THE ROLE OF FUNDAMENTALS,
VISUAL OBSERVATION AND STATE-OF-THE-ART
INSTRUMENTAL METHODS IN SOLVING PLASTICS FAILURES
Myer Ezrin
University of Connecticut, Institute of Materials Science
Storrs, CT 06269-3136
1.0 Abstract
While experience is a valuable asset, a novice failure
analyst can solve even complicated plastics failures using a
logical approach and utilizing the most appropriate analyti-
cal and test methods. By including old as well as state-of-
the-art new methods, the test results often point right to the
cause of failure. The eye, by itself, and with moderate
magnification early in the investigation can provide infor-
mation which helps point the way to solving the problem.
Observations and test results are assessed as to which of
the three major factors in plastics performance are involved
– design, material, processing or combinations thereof.
2.0 Introduction
Solving plastics failures requires knowledge of the
fundamental aspects of polymers and polymer additives
and how they respond to the stresses of processing and
environmental conditions in service. All materials, not
only plastics, succeed or fail depending on the appropriate-
ness of three major categories of materials technology –
material, design and processing.[1] Together they must
satisfy the service requirements of thermal, chemical, life-
time required, etc. and be cost effective and profitable.
Plastics are very complex, more so than many other mate-
rials.
While failure is often thought of as fracture, there are
many other types and causes of failure, including adhesion,
appearance, warpage, etc.[2] Methods available for failure
analysis include state-of-the-art methods. The eye can be
helpful in understanding how failure occurred. Examina-
tion of the failure and probing of the conditions which
caused failure can lead to the cause and to what methods of
failure analysis to utilize.
Several books and other publications on plastics fail-
ure are available.[3]
3.0 The Material
Included under material are the polymer and all its ad-
ditives. The fundamental variables involved are composi-
tion, molecular weight and intermolecular order.[4] Fail-
ure may be due to any one of these variables.
3062 / ANTEC 2002
3.1 Composition
The choices of polymers and additives are very broad,
and each must be chosen carefully to avoid failure due to
improper combinations. An incompatible plasticizer or
other additive, for example, could exude to the surface and
cause adhesive failure or environmental stress cracking.
Plastic clothes hangers caused colorant in the plastic to
migrate into jacket linings, making them unsaleable.[5]
The supplier had changed from impact polystyrene (IPS) to
polypropylene (PP). A dye had been used as colorant. It
was compatible in amorphous IPS, so that there was no
driving force for it to migrate. In PP, however, the highly
crystalline polymer could not accommodate the dye so it
migrated to the surface and into the jacket lining. A pig-
ment colorant would not have caused the problem.
An antioxidant (AO) or stabilizer may not be present
at a level adequate to protect the polymer from degrading
during processing and to leave enough to be protective in
service. Such additives have to be chosen well to provide
protection for various types of polymers. Fig. 1 [6] is the
oxidative induction time (OIT) tests by differential scan-
ning calorimetry (DSC)[7] for PP before molding, molded
into a rotor for steam service and a rotor degraded in ser-
vice. Most of the depletion of AO took place in process-
ing. Fracture occurred in short service time. OIT is a good
way of determining how much antioxidant is in the poly-
mer.
Some additives are not introduced intentionally, but
are present nevertheless. Water is a common example.
Even at levels as low as 0.01-0.02%, it can cause hydro-
lytic degradation in processing of condensation polymers
like polyamide (PA) and polycarbonate (PC) to lower mo-
lecular weight (MW), with possible failure due to lower
MW. Contaminants are another unintentional additive. PE
sheet delaminated on punching to size and shape (fig.
2)[8]. It was due to a thin plastic film which interfered
with fusion of the PE, so that separation occurred on
punching. Residual monomer or solvent are other uninten-
tional additives that can contribute to failure.
3.2 Molecular Weight (MW)
The long chain nature of polymers is what makes
polymers unique. MW is a controllable variable, as po-
lymerized. High MW is the most important feature of
polymers. The long chains exist entangled with them-
selves and with other chains. Polymer strength, toughness,
and chemical resistance depend on chain length and chain
entanglements. MW is the number of monomer units in a
chain times the MW of each unit. Fig. 3 [9] shows the
typical effect of MW on properties such as strength, tough-
ness and chemical resistance. Properties change rapidly
with increasing MW at low MW and tend to level off at
high MW. In most applications, only a small reduction in
MW is allowable without affecting properties.
The difficulty with MW is that it can be lowered in
processing, due to mechanical shear, oxidative degradation
or hydrolysis of groups like amide (-CONH-) or polycar-
bonate (-O-CO-O-). The effect is greater at lower MW
affecting polymers such as PC and PA severely. As little
as 5% lowering of MW can cause a high failure rate.
Polyolefins (PE, PP) are particularly susceptible to forma-
tion of free radicals, which may react with oxygen to form
peroxy radicals and ultimately chain scission to lower
MW. A practical test for antioxidant has been referred to
above. Ideally, OIT should be at least about 10-20 minutes
after processing. If it is known that the polymer will de-
grade in processing, even if properly processed and stabi-
lized, a higher MW may be a way of ensuring that MW
will be adequate for service life.
3.3 Intermolecular Order
Crystallinity is the ability of molecules to assemble
into highly ordered arrays which confer properties such as
thermomechanical resistance, chemical resistance, as well
as strength and toughness. If a crystalline polymer does
not crystallize to the degree required, failure may occur.
Processing conditions have to allow time for the crystalli-
zation to progress adequately. A nucleating agent may be
needed to accelerate the rate of crystallization. If a pig-
ment unintentionally acts as a nucleating agent, a high de-
gree of undesired crystallinity may result and cause failure.
PP caps of various colors included copper phthalocyanine
blue.[10] Blue caps broke when tightened on the tube
thread. The pigment had brought the degree of crystallin-
ity to a much higher level than in other caps of different
color. At higher crystallinity, cap dimensions were
smaller, causing fracture when screwed onto the tube. A
property like crystallinity can be desirable or undesirable,
contributing to failure. DSC showed the failed cap to have
high crystallinity.
Another example of how crystallinity may contribute
to failure is environmental stress cracking of high density
polyethylene (HDPE). Comonomers such as hexene re-
duce crystallinity and improve stress crack resistance.
Orientation and resultant frozen-in stress are present in
crystalline and non-crystalline (amorphous) plastics.
Molded parts and extruded shapes generally are cooled on
exiting the machines so that the plastic does not have time
to relax completely to an unoriented condition. This is a
problem especially for injection-molded parts, which have
high frozen-in stress, resulting in a lower level of environ-
mental stress crack resistance. Fig. 4 [11] shows a SAN
(styrene acrylonitrile) piano key as molded and after heat
reversion by heating above Tg (glass transition tempera-
ture). The high level of shrinkage and distortion are in-
dicative of the internal stress of the as molded part. Fig. 5
[12] is a close-up of the hook at the end of the piano key.
Fracture had occurred at the hook when a spring was at-
tached. The shrinkage of the hook is to about half its
original length, indicating very high internal stress at the
hook. Solvent used to remove stains from the back of the
piano condensed on the hooks, causing environmental
stress cracking of the hook. The solution was to mold so
as to have lower frozen-in stress. Just as unintentional
additives can cause failure, unintentional intermolecular
order can do so also.
4.0 Design
Chapters 3 and 4 of ref. [1] deal with design. Ref. [13]
has examples of failures due to design. Mold design was at
fault in syringe hub fracture in which stress from a metal
insert against a flat was exerted at a weld line.[13,14]
Some cavities in a 16-cavity mold placed the weld line at
the location of stress from the insert. Mold design was at
fault in a part which was slightly out of round, resulting in
fracture when an insert was welded into the out of round
top of the part.[13]
Part design can make it very difficult to mold a part,
resulting in unusually high frozen-in stress or degradation
due to higher temperature required to form the part.
5.0 Processing
Chapter 5 of ref. [1] deals with processing-related fac-
tors in failure. Ref. [15] focuses on processing for preve n-
tion of failure. Ref. [13] also includes processing as cause
of failure. The thermal, mechanical, oxidative and hydro-
lytic stresses on material during processing is as or more
severe than the plastic will experience in its total history,
including service. Temperatures well above the Tg or Tm
(melt temperature) are required to convert the solid poly-
mer to a liquid viscosity that the processing equipment can
handle. Some polymers such as PVC may decompose dur-
ing processing in the absence of oxygen, requiring stabili-
zation and/or use of lowest possible heat. But lower tem-
perature means higher melt viscosity, possibly incomplete
fusion or incomplete filling of the mold. Due to the high
temperature required in processing, all chemical reactions
causing degradation to lower MW are accelerated. Proc-
essing means walking a virtual “tightrope” to achieve a
well fused part without excessive frozen-in stress, without
degrading the polymer. Very little drop in MW can be
tolerated in many cases without significant failures. De-
sign must take into account the difficulty of achieving
proper melt flow in areas that may resist flow. Fig. 1 is an
example of processing of PP consuming most of the
antioxidant, leaving too little for service requirements. Use
of regrind without checking MW may contribute to failure.
ANTEC 2002 / 3063
 Failure may occur as damage to material other than the
plastic in a system, such as electrical insulation,[16] in
which the conductor or other component is discolored or
corroded by byproducts of processing such as hydrogen
chloride (HCl). The damage may be superficial but may
make the product unacceptable due to appearance. Unin-
tentional processing may be continued crystallization or
relaxation in service, causing shrinkage and possibly fail-
ure.
6.0 Failure Analysis and Test Methods [17,18]
6.1 “Low Tech” Methods
The eye is the first “instrument” that should be used in
diagnosing the cause of failure. It can be very informative
and can help focus attention on the area to be tested and
what to look for. Examples are degraded insulation in
electrical cable on which darkening started in the jacket
containing ground shield wires, instead of from the central
conductor (fig. 6)[19]. The jacket had overheated due to
high current in the ground wires, which normally carry no
current. Another example is fracture of pultruded polyes-
ter fiberglass guy strain insulator rod resulting in collapse
of an electrical transmission line (fig. 7)[20]. The metal
holder at the base of the rod where fracture had occurred
had a rust streak on one side where rain ran off from a well
at the top of the holder. Acid from the atmosphere that had
accumulated over time and concentrated in the well at the
top of the fitting had corroded the galvanized coating.
Failure of the rod was due to stress cracking of E glass
fiber by acid, based on the observation of the rust streak on
the metal fitting. Subsequently, rods were tested for resis-
tance to acid and water under stress. Rods were fractured
very readily by dilute acid under stress, confirming the
mode of failure. A third example was referred to above,
section 4, an out-of-round part. Visual observation was
confirmed by measurement.
The eye, with or without magnification, may locate a
fracture initiation site and the nature of fracture propaga-
tion, which may help determine the nature and cause of
fracture.[21]
Another useful “low tech” method is melt index (MI).
It is a very sensitive method of detecting change in MW,
usually to lower MW, by increasing MI. Melt viscosity is
a function of the 3.4 power of MW, and a higher power in
some cases [22]. MI should not increase more than about
10-30% from that of the starting resin in order to retain
control of MW.
Heat reversion [23] is a simple way of revealing how
highly stressed a plastic is due to frozen-in stress (fig. 4
and 5). Stress cracking by contact with a chemical is a
useful test.[24] For example, acetic acid causes ABS parts
to fracture in about a minute if internal stress is too high.
Birefringence [24] works well for transparent plastics, the
sample being observed between crossed polarizing sheets
(fig. 8).[25]
3064 / ANTEC 2002
When a part fails, the question may be is it due to the
starting resin being out of specification or to internal
stresses or degradation during molding. Compression
molding of the part into a flat sheet suitable for tensile or
other strength or chemical resistance testing relieves the
plastic of internal stresses. If the material then gives ac-
ceptable properties, comparable to the spec sheet for the
plastic or to tests on compression molding of original pel-
lets, the failure is probably due to internal stress or incom-
plete fusion.
6.2 “High Tech” Instrumental Methods
DSC is particularly useful among thermal analysis
methods. It is the method used for OIT to assess antioxi-
dant content and for degree of crystallinity. Micro infrared
spectroscopy (IR) and micro Raman spectroscopy can pin-
point contaminants, agglomerates of additives, etc., photo-
graph them and identify them. A dark stain was identified
as carbon by Raman. Peak ratios for different components
of a formulation, compared to a reference standard or start-
ing material, may indicate change in composition in proc-
essing or service or between lots. Additives in an electrical
cable jacket migrated into the insulation, causing insulation
resistance to drop.[16] This was first noted by IR and con-
firmed by thermal desorption gas chromatography/mass
spectroscopy (TD/GC/MS).
TD/GC/MS showed the cause of low adhesion of
bonded parts to be hydrocarbon contamination on the sur-
face, picked up from engine exhaust by unprotected, ex-
posed surfaces awaiting assembly.[26] A grease lubricant
for a metal screw in a small servo motor became stiff and
failed.[27] Triphenylphosphate plasticizer in the grease
had hydrolyzed to phosphoric acid, causing corrosion of
metal and hydrolysis of an ester type lubricant. The con-
version of esters to the corresponding acids was detected
by IR. TD/GC/MS showed the presence of the acids
formed by hydrolysis. Thermogravimetric analysis (TGA)
showed that metal had corroded, adding to the residue at
the end of the TGA run. Elemental analysis by scanning
electron microcopy/energy dispersive x-ray analysis
(SEM/EDX) detected corroded metal in the failed grease.
The solution was not to use an organo phosphate because it
might start a series of reactions ultimately causing failure.
X-ray photoelectron spectroscopy (XPS) determines
elemental composition and element combinations in sur-
faces of the order of 50 angstroms thick (5 nanometers),
the very uppermost layer of material. It can be useful in
adhesion problems and others where the problem may lie
in an extremely thin top layer. Color on the surface of an
electrical cable jacket was due to sulfur inadvertently in-
troduced in processing which had combined with lead in a
stabilizer to give black lead sulfide.[28]
Acoustic emission (AE)[29] is a non-destructive
method of detecting discontinuities, flaws, cracks, etc. in
plastics materials. It is particularly useful for large struc-
tures, such as fiber reinforced composite storage or reactor
vessels, tanks, pipes, etc. AE can be used in the field under
normal stress loading of structures. Potential failure sites
can be located before failure occurs. AE monitoring of
chemical storage tanks of fiber reinforced plastics resulted
in a major reduction in the number of tank failures. It can
be used as a failure analysis method and for QC (quality
control). AE is particularly useful for large volume appli-
cations by taking structures out of service, thus preventing
potentially catastrophic equipment damage and personal
injury. AE operates by detecting acoustic response to ap-
plied stress. The method locates the source of emission,
i.e., the site such as a crack or discontinuity, undergoing a
response to the imposition of stress. AE is also applicable
to the examination of small parts, and to particle reinforced
plastics as well as fiber reinforced composites.
Ultrasonic spectroscopy [30] also detects defects and
variations in material properties; it has been applied to
failed adhesive joints, among others. The rate and manner
in which sound transmits are functions of the medium’s
physical and mechanical properties. Three dimensional
maps of defects and/or material property variations are
obtained using recent advances in ultrasound equipment
and signal processing technology. The method has been
used to elucidate residual strength in polymer matrix com-
posites exposed to thermal loads.[31]
References
1. Ezrin, M., Plastics Failure Guide – Cause and Pre-
vention, Hanser (1996). 1a – p. 4.
2. Ref. 1, p. 7, table 1-1.
3. Books on plastics failure – 3a. Brostow, W. and R. D.
Corneliussen, eds., Failure of Plastics, Hanser (1986).
24 chapters by different authors, mainly on the science
of various aspects of plastics failure; 3b. Ref. 1; 3c.
SPE ANTEC Proceedings, papers of the Failure
Analysis and Prevention Special Interest Group since
1988; 3d. Hatch, B., Injection Molding Magazine, “On
the Road with Bob Hatch, 100 Injection Molding
Problems Solved by IMM’s Troubleshooter,” (1997);
3e. Wright, D.C., Environmental Stress Cracking of
Plastics, RAPRA (1996); 3f. Lewis, P.R., Polymer
Product Failure, RAPRA Review Report, 10:3, Re-
port 111 (1999); 3g. Scheiers, J., Compositional and
Failure Analysis of Polymers: A Practical Approach,
Wiley (2000); 3h. Wright, D.C., Failure of Plastics
and Rubber Products – Causes, Effects and Case Stud-
ies Involving Degradation, RAPRA (2001); 3i.
Wypych, G., ed., Plastics Failure Analysis and Pre-
vention, William Andrew Publ. (2001).
4. Ref. 1, p. 12, table 2-1.
5. Ref. 1, p. 23, sec. 2.3.2.2.1.
6. Ezrin, M. et al., Plastics Failure Due to Oxidative
Degradation in Processing and Service, SPE ANTEC,
3108-3112 (2000).
7. Ref. 1, p. 105, table 4-1.
8. Ref. 1, p. 30, fig. 2-13.
9. Ref. 1, p. 38, fig. 2-18.
10. Ref. 1, p. 26.
11. Ref. 1, p. 42, fig. 2-20.
12. Ref. 1, p. 43, fig. 2-20a.
13. Ezrin, M., G. Lavigne and J. Helwig, Product Failure
Due to Design, Material and Processing Problems,
SPE ANTEC, 3147-3150 (1998).
14. Ref. 1, p. 197.
15. Ezrin, M., Processing – Key to Preventing Plastic Part
Failure, Medical Design and Mfg. Conf., sec. 304
(1997).
16. Ezrin, M. and G. Lavigne, Case Studies of Failure of
Polymeric Electrical Insulation, SPE ANTEC, 2869-
2872 (2001).
17. Brown, R., ed., Handbook of Polymer Testing – Physi-
cal Methods, M. Dekker (1999).
18. Shah, V., Handbook of Plastics Testing Technology,
John Wiley & Sons (1998).
19. Ref. 1, p. 163, fig. 6-5.
20. Ref. 1, p. 178, fig. 6-18, and p. 344.
21. Ref. 1, chapter 7, sec. 7.5, p. 210.
22. Ghosh, T., J.C. Huang and M.C. Williams, Melt Index
for Molecular Weight Quality Control, Proc. XIIth Int.
Congr. on Rheology, 85-86 (1996).
23. Ref. 1, p. 41, sec. 2.6.3.
24. Ref. 1, p. 225, sec. 7.7.
25. Ref. 1, p. 226, fig. 7-24.
26. Ezrin, M. and G. Lavigne, Case Studies of Adhesive
Failure of Bonded Plastics, SPE ANTEC, 3936-3940
(1995).
27. Ezrin, M. and G. Lavigne, Case Studies of Plastics
Failure Related to Improper Formulation, SPE AN-
TEC, 3346-3349 (1999).
28. Ezrin, M. and G. Lavigne, Case Studies of Failure of
Polymeric Electric Insulation, SPE ANTEC, 2869-
2872 (2001).
29. Ref. 1, p. 214 and 230.
30. Cartz, L., Nondestructive Testing Radiography, Ultra-
sonics, Liquid Penetrant, Magnetic Particle, Eddy
Current, AMS International, Materials Park (1995).
31. Schueneman, G.T. et al., “Evaluation of Short Term-
High Intensity Thermal Degradation of Graphite Fiber
Reinforced Laminates via Ultrasonic Spectroscopy,” J.
Polym. Sci., Part B, Polym. Phys., 37, 2601-2610
(1999).
Keywords – plastics failure analysis, plastics analysis,
plastics testing
ANTEC 2002 / 3065
Figure 1 - DSC-OITime at 200°C of #1 unmolded PP pel-
lets (>50 min.); #2 unused molded rotor (4 min.); #3 rotor
Figure 5 - Close-up of hook end (left) of SAN piano key
degraded in service (1 min.).[6]
before (bottom) and after heat shrinkage (top).[12]

Delaminated
surface à

Figure 6 - Cross-section of degraded (right) and unde-

graded cable showing darkening of insulation started near
the outer part of the cable where the ground wires are lo-
cated.[19]

Figure 2 - 1/8” thick PE sheet delaminated at contaminant

site by punching to size.[8]

Figure 7 - Polyester fiberglass guy strain insulator rod frac-

tured in electric transmission service (mechanical support
rod) after 10-12 years. Cause of fracture is low resistance
Figure 3 - Generalized plot of properties vs. molecular of E glass to acid from snow and rain that accumulated and
weight.[9] concentrated in the top lip of the metal end fitting.[20]

Figure 4 - SAN piano key before (bottom) and after heat- Figure 8 - Photoelastic birefringence patterns in injection-
ing to release frozen-in stress.[11] molded polystyrene.[25]

3066 / ANTEC 2002