVID MERRICKS
Shipley Company Inc, Marlborough, MA, USA
7.1 INTRODUCTION
The ever increasing integration and sophistication found in nearly all of the
electronic devices produced over the past decade has led to tremendous
advances in information technology. The production of the next generation
of components such as microprocessors is becoming increasingly demand-
ing and manufacturers need to produce ever more complex products using
finer featured printed circuit boards (PCBs) and discrete and integrated
circuits (lCs), with increasing circuit densities. This in turn requires the use
of highly efficient manufacturing processes and short product development
times.
Figure 7.1 shows the trend in feature size reduction for both ICs and
PCBs. As a general rule, the smaller the features used the more functionality
the device contains and the faster it will be able to operate.
Feature size reduction is expected to continue within semiconductor
manufacturing for the foreseeable future as new processing techniques such
as electron-beam and X-ray lithography allow the industry to find ways of
concentrating progressively more sophisticated interconnection systems into
smaller spaces. In the PCB industry, the reduction offeature size is expected
to come to a natural limit using conventional technology, as it becomes
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
202 I I CHEMISTRY OF LITHOGRAPHY FOR ELECTRONICS PRODUCTION
1250
i~ 250
~ 25
!
2.5
0.75
0.25 +---r-----r---r-"---r---r--~
1970 1975 1980 1985 1990 1995 2000
Year In Production
physically impractical for feature sizes to get much smaller using traditional
technology. Consequently, new technologies will also be required here.
For the past 40 or so years, polymers have been a key element in the
growth of both the IC and PCB industries. New synthetic procedures and a
greater understanding of the relationship between the molecular structure
and the mechanical, physical and thermal properties of polymers has led to
the supply of an ever-increasing variety of materials to meet the needs of
both industries.
The circuit structures of both ICs and PCBs are produced by a series of
steps incorporating lithographic processes which:
1. Delineate the circuit pattern in a thin film of radiation-sensitive polymer
(photoresist), and
2. Transfer the pattern produced onto the substrate using appropriate
etching or deposition techniques.
Figure 7.2 shows a schematic representation of the lithographic process,
using both positive-working and negative-working photoresists.
If a photoresist is negative-working, the exposed areas become insoluble
in the developing solution. If it is positive-working, the exposed areas
become soluble and are removed by the developing solution. The photo-
resists which are used to produce the circuit pattern normally have to be
removed before further processing can continue. However, some photo-
resists are used as permanent coatings, such as liquid photo-imageable
soldermasks (LPISMs) and photo-imageable dielectrics. These are used
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __R_O_D_U_C_T_I_O_N________________~I
INT I 203
Exposing Radiation
Ht Ht Mask
____ Photoresist
----Film to be Patterned
~ Substrate
~m~'
Positive Image Negative Image
f CHz-J...lLCHZ-Jl
I Jx L I Jy
Negative-working Uquid,
ED or DF.
Positive-working ED.
COZRZ CO z R3 Positive e-beam, X-ray.
ACRYLIC
~'"+ CH 3
NOVOLAC
Positive-working liquid,
DF.
OH
~HCH2+-0-@+@-OCH2JHCH2+'O-@+@-OCIl2~\H2
EPOXY Negative-working LPISMI
Photodielectric
Positive-working
e-beam. X-ray.
POLYSULPHONE
Positive-/Negatlve-
working DUV resist.
POLY(HYDROXYSTYRENE)
photoresists used as etch resists or plating resists) and the second part with
secondary imaging (LPISMs used as permanent dielectric coatings). The
second section progresses from the optical photoresists currently used in Ie
production, through potential future technologies such as deep ultra violet
(DUV), to electron beam and X-ray sensitive photoresists.
L-_P_H_O_T_O_RE_S_I_ST_S_F_O_R_P_R_INTE
__D_C_IR_C_U_I_T_B_O_A_R_D_P_R_O_D_U_C_T_IO_N_-----'I I 205
7.2.1 Introduction
This section is mainly concerned with polymers used in the production of
PCBs, although the types of photoresists described here can also be used for
other applications, such as in the photochemical machining (PCM) industry
as print and etch resists or in the production of colour filters, where the resist
is permanently in place on the finished product. The first part of this section
is concerned with primary imaging which covers the use of polymers as print
and etch or plating resists on the inner and outer layers which comprise a
multilayer PCB, while the second part concentrates on secondary imaging,
describing polymer formulations used in solder masks.
• Polymer. This provides the physical strength of the photoresist film and
also determines the mode of developing and stripping.
• Monomer. This crosslinks on exposure by free-radical polymerization to
become insoluble in the developer.
• Photoinitiator. The source of free radicals during exposure. The choice of
photoinitiator depends on the principal wavelengths emitted from a
medium pressure mercury vapour lamp. The peak absorbance should be
on, or close to one of the major emission lines (usually 365 om).
Photoinitiators with absorbance wavelengths removed from those of the
lamp may also be used. In these cases photosensitizers, matched to the
principal lamp wavelengths are added which, after excitation, transfer
their energy to the photoinitiator.
• Plasticizers. These modify the physical and functional properties for a
given application.
• Dye. Normally a phototropic dye (usually blue or green) is incorporated,
which provides contrast between the exposed and unexposed areas. Dyes
are also used as anti-halation agents which prevent the undesired
reflection of UV light.
• Other additives. These can include adhesion promoters and thermal
polymerization inhibitors.
L-_PH_O_T_O_R_E_SI_S_T_S_F_O_R_P_R_INT_E_D_C_IR_C_U_IT_B_O_A_R_D_P_R_O_D_U_CT_IO_N_----'I I 207
Most dry films used today contain polymers formed from different types of
acrylate or methacrylate monomers. Figure 7.4 shows a variety of acrylates
used in negative-working dry film resists. The first two examples are aqueous
processable and also a crosslinking monomer is needed in the formulation
since there are no photopolymerizable groups present attached to the
backbone. The third example is a solvent processable resist and contains
pendent groups with unsaturation, which crosslink on exposure to form an
insoluble matrix.
Negative-working dry film systems based on polyfunctional alcohols with
acrylate end groups are also known 5 and a typical structure is shown in
Fig. 7.5.
Other additives are necessary in order for the above polymers to function
as dry film resists. Unsaturated monomers used are usually multiacrylates
and typical examples are shown in Fig. 7.5.
All negative-working photoresists contain a photoinitiator and also in
some cases a photosensitizer as well. The photoinitiator is a compound
which absorbs incident UV radiation of an appropriate wavelength from the
exposure lamp during the imaging step and splits into free radicals which
initiate polymerization of the unsaturated monomer (or unsaturated
polymer). Given an appropriate light source, efficient photoinitiation
Polymer Reference
2.
3.
fuz-iu-jfcuryu+ 5
C02CH2CHOH C0 2R
tH 20COCH=CH 2
R = Alkyl
Fig. 7.4. Acrylate dry film resists.
CH2-CHCO~OCH2CH2-r-O~O-+CH2CH20+-COCH=CH2
x ~ y
CHZ=CHCOZCHZCHZOCH2CH202CCH=CHZ
CH Z0 2 CCH=CH Z
"0 '"'"
.
..,'0
'0 0.15 i
0
~0 1:1
5'
"
0
Q.
;
e
~
0.10
1.0
0.05
rgQgr
+
Me
Polymer ·~-O
Me
O(CH2CH2OCOCH-C121
hv
Developed in
NazC03
Crosslinked Product
CH3 CH3
tCHz-t~CHz-t~t
COzNa b0 2cH3
+ C02 + H2 0
Polymer
Polymer
PAG
PhZI+ SbF 6-
This latter type of system has also been used with visible laser direct
imaging, using an argon ion laser at 488 nm and 514 nm. Visible laser
imaging of normal negative-working resists has been attempted but these
systems suffer from the same limitations as UV sensitive formulations
including swelling in the developer and the need for an oxygen barrier
during exposure. The use of a chemically amplified resist here would
212 I I CHEMISTRY OF LITHOGRAPHY FOR ELECTRONICS PRODUCTION
therefore have several advantages. These materials can be used as liquid
resists as well as in dry film applications.
+u,-t*u'-Htu,-p-
CH3 H 1H3
PAG
Polymer
Fig. 7.11. Chemically amplified waterborne resists.
conductive substrate. They are used primarily for high resolution selective
etching and pattern plating in the manufacture of PCBs (both inner and
outer layers), 3D moulded interconnect devices, multichip modules and
other interconnect level electronic devices.
These resists have the potential to displace solvent-based liquid resists as
well as dry film photoresists in most present-day applications over
conductive substrates, although ED resists are not yet that widely used.
The main advantages of ED resists over dry films are firstly, their superior
resolutions due to the ability to uniformly coat layers up to five times
thinner and secondly, their ability to conform to three-dimensional features.
Other advantages are the same as for liquid resists (see section 7.2.2.2).
A variety of electrodepositable film-forming resins are known, the most
often utilized to date being acrylates, epoxies and novo lacs, although the use
of polymers containing more diverse functionalities is gradually increasing.
In order for a polymer to be electrodepositable it must contain a distribution
of ionizable groups along its molecular chain. Polymers may be cationic,
containing basic sites such as amino groups, or they may be anionic
containing acidic sites such as carboxylic, sulphonic or phosphoric acid
groups. In some instances amphoteric polymers are known which have both
acidic and basic groups.
The polymer can also contain photosensitive functionalities such as sites
of unsaturation in the case of negative-working photoresists, or have
photo active groups such as diazonaphthoquinones (DNQ) attached, as in
positive-working photoresists. Although some polymers are formulated with
such functionalities present, in the most commonly used polymers such
L-_P_H_O_TO_R_E_SI_S_T_S_F_O_R_P_R_I_NT_E_D_C_IR_C_U_I_T_B_O_A_R_D_P_R_O_D_U_C_T_IO_N_----11 I 215
groups are absent and the main function of the polymer is as a carrier for the
other resist additives.
Another requirement of the polymer is that it must be able to form a
stable emulsion with suitable ionizers such as inorganic or organic acids
(which protonate the ionizable groups in cationic resins), or bases (which
deprotonate the ionizable groups in anionic resins) in the presence of other
resist additives such as photoinitiators, DNQ-photoactive compounds,
crosslinking agents and so on. Plasticizers are also added to control the
coating characteristics, to give good adhesion to the substrate and to ensure
that the coating is not brittle. The flexibility of ED resists is far superior to
that of dry film resists.
Negative-working and positive-working cataphoretic or anaphoretic resist
systems are known and examples of the resists used are shown in Fig. 7.12
and 7.13 (the numbers in parentheses indicate monomer parts by weight).
From Fig. 7.12 it can be seen that the polymers in examples 1,2 and 5 have
photosensitive groups along the polymer backbone, whereas crosslinking
monomers or photoactive compounps have to be added to the polymers in
examples 3, 4 and 6.
Resist additives such as monomers, dyes, photo initiators or photoactive
compounds are of the same basic types as those used in dry films (section
7.2.2.1) or DNQ photoresists (section 7.3.2.1).
Hydrophobic solvents, or plasticizers, are often added to electrodeposi-
table resists to lower the Tg of the polymer, enabling electrodeposition to
take place at lower temperatures and to control the film thickness more
easily. Plasticizers need to be chosen to maximize film coalescence and
flexibility and to minimize tack, which can sometimes be a problem during
contact exposure.
The coating mechanism, which is described briefly below, concentrates on
cataphoretic deposition for simplicity. The principles of anaphoretic coating
are basically the same, with only the charges on the micelles and the
electrodes being reversed.
In a cataphoretic emulsion, when an electric field is applied (approxi-
mately 10 volts/cm), micelles migrate by electrophoresis towards the cathode
at the rate of microns/second. In addition, all the water insoluble
components also migrate with the micelle. The conductivity of the solution
permits controlled electrolysis. Water decomposes to raise the pH at the
cathode and lower the pH at the anode, which is usually made of an inert
material such as stainless steel since, because it is the oxidizing electrode,
metal dissolution is otherwise possible.
When the micelles reach the cathode, their positive charge is neutralized
by hydroxide ions produced by the electrolysis of water. The micelles then
become destabilized and coalesce on the surface of the cathode to form a
self-limiting, insulating film that emerges nearly dry from the coating bath.
As the resistance of the film increases, the potential across the film increases
Type Polymer Reference
1. Negative!
I]
Anaphoretic
z. Positive!
Anaphoretlc 13
CH] CH] H f
-fCHz-jfCHz-j1fHz-j*HZ-,+
COZ COZDB" COZH
°~NCHZCHZ
NZ COZCH]
tllu I
]. Positive!
Anaphoretic 14
4. Negative!
Cataphoretic CH3 H 1H3 15
-fCHz-j1fHz-t*HZ-l+
COZDBu COZCHZCHZN(CH3)Z
COZCH3
5. Positive!
Cataphoretic 13
6. Amphoteric 16
1. Cataphoretlc/
Epoxy 17
?H OH OH
R2NCH2CHCH20~OCH2~HCH20~CH2~HCHZNR2
z. Cataphoretlc/ 18
Negative/Epoxy
OCOCH-CHZ
C~HCHZO~OCHZ!HCHZO~CH2~~HZ
MIXED WITH
~m N
tHzCHZCHZN(CH3)Z
3. Positive/
Anaphoretic/Novolac 19
4. Positive/
Cataphoretic/Novolac
16
~'"'~'"'t-
CHZ
I
tBu
N(CHZCHZOH)Z
causing water and occluded ions to be forced out of the coating by electro-
osmosis. Thus the combination of electrophoresis and electrolysis produces
a highly uniform and defect-free coating. The self-limiting nature of ED
coatings is mainly dependent on voltage, coating time and bath temperature
and conditions can be used which build up quite thick films. However, at
optimum conditions for thin coatings, film growth continues until film
resistance is so high that the electric field across the emulsion is too low to
induce delivery of the micelles, or the current flow drops low enough that the
cathodic pH is too low to induce coalescence. As a result, films should be
uniform even if they deposit at different rates across a part.
Meanwhile, the carboxylate anion, on reaching the anode, is neutralized
by hydrogen ions from the electrolysis of water. As resist solids are removed
from the bath at the cathode, there is a gradual build up of ionizer in the
bath. Therefore, to maintain bath chemistry, free acid must be removed by
ultrafiltration or by the use of semi-permeable membranes.
Figure 7.14 shows the cell diagram for cataphoretic deposition.
?H
A= ~CH_CHCO~OCH2CHCH2
~ ,H
B= 0~OCH2CHCH2
-O@CH-CHCO-
2 - - o@CH-CHCO-
hv
~
I I
-O@CH-CHCO-
OH
I I
-O-CHz -CH-CHZ-N-~-R
+ L:::,.. N
~ I
R-C-NH Z
-OCH Z-
1
H - CH z
"
OH
NH
Fig. 7.15. Photodimerization and thermal curing of a full-epoxy LPISM.
~
Ar2S + z· + HX
Ring Opening of Epoxldes
Thermal Cure
R - Epoxy polymer
Polymer
Styrene/Acrylic copolymer
Monomer
Trimethylolpropane triacrylate
Isopropylthioxanthone(ITX)
Fig. 7.18. Water-borne LPISM components.
7.3.1 Introduction
This section concentrates on polymers which are currently used (or are
planned to be used in the future) in photoresists for integrated circuit (IC)
production.
224 I I CHEMISTRY OF LITHOGRAPHY FOR ELECTRONICS PRODUCTION
The first part covers optical lithography, detailing near-UV systems which
utilize the g and i-line emissions of standard Hg-vapour sources and deep-
UV systems which operate at the 248 nm wavelength produced by an
excimer laser. The final part describes newer technologies, such as electron
beam and X-ray lithography, which will be necessary to produce the next
generation of IC devices.
The discussion here concentrates on single layer resists and multilevel
(bilayer and trilayer) systems are not covered.
A modem IC is a complex 3D structure of alternating patterned layers of
conductors, dielectrics and semiconductor films. The semiconducting
material is usually ultra-high purity silicon. The performance of the device
is, to a large degree, governed by the size of the individual circuit elements.
As a general rule, the smaller the elements, the faster the device and the
more operations it can perform. The device structure is produced by a series
of lithographic steps used to pattern each layer precisely.9
Forecasts of the future development of highly integrated silicon circuits
are focused on the dynamic random access memory (DRAM), which are the
devices with the highest complexity and integration density. DRAMs
currently represent the forefront of technology development.
Figure 7.19 shows the history of increasing complexity and simulta-
neously of decreasing smallest geometries on one chip.
It can be seen that dramatic advances are continually being made in
microelectronic device fabrication and especially in optical lithography, the
100
a!:l CONTACT PRINTING
......
..
~
::I
1\1
5.0
3.0 1:1 PROJECTION
J!
a
::I
2.0
a 5:1 PROJECTION
~ 1.0
0.5
0.3
0.2
0.1
o~----~----------~------------,-~
1978 1988 1998
Year
Fig. 7.19. Development trends of lithographic technology.
'--__
PH_O_TO_R_ES_I_ST_S_FO_R_I_NT_EG_RA_T_ED_C_I_R_CU_IT_P_R_O_D_U_C_T_IO_N_---'I I 225
~2_hV_'"
80 2 R
DNQ
~2H
80 2 R
Indene carboxyUc acid
Fig. 7.20. Photochemistry of a novolac-DNQ system.
:)9t16(:
OH
«x
o ~3
~
4
N2
Ballast-S02
2,I,S-PAC
«;~
o ~3
N2
5 4
S02-Ballast
2,1,4-PAC
Fig. 7.23. 2,1,5- and 2,1,4-DNQ PACs.
~-----1 SOFTBAKE
POST DEVELOP
BAKE
STRIP RESIST
_ fcH'Qi
+CH, Qi
+
OH
+ COl
+ CH1=C(CH3)1
OC01C(CH3)3 +
+H
/~ +CHZ OH
HO--©-t-<Q>-OH
+ ZCOZ + ZCHZ-C(CH3>Z
RN(CH20-Ph- )2
The impact of X-rays with the resist generates electrons which have enough
energy to initiate chain scission (positive-working resist) or crosslinking
(negative-working). Therefore, most polymers that are sensitive to electron-
beam lithography are sensitive to X-rays. These materials will be described
later in this section after a brief description of electron beam lithography.
The resolution capability of e-beam systems is very high, since the size of a
focused electron beam can be as small as 10 nm. Also, due to the electron
size, there will be no diffraction effects occurring. The technique eliminates
the use of a mask, as the computer-generated beam, which is emitted by a
tungsten filament or lanthanum hexaboride source, is scanned over the
surface of the resist. However, the exposure time is longer than when using
optical lithography since electron beam writing is a sequential process and
because computer time is required to correct for 'proximity effects'.14
While virtually no scattering in organic materials occurs with X-rays, e-
beams show substantial forward scattering which causes the beam to expand
with its passage through the resist. With low energy electrons the
penetration depth into the resist is limited, whereas when high energy
electrons are used to minimize forward scattering, backscattering from the
resist and substrate becomes important. E-beam lithography is therefore
limited to direct-write applications of low volume logic chips and also to the
fabrication of high quality masks for optical lithography.
At high accelerating energies, polymer scission and crosslinking can
simultaneously occur, but materials can be chosen where one of the
mechanisms predominates to form negative or positive-working systems.
The response of the material to radiation is measured by the number of
chain scissions (Gs) or crosslinks (GJ that are produced per 100 eV of
absorbed energy and if the value for Gs is greater than that of Gx , scission
will predominate and the polymer will be a positive resist system. G is
determined by subjecting the polymer to cobalt-60 gamma radiation and
measuring the change in weight and dispersivity of the polymer. The Gs
values for positive-resist systems can range from 1.3 for polymethyl
methacrylate (PMMA), the benchmark for standard e-beam systems, to
10 for a polysulphone system and the Gx values for negative resists can range
from 0.1 for polyethylene to 10 for epoxy polymers. 24 The vast majority of
polymer systems will crosslink. DNQ-type systems are not suitable for use
as e-beam or X-ray resists since, because exposure is carried out in a
vacuum, there will be no water to react with the ketene to give an indene
carboxylic acid.
Figure 7.29 shows the mechanism of chain scission in PMMA, which
shows extremely high resolution (due to no swelling in developing solution),
ease of handling and gives excellent film forming characteristics, but is
relatively insensitive. The sensitivity ofPMMA can however, range from 50
to 200 mC/cm2 depending on the molecular weight. Structures with line
widths down to 10 nm have been created when using PMMA.43
236 I I CHEMISTRY OF LITHOGRAPHY FOR ELECTRONICS PRODUCTION
CH3 CH 3
-fCH2-?-CH2-t~
yO {O n
OCH3 OCH3
Figure 7.30 shows modified PMMA systems which have been developed
for use as more sensitive positive e-beam resists.
Poly(butene-I-sulphone) which is an alternating 1:1 polymer of sulphur
dioxide with an alkene is a more sensitive positive resist with a Gs value of 10
and a sensitivity of 10 mC/cm2 (10 kV).44 This material has a rather low
stability and the stability has been improved by combining the polysul-
phone, as a sensitizer, with novolac-type resins. The coating capability and
thermal plasma resistance is improved in this system, which has a sensitivity
of below 5 mC/cm 2 (20 kV).45
A positive-working resist for X-ray lithography has been developed46
based on a 2:1:1 polymer of 4-t-butoxycarbonyl styrene (TBS) and sulphur
dioxide (Gs = 3.6). This material undergoes main chain scission on
irradiation followed by deprotection of the hydroxyl group catalysed by
the sulphonic acid produced from the interaction of the released sulphur
dioxide with water. This combination of scission and deprotection leads to a
high sensitivity (10 mJ/cm2 using a laser source at 1.4 nm). Positive-working
systems based on an alkenyl silane-sulphone copolymer, i.e. allyltrimethyl-
silane-sulphur dioxide, which enhances resistance to an oxygen plasma
without adversely affecting sensitivity (1.5 mCjcm2 , 20 kV) have been
formulated. 47
L--_ _ PH_O_T_O_R_E_SI_S_T_S_F_O_R_I_NT_E_G_RA_T_E_D_C_I_R_C_U_IT_PR_O_D_U_C_T_IO_N
__ --'I I 237
PMMA 40 1.5
(X-CH3,Y-COZCH3)
X-CH3,Y-COZH 35 Z.O
X-CH3,Y-CN 14 3.5
X-H,Y-CN lZ 3.5
1.
IH3
10zcHzc6'~Hz !oz~
Z. IH3
L2;H2ciCO~HZ
-f·,-It
3.
20pC/cm Z (20kY)
©D
-{c·,-It
4.
< SpC/cm2
©D
C1
-f·,Jt
S.
©D, ,
(R)J: S>=<S)
Fig. 7.31. Negative working e-beam systems.
7.4 CONCLUSION
lOS
jH3
-+.J}
CH3
N
I
~
...,
"-
a fHZ-C---]f cHz-L
"-
0 10 4 I
COZCH3
I m
COZH I D
COZCH3
~:;
I
'"»"
f
,:,
...0 10 3 1"3 ~• •
::f: fHZ-Ct
Lze"HzciCO~HZ
""
Z
.. -+'-~'O'l
10
10
10-8
both PCBs and ICs will greatly increase the power of this technology. The
development of new, specialized polymeric materials will continue to feature
highly in this success.
7.5 REFERENCES
1. Wallig, L. R., Image transfer. In Printed Circuits Handbook, 3rd edn, ed. C. F.
Coombs Jr, McGraw-Hill, New York, pp. 11.9-11.24, 1988.
2. Tilsley, G. M. and Roos, L., Circuit World, 13(1), 1986.
3. Sato, J., Netsu Kokasei Jushi, 1993, 14(2), 92-103.
4. DuPont, US Patent 3,469,982.
5. Allen, D. M., The Principles and Practice of Photochemical Machining and
Photoetching. Adam Hilger, Bristol, 1986.
6. Chang, C. H., Mar, A., Tiefenthaler, A. and Wostratzky, D., Photoinitiators:
mechanisms and applications. In Handbook of Coating Additives, ed. L.J. Calbo,
Dekker, New York, 1992, pp. 1-50.
7. IBM, US Patent 5,045,431 (1991).
8. Wallraff, G. M. et al., J. Imaging Sci. Technol., 36, 468, 1992.
240 I I CHEMISTRY OF LITHOGRAPHY FOR ELECTRONICS PRODUCTION
9. Steppan, H., Buhr, G. and Vollmann, H., Angewandte Chemie Int. Ed. Engl., 21,
455, 1982.
10. Rohm and Haas, US Patent 4,592,816 (1986).
11. Sugiura, T., Journal of the SID, 1, 341, 1993. (a) Barr, R. K., Morton
International Inc., US Patent 5,364,737 (1994). (b) Shida, N., Ushirogouchi, T.
and Nakase, M., J. Photopolymer Sci. Technol., 1995,8, 173-178.
12. Merricks, D., Electrodepositable resists. In Special Polymers for Electronics and
Optoelectronics, ed. J. Chilton and M. T. Goosey, Chapman and Hall, London,
1995.
13. Kansai Paint Co. Ltd, U.S Patent 4,965,073 (1990).
14. Kansai Paint Co. Ltd, European Patent 0,383,223 (1990).
15. Rohm and Haas Co., US Patent 4,592,816 (1986).
16. Ciba-Geigy Co., US Patent 5,080,998 (1992).
17. Nippon Paint Co. Ltd, US Patent 5,055,374 (1991).
18. Nippon Oil Co. Ltd, European Patent 0,441,308 (1991).
19. Ciba-Geigy Co., US Patent 4,632,900 (1986).
20. Kurisu, V., PC Fabrication, 1990,66.
21. Rembold, K.-H., Polymerisable epoxy resins and their application in the field of
printed circuits. In Proc. 1st Printed Circuit World Convention, London, 1978.
22. Shipley Co. Ltd, European Patent 0,502,382 (1992). (a) Knudsen, P., Brainard,
R. and Schell, K., In Proc. Surface Mount International Conference and
Exposition, San Jose, 1993, pp. 203-216. (b) Clark, E., Transitioning to MCM
production, Printed Circuit Fabrication, 16(20), 69--73, 1993. (c) Enomoto, R.,
Asai, M., US Patent 5,055,321 (1991).
23. Morton International, Inc., European Patent 0,546,768 (1992).
24. Shaw, J. M., Overview of polymers for electronic and photonic applications. In
Polymers for Electronic and Photonic Applications, ed. C. P. Wong, Academic
Press, New York, pp. 1--65, 1993.
25. Heuberger, A., Lithography production techniques for the next generation of
microelectronics. In Proc. 3rd European Particle Acceleration Conference, 1, ed.
H. Henke, H. Homeyer and C. Petit-Jean-Genaz, Frontieres, Gif-sur-Yvette,
1992, pp. 192--6.
26. Dammel, R., Diazonaphthoquinone-based Resists, SPIE Optical Engineering
Press, Washington, 1993.
27. Hanabata, M., Furuta, A. and Uemura, Y. Proc. SPIE Advances in Resist
Technology, 68, 76-82, 1986.
28. Templeton, M. K., Szmanda, C. R. and Zampini, A., Proc. SPIE 771, 1987,
136--47.
29. Reichmanis, E., The chemistry of polymers for microlithographic applications.
In Polymers for Electronic and Photonic Applications, ed. C. P. Wong, Academic
Press, New York, pp. 67-117, 1993.
30. Trefonas, P., Daniels, B. K. and Fischer, R. L., Solid State Techno/., 30, 131,
1987.
31. Reichmanis, E., Po/ym. Mater. Sci. Eng., 66,36-37, 1992.
32. Oowaki, Y. et al., IEEE-ISSCC Tech. Digest, paper 6.6, 1991.
33. Lamola, A. A., Szmanda, C. R. and Thackerey, J. W., Solid State Technol., 53,
1991.
"--_ _ _ _ _ _ _ _R_E_F_ERE_N_C_E_S_ _ _ _ _ _ _------', I 241
34. Ito, H. and Wilson, C. G., Polymers in Electronics, ACS Symp. Ser. Vol. 242, ed.
T. Davison, 11, 1984.
35. Schwalm, R., Polymeric material science and engineering, Proc. ACS Div.
Polymeric Mat. and Eng., 61, 278, 1989.
36. Frechet, J. M. J. et al., Polymers in Microlithography, eds. E. Reichmanis, S. A.
MacDonald, T. Iwayanagi, ACS Symp. Ser., 412, 74, 1989.
37. Schlegel, L. et aI., lpn. l. App. Phys., 28, 2114, 1989.
38. Feely, W., Imhof, J. C. and Stein, C. M., Polymer Eng. Sci., 26, 1101, 1986.
39. Neenen, TX. et al., Proc. SPIE, 2, 1086, 1989.
40. Bohland, J. F. et al., l. Photopol. Sci. Technol., 3, 355, 1990.
41. Kunz, R. R., Allen, R. D., Hinsberg, W. D. and Walraft', G. M., Proc. SPIE
Advances in Resist Technology and Processing, 10, 167-175, 1993.
42. Lawes, R. A., Submicron lithography for semiconductor device fabrication. In
Photochem. Process Electron. Mater., ed. I. W. Boyd and R. B. Jackman,
Academic, London, 1992.
43. Morgan, C. and Chen, G. S., Phys. World, 5, 28, 1992.
44. Bowmer, T. N., O'Donnell, J. H. and Well, P. R., Polym. Bulletin, 2, 103, 1980.
45. Bowden, M. J. et al., l. Electrochem. Society, 128, 1304, 1981.
46. Novembre, A. E. et al., Proc. SPIE, 1466,89, 1991.
47. Bowden, M. J. S. and Gozdz, A. S., European Patent 0,248,021 (1992).
48. Thompson, L. F., Feit, E. D. and Heidenreich,R. D., Polymer Eng. and Sci., 14,
259,1974.
49. Feit, E. D. and Stillwagon, L., Polymer Eng. and Sci., 20, 1058, 1980.
50. Hofer, D. C. et aI., Appl. Phys. Lett., 37, 314, 1980.
51. Allen, R., Semiconductor International, 10, 72-80, September 1997.
52. DeJule, R., Semiconductor International, 10,85-92, September 1996.
Ferroelectric Polymers 8
JOHN A. CHILTON
Bosworth College, Nazareth House, Barrack Road,
Northampton, UK
8.1 INTRODUCTION
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
244 II FERROELECTRIC POLYMERS
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~
8.1.1 Ceramics
PZT is derived from the solid solution of two perovskite materials,
tetragonally distorted lead titanate (PbTi03) and rhombohedral lead
zirconate (pbZr0 3). The anisotropic nature of this material, especially at
the morphotropic boundary between the tetragonal and rhombohedral
phases, makes it a particularly strong piezoelectric.
The performance of PZT as a transducer material is none the less limited
by the conflicting nature of its piezoelectric and dielectric properties:
1. In hydrostatic mode, the induced electric polarization per unit strain (dh )
is determined by the tensor elements d33 and d31 • The piezoelectric strain
coefficient is given by
(8.1)
and the large negative value of d31 (-170 pCN- I ) almost completely
cancels d33 (450 pCW I ).
2. The piezoelectric voltage coefficient (gh) is given by
dh
gh=- (8.2)
Sollr
This gives the electric field produced in the material, per unit stress
and the large dielectric constant of the material (1800) produces a very
low gh.
Further, the brittleness, high density and fabrication problems have limited
the exploitation of ceramic materials. Similar electrical properties in a
polymeric material would give considerable practical advantages for many
applications. The search for a superior material has considered both pure
polymers and polymer composites, where an active filler is used to produce
the electrical properties.
8.1.2 Polymers
Investigations into piezoelectric polymers extend as far back as the 1920s
when it was discovered that certain organic materials such as rubber and
celluloid could be rendered piezoelectric by cooling in an electric field. I
Investigations have also found detectable piezoelectricity in wood, bone and
tendon, DNA, polysaccharides and the polypeptide, poly-y-methyl-L-gluta-
mate. Small, unstable piezoelectric effects in commercially available
synthetic polymers such as polystyrene, polypropylene and polymeth-
ylmethacrylate were reported in 1968. However, the major breakthrough
came in 1969 with the discovery by Kawai 2 that stretched and poled
INTRODUCfION II
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~ 245
8.1.3 Composites
Piezoceramic/polymer composite materials are a means of overcoming the
limitations of both ceramic and polymer materials to provide complemen-
tary properties which produce a superior piezoelectric device. The polymer
phase lowers the density of the material providing a better acoustic coupling
to water as well as a more easily adjusted buoyancy than that obtained for a
high density homogeneous ceramic. The low dielectric constant of the
polymer phase effectively increases the gh coefficients and the figure of merit.
The high elastic compliance of the polymer provides a better resistance to
mechanical shock and exhibits the high damping required for a good passive
device. The polymer's flexibility allows the composite to form to any
reasonable profile. Finally, the piezoelectric properties of the ceramic are
easily adjusted within the composite as they may be tailored via the
connectivity patterns (discussed later) to impart the maximum piezoelectric
response to the unit.
246 I LI______________F_E_R_R_O_E_L_E_C_T_R_I_C_P_O_L_y_M__E_R_S____________~
8.1.4 Overview
This chapter will initially be concerned with intrinsic ferroelectric polymers
and then move on to ferroelectric polymer composites. Texts are available
for a more detailed discussion of these and ceramic materials. 7 Material
preparation and properties are discussed along with a comparison of the
electroactive properties. After applications, a final brief section introduces
so-called smart materials. These have the potential to revolutionize both the
properties of devices we presently manufacture and the philosophy we use in
the design process.
Key
® Carbon
® Hydrogen
® Fluorine
(a) (b)
Fig. 8.1. Schematic of the two most common chain conformations of PVDF: (a) IX-
phase (b) p-phase. Arrows indicate -CF2 dipole directions defined by the carbon
backbone.
rendering the PVDF film piezoelectric is known as poling and involves the
application of an electric field. This preferentially orients the crystallites in
the direction of the poling field, thus producing a net polarization.
In the copolymer of vinylidene fluoride and trifluoroethylene, the
increased number of comparatively large fluorine atoms prevents the
formation of the tg + tg~ conformation by replacing hydrogen atoms on
adjacent carbon atoms. This forces the molecular chains to crystallize
directly into the polar fJ-form. The copolymer film then undergoes a poling
stage similar to PVDF to make it fully piezoelectric.
8.2.3.1 Introduction
Both PVDF and its copolymer with TrFE are thermoplastic and as such can
be easily processed in conventional thermoforming equipment. PVDF has a
combination of processability, mechanical strength and chemical resistance.
Because of this it has found a niche in the speciality polymers market
ELECTROACTIVE POLYMERS
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ II
~ 249
Uniaxial Film
The most straightforward method of mechanically producing a thin film is
to stretch it in one direction only. In the laboratory this can be done by
applying a load to one edge of a preformed sheet held at elevated
temperature. The load and the temperature are adjusted to produce
elongation, typically about four times the original length.
Commercial production methods involve passing a continuous band of
polymer through two pairs of rollers rotating at different speeds, in some
variants of this process the polymer strip passes over two single rollers. If the
second pair (or single) roller is rotating faster than the first then a stretching
force is applied to the polymer, the amount of stretch achieved depending on
the relative rotational speed of the two rollers. As with laboratory
stretching, the polymer is held at a suitable elevated temperature during
the operation. Material produced by stretching along one axis is termed
uniaxial. The polymer is oriented by this stretching and the mechanical
properties in the stretch direction are very different from those in the in-
plane direction perpendicular to it. Properties, especially tensile strength, are
enhanced in the stretch direction although this is accompanied by a
transverse strength reduction.
Biaxial Film
Film which is as strong as possible in both directions is often required. This
has led to the development of techniques for stretching in two mutually
perpendicular in-plane directions. Typically, a combination of longitudinal
250 I IL-_____________F_E_R_R_O_E_LE_C_T_R_I_C_P_O_L_YM
__E_R_S____________~
stretching using rollers followed by tentering where the edges of the film are
gripped in clamps running in divergent tracks is used. Specialized machines
have also been made which stretch film in two directions simultaneously.
Film produced by stretching in two directions at right angles is said to be
biaxially stretched. Strength is enhanced in all in-plane directions, although
not to the extent achieved in a single direction with uniaxial stretching.
The salient feature of these two methods of film production is that both
the uniaxial and the biaxial films are oriented in the way necessary to allow
the production of electrically active ferroelectric material by poling, so that
a single process not only produces a thin film, but one which is suitable for
conversion into an electrically active form.
Film Casting
Thin films can also be made by flowing a solution of a polymer onto a
smooth surface. After the solvent has evaporated the film is stripped from
the casting surface. Film thickness, can be controlled by adjusting the
viscosity of the polymer solution (by varying its concentration) and by
defining the depth of the deposited solution through the use of a roller or
sharp movable edge (,doctor blade'). Both PVDF and its copolymer with
TrFE can be successfully cast into thin films from ketone solutions. In the
case of PVDF, the resulting film is not oriented although it has been
reported that oriented material can be produced under some conditions
using specific solvents. 14 The copolymer film produced from ketone solution
is oriented and needs no stretching stage before poling. This is a
considerable practical advantage over PVDF.
8.3 POLING
40
~ :
..... ..
.-.. 38
-:
Z
U
~ 36
::l
•
"C
34
32
•
10 20 30 40
PC?lymer
Grounded
healable
plale
Product Property
A product property utilizes different properties in the two phases to produce
a third property through the interaction of the phases. For example, a
magnetoelectric composite can be made from a ferroelectric and a
ferromagnetic phase. When a magnetic field is applied to the composite,
the ferrite grains change shape (due to magnetostriction) and the strain is
passed onto the piezoelectric grains, resulting in an electrical polarization. 16
8.4.2 Connectivity
The physical properties of a composite can change by many orders of
magnitude depending on the way in which phase connections are made. As
such, connectivity is a major feature in property development for multiphase
composites. Each phase may be self-connected in zero, one, two or three
dimensions producing the ten diphasic connectivities (illustrated in Fig.
8.4): 0-0, 1-0,2-0,3-0, 1-1,2-1,3-1,2-2,3-2 and 3-3. For example, a 2-1
connectivity pattern has one phase self-connected in two-dimensional layers
while the other is self-connected in one-dimensional chains (or fibres). The
connectivity patterns are not geometrically unique. In the case of a 2-1
pattern the fibres of the second phase may be perpendicular to the layers of
the first phase (as in Fig. 8.4) or they may be parallel.
By convention, in electro active composites, the first number refers to the
active phase and the second, the inactive phase.
Some of the more important diphasic connectivities will now be discussed.
•••
••••
Fig. 8.4. Connectivity patterns in diphasic composites.
volatilized out by very slowly heating the pressed pellet to 400°C. After
sintering, the cold highly porous pellets are vacuum impregnated with a
suitable polymer: either a flexible silicone rubber elastomer or a stiff
epoxy resin. The finished product contains PZT regions ranging from 20-
120 11m that are randomly interconnected within the composite.
Both these composite types and the ones discussed below are shown
schematically in Fig. 8.5.
D=2 (8.6)
A
Combining equations (8.4) and (8.5) gives
BoSrV
D=--= BoSrE (8.7)
t
In linear dielectrics, D is directly proportional to E, increasing and
decreasing linearly with E. In ferroelectric polymers, such as PVDF and
PROPERTIES OF FERROELECTRIC MATERIALS
~----------------------------------------------------~
II 259
-50 so
the VDF:TrFE copolymers, the surface charge density increases with field in
a non-linear manner and exhibits a hysteresis when the field is decreased,
with a significant amount of surface charge remaining when the field is
completely removed. Subsequent applications and reversals of the electric
field produce the characteristic D-E hysteresis loops shown in Fig. 8.6.
The amount of surface charge density that remains is termed the remnant
polarization P r and is caused by a net alignment of dipoles within the
crystalline phase of the material. When the direction of the electric field is
reversed the charge rapidly decreases and changes sign. The field at which
the polarization passes through zero is termed the coercive field Ee and the
time taken to switch from positive to negative polarization is called the
switching time "s.
Measurement techniques are based on a circuit developed by Sawyer and
Tower in 1930?2 For PVDF and the VDF:TrFE copolymers, typical values
for Pr and Ee are in the region of 5 JlCm-2 and 50 kVmm-I respectively.
Furukawa et al. 23 have shown that for PVDF, switching times of a few
microseconds can be obtained at room temperature with an applied field of
a few hundred kV mm- I . Further investigations into VDF:TrFE24 have
found that for a 65:35 molar composition at 20°C, "s decreases from I ms to
100 ns as field strength increases from 80 to 400 kV mm-I .
(8.l0)
Vo= (8.12)
(8.13)
kB is Boltzmann's constant (1.38 x 10-23 J K-1) and T the absolute
temperature of the material. Thus the value of tan(c5e) quantifies the
amount of electrical energy dissipated into thermal energy. This is an
important consideration when designing and evaluating piezoelectric
transducers and normally needs to be as low as possible.
Values of tan(c5 e) can be calculated from conductance measurements or
obtained directly using modern capacitance bridges. Polymers are compar-
atively lossy dielectrics. Typical room temperature values of tan(c5e) for
PVDF range from 0.015 to 0.020 at 1 kHz and for VDF:TrFE from 0.015 to
0.025, compared with values of 10-4 for some ferroelectric ceramics such as
lithium tantalate and lead germanate. Composite materials tend to have
tan(c5e ) values somewhere between the values for the two pure phases.
60 80 100 20 40 60 80 100
20 40
Temperlture ('C) Temperature ('C)
50
- - Dielectric constant
.M........ Dielectric losl
16
.or- _Heltlng 14
12
~ 30
~j f-~~ 80:20 10 g
I 20
VDF:TrFE ~.' j
.j! / I :S
j ~"./ j!
'V 10 ........ .......! ..)",......................... .91
............ 0
Fig. 8.7. Temperature dependences of Er and tan(b) for 56:44, 70:30 and 80:20 CDF:TrFE copolymers.
PROPERTIES OF FERROELECTRIC MATERIALS
L -____________________________________________________~ II 263
above, the polymers become flexible and elastomeric. Brittle, rigid polymers
have a Tg above room temperature, for example, polystyrene has a Tg of
100°C. Rubbery or elastomeric materials have a Tg below room tempera-
ture. This is the case for PVDF and the copolymers which have Tg in the
region of -40°c.
The mechanical properties of polymers show significant changes at
temperatures close to the glass transition. In some cases, this amounts to a
decrease in elastic modulus by a factor of over 1000 times. This effect on the
mechanical properties will be passed onto the piezoelectric properties of
PVDF and VDF:TrFE through the relationships described in later sections.
The glass transition temperature can be determined by measurement of
elastic constant as a function of temperature using dynamic mechanical
thermal analysis (DMT A).
Material p s. F. FD Ref
(p.G m-1JCi) (MJ m-3 JCi) (m1 ~i) (fm 1 !Vi)
d=!!.=~ (8.14)
T E
There are three additional piezoelectric constants; the g constant gives the
field produced by a stress and is useful in determining the voltage output of
a stressed material; the other constants, e which relates T to E and h which
relates S to E are only occasionally used. As T and S are related by the
elastic constants c of the material and E and D are related by the dielectric
constant, the piezoelectric constants are interrelated to one another.
Therefore
L-_______P_R_O_P_E_R_T_IE_S__
O_F_F_E_R_R_O_E_L_E_CT
__R_IC
__M_A_T_E_R_I_A_L_S______ ~I I 265
d e
- = - = BoBr (8.15)
9 h
e h
-=-=C (8.16)
d 9
The fonnal definition and derivation of these piezoelectric relationships may
be found in Berlincourt et al. 29
The piezoelectric constants are tensor quantities, hence it is necessary to
assign a coordinate system to produce a matrix of coefficients which relate to
the symmetry of the crystal axis.
Anisotropy is introduced into ferroelectric polymer films by the poling
process which aligns the dipoles in the direction of the poling field. By
convention, this axis is known as the '3' direction, as illustrated in Fig. 8.8.
For PVDF, an additional anisotropy is caused by the mechanical stretching
procedure as discussed in section 8.2.5. The 'I' axis is parallel to the stretch
direction and the '2' axis is perpendicular to this in the plane of the film.
Thus, for polymer films with symmetry type 2 mm (i.e. uniaxially stretched
PVDF) the components of the piezoelectric charge coefficient are:
J= 2 3 4 5 6
0 0 0 0 dl 5 0
dij = (8.17)
2 0 0 0 d24 0 0
3 d31 dn d33 0 0 0
J;
\O~
'0<:>
~~~~<t
~~J>~
#'...~
~2
Fig. 8.8. Conventional of axes for electrical and mechanical coefficients.
266 I LI______________F_E_R_R_O_E_L_EC_T_R_I_C_P_O_L_Y_M_E_R_S____________~
where the i subscript denotes the direction in which the charge is generated
and the j subscript denotes the direction of applied stress (1 to 3 for uniaxial
stresses and 4 to 6 for shear stresses). The matrix for ferroelectric polymers
with symmetry type 2 mm, for example VDF:TrFE copolymers and
biaxially stretched PVDF, is similar except the '1' axis is equivalent to the
'2' axis, thus d15 = d24 and d31 = d32 .
The coefficients diS and d24 associated with shear deformation are seldom
used as there is a lack of practical applications using the shear mode. The
most important piezoelectric coefficients for determining a material's
suitability for a particular device are the d3j and g3j components and the
hydrostatic coefficients dh and gh. The dh coefficient is a measure of the
charge generated due to application of a hydrostatic pressure. As the stresses
are uniform in all three directions dh is simply the sum of all d3j •
dh = d 31 + d 32 + d 33 (8.18)
The d33 coefficient describes the charge generated through the thickness of
the film ('3' direction) due to a stress applied along the '3' axis. As it is
difficult to stress a thin film in the thickness direction without contributions
from d31 and d32 , the calculation of d33 is often done using equation (8.18) and
measured values of d3h d32 and dh. However, it is possible to obtain d33 via the
converse piezoelectric effect by using optical techniques, e.g. interferometry30
to measure the minute changes in thickness (less than a micron).
The transverse coefficients d31 and d32 can be obtained using compara-
tively simple strain removal techniques, with generated charge passing
through a capacitor to allow the subsequent voltage to be measured on a
digital storage osciIIoscope. 31
The dh coefficient can be recorded by subjecting the sample to a
hydrostatic pressure. Methods of achieving this include enclosing the sample
in a uniform, low frequency (50 Hz 32 or 80 Hz 33) sound field and
application of a I atmosphere (0.1 MPa) pressure pulse by rapid air ingress
of an evacuated chamber. 31 In all cases, it is necessary to immerse the test
sample in silicone fluid or oil to stabilize the temperature and eliminate
compressional adiabatic heating effects. The gh coefficient, calculated using
equation (8.16), is important in comparing materials for hydrophones,
particularly in the hydrophone figure of merit product FOM h
FOMh = dhgh (8.19)
tan(be )
A comparison of piezoelectric coefficients between uniaxially and biaxially
stretched PVD F, the VDF:Tr FE copolymers and PZT 5 is shown in Table 8.2.
It is clear that the ceramic has very high piezoelectric charge coefficients
(d) when compared to the polymers, but as Er is over 100 times greater for
the ceramic (1700 compared to 12), g is an order of magnitude lower.
Therefore, ferroelectric polymers will produce a higher voltage output per
unit applied stress. However, this advantage is balanced out when figure of
Table 8.2 Comparison of piezoelectric properties
k _ d33
(8.21b)
33 - [ T E] 1/2
e s33
where s is the compliance, i.e. the ratio of strain in the direction annotated
by the first subscript to stress in the direction shown by using the convention
in Fig. 8.8. The superscript E indicates this is the value under short circuit
(constant electric field) conditions.
There are also another two coupling factors, kp, for the radial vibration in
the plane of the sample, and k t for the vibration of thin specimens along the
polarized direction, where
2
kp = k31 1/2 (8.22a)
[1 - CTE]
L -_ _ _ _ -..l1 I
P_R_O_P_E_R_T_IE_S_O_F_F_E_R_R_O_E_L_E_CI'_R_IC_M_A_T_E_R_I_A_L_S_ _ _ 269
5
kt = h33 B~3 (8.22b)
c33
and (J is the Poisson's ratio under short circuit conditions, h33 is the
piezoelectric constant in equation (8.16) and C33 is the ratio of stress to strain
under open circuit (constant electric displacement) conditions.
Values of the coupling factors are determined by the piezoelectric
resonance method. This analyses the impedance Z of the material when it is
excited by a voltage source and uses the resonant frequency at which Z = 0
and the anti-resonant frequency for which Z is infinite, to calculate the
appropriate k.35 This method is satisfactory for ceramics, but large errors
occur when applied directly to the polymers because of their large
mechanical losses and subsequent high damping effects. It is necessary to
use equivalent circuit models to select the parameters that produce
resonance curves that are best fitted to observed ones. 36
The most practical coupling factor is kt> which is important in evaluating
a material for applications such as ultrasonic imaging that exploit the
converse piezoelectric effect to produce sound waves. The values of k t for
PVDF (0.2) and VDF:TrFE (0.3)37 are approximately half that of PZT
(0.55). Thus for the same driving voltage, the ceramic will be a superior
transmitting material. However, the peak driving voltage that can be
applied without damaging the material is 25 V pm-I for the polymers and
0.3 V pm-I for ceramics, almost 100 times greater. The smaller coupling
factor can therefore be compensated for by using a larger driving voltage,
although in some applications, high voltages are undesirable.
R - YIP (8.31 )
v - Hvt(CE + CA)W
If the capacitance of the element is larger than the amplifier's capacitance
(CE > CA ) then
(8.32)
*vA
D =NEP
(units mHzl/2 W- I ) (8.37)
allows a comparison of detectors with different areas when the dielectric
noise dominates, which is true for many devices. The D* will then become
272 I ~I______________F_E_R_R_O_E_L_EC_T_R_I_C_P_O_L_YM
__E_R_S____________~
D* =RvVA (8.38)
Vn
On substitution of Rv and Vo from equations (8.32) and (8.12) respectively,
a further figure of merit Fo is derived, showing that D* will be optimized by
maximizing Fo where
Fn= p (8.39)
BvVereo tan{be )
However, this does not apply for small detectors operating at very low or
very high frequencies, as other noise sources dominate. In this case, the
detectivity will be maximized by maximizing Fv in equation (8.33).
All pyroelectric materials are piezoelectric. Thermal expansion of the
ferroelectric polymer induces a small secondary pyroelectric effect due to the
piezoelectric coefficients coupling to the strains produced on expansion.
Another undesirable signal is generated when the pyroelectric detector
operates in an environment that is acoustically noisy or highly vibrational.
This microphonic signal can be significantly reduced by mounting the
detector in a rigid, mechanically isolated package.
A typical technique for obtaining values of p involves the use of a
chopped radiation source, such as a laser beam or an infra red lamp. The
measured charge is then compared with a known calibration sample.
Values of p, Bv and the figures of merit Fv and Fo for the ferroelectric
polymers are shown in Table 8.1 against values for a typical pyroelectric
ceramic based on doped lead zirconate. 40 The pyroelectric constants of the
polymers are at least an order of magnitude lower than the ceramic and the
smaller Fo indicates an inferior detector performance. However, in terms of
cost, the polymers offer an advantage as they are readily available in large
area, thin sheets, eliminating slicing, lapping and polishing required by the
ceramics. It is interesting to note that the increase in piezoelectric coefficients
exhibited by the copolymers does not manifest itself in the pyroelectric
properties, although a room temperature value of 90 p.C m-2 K- 1 has been
reported for Japanese 51:49 VDF:TrFE copolymer. 41
28
24
20 - - - - - 80:20
---------- 70:30
---56:44
16
12
- ...... -- .............. -_ ............ -_ ......... -_ ............................. -_ ..............................
13 % loss
...
8 _______________________ _
13 % loss
6S % loss
2 6 10 14 18 22
(8.42)
genous stress and strain. This condition is satisfied when the sample length is
much larger than its width and thickness.
The simplest technique for measuring Yll is to apply a tensile force F
parallel to the '1' axis to a specimen of areas A and length I. The Young's
modulus is proportional to the initial gradient of the resultant stress-strain
curve. Thus
F/A
Y= M/I (8.43)
S31
0-31 =-- (S.47b)
Sl1
S32
0-32 =-- (S.47c)
S22
8.6 APPLICATIONS
8.6.1 Introduction
The discovery by Kawai 2 in 1969 of the strong piezoelectric effect in PVDF
and two years later by Wada45 and Bergman46 of a corresponding pyro-
electric response were important developments for the field of electrome-
chanical and pyroelectric transducers. The enormous potential for a light,
rugged, large area flexible material with piezo- and pyroelectric properties
was quickly realized and it inspired a world-wide effort on applications. The
low mass of these materials is particularly important in aerospace and space
applications, especially with the constraints placed on desigOers of satellites,
probes and a possible space station.
The need for a combination of desirable properties that often cannot be
obtained in single-phase materials has driven the development of composite
materials. For example, in an electromechanical transducer, one may wish
to maximize the piezoelectric voltage sensitivity, minimize the density (to
obtain good acoustic matching), and also make the transducer mechanically
flexible to conform to a curved surface. These properties are partially
conflicting in nature, and a single-phase material that simultaneously
'----_ _ _ _ _ _ _ _AP_P_L_IC_A_TI_O_N_S_ _ _ _ _ _ _ ----'I I 277
(8.48)
~= 1 (8.52)
1+ C,+Cc
Ce(l-jlan(b))
The predicted noise spectra for a hydrophone is chosen in Fig. 8.11. It can
be seen that the sea state zero noise dominates at low frequencies and the
amplifier noise dominates at high frequencies. So for most practical
applications gh is the best figure of merit.
10-6
-7
10
10-8
~ ro
10 ~....1-L..L.LLI.I.Ir---1.--L.L.LLLLL,--..L...l....L.L.1.ICJJ.,. 01}
104
Frequency (Hz)
Fig_ 8.11. Predicted hydrophone noise spectra: (I) sea-state zero; (II) Johnson noise
of dielectric loss; (III), (IV) voltage noise of two existing pre-amplifiers. Materials
constants were as follows: Ilr = 44, tan(l5) = 0.032 + 0.47Vl + 1.4/- 1(f = 20 Hz-
10 kHz), dh = 13.8 pCN- 1 , gh = 0.035 V mM- 1 •
(8.53)
This final section discusses what have been described for aerospace and
space applications as the most important emergent materials technology
area. SMART materials are more like a device than an intrinsic material.
The term 'SMART' comes from the fact that 'these materials sense some
effect and respond in some way as a result. In this way, they have both
'sensor' and 'actuator' properties. They are often modelled on biological
systems where the sensor/actuator/process parts are likened to the nervous
system/muscles/brain. This is a graphic analogy but not always very useful.
There is a suggestion that the 'brain' of the SMART structure is some form
of processing unit. Although this may be true, the most interesting area for
materials is where all three functions are performed intrinsically by the
material itself. Certainly, if a SMART structure consists of a sensor
connected to an actuator by some electronic processing unit, then the
structure is better described as a 'device' rather than a pur,e material.
Perhaps the simplest SMART materials can be used for vibration control.
This is very important in space applications where it is important to keep
satellite devices under tight control in order to work correctly. The SMART
structure can be a simple sensor adhered to a structure and electrically
connected to an actuator. The sensor could be a piece of 0--3 composite
(with lead titanate as the active material) and the actuator could be a PZT
disc. Any vibration produces an electrical signal in the composite and this is
fed to the sensor through an inverting amplifier circuit so that the PZT is
driven in antiphase. With the correct gain set, the vibration can be removed.
L -_ _ _ _ _ _ _ _ _ S_M_A_R_T_M_A_T_E_R_I_A_L_S_ _ _ _ _ _ _ _ -----'I I 285
Such a system as the one described above would have problems operating
over a wide range of frequencies because the composite, PZT and amplifier
circuit all have a frequency response and so the optimum gain would depend
on the type of frequencies expected for a particular operation.
The above principle is the basis of active vibration cancelling systems.
Such systems have been researched by many automotive companies for the
reduction of noise in the cavity (i.e. where the driver sits) and for the
production of a superior active suspension system.
One of the problems of such systems is that, the more complicated the
structure, the more care has to be taken to ensure that it is secure and
reliable over an extended period of time. For example, military specifica-
tions can often require 20-30 years in-service lifetimes. Reliability may
increase if the sensor, actuator and interface are intrinsically part of the
same piece of material. This could be achieved easily with 0-3 composite
material in the form of a disc. The central area would be the sensor (using a
lead titanate ceramic) and a ring of active material would surround this
(using a PZT ceramic). If the polymer matrix were the same for both
powders then the two regions of the material would be compatible and
integrity should therefore be high. Such a structure has been used to actively
cancel noise for underwater military applications. In such a system, the
tuning of the sensor and actuator parts would be carried out by controlling
the poling efficiency for the two different parts of the structure.
The above describes a basic form for SMART materials. A sensor and
actuator are coupled together either via electronics or through the material
itself. For the interested reader, there are many papers discussing specialized
materials incorporating different aspect of intelligence, most notably in
Japan and America. 70
Sensors can be based on acoustic waveguides,71 the piezoelectric effect72
or optical fibres. 73 The mechanism depends on the input to be sensed. The
actuator mechanisms fall within five areas:
1. Shape memory alloys. These are metals that revert to a predefined shape
above a certain temperature. In doing so, they can generate large forces
for actuation. The most prominent material is possibly nickel-titanium
alloys. The main drawback is the slow response of these materials. The
response depends on heating and cooling and so, the response depends
on how fast the temperature can shift.
2. The piezoelectric effect. These materials produce a much smaller stress
than shape memory alloys but they react much quicker. One of the
problems is that piezoelectric actuators are non-linear and hysteretic.
3. Electrostrictive materials. The electrostrictive effect is a result of the
quadratic dependence of the material permittivity with electric field. This
is typified by lead magnesium niobate - it is non-hysteretic but also non-
linear. No poling is required but a bias voltage may be necessary.
286 II FERROELECTRIC POLYMERS
~----------------------------------------------------~
8.8 CONCLUSIONS
This chapter has provided the reader with an introduction to a unique class
of materials. From the discussions of properties and applications, it can be
concluded that electro active materials offer advantages over many alterna-
tive technologies, including flexibility, robustness and availability in large
areas, thin sheets or unusual geometries.
However, it is widely believed that these materials will become established
in applications that exploit their unique features, rather than through the
substitution of existing transducer materials. This requires an innovative
approach, which on examination of recent literature, appears to have
already begun and looks set to continue.
Table 8.4 gives some of the points to consider when designing a system
which will utilize an active material. It is observed that no one material is
suitable for all applications. Further, many disadvantages can be engineered
around with good system design.
The present position is such that technological and industrial advances
demand more versatile and responsive transducer devices. In the develop-
ment of such, a particular technology may well be used over others but each
individual application is considered on its own merits and in some cases, a
mixture of materials may well suit.
Table 8.4 Comparison of different electroactive materials
Electroactive material
Very high active performance (d 33 ) Low d 33 produces a low active Reasonable acoustic coupling
performance
Very poor hydrostatic activity Large d 31 produces only a Adjustable buoyancy
reasonable hydrostatic
performance
Large e,. reduces gh Low T c leads to a strong Low Gr leads to a high gh
temperature dependence and
depoling problems
Brittle Low density leads to a good Good shock resistance
acoustic coupling
High density leads to poor Low Gr leads to a high gh Good formability (e.g. domes)
acoustic coupling
Fabrication problems Good shock resistance Piezoelectric properties are
easily adjusted
Shock resistance is very Very good formability - can be Reasonably low dielectric loss
dependent on microstructure formed into exotic shapes (e.g. (5%)
domes)
Only simple structures possible Stretching can be problematic er around 100-200 - good for
(e.g. discs) for PVDF many applications
Very low dielectric loss (~3%) Relatively high dielectric loss Interphase interactions and
(15%) defects can cause problems
High Gr leads to a high Low Gr (",10) which is bad for
capacitance - good for some applications
hydrophone applications
288 II~______________F_E_R_R_O_E_L_EC_T_R_I_C_P_O_L_Y_M_E_RS______________~
8.9 ACRONYMS AND SYMBOLISM
8.9.2 Symbolism
C = capacitance
8 = permittivity
8 0 = permittivity of free space
Q = charge
E = electric field strength
tan( c5e} = dielectric loss
w = angular frequency
kB = Boltzmann's constant
i = current
S = strain
c = modulus
s = elastic compliance
p = pyroelectric constant
t = time constant
R = resistance or responsivity
F = figure of merit or force
Y = Young's modulus
p = density
S = noise
A = surface area
8 r = relative permittivity
t = thickness
V = voltage
D = electric displacement or detectivity
G = conductance
f = frequency
T = absolute temperature or applied tress
g = piezoelectric voltage coefficient
d = piezoelectric strain coefficient
k = electromechanical coupling factor
(J' = Poisson's ratio
L -___________________R_E_F_E_R_E_N_C_E_S__________________ ~I I 289
P = polarization
H = heat capacity
W = power
'1 = emissivity
I = length
M = sensitivity
OJ . = volume flow rate
Note:
T is used for both absolute temperature and applied stress
R is used for both resistance and responsivity
F is used for both figure of merit and force
D is used for both electric displacement and detectivity
It is hoped that this will not confuse the reader.
8.10 REFERENCES
1.
Brain, K. R., Proc. Phys. Soc. (London), 36,81, 1924.
2.
Kawai, H., Jap. J. Appl. Phys., 8, 975, 1969.
3.
Yamada, T., Veda, T. and Kitayama, T., J. App. Phys., 52, 948, 1981.
4.
Lando, J. B. and Doll, W. W., J. Macromol. Sci. (Phys.), H2, 205, 1968.
Mather, S. c., Scheinbeim, J. I. and Newman, B. A., J. App. Phys., 56(9), 2419,
5.
1985.
6. Miyata, S., Yoshikawa, M., Tasaka, S. and Ko, M., Polymer J., 12(12), 857,
1980.
7. Special Polymers for Electronics and Optoelectronics, Chilton and Goosey,
Chapman & Hall, 1995.
8. Brydson, J. A., Plastics Materials, Butterworths, 1982.
9. Young, R. J., Introduction to Polymers, Chapman and Hall, 1983.
10. Lovinger, A. J., Science, 220 (4602), 1115, 1983.
11. Dukert, A. A., Soc. Plast. Eng. 18th Ann. Nat. Tech. Con!, Pittsburgh 1962.
12. Gebauer, P., Kunstoffe, 65, 356, 1975.
13. Stallings, J. P. and Howell, S. G., Poly. Sci. Eng., 11, 507, 1971.
14. Prest, W M. and Luca, D J., J. App. Phys., 49, 5042, 1978.
15. Bamji, S. S., Kao, K. J. and Perlman, M. M., J. Poly. Sci. (Phys.), 18, 1945,
1980.
16. vanden Boomgaard, J. et al., J. Mat. Sci., 9, 1705, 1974.
17. Safari, A. et al., Ferroelectrics, 41, 197, 1982.
18. Harrison, W. B. Proc. Wkshop Sonar Mat., NRL 1976.
19. Savakus, H. et aI., Mat. Res. Bull., 16, 677, 1981.
20. Williams, J. C., Doctor-blade process, in Treatise on Materials Science &
Technology, ed. Wang, F. Y., vol 9, p.l73, Academic Press 1976.
21. Yamashita, Y. et al., Jap. J. App. Phys., 20, 183, 1981.
22. Sawyer, C. B. and Tower, C. H., Phys. Rev., 35, 269, 1930.
23. Furukawa, T. and Johnson, G. E., App. Phys. Lett., 38, 1027, 1981.
290 I LI______________F_E_R_R_O_E_L_E_CT_R_I_C_P_O_L_Y_M__ER_S____________~
24. Tajitsu, Y., Masuda, T. and Furukawa, T., Jap. J. Appl. Phys., 26 (10), 1749,
1987.
25. Scaife, B. K. P., Complex Permittivity (ch 3), Hodder 1971.
26. Higashihata, Y., Sako, J. and Yagi, T., Ferroelectrics, 32, 85, 1981.
27. Yamada, T. and Kitayama, T., J. App. Phys., 52 (11), 6859, 1981.
28. Choy, C. L., Chen. F. C. and Luk, W. H., J. Poly. Sci. Poly. Phys., 18, 1187,
1980.
29. Berlincourt, D. A., Curran, D. R. and Jaffe, H., Physical Acoustics, Principles
and Methods, ed. Mason, W P., Academic Press, New York 1964.
30. Kepler, R G. and Anderson, R. A., J. App. Phys., 49, 4490, 1978.
31. Humphrey, K. J., Gamer, G. M., Shorrocks, N. M. and Whatmore, R. W., Proc.
6th IEEE Int. Sym. App. Ferr., 543, 1986.
32. Tancrell, R H., Wilson, D. T. and Rickett, D., IEEE utt. Sym., 624, 1985.
33. Humphrey, K. J., Gamer, G. M., Goosey, M. T., Whatmore, R W. and Ainger,
F. W., 1st Int. Con! Elec. Opt. Acoust. Prop. Poly., RAPRA, London, 1, 1988.
34. Yamada, T., J. App. Phys., 53(9), 6335, 1982.
35. IRE Standards on Piezoeletric Crystals, Proc. IRE,49, 1161, 1961.
36. Ohigashi, H., J. App. Phys., 47(3), 949, 1976.
37. Murayama, N. and Obara, H., Jap. J. App. Phys., 22 (22-3), 3, 1983.
38. Whatmore, R. W., Rep. Prog. Phys. (GB), 49 (12),1335,1986.
39. Porter, S. G., Ferroelectrics, 33, 193, 1981.
40. Whatmore, R. W. and Ainger, F. W., Proc. SPIE, 395, 261, 1983.
41. Yamazaki, R., Ohwaki, J., Yamada, T. and Kitayama, T., App. Phys. Lett., 39
(9), 772, 1981.
42. Gamer, G. M., Allen Clark Research Annual Review, 62,1975.
43. Humphreys, J., Lewis, E. L. V., Ward, I M., Nix, E L. and McGrath, J c., J.
Poly. Sci. (B) Poly. Phys., 26,141, 1988.
44. Lerch, R., J. Acoust. Soc. Am., 67, 1229, 1981.
45. Wada, Y. and Nakamura, K., J. Poly. Sci., A29, 161, 1971.
46. Bergman, J G., McFee, J H. and Crane, G R, App. Phys. Lett., 18, 203, 1971.
47. Tamura, M., Yamaguchi, T., Oyaba, T. and Oshimi, T. Y., J. Audio Eng. Soc.,
23,21 1975.
48. Lerch, R and Sessler, G. M., J. Acoust. Soc. Am., 76, 1379, 1980.
49. Gamer, G. M., Systems Technology, 27, 22,1977.
50. Micheron, F. and Lemonon, C., J. Acoust. Soc. Am., 64, 1720, 1978.
51. Sullivan, T. D. and Powers, J. M., J. Acoust. Soc. Am., 63, 1396, 1978.
52. Holden, A. J., Parsons, A. D. and Wilson, A. E. J., J. Acoust. Soc. Am., 73, 1858,
1983.
53. Ricketts, D., J. Acoust. Soc. Am., 64,555, 1978.
54. Gamer, G. M. et al., Ferroelectrics, 93, 169, 1989.
55. Ting, R. Y. et al., Proc. 10 Acoust., 9 (2), 153, 1987.
56. Pearman, G. T., Hokanson, J. L. and Meeker, T. R., Ferroelectrics, 28, 3111980.
57. Toda, M. and Osaka, S., Proc. IEEE, 67, 1171, 1979.
58. Toda, M., Fe"oeiectrics, 32, 127, 1981.
59. Starita, A., Basta, F., Carbone, B., Dario, P. and Perfetti, C. C., Medical and
Biological Eng. and Comp., 23 Suppl. Pt 1,420, 1985.
~_________________R_E_F_ER_E_N_C_E_S________________~I I 291
60. Long Shen, Y., De Rossi, D., Dario, P. and Galletti, PM., Life Support Systems,
4,239, 1986.
61. Vematsu Y. and Suzuki, M., Proc 10th ISIR, 571, 1980
62. Fukada, E., Mechanisms of Growth Control, ed. Becker, R. O. and Thomas, C.
C., Springfield 1981.
63. Ficat, F. F., Durroux, R., Faruan, M. J., Escourru, G., Fiscat, P., Micheron, F.
and Lacabanne, C., Current Advances in Skeletogenesis, ed. Silberman, M. and
Sierkin, M. C., Exerpta Medica, Amsterdam, Int. Congr. Ser. 1982.
64. Sharpe, R., Phil. Trans. R. Soc. Lond., A292, 163, 1979.
65. Murali Dhar, C. et al., J. Mat. Sci. Lett., 6, 33, 1987
66. Glass, A. M., McFee, J. H. and Bergman, J. G., J. App. Phys., 42, 5219, 1971.
67. Pfister, G. and Abkowitz, M. and Crystal, R. G., J. App. Phys., 44,2064,1973.
68. Burkard, H. and Pfister, G., J. App. Phys., 45, 3360, 1974.
69. Southgate, P. D., App. Phys. Lett., 28, 250, 1976.
70. Rogers, C. A., Sci, Am., Sept, 154, 1995.
71. Harold, T. et aI., SPIE, 1170, 1989.
72. Crawley, E., SPIE, 1170, 1989.
73. Udd E. et al., SPIE,838, 1987.
74. Humphrey, K J., Gamer G M. and Whatmore, R W., Ferroelectrics, 76 (3-4),
383, 1987
Polymers in Printed Circuit 9
Board (PCB) and Related
Advanced Interconnect
Applications
MARTIN GOOSEY
Shipley Europe Ltd, Coventry, England
9.1 INTRODUCTION
The first printed circuits were developed over 50 years ago in the United
States during the Second World War and were designed to reduce weight,
save space and increase reliability in military equipment. This new
technology replaced discrete wiring of electrical components which was
previously accomplished by the use of terminal boards and point-to-point
wiring and soldering. With the widespread commercialization of the
transistor from the 1950s onwards the demand for printed circuit boards
began to expand significantly. The world consumption of printed circuit
board laminate materials for 1998 has been estimated at around 225 million
square metres!
In recent years the term printed wiring has been used synonymously with
printed circuit and is technically more correct since a circuit also includes
various components and devices ('printed wiring board' is the term used in
the USA whilst in Europe 'printed circuit board' is preferred). A printed
circuit is a pattern of electrical conducting paths reproduced onto an
insulating medium designed to interconnect various devices (Fig. 9.1). There
Plastics for Electronics, 2nd edn. Edited by Martin Goosey_ Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
294 I ~I__________P_o_L_Y_M_E_R_S__IN__P_R_INT__E_D__c_IR~C_U__IT__BO__A_R_D________~
Fig. 9.1(a). Printed circuit board for a mobile telephone application (courtesy
Nokia).
Fig. 9.1(b). An assembled printed circuit board with both through hole leaded and
surface mount components.
are many methods used for defining these conductive patterns but they all
involve the definition of a circuit pattern, usually in copper, on a dielectric
material such as glass fibre reinforced epoxide or paper-loaded phenolics. In
this chapter the basic approach to printed circuit board manufacture and
testing will be briefly outlined before details are given of the specific
polymeric substrates commonly used in current PCB manufacturing
operations around the world. The final part of the chapter will then address
L -________ P_RI
__NT
__E_D_C_I_R_C_U_IT
__B_O_A_RD
___FA_B_RI __________~I
__C_A_T_IO_N I 295
some of the new advanced interconnection methodologies that are begin-
ning to be forced into existence by the emergence of increasingly
sophisticated and highly integrated semiconductor devices. These new
routes are particularly interesting in the context of this book because they
involve polymeric materials throughout the fabrication process and as
substrates and inter-metal dielectrics.
Printed circuit boards can vary in complexity from the most basic single-
sided boards used in low-cost consumer electronics to sophisticated fine
pitch multilayer boards sometimes containing up to 40 or more interconnect
layers. There are often many possible process variations for the fabrication
of the same board and the exact choice will be determined by numerous
factors such as cost limitations, production volumes, the manufacturing
facility and the specific application. The following text outlines a typical
route used but it is not intended to provide an exhaustive account. The
reader is therefore recommended to refer to more comprehensive texts on
printed circuit board processing such as that by Leonida J for a more
detailed description.
In the simplest case, a printed circuit board may have a single layer of
interconnection tracks defined on only one side of a suitable substrate. This
is known as a single-sided board and is typical of what would be found in
basic consumer electronics, for example, transistor radios and remote
control units. The circuit is defined in a thin layer of copper laminated onto
a suitable low-cost dielectric material such as a paper-filled phenolic using a
process known as print and etch. In this process, pieces of laminate that
have been cut to an appropriate size for processing are cleaned and coated
with an etch-resistant ink that will protect the areas where the copper tracks
are to remain. This resist can be screen printed onto the substrate or
alternatively a liquid or dry film photoresist could be used. Having cured the
ink or patterned the resist, the exposed copper is then etched away leaving a
pattern of copper under the protective resist. The organic resist is then
removed using a suitable solvent to expose the copper circuitry. In this case
the conductors are normally left as bare copper with no additional metal
plated on top. If there are special requirements for the board or if it needs to
have components reftow soldered onto it, an alternative method may be
used which allows the electrodeposition of a tin-lead based finish.
Additionally, after further cleaning of exposed copper in areas where there
will be components, a solderability preservative may be applied and the
holes, if required, for through hole leaded components, punched using a
suitable die. Although this is probably the most basic printed circuit
fabrication process, even here there are many process variables possible. For
296 I I~_________PO_L_YM
___E_R_S_IN
__P_R_INT ________~
__E_D__C_IR_C_U_I_T_B_O_A_R_D
example, there are options on when the holes are made, whether they are
punched or drilled and in the type and application methods of the final
solder mask protective coating and solderable finish, to name but a few.
The next type of board in terms of complexity is the double-sided board in
which the circuitry is confined to the outer surfaces of a single piece of
laminate (Fig. 9.2). In this process an additional procedure is normally
required to form interconnecting holes that connect the two sides of the
board. These are known as plated through holes (PTH) and are typically
provided by a combination of electroless copper followed by electroplated
copper to build up the required thickness. Plated through hole fabrication
involves drilling or punching the holes in the copper clad board at the
locations where the interconnections are required. This is followed by a
number of chemical processing stages that clean the hole walls and
condition them in preparation for metallization. Because the hole walls are
non-conducting they cannot be electroplated with copper directly and,
traditionally, an e1ectroless copper deposit is used to enable electroplating.
Recently, alternatives to electroless copper known as direct plate processes
have become popular and these employ a variety of conductive species such
as carbon and palladium sulphide to provide a conductive layer which then
facilitates direct electroplating of copper. Most double-sided boards have all
their conductors and hole walls made from copper with an additional thin
metallic solderable coating such as tin, tin-lead, nickel-gold or palladium.
Some of these coatings can also actually be used as etch resists in place of an
organic resist during manufacture.
Fig. 9.2. An actual section through a hole on a double·sided board. (Note the glass
fibre bundles in the central part of the section.)
L -_ _ _ _ -'I I
L_A_M_I_N_A_T_E_S_F_O_R_R_IG_I_D_P_R_I_NT_E_D_W_IR_I_N_G_ _ _ _ 297
Nearly all (rigid) printed wiring boards in current use are laminates made by
bonding together several sheets of fibrous reinforcing material which are
then surface clad with copper foil on one or both sides. The fibrous sheets
are usually in the form of woven fabric, mat or paper and made of glass,
often an electronic grade (E-glass), cellulose or synthetic polymers. These
are frequently referred to as base materials. The binding agents have
the use of a drill entry material. Another recent innovation in copper for
laminates is the production of so-called double-treated foils where both
sides of the foil have optimized morphologies. In these the shiny side of the
laminate is modified such that it is claimed to obviate the need for the
traditionally used oxide treatment which ensures good bond integrity. Other
recent innovations in copper for laminates include high ductility coppers
optimized for use in flex circuit applications and very low profile coppers
with improved elongation values and reduced roughness for fine inner layer
production (0.1 mm). In addition to the many different thicknesses of
copper available there is also a range of resin contents possible in the pre-
pregs used.
Multilayer boards are also assembled in a process very similar to that used
for making laminate where the individual layers of the board and prepreg
are heated under pressure in a press to give the final cured and bonded
assembly. In a recent development, the copper foil used in the board itself
has been employed to carry a high electrical current which causes heating in
the lay-up structure and curing of the assembly. Also, it is worth noting that
some companies use an autoclave, rather than a bonding press, to
manufacture multilayer boards. The vacuum autoclave offers a number of
advantages for certain specific applications. It is an isostatic vacuum press in
which nitrogen or carbon dioxide is used for both temperature and pressure
transmission.
Method ASTM
test no.
Electrode 1 Electrode 2
Electrode 3
Fig. 9.4. Electrode arrangement for surface and volume resistivity measurements.
304 I ~I__________PO_L_YM
___E_RS__IN
__P_RI_NTE
___D __B_O_A_R_D________~
__C_IR_C_U_IT
N
6
== normal range
;§ for laminates
'iii
;:
!! 5
;!l
8
j
is 4
3 -L------r------r------r------.----~
o 20 40 60 80 100
Resin content (wi. %)
b ......t - - - - - E -----t~
..
Fig. 9.6. Electrode arrangement for (a) dielectric strength (field perpendicular) and
(b) dielectric breadown measurements (field parallel).
LOAD
.1r---------------------------------------~~
.08
.02
o~~~~----~----~----~~--~----~
20 40 60 80 100 120 140
Temperaturef'C
300
U
]
,eo
a:>o 200
..
oW
1:1
.S:!
15
~
U 100
8
1
~
0
0 40 80 120 160 200
Temperaturef'C
Fig. 9.9. Z axis expansion coefficient versus temperature for a typical multilayer
epoxy laminate.
'--_ _ _P_O_L_y_M_E_R_S_U_SE_D_F_O_R_L_A_M_IN_A_T_E_M_A_NU_F_A_C_T_U_R_E_ _ _-'I I 309
Nearly all laminates suitable for printed circuit applications are based on
thermosetting polymer chemistry. Compared with the commonly available
thermoplastics, thermosets possess improved temperature performance and
chemical stability due to the crosslinking which occurs on curing to produce
a three-dimensional rigid chemical structure. The most important polymers
used for the fabrication of printed wiring laminates are now described.
C6HCH20!X-OCH2&CH2-JX-OCH2C~H2
X= -50H
y
Y
Hc~~
0 : 0
H
Y
CH3
H Y
where Y = H or Br. A more detailed description of the chemistry of these
and related epoxy compounds is given in Chapter 4.
The choice of DGEBA resins for use in laminates is based on the higher
thermal and chemical stabilities which characterize aromatic systems,
coupled with the relatively low cost of raw materials. The resins are used
with n values in the range 0-2. A typical formulation would have an average
molecular weight of 900 (n = 2 for the non-brominated DGEBA) and a
solids content of 20% in a solution of acetone or cellusolve type solvent.
Aromatic polyamines are commonly used as curing agents, particularly,
metaphenylenediamine (MPDA), 4,4'-methylenedianiline (MDA) and di-
aminodiphenylsulphone (DADS), the formulae of which are as follows:
For FR4 grade laminates and prepregs, the most commonly used curing
agent is dicyandiamide (DICY).
fI2
H 2 N-C=N-C=.=N Melting point. 208"C
9.5.3 Polyimides
Polyimides were developed during the 1960s and early 1970s in response to
the demands of the military and aerospace industry for high temperature
performance polymers, as matrix materials for laminates and other
composites. Most aromatic/heterocyclic polymer systems that have a small
number of oxidizable C-H bonds per molecule exhibit excellent oxidative
314 I ~I__________PO_L_YM
___E_R_S_IN
__P_R_UNT
__E_D
__ ________~
C_IR_C_U_I_T_B_O_A_R_D
Table 9.4 Comparison of cyanate ester and FR4 laminate properties
Property FR4 Cyanate ester
Glass transition temperature (TgrC) 135 245
Thermal Expansion Coefficient (ppm 0C)
x- and y-axis 16 12
z-axis 80 45
Peel strength Obs/inch) 10 7
Moisture absorption (%) (0-24/23) 0.9 0.7
Dielectric constant @ 1 MHz 4.3 3.5
Dissipation factor @ I MHz 0.020 0.005
dl)- o
These units may be linked through alkyl or aryl groups to form the main
polymer chain, the latter generally giving higher temperature performance.
The most general method for the preparation of high temperature
polyimides is based on the reaction between an aromatic diamine and an
aromatic dianhydride or tetracarboxylic acid. This proceeds via the
formation of the intermediate polyamic acid which in further heating
undergoes a ring closure condensation reaction leading to imide formation.
The reaction is as follows:
o
II
HOOC C NHR-
NHRNH~H
o
Polyamic acid
L - -_ _ _ ------'I I
P_O_L_YM_E_R_S_U_S_E_D_F_O_R_L_A_M_I_N_A_T_E_M_A_NU_F_A_CT_U_R_E_ _ 315
The fact that water is eliminated during the final cure to produce a highly
condensed polymer with rapidly increasing viscosity (low flow) results in a
relatively high concentration of voids, removal of which requires heating
above Tg (typically > 290°C) under pressure for several hours. For this
reason, the condensation polyimides have not been successfully applied to
the fabrication of printed wiring and laminates, although polyimide films
prepared by this route are used in other electronic applications, including
flexible film substrates (as described in section 9.6).
In the late 1960s another class of polyimides was developed by Rhone
Poulenc in which the anhydride or acid is replaced by a bismaleimide and
the resulting polyimide is formed without elimination of water. The reaction
illustrated schematically below proceeds via amine addition about the
double bonds of bismaleimide to form a linear pre-polymer.
tJi! .20
§
"!
1i;I;l.10
which can add (electrophilically) to either of the two ortho or the para
positions of a phenol:
318 II POLYMERS IN PRINTED CIRCUIT BOARD
~----------------------------------------------------~
~
LQJ + CH2=O ~ LQJ
L C H2-OH
+ ¥
~
CH2 0H
Ortho Para
Since the linking of the phenolic groups may also occur at either the ortho
or the para positions, branching occurs as the polymerization proceeds:
f1: 11:
OH
rQr'CH'OH +© ~ ~,~
OH
6
condensation primarily occurs through formation of ether linkages:
~CH'--oH + + CH,O
~ r 1:1 H~
t OCH,
~ CH2 0H
The resulting resole resins can undergo further (thermal) reaction with
elimination of water to produce a highly crosslinked system. By contrast,
curing of the novo lac resins requires the addition of a curing agent capable
of forming additional methylene groups. Curing agents most commonly
used are resole resins or hexamethylene tetramine. In the case of the latter,
ammonia rather than water is liberated in the condensation reaction. Since
novo lacs require a second stage to fully cure, they are frequently referred to
as two-step phenolics, whereas the resoles are referred to as one-step
phenolics.
Phenolic resins used in lamination are primarily the one-step resoles,
although novolacs may be added to improve resistance to water. Further
improvements in water resistance and a reduction in brittleness are achieved
by using para-cresol in place of phenol. This inhibits branching since the
para position is blocked by the methyl group.
Phenolics are the lowest-cost resins used in printed wiring substrate
fabrication, but in the USA and Europe they account for only a relatively
small percentage of the volume due to their replacement by higher
--'I I
'--_ _ _P_O_L_y_M_E_R_S_V_S_E_D_F_O_R_L_A_M_I_N_A_T_E_M_A_NV_F_A_C_T_V_R_E_ _ 319
performance epoxies. In Japan and the Far East, phenolic resins represent a
significant proportion of laminate consumption in printed wiring manufac-
ture due to the wide acceptance of XXXP and FR2 grade laminates for
consumer electronics. The properties of typical materials are shown in Table
9.5.
The major disadvantage of phenolics is their low insulation resistance.
Phenolics generally exhibit greater water absorption and sensitivity of the
electrical properties to humid environments than epoxy resins, although
only the so-called low-loss grades are used for printed wiring substrates. For
these reasons, the use of phenolics in laminate fabrication has been
restricted to the lower cost paper-based reinforcements.
9.6.2 Polyimide
For many years only one type of electrical grade polyimide was available in
film form and this was manufactured by DuPont under the trade name
Kapton®. Kapton polyimide is formed by the polycondensation of
pyromellitic dianhydride and 4,4'-diamino diphenylether. The reaction
proceeds via the intermediate formation of the polyamic acid as described in
section 9.5.3. More recently, however, other grades of polyimide have now
become available and these have trade names such as Apical, Novax, Upilex
and Espanex.
The polymer is melt-processed into the film form, the properties of which
are shown in Table 9.6. Polyimide films possess excellent dielectric and
insulation properties which remain constant over a wide range of temper-
ature and frequency. An outstanding feature of polyimides is their high
Table 9.6 Properties of high temperature thermoplastics
Property Polysulphone Udel Polyether Polyphenylene Polyetherimide Ultem + 20%
suIphone sulphide Ultem glass reinforced
Victrex Ryton
Flexural strength
(x 10-3 psi) 15.4 18.6 25 21 30
Flexural modulus
(x 10-6 psi) 0.39 0.37 1.7 0.48 0.9
Tensile strength
(x 10-3 psi) 10.2 12.2 16.2 15.2 20
Dielectric constant
(l MHz, 25°C) 3.03 3.45 3.8 3.1 3.5
Dissipation factor
(1 MHz, 25°C) 0 0.008 0.0014 0.006 0.0015a
Volume resistivity
(x 10- 15 n cm) 50 100 45 6.7 0.7
Electric strength
(kVjmm) 17 16 17.7 28 26.5
Deflection temperature
eC) (264 psi) 174 202 243 200 209
Thermal expansion coefficient
(ppmrC) 56 55 40 56 25
Water absorption
(%) (24 h) 0.3 0.4 0.05 0.25 0.26
Maximum continuous
temperature (0C) 160 170-200 170 170 170
a 1 kHz.
P_O_L_y_M
L -_ _ _ _ _ _ _ __E_R_S_F_O_R__F_L_EX
__ __IR_I_N_G________~I
IB_L_E_P_R_I_N_T_E_D_W I 323
thermal stability and ability to retain good physical properties over a wide
temperature range (-250 to + 300°C or even higher). The main disadvantage
of polyimide film is its relatively high water absorption which results in an
increase in the dielectric constant and dissipation factor and a reduction in
the resistivity and dielectric strength on exposure to high-humidity
environments. Moisture must be removed from Kapton films prior to
lamination with copper foil to avoid blistering and delamination.
Since polyimide film offers excellent dimensional stability, high tensile
strength and high temperature resistance, its use in more sophisticated
flexible circuit applications is growing rapidly. Polyimide film is also used
for the fabrication of a hybrid type of wiring board referred to as 'flex-rigid'
in which conventional rigid glass reinforced laminate wiring boards are
combined in a multilayer package with integral printed flexible cables that
are electrically connected by means of plated through holes. This type of
circuit board is used mainly for military and high reliability systems. A
major development in polyimide foils occurred in the last few years with the
introduction of adhesiveless flex materials (see below). These materials are
popular with producers because they offer better dimensional stability,
potential for higher layer counts, and a greater degree of flexibility due to
the absence of an adhesive layer. An example of this type of material is
Pyralux AP from DuPont which has a glass transition temperature of
220°C, a dielectric constant of 3.24 and an in-plane expansion coefficient of
40 ppm °C.- 1
superseded by ball grid arrays in which the connections between the package
and the substrate are normally made by arrays of metal balls or bumps
covering a large part of the underside surface of the package. Further
evolution of this type of packaging involves the use of increasingly finer
pitches and reductions in the overall size of the package. The latest
manifestation of this type of packaging is represented by the so-called chip
scale package (CSP) where the package is only slightly larger than the device
it encapsulates. All predictions from the semiconductor industry indicate
that this inexorable increase in the number of l/Os on state-of-the-art silicon
devices will continue and therefore there will be increasing pressure to
continue the development of advanced packaging solutions incorporating
finer pitch area interconnects.
Printed circuit board makers have, up to now, been able to accommodate
these packaging innovations by making the features on their printed circuit
board smaller and by adding additional layers to their boards. However, the
industry is now reaching the limit of possible improvements that can be
made using conventional technology. For example, although leaded
components are usually no longer required, board makers still have to
provide holes in their boards which can be metallized to allow interconnec-
tion of the various layers of circuitry in a multilayer board. These holes are
traditionally mechanically drilled and it is desirable to make them as small
as possible. When hole diameters were 0.5 rom these could be routinely
drilled at high speed and with good accuracy but as the requirement for
much smaller holes has increased the difficulty of mechanically drilling them
has also increased. It is generally agreed that mechanical drilling of holes
with diameters below '"" 100 J.Lm is not practically possible for volume
production. Similarly, since the late 1960s the PCB industry has routinely
used dry film photoresist in the patterning process for making PCBs.
Although this is a somewhat contentious issue, it is generally acknowledged
that the volume production of fine line interconnects becomes increasingly
difficult below about 100 J.Lm using conventional dry film imaging technol-
ogy.
Consequently, there is a wide acceptance that in order to produce the sub-
100 J.Lm lines and spaces and small diameter via (holes) that will be needed to
provide acceptable interconnects for the high lead count area array packages
now beginning to be produced, an alternative method of making these
interconnects will be required. One popular route which has a number of
possible variants is the so-called sequential build-up (SBU) approach. This
method relies heavily on the use of polymeric dielectrics. In the sequential
build up route additional layers of fine line interconnects are built up on top
of a conventional substrate. The conventional substrate would typically be a
4--6 layer multilayer printed circuit board manufactured using conventional
processes. Extra layers of interconnect, typically one, two or three layers,
would then be added sequentially to both sides of the board. These
328 I ~I__________PO_L_y_M__E_R_S_IN__P_R_INT
__E_D__C_IR_C_V_I_T_B_O_A_R_D________~
additional layers would be sufficiently fine to enable chip scale packages to
be assembled and to provide the interconnect pathways required for the
large numbers ofl/Os. This is shown schematically in Fig. 9.11.
Many potential routes for providing this type of sequentially built-up
circuitry have been investigated and currently they can be divided into three
principal basic routes that have been introduced commercially. Of these
three routes, the first utilizes plasma etching in a mass via formation
method, a good example being the Dycostrate technology developed by the
Swiss company Dyconex. This method uses copper coated polyimides which
may ultimately be laminated on to an existing substrate. The copper is
patterned using conventional techniques and the via holes are then etched
through the polyimide using a mixed oxygen and fluorocarbon based
plasma. The organic layer cannot contain any glass reinforcement and base
materials containing Thermount reinforcement have been produced by the
company Isola for this type of process. Resin coated copper foil can also be
used to build up purely organic insulating layers and resin systems based on
epoxy and polyphenylene ether (PPE) have also been evaluated. Subsequent
conventional PCB metallization processes are then used to plate the vias
thereby forming the connection between the two levels of metallization. This
process is particularly elegant in that the vias are all formed in a single
plasma etching stage. There are a number of process variations on the basic
theme and boards produced in this way have found particular favour for
high reliability military and aerospace applications.
The other two SBU processes that are currently very popular use
alternative techniques to form vias in a suitable dielectric. One uses a
specially developed photo-imageable permanent dielectric and the other
Sequential
lamination
Conventional PCB
Technology
Fig.9.11. Typical application for sequential build-up technology with two additional
interconnect layers being deposited on to a conventional multilayer printed circuit
board.
,--_P_O_L_Y_M_E_R_S_IN_A_D_V_A_N_C_E_D_I_NT_E_R_C_O_NN_E_Cf
__ A_PP_L_I_C_A_T_IO_N_S-,-------'! I 329
brought about by the use of an aryl phosphonium salt that generates an acid
group during exposure. This in tum is capable of catalysing the epoxide
curing reaction. Final curing is achieved thermally.
With these dielectrics one of the key requirements is good adhesion since
the dielectric will be in contact with a number of interfaces and different
materials. Any thermal expansion coefficient mismatch between these
materials can lead to reliability problems occurring during any thermal
excursions experienced in assembly or service. The dielectrics need to have
good adhesion to underlying copper and substrate material, typically FR4.
In addition, any subsequently deposited dielectric and copper metallization
must also adhere well to the previously deposited layer. Good adhesion of
these materials is usually achieved through the use of special formulation
additives such as flexibilizers and optimized surface modification stages
during the processing.
With the laser 'drilling' approach to via generation the dielectric does not
need to be photo-imageable and this enables a wider choice of chemistries to
be considered. The vias are 'drilled' (i.e. ablated) in the dielectric using either
a CO2 laser operating in the infrared (10.6 or 9.4 nm) or an excimer or Nd/
YAG laser with an output in the UV (248 and 355 nm respectively). Until
recently, UV lasers were preferred because of their ability to form better
quality via holes but the recent CO2 lasers have been much improved such
that they now can also produce good quality vias. Drilling rates of up to 280
holes per second, independent of hole diameter up to 0.35 mm, have been
achieved with a TEA CO2 laser. Unlike the photo-imageable dielectric
method where all vias are formed in one exposure, the laser method is a
sequential process and therefore the speed of the laser is particularly
important. The different lasers also have varying material drilling capabil-
ities, for example the Nd/YAG lasers are able to drill copper, glass and resin
whereas the excimer lasers cannot drill copper and glass. This means that
different approaches and materials have to be selected depending upon the
laser choice.
Laser drilling speeds have been increasing significantly over the last two
years and it is estimated that laser drilling now accounts for over 70% of
current commercial SBU board production as opposed to less than 30%
using a photo-imageable dielectric. The photo-imaging route is however
currently dominant in the production of advanced package substrates.
The most widely used polymeric materials in the printed circuit board
industry are epoxy resins based on bisphenol A. Epoxy-glass fabric
laminates account for the vast majority of the consumption of printed
circuit substrates. In particular, laminates with specifications equal to or
L -_ _ _ _ _ _ SD_M_M_A_R_y_A_ND
__C_O_N_C_L_D_S_IO_N_S_ _ _ _ _ _ ---'I I 331
better than NEMA grade GlOor FR4, which combine very good
dimensional stability with excellent electrical properties, are suitable for
the fabrication of high density wiring that is required for computers and
telecommunications equipment.
In much of the consumer electronics field where interconnection density is
generally low, cheaper substrates such as paper-based phenolic resins,
epoxy-glass mat composites and polyester or nylon-based flexible films are
employed. However, it is worth noting that with the increasing sophistica-
tion of consumer electronic goods such as mobile phones, camcorders and
palm top computers, the need for high density interconnects is growing
rapidly and these need higher quality substrate materials.
Polyimides which are available as both laminating resins and flexible films
are used in military and aerospace applications where high temperature
performance is critical. High temperature thermoplastics such as polysul-
phones and polyetherimides offer the potential for production of relatively
low-cost moulded substrates with excellent electrical characteristics and
thermal stability (see Chapter 11). The low dissipation factors characteristic
of these materials have generated considerable interest in their potential for
use in high-frequency circuits, as a substitute for the much more expensive
teflon-based substrates. Much of the demand for new or improved substrate
materials for printed wiring is generated by the ever-increasing density of
silicon microcircuits. As the chips become larger with more functionality the
number and density of interconnections increases. This requires the use of
finer traces, smaller pads and plated through holes, in conjunction with an
increasing employment of multilayer wiring. Consequently, substrates with
high dimensional stability are required to facilitate the precise registration of
circuit layers and to eliminate failures due to the cracking of plated through
holes. This latter problem results from excessive expansion and contraction
of the substrate that occurs during thermal cycling in processing steps such
as lamination and soldering. Furthermore, the higher current densities
associated with the current range of microprocessors gives rise to a greater
level of heat dissipation and hence thermal stress which add to the problems
of substrate thermomechanical stability.
Today, most of these problems are being resolved by the use of rigid
substrates based on high glass transition temperature resins such as
polyimides and modified epoxides. The reduced thermal expansion of these
materials at the normal processing or operating temperatures provides for
significant increases in dimensional stability compared with GIO or FR4
laminates. As device operating speeds increase, there is, however, still a need
for materials with reduced dielectric constants and improved loss factors.
In addition, composites are being developed in which the thermal
coefficients of expansion are matched more closely to those of the chip
substrates and copper foils to reduce thermal stress. Many of these
developments at the so-called leading edge of the technology are heralding
332 I ~I__________PO__L_YM__E_R_S__IN__P_RI_NT__E_D__C_IR_C_U__IT__B_O_A_RD__________~
the future for new approaches to providing advanced interconnect struc-
tures. The use of sequential build-up technology is one example of how a
radically different approach to providing interconnects is being utilized to
enable the increasingly sophisticated semiconductor packages with hundreds
or even thousands of I/Os to be accommodated. What is certain is that no
single polymeric material or composite will provide the necessary cost/
performance capability for all of the increasingly diverse interconnect needs
of the electronics industry.
9.10 ACKNOWLEDGEMENT
In the previous edition of this book the chapter detailing polymers used in
printed circuit board applications was written by Dr Rodney Hurditch.
Although there have been many significant new developments in the
materials used in this area much of the information presented in the original
chapter remains valid today and has. been included in this chapter. I would
therefore like to acknowledge the important contribution made to this
chapter by Dr Hurditch and to thank him for his valuable input.
9.11 REFERENCES
10.1 INTRODUCTION
The principal drivers for the development of new polymers for electronic
applications include:
• device, substrate and system miniaturization impacting operational speed
and product weight (e.g. portability),
• coefficient of thermal expansion (CTE) mismatch of semiconductor and
related packaging and interconnecting materials,
• increase in data signal processing speeds and need for dielectric constant
(&') and dielectric loss (&'') reduction,
• integration of low cost, production efficient, high performance materials
and enabling technologies into consumer products,
• improved product reliability.
With these considerations the global electronics business is continually
challenging the cost/performance balance of materials employed by the
industry. In order to get 'more-for-less' electronic designers, material
developers and manufacturers have spent considerable effort to investigate
the potential of new polymers, particularly ones that enable new technol-
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
334 !NEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
Conventional
Pad-Ies~ microvias
•
•
(a)
Fig. 1U.l.
336 fEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
H H-o- Hie
A.KEVLARTM B. NOME)(TM
NlCJN'
~-o-'\ ~-c
- II
o
'I
-
'\ c
II
0
I
I
-...::
~
c
II
0
I
~ c
0
II
n n
binder, and it is claimed that employing Speedboard can reduce the size of a
PWB by over 50%, and bring a gain of 20% in signal speed over standard
FR4 constructions. 3 Given that its dimensional stability is not as good as
laminates containing glass or aramid reinforcements, Gore promotes the use
of Speedboard via a so-called mixed dielectric concept, Fig. 10.4. The
Speedboard layers conform to the topography of the core dielectric, for
L -_ _ _ ---'I I
P_W_B_SV_B_S_T_RA_T_E_S_W_I_T_H_O_V_T_R_EI_N_F_O_R_C_E_M_E_NT_S_ _ _ 337
(a
Fig. 10.4. (a) Micrograph of ePTFE matrix; (b) cross-section of mixed dielectric
PWB (courtesy of W. L. Gore & Associates).
example FR-4, and also assume the core's improved thermal expansion
coefficient (CTE) characteristics.
Films produced from polyimides have been used for over twenty years in
flexible PWB applications (see Chapter 9). More recently they, and films
containing epoxy, bismaleimide (BT) or poly(phenylene oxide) based resins,
have been used for the sequential build-up of circuitry on glass reinforced
substrate cores4 in so-called MCM-L technology. Indeed well over a dozen
different SBU methodologies, including DYCOstrate™ and the use of
photoimageable dielectric films, have been cited. 4 However, few of them
have sufficient dimensional stability to be employed in a solo form as a
substrate dielectric. Liquid crystalline polymers (LCPs), on the other hand,
possess a significant degree of molecular ordering and have correspondingly
low CTEs. Of the many main-chain thermotropic LCPs that have been
synthesized to dateS several have been commercialized.6 Far fewer, however,
have been utilized for electronic interconnect applications. Figure 10.5 and
Table 10.3 highlight the chemistry and properties of one such example.
Liquid crystal polymers offer additional advantages in that most of them
include inherent flame retardancy and recyclability. One of the regularly
quoted difficulties with LCP technology, i.e. the ability to produce a
dielectric film with uniform properties in the x- and y-directions, is claimed
to be avoided through the use of a counter-rotating die fabrication
process.? Further potential applications of LCPs in the electronics industry
that are currently being investigated include dimensionally stable fibres and
micro-foamed films with extremely low dielectric constants and losses.
338 fEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
y
Fig. 10.5. Chemical structure of Vectran™ LCP film.
Property Vectran™
o
o~ ft II
:~<C}-Ar
C
II
o
-f > C
c
II
o
n
Fig. 10.6. Chemical structure of PiBO.
Property Value
Tg eC) >500
r/ at 1 MHz 2.6
E" at 1 MHz 0.004
Tensile strength (Mpa [Ksi)) 393 [57]
Tensile modulus (GPa [Msi]) 11 [\.6]
Water abs., 50% RH, 23°C (%) 1.3
Dielectric strength (V/mil) 8000
x,y-CTE (ppm/K) 3-5
with much lower stresses than the inorganics and can be applied by spin
coating rather than the typical chemical vapour deposition (CVO) route
required for materials such as Si02 and Si3N4 • The organic materials also
help to planarize topography rather than conformally coat as does deposited
Si02 • Some polymers also have lower dielectric constants, lower water
absorptions and are resistant to copper migration. On the other hand Si02
unquestionably offers better thermal conductivity, thermal stability and
chemical inertness.
The suitability of a specific polymeric material is highly dependent on the
intended application, the process chosen to fabricate the structure and the
environment that the device will be expected to operate in. Commercially
available materials are usually categorized by the via formation technique
that they are compatible with, i.e. dry etching (plasma, RIE or laser), wet
etching (usually aqueous base) or photosensitive (features created in a
process similar to negative working photoresists).
In general, polyimides (PIs) and benzocyclobutenes (BCBs) are the
polymeric materials being used commercially to fabricate polymeric thin
film devices for the applications detailed below. A variety of polymeric
materials are being examined for use as inner layer dielectrics on intergrated
circuits (ICs, see section 10.4.3.2). Generally, however, polymeric materials,
340 fEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
Vendor Product Photo E' Tg (0C) CTE Tensile Elongation Water Stress Planar-
designation (ppm) strength (%) absorption (Mpa) izationt
(MPa) (%;*
• Water absorption values are reported under various humidity conditions, see ref. 2l.
t Planarization of an isolated 20--30 micron feature. See ref. 22.
342 !NEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
o 0
o~o +
o
PMDA
0
! ODA
! -2HzO
tN~-&t Pol~.
o 0 n
Fig. 10.7. Generalized PI synthesis.
the polyamic acid structure except that the alcohol group (OH) has been
replaced by an ester group (OR) where the R group contains a photo cross-
linkable species. The salt-based materials are derived from the reaction of
polyamic acids with bases containing the photocrosslinking species. I 4-16
DRY ETCH
POLYMER PHOTOPOLYMER
iIIIIIIIi
COAT DIELECTRIC
POLYMER
SUBSTRATE
~
RESIST
COAT & DRY
RESIST
-..
EXPOSE &
DEVELOP
RESISTOR
POLYMER
TRANSFER
PATTERN TO
POLYMER
~
STRIP
RESIST
~
CURE
POLYMER
l.FLAMETM 2. Parylene-AF4
F F
{_R_O :
F
F F
. jt 1
3. F-polyimide (FPI- 136M from DuPont)
~III
o
r~
II
0
1-%
H
II
0
CJ'. CJ'.
II
0
%
OCJ'.
ocr
,
1-Y
--0--
Y
n n
6.PFCB
F F F
F~O O>=<F
10.5 TRADENAMES
+-- Polarizer
~ Alignment layers
. ~
(not to scale)
Fig. 10.14. LCD cross-section.
10.6 REFERENCES
1. Hirakawa, T. and Kato, c., A new laminate for MCM-L, Proc. [SHM'93, p.l99,
1993.
2. Homan, G. E. and Powell, D. J., Fabricating PWBs and MCM-Ls with a new
nonwoven aramid reinforcement, 7th International SAMPE Electronics Confer-
ence, p.l77, 1994.
3. Haezebaert, G. and Durigon, A., Lower dielectric constant prepreg with
standard FR4 processing for high performance digital PWBs, Printed Circuit
Europe, July-August, 23,1996, and Snyder, W. W., A new low-dielectric material
for PCBs, Printed Circuit Fabrication, March, 41, 1990.
4. Novak, A., Lindahl, T. and Hellgren, E., Study of different techniques for
building MCM-L substrates, Proceedings PCWC VII, paper 20-1/7, Basel 1996,
and Aday, J., Tessier, T., Crews, H. and Rasul, J., A comparative analysis of high
density PWB technologies, [MAPS Journal, 19 (4), 404, 1996.
5. Perec, V. and Tomazos, D., in Comprehensive Polymer Science, ed. G. Allen,
First Supplement, pp. 300 and 701, 1992, and Weiss, R. A. and Ober, C. K.,
Liquid crystalline polymers, ACS Symposium Series, No. 435, 1990.
6. Editorial, High Performance Plastics, pp. 2 and 6, January, 1996.
R_EF_E_R_E_N_C_ES________________
L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ~I I 351
7. Jayaraj, K., Noll, T. E. and Blizard, K., Controlled thermal expansion PWBs
based on liquid crystal polymer dielectrics, Circuit World, 22 (2),1996, and Noll,
T. E., Jayaraj, K., Farrel, B. and Larmouth, R., Low cost. Near hermetic MCM
based on liquid crystal polymer dielectrics, IMAPS Journal, 19 (4),469, 1996.
8. Lisignea, R., Piche. J. and Mathisen, R., Ordered polymers for interconnection
substrates, ACS Symp. Series, No. 407, 446, 1989.
9. Adema, G., Hwang, L., Rinne, G. and Turlik, I., Passivation schemes for CuI
polymer thin film interconnections used in multichip modules, IEEE Trans.
CHMT, 16, 53,1993.
10. Ferger, C., Selection criteria for MCM dielectrics, in MCM Technologies and
Alternatives, eds D. Doane, and P. Franzone, Van Nostrand Reinhold, 1993, p.
311.
11. Ferger, C. et al., Polyimides: Materials, Chemistry and Characterization,
Elsevier, 1989.
12. Mittal, K. L. (ed.), Polyimides, Plenum Press, 1984.
13. Deutsch, A. et al., Measurement of Dielectric Anisotropy of BPDA-PDA PI in
Multilayer Thin Film Packages, IEEE Trans. CHMT, 17,486, 1994.
14. Volksen, W., Yoon, D. Y. and Hendrick, J., Polyamic alkyl esters: Versatile PI
precursors for improved dielectric coatings, Proc. ECTC, 572, 1991.
15. Rubner, R., Siemans Forsch-u. Entwicki-Ber., 5, 232, 1976.
16. Hiramoto, H., Photosensitive polyimides Proc. MRS, Advanced Electronic
Packaging Materials, 167, 87, 1990.
17. Kirchoff, R. and Bruza, K., Benzocyclobutenes in polymer synthesis, Progress in
Polymer Science, 18, 85, 1993.
18. Garrou, P. et al., Rapid thermal curing of BCB dielectric, IEEE Trans. CHMT,
16, 46, 1993.
19. Moyer, E. S. et al., Photodefinable BCB formulations for microelectronics
applications, part II, Proc. IEPS, 37, 1992.
20. Strandjord, A., Ida, Y., Rogers, B., Garrou, P., Cummings, S. and Kisting, S.,
Photo BCB, MCM-D processing and area array assembly, Int. J. Microcircuits
and Microelectronic Packaging, 19, 260, 1996.
21. Garrou, P. E., Polymer dielectrics for multichip module packaging, Proc. IEEE,
SO, 1942, 1992.
22. Beyne, E., Van Hoff, R., Pieters, F., Lenaerts, S. and Achen, A., The use ofBCB
and photo BCB dielectric in MCM-D for high speed digital and microwave
applications, Proc. Int Con! MCMs, 513, Denver, 1995.
23. Wilson, A., Use ofpolyimides in VLSI fabrication, in Polyimides, ed. K.L. Mitel,
Plenum Press, 715, 1984.
24. The National Technology Roadmap for Semiconductors, Semiconductor Industry
Association 1994.
25. Samuelson, G., Polyimide for multilevel VLSI, in Polymers for Electronic
Applications, ed. Feit and Wilkins, Amer. Chern. Soc. Symp. Series, 184, 93,
1982.
26. Bothra, S., Rogers, B., Kellam, M. and Osborn, C., Analysis of the effects of
scaling on interconnect delay in ULSI circuits, IEEE Trans. Electron Devices, 40,
591, 1993.
352 !NEW POLYMERS FOR EMERGING INTERCONNECTION APPLICATION~
11.1 INTRODUCTION
During this decade, the competitiveness of the electronics industry has been
determined by the following six aspects: cost reduction, miniaturization of
devices, function integration, product reliability, design freedom and
improved environmental compatibility.
Plastics for Electronics, 2nd edn. Edited by Martin Goosey. Published in 1999 by Kluwer
Academic Publishers, Dordrecht. ISBN 978-90-481-4018-3
356 II MOULDED INTERCONNECT DEVICES (MIDs)
~----------------------------------------------------~
Fig. 11.1. Transition from conventional printed circuit board to moulded inter-
connect device.
Molded
Interconnect
Devices
Household MultimediB
Appliances
Telecommunication
Passenger Comp. Office Automation
Engine Comp. Systems
TV, HiFi, Video Automation
Skatchof
Dimension "tYpe Features Applications
--
onsartlon process
2D 0
planar process
plane ..• conventional circuit
board
planar process
1A plane, 3D-elements
at the opposite side
U- - ffl simple casings
planar process
2"hD
18 plane, 3D-elements
at the opposite side
f1-u--Jl easy assembly,
modular structure
side of it. Type IA MIDs are relatively widespread, because components can
be assembled onto them using conventional off-the-shelf onsertion equip-
ment, whilst still providing additional three-dimensional mechanical fea-
tures, e.g. fasteners, that help reduce assembly cost. Type IB and Ie MIDs
require modifications in the assembly equipment, at least an enlarged z-
stroke of the placement head. This limits their use in mass production. If no
components are necessary, and there are many examples of this situation,
MIDs with circuitry and mechanical features on the same side can be very
useful.
N-times 2-D MIDs have several process planes which are situated at an
angle to one another, whilst 3-D types have regular or even free-form
surfaces as process surfaces. These types increase the freedom of design
although they can also present further challenges for assembly equipment.
The more complex the type of MID, the more complicated the
downstream manufacturing processes such as assembly, connection and
test become. Therefore, any designer should try to keep the part as simple as
possible.
'--_ _K_E_y_A_D_V_A_NT_A_G_E_S_A_ND
__ -,I I
INT_R_O_D_V_CT_I_O_N_S_T_RA_T_E_G_I_E_S_ _ 359
Freedom of Design
o Integration
electronics - mechanics
o Miniaturization
New functions
Rationalization
o Less part count
o Shorter process chains
o Reduced use of material
o Increased reliability
Environmental Compatibility
Socket
Snap fits
Shielding walls ---tf~~~'~~
Switches
total cos1s
conventional
Break·Even-Point
production number
Fig. 11.6. Cost comparison between a typical MID and a conventional solution.
~_____________M_A_N_U_F_A_C_TU
__~__N_G_M
__E_TH
__O_D_S____________~I I 363
included in the cost calculation of a single part or whether investment is
calculated on a separate basis.
So far, cost savings of up to 40% have been achieved in MID products by
integration depending, of course, on the specific product, lot sizes and other
boundary conditions. In one project that achieved savings of 24% in total, it
should be noted that only about half of the total cost could be influenced by
the MID itself. In these areas, costs were reduced by 50-70%.
After the basic decision to employ MID technology for a specific product
has been made, the manufacturing method and suitable materials need to be
selected. MIDs can be produced in a variety of ways. The most common
ones are explained below and they can be structured according to Fig. 11.7.
Basically, there is a distinction according to the moulding method
between one-shot, two-shot and insert moulding processes. After the
moulding, one-shot parts can either be metal plated by a hot embossing
process, in which the structure is formed by an engraved die, or can be
plated galvanically (in the following, the term 'galvanic' includes electro-
chemical as well as autocatalytic chemical plating processes). After galvanic
plating, the structuring is done by an imaging process, either by means of a
mask as in conventional circuit board manufacture or by direct laser
imaging.
In two-shot moulded parts, plating is always done galvanically, using the
different adhesion properties of the two plastic materials used for moulding
to achieve selective plating.
To achieve insert-moulded parts, a flexible foil carrying the circuitry is
inserted in the mould tool and subsequently over-moulded with a suitable
I
EMBOSSING GALVANIC FOIL
11.3.1 Photo-imaging
The photo-imaging process is the oldest of all MID manufacturing
processes. It is quite similar to the traditional semi-additive process for
conventional printed circuit boards. Most adhesion promotion solutions for
thermoplastics were developed with this process in mind, and a compar-
atively large number of them are available for solderable thermoplastics
such as PEl and LCP.
Firstly, the surface of the plastic blank obtained by one-shot moulding is
activated using a chemical process suitable for the material used. The next
step is the application of a thin so-called flash layer of chemically deposited
copper. The thickness of this layer is usually no more than a few
micrometres.
There are then several ways of proceeding, the most common being a
semi-additive process. The part is covered in a photosensitive resist which is
then selectively exposed with UV radiation through a suitably patterned
mask, as in the conventional printed circuit board manufacturing process. A
planar film type mask, such as would be used on two-dimensional boards, of
course, has limitations when one is considering the imaging of three-
dimensional structures by omni-directionallight, as stray light will blur the
contours in recessed areas. This can sometimes be tolerated to a certain
extent depending on the application, but will usually have to be catered for
M_A
L -_ _ _ _ _ _ _ _ _ _ _ _ _ __ __E_T_H_O_D_S____________~I
N_U_F_A_C_T_U_R_I_N_G_M I 365
injection moulding
D
SUrfacDvation
electroless copper
D
apply photoresist
D
expose with 3-D mask
D
develop photoresist
D
electr
D copper
remove photoresist
D
etch away copper
L--_--li
injection moulding
{} 1 - . 1_ - - '
surface activation
1....-------11 1.-1_----'
{}
electroless copper
L-_---.JIII-._--'
{}
full-build
electroplate copper
{}
plate etch resist 1 I iI
{} ~~.
laser structuring
of etch resist c=Jc:J
• •
" --
1electroplate copper
etch copper
1 differential etch
injection mould
1. shot
~
catalyse
D I~I I m m I 12!!::!51
non-platable resin pial able resin non catalytic resin
m etal plate
-
Mould plastic substrate L I_ _ 1t
- - JI -...J
D
Place substrate
and foil in press
lEis:
Through-hole contact
-: :E.
Emboss
manufacture
plastic film
film
structure film
thermoform film
overmould film
that is inserted into the mould tool in a very similar manner to that
described above, but with the circuit film facing the tool wall and the traces
facing the cavity. When the resin has been injected and cooled down, the
part is removed from the cavity, with the carrier film covering the substrate
and the circuitry. The film is peeled off the part to uncover the circuit traces,
and disposed of.
Another variation of the in-mould technique developed in England by
AIT Ltd uses an aluminium--copper double layer foil. The copper layer is
structured before moulding and the aluminium foil inserted into the mould
tool in the same manner as with the transfer decal process. After moulding,
the aluminium layer is removed from the substrate by an etching step, rather
than by peeling. This method allows the use of polyphenylene sulphide
(PPS) as the substrate material and gives very good adhesion properties.
MANUFACTURING METHODS
L -____________________________________________________~ II 373
Aluminium foil ~
circuitry substrate
Fig. 11.14. Aluminium foil process.
Both the transfer decal and the AIT processes yield a completely level
surface, as the circuit traces are actually embedded in the moulded plastic
part. This can be advantageous when designing switches with sliding
elements or potentiometers etc.
The basic characteristics of the in-mould processes are as follows:
• there is planar structuring even before moulding;
• there are few process steps;
• there is a short cycle time;
• they are suitable for glossy surfaces;
• multilayers are possible.
The foil or film is structured in its planar state before moulding and addition
of the three-dimensionality.to the circuit structure. This makes it possible to
apply conventional structuring technology, such as photo-imaging and
screen printing of conductive paste, etc., without the need to purchase new
equipment and having to introduce unproven processes. The number of
process steps in the moulding itself is very limited, yielding short cycle times.
The fact that moulding is basically done after structuring leaves various
possibilities open for the type of surface quality required, thus, for example,
enabling in-mould labelled parts to be used for glossy surfaces. A last, but
very important advantage of the in-mould processes is the ability to
structure the foil or film in multiple layers, thus enabling the manufacture of
three-dimensional multi-layer circuitry.
D
Metalplating [Lr1-IJ I I
Milling of Isolation Channels
~b
Fig. 11.15. Mechanical structuring processes.
Injection Moulding
{}
Apply Catalyst
{}
Exposure by LASER
(or UV-Light and mask)
{} ==
Remove unexposed catalyst
{}
Electroless copper
....
Fig. 11.16. Selectively activated catalyst.
Surface activation
Electroless copper
{}
Apply etch resist
Etch copper
Injection Moulding
Print Primer
Electroless Copper =
the advantages of the screen printing process and the selectively activated
catalyst, again drastically reducing process steps.
As MID technology is still very new, all limitations are constantly being
altered. When considering a technology, it is highly recommendable to
discuss the specific requirements of the project in question. For example,
minimum track widths depend on the length in general and especially on the
parallel length of those tracks, and on the flow characteristics of the
thermoplastic used.
Not mentioned in the table are the possible thicknesses and compositions
of metallization layers. As a rule of thumb, every surface available for
conventional printed circuit boards is also feasible for moulded circuit
carriers. This includes gold for bonding, tin/lead for soldering and other
surface coatings such as nickel. Thicknesses are usually derived from those
common for planar interconnects, e.g. 18, 25, and 35 microns. These apply
for all methods and in some cases 70 IIDl thicknesses are also used. Hot
embossing foil is available in thicknesses up to 150 1IDl, thus enabling
relatively high currents to be carried. A survey showed that moulded
interconnect devices are at least equal and often even superior to
conventional circuit boards in current carrying capacity.
After manufacture of the circuit carrier, the next step usually is assembly. In
many cases, however, onsertion of components is not necessary to exploit
378 I I~________M__O_U_L_D_E_D_I_NT__E_R_C_O_NN__E_CT__D_E_V_I_C_E_S_(M__ID_s_)______~
the advantages of MID technology. Of 100 or so MID applications realized
so far, only about half are equipped with components in a conventional
way. The other half are used to fulfil important connection tasks such as
replacing connectors and wire harnesses.
The entire system is more complex than a conventional pick and place
machine, as are the placement operations. In order to avoid collisions, a
cinematic simulation of the placement process has been introduced. As a
general rule, the process surfaces of an MID should be kept to a minimum,
as plain as possible and without too many surrounding walls.
methods have been utilized. The part comprises the functions of the circuit
board, display unit in the upper portion, battery holder and battery contacts
in the lower part as well as the function of a support to place the
thermometer on a table.
The components have been placed on the product using the above-
mentioned placement system and soldering was performed in an infrared
oven. The material used was PEl for the photo-imaged and laser structured
parts with polybutyleneterephthalate (PBT) used for the hot embossed
variation.
11.7 SUMMARY
11.8 LITERATURE
Burkner, G., Feldmann, K., Romer, M. and Franke, J. (eds), Proc. 1st International
Congress on Molded Interconnect Devices, MID 94, Meisenbach, Bamberg, 1994.
Burkner, G., Feldmann, K., Pohlau, F. and Romer, M. (eds), Proc. 2nd International
Congress on Molded Interconnect Devices, MID 96, Meisenbach, Bamberg, 1996
Franke, J., Integrated Product and Process Development for Three-Dimensional
Electronic Assemblies, Hanser, Munchen, 1995.
Hunter, P., MIDs-PCBs move into the third dimension, Electronic Production, June
1995.
Pohlau, F., Decision fundamentals for the introduction of MID technology. Proc.
VDI-ADB-Seminar Electronics Production, Munchen, 1995.
Index
Figure locations are indicated by bold type and tables by italic.