Performance Evaluation of Unexposed and Field-exposed Thermoplastic Polyolefin (TPO) Roof Membranes

Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 59
Performance Evaluation of Unexposed
and Field-exposed Thermoplastic
Polyolefin (TPO) Roof Membranes
Roof Consultants Institute
Jim Carlson1, Ana H. Delgado2, Ernie Rosenow1, K.C. Barnhardt1, and
Ralph M. Paroli2
1Western States Roofing Contractors Association
2National Research Council of Canada, Institute for Research in
Construction
ABSTRACT
Four test roofs were installed across the Western U.S. by the WSRCA in 2001
as part of its TPO Weathering Farm Project. An overview of the findings
gleaned from the field, as the membranes now progress into their fifth year
of weathering, will be presented. This will include field observations and field
solar reflectivity testing data. In addition, laboratory results such as: breaking
strength, elongation, thickness, glass transition temperature, thermogravimetry,
and spectroscopy will be presented.
Overall, this presentation is intended to provide roof consultants, roofing
contractors, manufacturers, roof designers, and specifiers with updated technical
information regarding TPO roof membranes. Attendees are expected to
depart with an understanding of the WSRCA TPO Weathering Farm Project,
TPO roof membranes in general, the problems with some of the earlier formulations,
the successes regarding reflectivity, and other current pertinent
information.
SPEAKERS
JIM CARLSON, technical director/principal consultant and founder of Building Envelope
Technology & Research in Seattle, Washington, has been a member of RCI since 1988. Jim previously
worked with the NRCA as its deputy director of technology and research. While with
NRCA, Jim conducted the re-write of the NRCA Roofing & Waterproofing Manual, of which he
authored several sections and co-authored others. Jim also co-authored the Repair Manual for
Low-Slope Membrane Roofing Systems, conducted numerous roofing research projects, including
joint projects with SPRI, ARMA, MRCA, and WSRCA, and authored many technical papers to benefit
the roofing and waterproofing industry. Jim began his career in the roofing industry in the
early 1970s working as an apprentice for a roofing contractor. Prior to his time with NRCA, Jim
served as senior field engineer/roofing consultant for Wetherholt & Associates, a consulting engineering
firm that specializes in roofing and waterproofing projects. Jim also serves as an officer
with ASTM, technical advisor for WSRCA, co-advisor for the Midwest Roofing Contractors
Association, and is an accomplished author and speaker for the industry.
RALPH M. PAROLI is the director of the Building Envelope and Performance Program at the
Institute for Research in Construction (IRC), National Research Council of Canada. Prior to this,
he led the Roofing Materials and Systems Performance activity at IRC. He has been involved with
spectroscopy and thermal analysis for nearly 20 years. As a researcher, Paroli has participated in
and led government research projects in the characterization of roof system performance. Paroli
is a chartered chemist with the Chemical Profession of Ontario. He is a member of the American
Chemical Society, the Chemical Society of Canada, the International Union of Pure and Applied
Chemistry, the Canadian Society for Analytical Sciences and Spectroscopy, the Society for Applied
Spectroscopy (USA), ASTM, and RCI. In addition, he is a past president of the Spectroscopy
Society of Canada and of the Canadian Thermal Analysis Society. He is a member of the Canadian
Roofing Contractors Association’s National Technical Committee and the vice-Chair of ASTM D-
08. Paroli received his B.S. degree in analytical chemistry from the University of Concordia and
his Ph.D. in physical/inorganic chemistry from the University of McGill, Montreal, Canada.
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 60 Proceeedings of the RCI 21st International Convention
SUMMARY
Western States Roofing Contractors
Association’s (WSRCA)
TPO Weathering Farm Project was
initiated to “Examine weathering
characteristics of thermoplastic
polyolefin (TPO) roof membranes
by exposing them to various
weather conditions of four distinctly
different climatic regions in
the Western United States.” The
project’s goal is: “To provide pertinent
technical and performance
information to the Western States
Roofing Contractors Association
membership, and the North
American roofing industry at
large, regarding TPO roof membranes’
attributes, their performance
properties, and resistance
to degradation due to the effects
of weathering.” Although it is a
western states focus, the results
will be applicable to other regions
in North America.
Four different roofing manufacturers
supplied the TPO roofing
membranes (blind-labeled as
Nos. 1 through 4). Sixteen out of
the 32 membranes were unexposed
and were used as a control.
The other 16 (one from each manufacturer)
have been exposed to
weather on full-scale test roofs in
Anchorage, Alaska; Las Vegas,
Nevada; Seattle, Washington; and
San Antonio, Texas. The thickness
(sheet overall and coating
over scrim), dimensional stability,
water absorption, tensile test,
elongation, glass transition temperature
(Tg) and thermal stability
(weight loss) of the TPO roof membranes
were evaluated by
mechanical testing, dynamic
mechanical analysis (DMA), and
thermogravimetry (TG). Results of
the analyses showed that the
mechanical and chemical properties
investigated did not undergo
significant changes after threeyear
exposure. The five-year data
is still being collected and analyzed,
to be reported in the near
future.
BACKGROUND
Forty-nine polyester/fiberreinforced
TPO roof membrane
samples were received from the
WSRCA. They were obtained from
exposure sites in four regions in
the U.S., labeled as Alaska (AK),
Las Vegas (LV), Seattle (SEA), and
Texas (TX). Sixteen of the 32 samples
were new and the rest had
been exposed for up to three
years. The suffix “u” was used for
unexposed or new samples,
whereas the suffix “e” was used
for the exposed samples. One
sample per manufacturer and
exposure location was from a
seam test cut. The top ply (weathered
surface) was white for all
samples. Note that the roll direction
(machine and cross directions)
was not specified on the
samples when received. Therefore,
using an optical microscope,
the directions were identified in
the laboratory by the shapes of
the fiber bundles on the crosssection
of the samples.
The test site locations were
chosen because they represent
the four more diverse climatic
regions of western North America
(see Map 1). The locales and climates
are: Las Vegas, Nevada,
generally hot and dry; San
Antonio, Texas, generally hot and
humid; Anchorage, Alaska, generally
cold and damp; Seattle,
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 61
Performance Evaluation of Unexposed
and Field-exposed Thermoplastic
Polyolefin (TPO) Roof Membranes
Map 1 – Test site locations were chosen to represent four
diverse climatic regions of western North America.
Washington, moderate with wet
and dry periods.
The on-roof weathering and
testing is unbiased, with all TPO
membranes blind-labeled and laid
out similarly in each of the four
test roofs (see Map 2).
FIELD OBSERVATIONS
This TPO roofing research
project has now entered into its
fifth year of weathering, originally
beginning the test roof installations
with the more moderate of
the four different climatic locations,
in Seattle, Washington.
Based on the visual field inspections
(Photos 1A-D), all of the roofs
appear to be performing well,
despite the rather harsh exposure
extremes.
For example, the Anchorage,
Alaska roof has experienced low
temperatures in excess of -30°F
(-34°C), and weeks of being covered
with snow and ice, to highs
in the +80°F (27°C), with sun; but
not as much sun load as the Las
Vegas, Nevada site that normally
has 210 very sunny days per year
and experiences rooftop temperatures
in excess of +130°F (54°C).
All seams remain welded, and
all roofs are leak free, despite the
maintenance traffic to which they
are subjected, including one on a
cabinet shop in Las Vegas and one
on a hospital in Anchorage. The
Las Vegas roof has swamp coolers
mounted on curbs, which regularly
need their filters changed due
to the heavy dust from the saws
and sanders inside the building
and the volatile stain and paint
fallout from the finishing shop
area inside that exhausts various
petroleum distillate fumes from
power exhaust vents mounted on
the roof. From the exposure to
sun (e.g., ultraviolet radiation)
and heat loads, combined with
their thermal expansion and contraction
differentials; to the ice,
snow and rain, including wet-dry
cycling, the maintenance traffic,
and the air-borne dust and fallout
– these 60-mil, white, TPO membranes
are so far showing good inservice
performance.
Seam Integrity
In general, all of the TPO
membranes examined in the field
to date have maintained their
seam quality. All hot-air welded
seams, which are randomly
probed during the yearly inspections,
are proving to have great
weld integrity. Initially, a few
seams on a couple of the roofs
were thought to have opened
slightly (approximately 1/16 to
1/4-inch) between the time the
roofs were initially installed and
the first years’ follow-up inspection.
Transitions adjacent to a
drain in Las Vegas were suspect
last year. However, once the
seams were inspected and
touched-up (i.e., reheated and
welded closed along the outer
seam edge), the welds stayed
closed. Therefore, it is now
thought that these few initial
seam edges were not originally
welded as thoroughly (i.e., cold
welds) as needed.
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 62 Proceeedings of the RCI 21st International Convention
Map 2 – Roof layout.
Seattle Alaska
Photo 1 – Test roofs in order of installation: Seattle (A),
Alaska (B), Las Vegas (C), and Texas (D).
Las Vegas Texas
Membranes Initially
Loose, But Have Since
Tightened
Judging from the
yearly field inspections
and data collection, there
has been some tightening
of the sheets, which is
visible at a few locations
as bridging at roof-towall
intersections and
some roof-to-area divider
transitions. The tightening
is not unique to
TPOs. It has also been
observed over the years with PVC
and EPDM roof membranes,
which can tighten as they age.
The cause of tightening, however,
is different.
Initially, and up through the
third year of weathering, some of
the TPO membranes were somewhat
loose or baggy with some
very apparent wrinkling (this
could be due to time of inspection
– i.e., morning). The Texas roof,
which was installed slightly differently
than the others (by first
welding neighboring sheets, then
installing the mechanical fasteners),
did not exhibit the looseness
that the other three roofs presented
during their first, second, and
to some extent, the third years of
service. This would indicate that
installation techniques and workmanship
can significantly affect
the overall appearance of the
membrane. However, as all of the
roofs have aged through their
third and fourth years, most have
tightened and are now laying
quite smooth. There are isolated
areas of bridging that have been
identified and marked, and they
will be monitored as this research
and testing project progresses
(See Photos 2 and 3).
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 63
Photo 2 (left) – Seattle test roof before three years of installation. Photo
3 (right) – Seattle test roof after three years of installation.
Table 1 – Reflectivity values for TPO membranes before and after cleaning.
Surface Characteristics and
Color or Hue
Another unique quality that
has been observed is the surface
condition of the TPO roof membrane
sheets. Some of the sheets
are noticeably dirtier than others,
apparently accumulating dust
and airborne particulate differently
than their neighboring sheets.
Research will hopefully determine
if static charge(s) are different,
potentially due in part to different
polyisocyanurate insulation facers.
Or it may show, if membrane
surface roughness increases over
time (for example, developing a
rougher surface characteristic
such as PVC, which may change
from almost slick to rough – e.g.,
with surface crazing), thus making
the membrane more prone to
holding dust and airborne particles
and visibly accumulating dirt,
as all other roof membranes do
over their years of service.
Solar Reflectivity
Reflectivity measurements
shown in Table 1 were obtained
on the roofs during the third-year
(fourth-year in Seattle) inspections.
Solar reflectivity measurements1
were conducted on both
weathered areas and areas that
had been cleaned (Photos 4 – 5). A
comparison of two common cleaning
methods was tested. Areas
were wetted, scrubbed with stiffbristle
push brooms or brushes,
then rinsed, along with a side-byside
comparison of pressure rinsing/“
washing.” In nearly all cases,
the scrubbed areas produced a
visually cleaner and a slightly
more reflective surface.
However, even before cleaning,
all of the TPO roofs far surpassed
the ENERGY STAR® reflectivity
requirements for three-year-old
roof membranes. At all the areas,
solar reflectivity improved after
cleaning, though no membrane
returned to its original whiteness.
It should be noted that mild
cleaning procedures were used for
testing, and that no strong detergents
or cleaning chemicals were
used to clean the TPO surface.
Fifty-eight percent of the reflectivity
component occurs in the nonvisible
region of the solar spectrum.
EXPERIMENTAL METHOD
Mechanical Properties
Visual inspection showed that
the weathered surface of the samples
was in good condition with
neither pinholes nor crazes. Some
of the field samples were dusty, so
a wet paper towel was used to
wipe the surface before testing.
The following physical and
mechanical properties of all the
samples were measured.
Thickness, Sheet Overall
The overall sheet thickness of
the polyester-reinforced roofing
membranes was measured using
a digital micrometer (ASTM D751-
00 [1]). Five measurements were
taken at randomly-selected points
over the sample surface to obtain
a representative average.
Thickness of Coating Over
Scrim
The thickness of coating over
scrim was measured by modifying
the optical method described in
Annex A1 of ASTM D6878-03 [2].
A rectangular specimen of approximate
dimensions of 50 mm x
5 mm (2 in. x 0.2 in.) was cut from
each sample. A clean cut through
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 64 Proceeedings of the RCI 21st International Convention
Photo 4 – Membrane after cleaning.
Photo 5 – Reflectivity measurements.
1The task group wishes to acknowledge Ross Roberts and Firestone for their extra contributions of expertise, testing equipment, and travel expenses.
the sheet was made on
the long dimension of the
specimen by cutting along
the center of a fiber
bundle using a sharp
utility knife. The crosssection
was placed under
an optical microscope
equipped with a CCD
camera. Micrographs at
50X magnification were
digitally captured by the
camera and stored on a
computer. The thickness
of the weather surface at
each fiber intersection
was measured on the
digitized micrograph, using
image-analysis software.
Linear Dimensional Change
The linear dimensional change
of the samples was measured
based on ASTM D1204-02 [3]. The
specimens were conditioned at
70 ± 2°C (158 ± 3°F) for six hours
as specified by ASTM D6878-03
[2]. Only one specimen of dimensions
250 mm x 250 mm (10 in. x
10 in.) was cut from each sample.
Before conditioning, each specimen
was marked and measured
at three equally spaced points
along each side in both the
machine (MD) and cross (XD) directions,
using a digital calliper.
The specimens were dusted with
talc and sandwiched between two
pieces of silicone-coated paper
that were secured together with
paperclips. The assemblies were
placed in a convection oven at
70 ± 2°C (158 ± 3°F) for six hours.
They were then removed from the
oven and reconditioned in the laboratory
at 23 ± 2°C (73 ± 4°F) and
50 ± 5% relative humidity for at
least one hour before final measurements
were made.
The linear dimensional change
is the change in dimension expressed
as a percent of the original
dimension. A positive linear
dimensional change indicates
expansion while a negative value
denotes shrinkage.
Water Absorption
The resistance to water
absorption was measured according
to ASTM D471-98 [4] on the
weathered surface of the samples
only. The samples were cut into
disks of 60 mm (2.4 in.) in diameter
and initial weights recorded.
The specimen was placed in the
apparatus, weather-side up. The
chamber was then filled with
deionised water to a depth of
approximately 15 mm (0.6 in.),
which was equivalent to about 30
mL (1.8 in.3) in volume, and tightly
sealed by a stopper. The assembly
was placed in an oven at 70 ±
2°C (158 ± 3°F) for 166 hours. At
the end of the immersion period,
the apparatus was allowed to cool
to room temperature in the laboratory
for two to three hours. The
specimen was released from the
apparatus and excess surface
water was removed by blotting
with a filter paper. The final
weights of the specimens were
immediately measured.
Tensile Properties
The tensile properties of the
samples were measured according
to ASTM D751-00 [1],
Procedure B – Cut Strip Test
Method, instead of the Procedure
A – Grab Test Method as specified
by ASTM D6878-03 [2] due to the
limited materials available. The
samples were cut into specimens
of 25 mm x 150 mm (1 in. x 6 in.)
using a pneumatic die cutter. The
specimens were tested in a universal
testing machine at a constant
crosshead speed of 5 ± 0.2
mm/s (12 ± 0.5 in./min). The
gauge length of the specimens
was 75 mm (3 in.). A minimum of
five specimens was tested in both
machine (MD) and cross (XD)
directions except for some samples
that had limited material.
Chemical Properties
Dynamic mechanical analysis
(DMA)
Sample preparation
After cleaning the samples
with wet paper towel, a piece of
approximately 40 x 20 mm (1.6 x
0.8 in.) was cut from each of the
“as received” samples. One exposed
sample from each manufacturer
was from a seam cut; therefore,
a specimen was cut from top
and bottom sheets and labeled as
‘t’ and ‘b’ (Figure 1). Unexposed
and exposed specimens were conditioned
in an air-circulating oven
for one hour at 80 ± 2°C (176 ±
36°F) (as per ASTM D 6382-99) [5] to provide a consistent thermal
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 65
Figure 1 – Picture of the top and bottom sheets from a seam cut of
exposed samples.
history before testing.
They were removed
from the oven, stored
in a desiccator, and
allowed to relax for at
least eight hours before
conducting the analysis.
A Rheometric Scientific
RSA3 (now TA
Instruments) Solid Analyzer
equipped with
an environmental controller
and a dual cantilever
geometry was
used to measure the
glass transition temperature
(Tg) of the unexposed
and exposed
samples. Rectangular
strips were cut with a
utility knife and placed
in the dual cantilever
fixture. The specimen
was cooled to -80°C
(-112°F), the temperature
was allowed to stabilize
for five minutes
and increased at
2°C/min (4°F/min) to
the final temperature
program. All pre-conditioned
(control) specimens
were run at least
in duplicate. The test
was conducted as per
ASTM D 5418-01 [6].
Experimental parameters
are given below:
Geometry: Dual cantilever
Sweep type: Dynamic Temperature
Ramp
Dimensions: Length: Fixed at 36.67
mm (1.44 in.)
Frequency: 1 Hz (6.28 rad/sec)
Temperature program: -80°C to 50°C (-112 to
122°F)
Ramp Rate: 2°C (4°F)
Soak Time: 1 minute
Time per Measure: 1 minute
Strain: 0.01-0.02%
Autotension Mode: Off
Autostrain Mode: On
Strain Adjustment: 25-30% of current
strain
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 66 Proceeedings of the RCI 21st International Convention
Exposure Thickness Overall** Exposure Thickness Overall**
Sample Time ThicknessChanges Sample Time Thickness Changes
ID* (yrs) (mm)+ (mm) ID* (yrs) (mm)+ (mm)
AK1 0 1.4 – AK3 0 1.5 –
1 1.4 0.0 1 1.5 0.0
3 1.5 0.1 3 1.5 0.0
AK2 0 1.4 – AK4 0 1.4 –
1 1.5 0.1 1 1.4 0.0
3 1.5 0.1 3 1.5 0.1
LV1 0 1.5 – LV3 0 1.4 –
1 1.5 0.0 1 1.4 -0.0
3 1.6 0.1 3 1.4 0.0
LV2 0 1.4 – LV4 0 1.5 –
1 1.4 0.0 1 1.4 -0.1
3 1.5 0.1 ` 3 1.4 -0.1
SEA1 0 1.4 – SEA3 0 1.5 –
1 1.4 0.0 1 1.5 0.0
3 1.4 0.0 3 1.5 0.0
SEA2 0 1.4 – SEA4 0 1.4 –
1 1.4 0.0 1 1.5 0.1
3 1.4 0.0 3 1.5 0.1
TX1 0 1.4 – TX3 0 1.5 –
1 1.4 -0.0 1 1.4 -0.1
3 1.4 -0.0 3 1.4 -0.1
TX2 0 1.5 – TX4 0 1.4 –
1 1.5 -0.0 1 1.4 -0.0
3 1.5 0.0 3 1.4 -0.0
Table 2 – Average thickness, sheet overall of unexposed, 1- and 3-year
exposed TPO roof membranes.
*AK = Alaska; LV = Las Vegas; SEA = Seattle; TX = Texas
(1-4) = Manufacturer #
+1 inch = 25.4 mm
ASTM D 6878-03 specification: 1.0 mm (0.039 in)
RESULTS AND DISCUSSION
Thickness, Sheet Overall
The overall sheet thickness of
the samples is summarized in
Table 2 and Figure 2. Each entry
is the average and standard deviation
of five individual measurements.
The overall sheet thickness
of the samples ranged from
1.35 -1.55 mm (0.053 -0.061 in.),
which significantly exceeded the
minimum requirement of 1.0 mm
(0.039 in.) as specified in ASTM
D6878 for new TPO membranes.
The overall sheet thickness
changed less than ±0.10 mm
(0.004 in.) after exposure.
Thickness of Coating Over
Scrim
The thickness of the top coating
over the fiber intersection of
the samples is summarized in
Table 3 and Figure 3. Each entry
is the average and standard deviation
of six individual measurements.
The thickness over scrim
value represents the minimum
coating thickness on the weathered
surface of the samples. The
coating thickness for all samples
exceeded the minimum requirement
of 0.305 mm (0.012 in.) as
specified in ASTM D6878 for new
TPO membranes. It ranged from
0.470 – 0.650 mm (0.018 – 0.026
in.) and 0.390 – 0.590 mm (0.015
– 0.023 in.) for the unexposed and
exposed samples, respectively.
The coating thickness reduced
by 0.005 – 0.170 mm (0.0002 –
0.007 in.) after one year of service
in the field. The change in top
coating thickness was within the
experimental error of the measurement
method and thus considered
insignificant for all samples
but Sample LV1. Its top coating
thickness over scrim dropped
0.167 mm (0.007 in.) after one
year of exposure. In some cases,
this variance may be due to the
manufacturing tolerances. However,
more data is required to see
if this trend continues
Linear Dimensional Change
The linear dimensional
change of the samples is summarized
in Table 4. Each entry is the
average and standard deviation of
three individual measurements
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 67
Figure 2 – Thickness, sheet overall of TPO samples: Alaska, Las Vegas, Seattle, and Texas.
on one specimen. The
linear dimensional
change of all samples
was less than ±0.4%,
which was within the
maximum allowable
limit specified in
ASTM D6878 for new
TPO membranes. No
significant difference
was observed between
the unexposed
and the exposed samples.
This indicated
that the samples remained
dimensionally
stable after one year
of service in the field.
Water Absorption
The amount of
water absorption of
the samples is summarized
in Table 5.
Each entry is the
average and standard
deviation of three
specimens unless
otherwise specified.
The water absorption
changes by weight of
all samples (see
Figure 4) was less
than ±0.4%, which
was well within the
maximum allowable
limit of ±3% as specified
in ASTM D6878
for new TPO mem-
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 68 Proceeedings of the RCI 21st International Convention
Exposure Thickness Over Scrim** Exposure Thickness Over
S c r i m * *
Sample Time Thickness Changes Sample Time Thickness Changes
ID* (yrs) (mm)+ (mm) ID* (yrs) (mm)+ (mm)
AK1 0 0.529 – AK3 0 0.537 –
1 0.494 0.035 1 0.451 -0.086
3 0.461 0.068 3 0.445 -0.092
AK2 0 0.507 – AK4 0 0.555 –
1 0.475 -0.032 1 0.539 -0.016
3 0.388 -0.119 3 0.430 -0.125
LV1 0 0.643 – LV3 0 0.623 –
1 0.476 -0.167 1 0.501 -0.122
3 0.511 -0.132 3 0.497 -0.126
LV2 0 0.574 – LV4 0 0.551 –
1 0.536 -0.038 1 0.491 -0.060
3 0.519 -0.055 ` 3 0.458 -0.093
SEA1 0 0.492 – SEA3 0 0.584 –
1 0.488 -0.004 1 0.526 -0.058
3 0.430 -0.062 3 0.500 -0.084
SEA2 0 0.476 – SEA4 0 0.501 –
1 0.393 -0.083 1 0.477 -0.024
3 0.383 -0.093 3 0.400 -0.101
TX1 0 0.555 – TX3 0 0.552 –
1 0.486 -0.069 1 0.446 -0.106
3 0.554 -0.001 3 0.414 -0.138
TX2 0 0.616 – TX4 0 0.508 –
1 0.589 -0.027 1 0.498 -0.010
*AK = Alaska; LV = Las Vegas; SEA = Seattle; TX = Texas; (1-4) = Manufacturer #
**Fiber (cut along XD fiber)
+1 inch = 25.4 mm
ASTM D 6878-03 specification: 0.305 mm (0.012 in)
Table 3 – Average thickness, coating over scrim of unexposed, 1- and 3-year
exposed TPO roof membranes.
Figure 3a – Thickness, coating over scrim of TPO samples: Alaska and Las Vegas.
branes. This indicated
that the
weathered surface
of the samples was
resistant to water
absorption.
It is interesting
to note that all the
exposed samples
had positive water
absorption values
while some of the
unexposed samples
had negative
water absorption
values up to -0.15%;
i.e., they lost
weight after being
in contact with water
at 70°C (158°F)
for seven days. It is
suspected that the
weight loss was
due to heat exposure
in the oven.
To estimate the
weight loss due to
the heat exposure,
two small rectangular
specimens of
dimensions 25 mm
x 50 mm (1 in. x 2
in.) were cut from
each sample and
placed in a convection
oven at 70 ±
2°C (158 ± 3°F) for
seven days, the same temperature
and exposure period as the water
absorption test. The weight loss
due to heat exposure only was
calculated from the weights of the
specimens before and after heating.
The weight loss ranged from 0
– 0.25% with an average of about
0.1%. The measurement error of
the electronic balance was ±0.001
g, which corresponded to about
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 69
*n/a = not enough material for testing
ASTM D 6878-03 specification for membranes new: ±1%
Table 4a – Linear dimensional change of unexposed, 1- and 3-year exposed
TPO roof membranes (continued on next page).
Figure 3b – Thickness, coating over scrim of TPO samples: Seattle and Texas.
Exposure Linear Dimensional Change(%) Linear Dimensional Change (%)
Sample Time (yrs) Machine Direction Change (%) Cross Direction Change (%)
AK1 0 -0.10±0.04 – -0.09±0.02 –
1 n/a* – n/a* –
3 -0.05±0.03 0.05 -0.03±0.01 0.06
AK2 0 -0.23±0.04 – -0.26±0.06 –
1 -0.16±0.00 0.07 -0.12±0.02 0.14
3 -0.16±0.02 0.07 -0.07±0.01 0.19
AK3 0 -0.06±0.06 – -0.06±0.05 –
1 -0.09±0.00 -0.03 -0.03±0.00 0.03
3 -0.06±0.01 0.00 0.00±0.00 0.06
AK4 0 -0.16±0.06 – -0.04±0.02 –
1 -0.06±0.06 0.10 -0.05±0.00 -0.01
3 -0.07±0.01 0.09 0.00±0.00 0.04
LV1 0 0.18±0.01 – -0.05±0.02 –
1 -0.03±0.01 -0.21 -0.08±0.01 -0.03
3 -0.03±0.01 -0.21 0.04±0.00 0.09
LV2 0 -0.05±0.01 – -0.10±0.07 –
1 -0.06±0.02 -0.01 -0.13±0.00 -0.03
3 -0.11±0.02 -0.06 0.01±0.01 0.11
LV3 0 n/a* – n/a* –
1 -0.03±0.01 – -0.04±0.00 –
3 -0.02±0.00 – 0.03±0.01 –
LV4 0 -0.12±0.04 – -0.03±0.01 –
1 -0.05±0.01 0.07 -0.02±0.01 0.01
3 -0.04±0.01 0.08 0.03±0.00 0.06
0.02% of the weight
of the specimen
(about 4.5 g).
Therefore, within
measurement error
of the balance, this
confirmed that the
negative water absorption
data came
from weight loss of
the samples due to
heating.
Since the samples
lost weight due
to heat exposure in
the oven while
gaining weight
from water absorption
at the same
time, the final
weight measurement
in the water
absorption test underestimated
the
actual weight gain
due to water absorption.
The actual
water absorption
values should be
the sum of weight
loss due to heat
exposure and the
water absorption
data, based on final
weight measurement
(Table 5). The
average weight
losses due to heat exposure in the
oven are low: 0.11% and 0.18%
for the unexposed and the
exposed samples, respectively.
The measured water absorption
data were less than ±0.4%. After
adjusting the weight loss due to
heat exposure, the actual water
absorption would still be within
the maximum allowable limit of
±3% for the samples tested. In
other situations where the samples
are sensitive to losing weight
due to heat exposure, it is important
to adjust the water absorption
data accordingly.
Tensile Test
Tensile Response and General
Observations
The mechanical response of
the samples under tensile loading
was measured and displayed as
force-displacement curves. They
varied with manufacturers, test
direction, and exposure. There
were two types of samples made
by manufacturer #2: AK2 and
SEA2 consisted of black and
white polymer matrix; while LV2
and TX2 had grey and white polymer
matrix. They responded differently
to tensile loading.
For AK2 and SEA2 (black and
white), after the reinforcement
broke, the polymer matrix bore
the load. Stress concentration
was observed at fiber bundles
running across the width of the
specimen and cracks started to
form and propagate in these areas
until the specimen failed. The
polymer plies remained intact
where no fiber was present.
The same failure mode was
observed for the exposed samples
in XD. However, the samples in
MD changed failure mode after
one year of field service. The specimen
broke in the top ply, and the
bottom ply continued to stretch
and peel away from the top ply.
The change in failure type possibly
indicated that the bonding
between the polymer plies was
weakened after exposure.
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 70 Proceeedings of the RCI 21st International Convention
Exposure Linear Dimensional Change(%) Linear Dimensional Change (%)
Sample Time (yrs) Machine Direction Change (%) Cross Direction Change (%)
SEA1 0 -0.06±0.01 – -0.12±0.02 –
1 -0.04±0.02 0.02 -0.09±0.02 0.03
3 -0.04±0.01 0.02 0.02±0.01 0.14
SEA2 0 -0.36±0.02 – -0.22±0.01 –
1 -0.08±0.06 0.28 -0.29±0.02 -0.07
3 -0.21±0.02 0.15 0.10±0.00 0.32
SEA3 0 n/a* – n/a* –
1 -0.05±0.03 – 0.00±0.02 –
3 -0.03±0.01 – 0.09±0.02 –
SEA4 0 -0.19±0.04 – -0.06±0.03 –
1 -0.15±0.04 0.04 0.00±0.01 0.06
3 -0.07±0.01 0.12 0.05±0.03 0.11
TX1 0 0.05±0.02 – -0.03±0.04 –
1 -0.01±0.01 -0.06 -0.02±0.02 0.01
3 -0.01±0.00 -0.06 -0.04±0.01 -0.01
TX2 0 -0.15±0.04 – -0.04±0.02 –
1 -0.09±0.01 0.06 -0.08±0.04 -0.04
3 -0.07±0.01 0.08 -0.03±0.01 0.01
TX3 0 -0.09±0.03 – -0.05±0.02 –
1 -0.06±0.01 0.03 -0.04±0.01 0.01
3 0.00±0.01 0.09 -0.02±0.01 0.03
TX4 0 -0.13±0.00 – -0.09±0.02 –
1 -0.04±0.01 0.09 -0.10±0.02 -0.01
3 -0.03±0.02 0.10 0.00±0.01 0.09
Table 4b – Linear dimensional change of unexposed, 1- and 3-year exposed
TPO roof membranes (continued from previous page).
*n/a = not enough material for testing
ASTM D 6878-03 specification for membranes new: ±1%
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 71
Measured Corrected
Sample Exposure Water Change Water Change
ID Time (yrs) Absorption % % Absorption % (%)
AK1 0 -0.05 – 0.01 –
1 0.07 0.12 0.30 0.29
3 0.09 0.14 0.29 0.28
AK2 0 -0.12 – -0.09 –
1 0.05 0.17 0.19 0.28
3 0.10 0.22 0.19 0.48
AK3 0 -0.07 – 0.06 –
1 0.05 0.12 0.24 0.18
3 0.07 0.14 0.36 0.30
AK4 0 0.01 – 0.09 –
1 0.07 0.06 0.18 0.09
3 0.06 0.05 0.27 0.18
LV1 0 -0.15 – -0.05 –
1 0.08 0.23 0.13 0.18
3 0.23 0.38 0.31 0.36
LV2 0 -0.02 – -0.02 –
1 0.17 0.19 0.23 0.25
3 0.30 0.32 0.50 0.52
LV3 0 0.35 0.38 –
1 0.36 0.01 0.47 0.09
3 0.50 0.15 0.61 0.23
LV4 0 0.02 – 0.15 –
1 0.14 0.12 0.27 0.12
3 0.36 0.34 0.46 0.31
SEA1 0 0.11 – 0.17 –
1 0.08 -0.03 0.08 -0.09
3 0.24 0.16 0.39 0.22
SEA2 0 0.07 – 0.16 –
1 0.03 -0.04 0.15 -0.01
3 0.28 0.21 0.58 0.42
SEA3 0 0.24 – 0.27 –
1 0.04 -0.20 0.18 -0.09
3 0.18 -0.06 0.47 0.20
SEA4 0 0.06 – 0.09 –
1 0.08 0.02 0.19 0.10
3 0.22 0.16 0.56 0.47
TX1 0 -0.10 – -0.04 –
1 0.12 0.22 0.15 0.19
3 0.24 0.34 0.46 0.50
TX2 0 0.01 – 0.04 –
1 0.09 0.08 0.12 0.08
3 0.27 0.26 0.44 0.40
TX3 0 0.13 – 0.21 –
1 0.19 0.06 0.27 0.06
3 0.41 0.28 0.52 0.31
TX4 0 0.01 – 0.01 –
1 0.14 0.13 0.24 0.23
3 0.22 0.21 0.51 0.50
Table 5 – Water absorption values for unexposed, 1- and 3-yr
exposed TPO roof membranes.
Note: Negative weight change denotes weight loss.
ASTM D 6878-03 specification: ±3%
Figure 4: Water absorption
changes of TPO samples.
Water Absorption ±3% Allowed
For LV2 and TX2 (grey
and white), in MD, after
the reinforcement broke,
the polymer matrix did
not stretch much before
failure, with ultimate
elongation around 100%.
Exposure reduced the
ultimate elongation further.
In XD, the polymer
matrix elongated over
200% before failure. It is
suspected that the polymer
matrix might have
been strengthened or
weakened preferentially
during the manufacturing
process of the membranes.
The four failure types
were the same at the
beginning and the differences
occurred after the
reinforcement break, i.e.
the differences reflected
the tensile behavior of the
polymer matrix. Weathering
also affects polymer
matrix predominantly as
it was directly exposed to
the elements. On the
other hand, the tensile
properties such as breaking
strength and elongation
at break are mainly
controlled by the reinforcement.
Therefore, outdoor
exposure could have
affected the polymer
matrix (thus different failure
types) in this case but
the tensile properties remained
relatively unchanged.
However, if the
polymer matrix continues
to degrade, as it would be
expected due to weathering
effects, the reinforcement
underneath the
polymer ply will eventually
be exposed and this
may affect its performance.
The tensile breaking
strength and the elongation
at break are summar
i z e d
i n
Tables
6 and
7. Note
t h a t
procedure
B
– strip
t e s t
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 72 Proceeedings of the RCI 21st International Convention
Table 6 – Breaking strength of unexposed, 1- and 3-year exposed
TPO roof membranes.
*(MD) = Machine direction **1 Newton (N) = 0.2248 lbs +(XD) = Cross direction
ASTM D 687803 specification: 976 N (220 lbf) is for grab method and it does not apply to this
data, which is for the strip method.
Sample Exposure Breaking Strength (MD)* Breaking Strength (XD)+
ID Time (yrs.) N** Change N** Change
AK1 0 956 ± 38 – 922 ± 21 –
1 995 ± 13 39 904 ± 37 -18
3 985 ± 14 30 903 ± 19 -19
AK2 0 883 ± 36 – 735 ± 39 –
1 801 ± 30 -82 747 ± 25 12
3 715 ± 19 -168 682 ± 28 -53
AK3 0 896 ± 17 – 736 ± 50 –
1 919 ± 60 23 733 ± 16 -3
3 926 ± 21 30 701 ± 21 -35
AK4 0 1055 ± 27 – 849 ± 30 –
1 969 ± 18 86 838 ± 11 -11
3 974 ± 37 -81 853 ± 4 4
LV1 0 970 ± 40 – 808 ± 40 –
1 950 ± 41 -20 806 ± 29 -2
3 976 ±10 6 775 ± 18 -33
LV2 0 862 ± 30 – 748 ± 30 –
1 884 ± 40 22 742 ± 30 -6
3 913 ± 6 51 815 ± 23 67
LV3 0 876 ± 40 – 709 ± 10 –
1 815 ± 21 -61 773 ± 32 64
3 811 ± 28 -65 746 ± 28 37
LV4 0 951 ± 40 – 767 ± 10 –
1 964 ± 11 13 740 ± 17 -27
3 913 ± 24 -38 753 ± 4 -14
SEA1 0 958 ± 31 – 846 ± 25 –
1 995 ± 32 37 873 ± 39 27
3 929 ± 40 -29 857 ± 8 11
SEA2 0 801 ± 10 – 596 ± 17 –
1 827 ± 18 26 705 ± 29 109
3 857 ± 33 56 622 ± 32 26
SEA3 0 865 ± 20 – 677 ± 32 –
1 857 ± 29 -8 809 ± 20 132
3 897 ± 33 32 827 ± 8 150
SEA4 0 968 ± 40 – 827 ± 42 –
1 975 ± 51 7 771 ± 30 -56
3 1025 ± 36 57 842 ±13 15
TX1 0 961 ±29 – 819 ± 31 –
1 909 ± 33 -52 780 ± 19 -39
3 945 ± 5 16 746 ± 31 -73
TX2 0 823 ± 28 – 727 ± 31 –
1 858 ± 25 35 655 ± 36 -72
3 870 ± 35 47 731 ±24 4
TX3 0 912 ± 22 – 720 ± 25 –
1 831 ± 15 -81 740 ± 24 20
3 808 ± 7 -104 741 ± 25 21
TX4 0 962 ± 19 – 749 ± 31 –
1 964 ± 6 2 766 ± 18 17
3 940 ± 26 -22 680 ± 20 -69
method was used instead of procedure
A – grab test method in
ASTM D751 due to the lack of
materials. The ASTM specification
for tensile breaking
strength of 976 N (220 lbf) was
based on the grab test method;
therefore, the value was put on
the figures for reference only.
However, samples were tested
before and after exposure.
Therefore, if the membrane
material undergoes any changes
after exposure, it is possible to
compare the results obtained by
the same method.
The tensile breaking
strengths of the samples were
800 – 1050 N (180 – 235 lbf) and
650 – 920 N (145 – 205 lbf) for
MD and XD, respectively. They
were generally lower than the
ASTM specification of 976 N
(220 lbf) because strip test
method was used instead of the
grab test method. As expected,
the breaking strength in the MD
was equal or stronger than the
XD. In general, there was no significant
difference between the
unexposed and the exposed
samples. One year of service in
the field did not affect the tensile
breaking strength of the
samples. The breaking strength
between the samples made by
different manufacturers was
similar, mostly within ±15%.
The tensile elongations at
break of the samples are 29 –
44% and 26 – 46% for MD and
XD, which significantly exceeded
the ASTM specification of
15%. They were similar for MD
and XD. The difference between
the unexposed and the exposed
samples was generally small,
less than ±20%. One exception
was Sample SEA2u, whose elongation
at break increased by
about 45% after exposure.
Unfortunately, due to the lack of
samples, only one specimen was
tested in the unexposed samples
so its statistical validity is questionable.
The elongation at break
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 73
Table 7 – Elongation at break of unexposed, 1- and 3-year
exposed TPO roof membranes.
*(MD) = Machine direction +(XD) = Cross direction
ASTM D 687803 specification: 15%
Sample Exposure Elongation at Break (MD)* Elongation at Break (XD)+
ID Time (yrs.) % Change % Change
AK1 0 38.5 ± 4.7 – 39.0 ± 0.7 –
1 38.9 ± 0.9 0.4 34.2 ± 1.4 -4.8
3 39.9 ± 1.5 1.4 34.0 ± 1.6 -5.0
AK2 0 33.5 ± 3.2 – 27.3 ± 4.8 –
1 32.3 ± 2.3 1.2 32.4 ± 1.2 5.1
3 31.9 ± 2.5 1.6 27.3 ± 0.9 0.0
AK3 0 33.9 ± 4.3 – 38.9 ± 1.2 –
1 31.0± 0.8 -2.9 36.9 ± 1.2 -2.0
3 35.1 ± 0.4 1.2 32.1 ± 0.6 -6.8
AK4 0 32.9 ± 2.0 – 41.7 ± 2.0 –
1 30.5 ± 1.0 -2.4 42.9 ± 0.4 1.2
3 33.6 ± 1.5 0.7 40.8 ± 0.5 -0.9
LV1 0 39.4 ± 2.2 – 33.8 ± 1.1 –
1 38.1 ± 1.1 -1.3 30.9 ± 2.2 -2.9
3 36.8 ± 0.7 2.6 28.9 ± 0.6 -4.9
LV2 0 36.1 ± 1.4 – 26.8 ± 1.0 –
1 33.7± 2.3 -2.4 26.3 ± 1.0 -0.5
3 33.7 ± 1.0 -2.4 25.8 ± 0.8 -1.0
LV3 0 36.3 ± 1.1 – 29.6 ± 0.1 –
1 35.0 ± 1.2 -1.3 29.2 ± 1.8 -0.4
3 32.5 ± 1.1 3.8 25.7 ± 0.8 -4.0
LV4 0 32.1 ± 0.5 – 32.9 ± 1.4 –
1 30.8 ± 1.1 1.3 30.3 ± 1.7 -2.6
3 28.0 ± 0.7 4.1 28.9 ± 0.7 -4.0
SEA1 0 38.8 ± 3.0 – 41.3 ± 0.3 –
1 36.8 ± 2.7 -2.0 34.2 ± 1.1 -7.1
3 44.6 ± 0.9 5.8 33.9 ± 0.3 -7.4
SEA2 0 40.4 ± 1.2 – 47.0 ± 3.1 –
1 38.0 ± 1.9 -2.4 44.7 ± 2.0 -2.3
3 38.5 ± 1.0 -1.9 37.6 ± 0.8 -9.4
SEA3 0 43.8 ± 2.3 – 36.8 ± 1.4 –
1 40.4 ± 1.7 -3.4 41.3 ± 1.6 4.5
3 41.2 ± 0.6 -2.6 36.5 ± 0.8 -0.3
SEA4 0 34.2 ± 1.7 – 42.4 ± 1.0 –
1 32.2 ± 1.1 -2.0 41.3 ± 1.2 -1.1
3 32.3 ± 0.6 -1.9 41.1 ± 0.9 -1.3
TX1 0 39.5 ± 2.2 – 33.8 ± 1.7 –
1 40.2 ± 2.2 0.7 34.9 ± 1.7 1.1
3 39.8 ± 0.6 0.3 31.5 ± 0.7 -2.3
TX2 0 39.8 ± 0.7 – 32.5 ± 1.3 –
1 37.0 ± 1.8 -2.8 27.2 ± 1.8 -5.3
3 37.0 ± 0.8 -2.8 28.7 ± 1.7 -3.8
TX3 0 34.9 ± 1.0 – 39.3 ± 1.1 –
1 36.4 ± 0.5 1.5 36.8 ± 0.9 -2.5
3 32.9 ± 0.3 -2.0 29.2 ± 1.1 -10.1
TX4 0 41.4 ± 1.6 – 32.5 ± 1.4 –
1 33.8 ± 0.4 -7.6 31.1 ± 2.9 -1.4
3 44.1 ± 1.2 2.7 26.5 ± 0.8 -6.0
value between samples made by
different manufacturers was similar,
mostly within ±15%.
Chemical Analysis
Dynamic Mechanical Analysis
(DMA)
Figure 5A displays typical
DMA curves for TPO. The figures
shows E’ and E” are the storage
and loss modulus, respectively
and tan or the damping factor.
Figure 5B shows how the Tg was
obtained from the E’ and E”
curves. The values were obtained
from the intersect of two tangents
on the storage modulus (E’int)
curve as well as from the maximum
of the loss modulus (E”max)
curve.
Due to the large amount of
data, the data is not presented
here. The Tg obtained storage
modulus (E’int) and the maximum
of the loss modulus (E”max) curves
ranged from -48°C (-54°F) to -42°C
(-44°F) and from -35 °C (-31°F) to
-32°C (-26°F), respectively. The Tg
obtained from the E’int follows a
similar trend as that of the E”max;
hence, only the latter will be discussed.
Regardless of the manufacturer,
the glass transition temperature
Tg (E”max)] of the unexposed
samples and those exposed in
Alaska for one year is very similar.
The Tg between the unexposed
and exposed samples ranges
between 0 and +2°C (0 and +4°F).
Such Tg is well below the 8°C
(16°F) limit recommended by the
CIB/RILEM Committee [7] and
suggests that one-year exposure
to Alaska weather did not have a
significant effect on the Tg of the
TPO membranes.
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 74 Proceeedings of the RCI 21st International Convention
Figure 5A (left) shows typical DMA plot. 5B (right) shows how the Tg was obtained from E’ and E” curves.
Figure 6 (left) shows E” curves for an unexposed specimen from Manufacturer #1, while Figure 7
(right) shows the same for Manufacturer #2.
Figure 6 shows the E” curves for specimens
AK1Au, LV1Au, SEA1u, and TX1Au from Manufacturer
#1. The curves for specimens LV1Au
and TX1Au show the same maximum [(Tg of -32°C)
(-26°F)], which is °C (8°F) higher than that from
AK1Au and SEA1Au. The modulus value is slightly
higher for the LV1A. The overlay of the E”
curves for unexposed specimens from Manufacturer
#2 shown in Figure 7 indicates that the Tg
(E”max) is very similar for all specimens (-35 to -34°C)
(-31°F to -29°F). The curves for AK2Au and
SEA2Au specimens are almost identical throughout
the temperature range of analysis but this is
not the case for the LV2Au and TX2Cu specimens,
especially above the Tg region where LV2Au
and TX2Cu deviate from the other two curves.
The E” curves for unexposed specimens from
Manufacturer #3 (Figure 8) have their maximum
at -35°C (-31°F). The curves are grouped in sets
of two. One is AK3 and LV3 and the other is SEA3
and TX3. On the other hand, specimens from
Manufacturer #4 (Figure 9) show similar curves,
with exception of SEA4Cu, which shows slightly
higher E” modulus value.
The differences observed in the curves of the
unexposed samples within the same manufacturer
may be indicative of some changes during the
manufacturing process or they are from different
batches. It could also be due to storage location.
Although the Tg of the membranes is very similar,
it is possible that somehow the reinforcement
was placed differently on the membrane. In fact,
the reinforcement on the gray, black, and green
side (ply) of the membranes does not look the
same to the naked eye. For example, the weaving
of the reinforcement on the AK1u membrane from
Manufacturer #1 is different from that of SEA1u,
LV1u, and TX1u membranes.
In general, the bottom sheet (underlap) of an
exposed roof membrane sample has a Tg closer to
the unexposed sample because it is relatively shielded
from the environmental factors than the top (exposed)
sheet. If no chemical breakdown of the polymer
chains (chain scission) occurs due to exposure,
it would be expected that the unexposed or new
sample(s) have a lower Tg value than the exposed
sample. Changes in Tg lower than 5°C (9°F) after
exposure may be considered not significant because
the standard deviation of the individual measurements
is ±2 (±4°F) experimental error, which corresponds
to the temperature ramp of 2°C (4°F) used in
the analysis. Specimens taken from the top sheet of
pieces 2A (LV2Aet) and LV2D (LV2Ce) from
Manufacturer #2 show a similar behavior. However,
the change in Tg (-1°C) (-2°F) observed for these two
specimens is even lower than that observed for those
from Manufacturer #1. Although not shown, the E”
curves for the unexposed specimens from these
manufacturers overlay quite well that of the unexposed
specimen.
The DMA results showed, in general, that exposure
at the four different locations did not affect the
membrane material significantly. Regardless of the
manufacturer and exposure location, the Tg of the
samples is below the recommended 8°C (16°F) limit.
These results are consistent with tensile test results.
They also showed that membranes from the four
manufacturers were not significantly affected by the
Proceeedings of the RCI 21st International Convention Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 75
Figure 8 (top) shows E” curves for unexposed specimens
from Manufacturer #3, while Figure 9 (bottom)
shows an unexposed specimen from
Manufacturer #4.
exposure, with exception of LV1,
which shows a reduction in thickness.
CONCLUSION
The test results have been
presented in great detail. There
are some items that are being
closely watched, such as membrane
tightening, surface color or
hue, reflectivity, membrane thickness
overall, thickness of the
coating over the scrim, etc.
WSRCA’s Low-slope Committee,
the assigned TPO Task Group,
and WSRCA in general intend to
continue monitoring the findings
from the field and the laboratory
testing through the next round of
test cuts and testing. However, it
appears too early in the project for
anyone to make predictions as to
the life expectancy of the current
generation of TPO roofing membranes,
which appear to be
weathering quite well on the four
WSRCA TPO Weathering Farm
test sites.
REFERENCES
1. ASTM D751-00 Standard
Test Method for Coated
Fabrics
2. ASTM D6878-03 Standard
Specification for Thermoplastic
Polyolefin Based
Sheet Roofing
3. ASTM D1204-02 Standard
Test Method for Linear
Dimensional Changes of
Nonrigid Thermoplastic
Sheeting or Film at
Elevated Temperature
4. ASTM D471-98 Standard
Test Method for Rubber
Property – Effect of Liquids
5. ASTM D 6382-99 Standard
Practice For Dynamic
Mechanical Analysis and
Thermogravimetry of
Roofing and Waterproofing
Membrane Material
6. ASTM D 5418-01 Standard
Test Method for Measuring
The Dynamic
Mechanical Properties of
Plastic using a Dual Cantilever
Beam
7. Thermal Analysis Testing
of Roofing Membranes Materials.
Final Report of the
Thermal Analysis Task
Group RILEM 120-
MRS/CIB W.83 Joint
Committee on Membranes
Roofing Systems, December
1995.
ACKNOWLEDGEMENTS
The authors wish to thank all
those helping with this project.
Participating contractors:
Snyder Roofing, Kyle King and
Tim Gardner of Snohomish,
Washington; Industrial Roofing,
Jesse Martin and Bill Kramer (former
owner) of Anchorage, Alaska;
American Roofing, Richard Higgs
and Eddie Spalten of San Antonio,
Texas; and Commercial Roofing,
Dennis Conway and Bruce Martin
of Las Vegas, Nevada.
Participating manufacturers
(in alphabetical order): Carlisle
Syntec Inc., with project representative
Randy Oberg; Firestone
Building Products, represented by
Jim Jannasch; Genflex Roofing
Systems, represented by Mike
Hubbard; and Stevens Roofing
Systems, represented by Steve
Moskowitz.
WSRCA’s Low-Slope Technical
Committee: Tom Bradford, chairman;
Greg Bolt, co-chairman; Bill
Baley; KC Barnhardt, executive
liaison; Chuck Chapman, Chris
Dean, Don Fry, Tom Johns,
Christian Madsen, Gene Meyer,
David Montross, Dennis Ryan.
Dr. Karen Liu, Sebastian
Evoniak, National Research
Council of Canada.
Carlson, Delgado, Rosenow, Barnhardt, and Paroli – 76 Proceeedings of the RCI 21st International Convention