Low-Rise Foam Adhesive Research Project

IIBEC Interface December 2022
Low-Rise Foam Adhesive
Research Project
By Randy Adams and Richard S. Koziol
This article describes a testing
and research project to determine
how the bond capacity
of low-rise foam adhesive
between insulation panels at
varying adhesive ribbon spacings
was affected when the adhesive was applied
to both fiberglass- and organic-faced polyisocyanurate
insulation boards. The project scope
included design and fabrication of custom 4 ×
4 ft aluminum frames; testing of eighteen 4-ft
square specimens in direct tension until failure;
testing six companion, small-scale (12 ×
12 in.) specimens in direct tension until failure;
evaluation of test results; and development of key
observations from the test program.
The research and testing program outlined
in this article was developed based on
a request for proposal (RFP) titled “Low Rise
Foam Research Project,” which was issued by
the Midwest Roofing Contractors Association
(MRCA) Technical Research Committee in May
2019. It has been recognized in the industry that
the handheld wand and cartridge application
technique for field installation of low-rise foam
adhesives is unlikely to produce consistent ribbon
spacing. Thus, the problem statement presented
in the MRCA RFP states, “What impact
does variation in foam ribbon spacing have on
ultimate roof uplift capacity?”
The problem defined in the MRCA RFP
states that typical manufacturer installation
instructions provide for low-rise foam adhesive
ribbons to be applied at 6 and 12 in. spacings.
Accordingly, the program was ultimately refined
to include testing of insulation adhesion at these
spacings, as well as ribbons spaced at 18 in. on
full-size, 4 × 4 ft square specimens. In addition,
investigators performed supplemental testing of
adhesion on 1 × 1 ft square companion specimens,
as well as tensile testing of adhesive-only
samples.
OBJECTIVE
The objective of this research project was to
determine the effect on bond capacity of lowrise
foam adhesive between insulation panels
at varying adhesive ribbon spacings, with the
adhesive applied to both glass fiber- and organicfaced
polyisocyanurate insulation boards. The
project scope included:
• Development of a testing plan and protocol
in collaboration with the MRCA Technical
and Research Committee.
• Design and fabrication of custom aluminum
loading frames to apply uniform forces to the
4 × 4 ft square test specimens.
• Preparation of 18 specimens, each composed
of two 4 × 4 ft square, 2-in.-thick polyisocyanurate
insulation boards adhered together with
low-rise foam adhesive ribbons. Three sets of
six panels were prepared with ribbons applied
at spacings of 6, 12, and 18 in. centers. Within
each set of six specimens, three specimens
featured glass fiber-faced insulation boards
and three featured cellulosic-felt-faced insulation
boards. The insulation boards were
attached to nominal 3/4-in.-thick plywood base
layers fitted with special tee-lock connectors
to accommodate anchorage of the specimens
to the test frames.
• Testing of the 18 specimens in direct tension
until failure, while simultaneously monitoring
loads, measuring specimen elongation, and
recording relative displacements at edges of
specimens.
• Testing of six companion, small-scale (12 ×
12 in.) insulation specimens in direct tension
until failure. These specimens—three
each for glass fiber-faced and organic-faced
insulation—were fabricated with a single foam
adhesive ribbon.
• Testing two companion, small-scale (12 × 12
in.) insulation specimens fabricated with fullcoverage
foam adhesive between insulation
boards—one each for glass fiber-faced and
organic-faced insulation.
• Direct tension testing of the low-rise foam
adhesive. Tests were conducted of cured foam
specimens that were either approximately
⅛-in. thick or approximately 1/32-in. thick.
• Evaluation of test results and development of
key observations from the test program.
MATERIALS
The polyisocyanurate insulation boards
and foam adhesive cartridges used for the
testing were supplied to Wiss, Janney, Elstner
Associates Inc. (WJE) by an MRCA member
roofing contractor. Four pallets of insulation
with factory-protective wrapping were
received approximately 30 days before testing
and stored in a warehouse building with the
wrapping removed. Two types of insulation
were used for the 18 full-size tests: nine of the
specimens used ASTM C12891 Type II, Class 1,
Grade 2 polyisocyanurate boards (glass-fiberreinforced
cellulosic felt facers and 20 psi compressive
strength), and nine specimens used
ASTM C1289 Type II, Class 2, Grade 2 boards
(coated polymer-bonded glass fiber mat facers
and 20 psi compressive strength). The dimensions
of the insulation boards for all full-size
specimens were nominally 48 × 48 in. with a
thickness of 2 in. The insulation boards were
The objective of this research project was to
determine the effect on bond capacity of lowrise
foam adhesive between insulation panels
at varying adhesive ribbon spacings, with the
adhesive applied to both glass fiber- and organicfaced
polyisocyanurate insulation boards.
This paper was originally presented at the 2022 IIBEC International Convention and Trade Show.
December 2022 IIBEC Interface • 27
delivered to our lab with these dimensions; no
larger boards were cut to size. For this test program,
polymer-coated glass fiber- and cellulosic
felt-faced insulation specimens were designated
Type A and Type B, respectively.
A two-component, low-rise polyurethane
foam adhesive was used to adhere the insulation
boards. The adhesive was provided in dualchambered
cartridges with attached mixing tips
that combined the two components. The adhesive
was applied with an electrically powered
applicator in beads spaced as required by each
test series. To create a base layer for specimens,
the same two-component low-rise adhesive was
used to attach insulation boards to 3/4-in.-thick,
48 × 48 in. sanded BC plywood panels.
FABRICATION OF SPECIMENS
Full-Size Specimens
Low-rise foam adhesive was applied by a
skilled roofing tradesperson to all plywood panels
and insulation boards using a power-actuated
applicator inside the conditioned testing facility.
Each of the plywood base-layer panels was
drilled with 29/64-in.-diameter holes at 26 locations
to match the locations of the attachment
points on the top and bottom of the customfabricated
aluminum stiffening frames. Into each
of these holes, a nominal ⅜-in.-diameter twistresistant
tee nut with a 1-in.-diameter flange
was inserted and embedded into the plywood.
Polyisocyanurate insulation specimens were
sorted to separate the Class 1 boards from the
Class 2 boards and put into six groups of three.
The boards were inspected for labeling that
read “This Side Up” or “This Side Down” and
were placed appropriately. A permanent marker
was used to highlight the applicable insulation
board facer surfaces with line markings at the
test spacings of 6, 12, or 18 in. (Fig. 1). The “knit
line” orientation on the insulation boards was
Figure 1. Lines were drawn on the insulation
boards (with spacings centered on boards)
to guide the application of the low-rise foam
adhesive ribbons. The lines were centered
so that application of the adhesive would be
symmetrical. The photo shows the line layout
for a 12 in. ribbon spacing.
Figure 2. Grid of low-rise
foam adhesive being applied
to plywood panels to achieve
an approximate 6 in. grid onto
which one of the insulation
boards would be placed.
Figure 3. Typical low-rise foam adhesive bead width being applied to two separate insulation
boards with 18 in. ribbon spacing lines.
28 • IIBEC Interface December 2022
identified and placed in positions so that the knit
lines of the top insulation board were oriented
in perpendicular position relative to the bottom
insulation board for each pair of boards comprising
a test specimen.
The two-part polyurethane foam adhesive
was initially installed onto the plywood board
surfaces in ribbons approximately 6 to 9 in. on
center in each direction (Fig. 2). Immediately
after placement of the adhesive, insulation
boards were installed on the prepared plywood
boards. Five weighted buckets, approximately 26
lb each, were placed on the insulation boards,
one at each corner and one in the center. The
buckets remained in place for approximately
10 to 15 minutes while the adhesive cured. The
adhered plywood and insulation halves for each
of the specimens were allowed to set up and cure
between 12 to 24 hours before final assembly.
After setup and curing, the two-part polyurethane
adhesive was applied in ribbons onto
half of the specimen along the
highlighted markings at designated
ribbon spacing. The foam adhesive
ribbons were applied at an initial
application width of approximately
3/4 in. (Fig. 3). The foam ribbons
were installed in straight, parallel
lines from one end of the insulation
board to the other, in lieu of a serpentine
pattern. Effort was made to
keep the bead width as uniform as
possible; however, some variation
in adhesive bead width did occur.
In addition, in some instances, a
small amount of adhesive accumulated
at the ends of the boards
because of the nature of the applicator
tool and the handheld process
used.
Immediately after placement of the adhesive
ribbons, the top insulation board was
positioned on top of the ribbons, and weighted
buckets (approximately 26 lb each) were placed
on the plywood panels at the corners and in
the center (Fig. 4). The buckets remained
in place for approximately 10 to 15 minutes
while the adhesive cured. Figure 5 presents
a cross-sectional view of a typical completed
specimen assembly.
Companion Specimens
Companion tests were performed on two
small-scale specimen configurations to provide
supplemental information of bond strength of
the adhesive between the polyisocyanurate insulation
layers as well as the tensile strength of the
two-part polyurethane foam adhesive.
1 × 1 ft Specimens
Eight 1 × 1 ft square companion specimens
were made with the same plywood and insulation
materials as the full-sized specimens
(Fig. 6). Six test specimens were made by
adhering insulation with a centrally posi-
Figure 5. A cross section of
an insulation and plywood
specimen.
Figure 4. Typical weights used as ballast
until the foam adhesive set up, typically for
10 to 15 minutes per specimen.
Figure 6. A 12 × 12 in. insulation board specimen with a single ribbon of adhesive.
Threaded Rods
Connected to
Concealed Tee Nut
Connectors
Plywood
Panel (Top
and Bottom)
Plywood
Panel (Top)
Plywood
Panel (Bottom)
2 Layers of 12 in. by
12 in. by 2-in.-thick
Polyisocyanurate
Insulation
Single Adhesive
Ribbon
Two Layers of 2 in.
Polyisocyanurate
Insulation Adhered
with Low-Rise
Foam Adhesive
Ribbons
December 2022 IIBEC Interface • 29
tioned single 12-in.-long ribbon of foam
adhesive. Three of the six specimens were
made with coated glass fiber facers (Type A)
and three with cellulosic felt facers (Type B).
The remaining two specimens, one made with
Type A insulation and the other made with
Type B, were fabricated with a continuous
film of foam adhesive to effectively provide
full coverage over the 1 ft2 surface area.
Adhesive-Only Specimens
To assess the tensile strength of only polyurethane
foam adhesive, we used a controlled,
thin-layering methodology in which the material
was placed and adhered between two aluminum
pucks (Fig. 7). The pucks have a machined
surface on one side and a threaded hole on the
other side to receive a threaded rod used to apply
tensile test loads. Six specimens were prepared
by initially cleaning and abrading the flat face
of the aluminum puck. A layer of the project’s
two-part polyurethane foam was then applied
to the face and inserted into
a fabrication jig that holds
and secures both pucks
while maintaining a ⅛ in.
gap between their planar surfaces.
Subsequent expansion
of the foam produces a controlled
⅛ in. separation of the
pucks, ensuring appropriate
resistance and simulating the
effects of restraining ballast
for actual installations. These
specimens were intended to
determine tensile strength of the foam adhesive
at a thickness judged to be representative
of typical use.
Three additional specimens were fabricated
in a similar configuration as the first series,
except that the pucks were adhered with a minimal
space between the two puck surfaces (1/32 to
1/16 in. thickness). This specimen was intended
to evaluate tensile strength of the foam adhesive
in a thin-film configuration, possibly representing
a condition of maximum confinement and
restraint.
DESIGNATION AND
IDENTIFICATION OF SPECIMENS
Full-Size Specimens
To provide unique identification of the 18
full-size test specimens, investigators developed
an alphanumeric system consisting of letters
to represent the facer type—polymer-coated
glass fiber (A) and cellulosic felt (B)—followed
by the ribbon spacing measurement, and the
specimen test number (Fig. 8). For example,
specimen number B-18-3 indicates a cellulosic
felt-faced insulation board with 18-in. adhesive
ribbon spacing, and specimen no. 3. Table 1
summarizes the test matrix for all 18 specimens.
Companion Specimens
Of the six 1 × 1 ft square companion specimens
fabricated with a single ribbon of adhesive,
three were made with polymer-coated glass fiber
Figure 7. Two pucks with a ⅛-in.-gap
space filled with foam adhesive.
Figure 8. Example of the alphanumeric system and label used to
identify the insulation specimens. B-18-3 indicates a cellulosic felt
facer (B) with 18 in. ribbon spacing, and specimen number 3.
Table 1. Test matrix for the full-size specimens
No. of specimens
6 3
12 3
18 3
6 3
12 3
18 3
30 • IIBEC Interface December 2022
facers (A) and three with cellulosic felt facers
(B). The three specimens with polymer-coated
glass fiber facers were numbered A-1, A-2, and
A-3, and the three with cellulosic felt facers were
numbered B-1, B-2, and B-3.
The two additional 1 × 1 ft square companion
specimens with full-coverage foam adhesive
(FCA/FCB) applied between insulation boards
were numbered FCA-1 and FCB-1.
Adhesive Specimens
The six puck specimens fabricated for tensile
strength testing of the foam adhesive were
identified as specimens TBS-1 through TBS-6.
All six specimens were fabricated in a manner
that allowed for a ⅛ in. gap between the pucks.
Specimens TBS-1, TBS-2, and TBS-3 were
allowed to cure for two days, and specimens
TBS-4, TBS-5, and TBS-6 were allowed to cure
for five days before testing.
Three additional pucks fabricated with the
thin-film configuration were identified as ABS-1
through ABS-3.
AGE AND CONDITIONING OF
SPECIMENS
The insulation boards and polyurethane
foam adhesive were delivered to the laboratory
on September 17, 2019, by an MRCA member
contractor. The materials were delivered
in unopened packaging, and after they were
relocated into a warehouse building, the coverings
were initially cut to allow for air movement.
The materials remained in this location until
assembled into the test specimens (approximately
30 days).
The environmental conditions in the warehouse
during the time that the 4 × 4 ft square
specimens were adhered with low-rise foam
adhesive were 55°F to 67°F with relative humidity
(RH) varying between 42% and 50%. The
adhered specimens were kept in the warehouse
building for two days; then they were moved to
the conditioned laboratory space and allowed
to acclimate to space environment of 70°F temperature
and 30% RH for three additional days
(five days total), prior to the start of testing on
October 22, 2019.
The six 1 × 1 ft square insulation specimens,
A1 through A3 and B1 through B3, and six
pucks, TBS-1 through TBS-6, were fabricated
in the warehouse building under similar conditions
as full-size specimens. A single cartridge
of low-rise foam adhesive was used to adhere
the insulation boards to the plywood panels,
and a separate cartridge was used to adhere the
insulation boards together into adhered layered
specimens for testing. This same cartridge was
also used to adhere pucks TBS-1 through TBS-6.
After setup, these samples were then relocated
to the laboratory building and allowed to cure
for five days. Pucks TBS-4 through TBS-6 were
fabricated in a similar fashion but were allowed
a cure time of two days before testing.
TEST APPARATUS AND
INSTRUMENTATION
It was the intent of the program to subject
the full-size adhered insulation specimens to
direct axial tension force to effectively determine
ultimate bond strength without inadvertent
introduction of eccentric loading, prying,
or peeling actions. This was accomplished by
fabricating a custom test frame that secured the
test specimen to an upper and lower aluminum
plate system. These aluminum frames served as
stiffening elements to the insulation specimens
and ensured near-uniform axial loading to the
adhered, layered specimen during testing.
The frames consisted of an upper and lower
grillage of orthogonally oriented 6-in.-deep
aluminum plates welded to a 4 × 4 ft square,
⅜-in.-thick aluminum base plate. Refer to Fig.
9 for an overall schematic of the test frame. The
lower frame was anchored to and supported by
four steel tube legs on 3 ft spacings, while the
upper aluminum frame was fitted with a steel
yoke that served to transfer a single vertical
force to the center of the 4 × 4 ft square grillage.
Attachment of the assembled insulation test
specimen to the frames was accomplished by
a series of 26 anchor bolts each at the top and
bottom sections of the specimen. The ⅜ inch
studs were threaded into the tee nuts installed
in the upper and lower plywood backers and
passed through mating holes within the aluminum
base plates and secured uniformly with
wing nuts. Refer to Fig. 10 for an overall view
the upper test frame with an attached insulation
specimen.
The aluminum test assemblies were positioned
in an existing load reaction frame within
the laboratory. The load reaction frame is a steel
assembly used to support and apply loads and
consists of steel columns and overhead steel
back-to-back wide-flange beams secured to the
laboratory’s reinforced concrete reaction floor
system. A 20,000-lb-capacity hydraulic actua-
Figure 9. Schematic of stiffened upper aluminum
test frame. All plates are ⅜ in. thick.
Figure 10. A stiffened aluminum test frame and secured insulation board test specimen.
Plan View
End View
December 2022 IIBEC Interface • 31
tor was positioned on top of the reaction frame
and connected to the captured test specimen
via a high-strength coil rod with end swivels.
For schematic and overall views of the existing
reaction frame with aluminum test frame setup,
see Fig. 11 and 12.
A key performance parameter to evaluate relative
behavior of the full-size test specimens is
the load-deformation relationship. Accordingly,
displacement of the specimens during tension
loading was measured by four discrete displacement
transducers (string potentiometers)
positioned at the midpoint of each of the four
edges of the specimens to measure movement
between the upper and lower aluminum test
frames (Fig. 13). The displacement transducers
measured the combined effects of axial strain
(stretching) of the two 2-in.-thick insulation
layers, as well as separation and elongation of
the low-rise foam adhesive during loading. The
displacement transducers at each edge location
had a total stroke of 2 in. and an accuracy of
0.001 in. For overall test fixture monitoring and
measurement of any flexural deformation of the
upper aluminum load frame, displacements
near the center of the frame were additionally
measured relative to the reaction frame at two
locations.
Applied load provided by the hydraulic ram
was monitored by a 20,000-lb-capacity electronic
load cell. Output from all six displacement
transducers and the load cell were captured by
a computer-controlled data acquisition system
that scanned sensors at approximately 1-second
intervals. Displacements and loads were visually
displayed on a large light-emitting diode screen
to facilitate monitoring during testing.
TESTING PROCEDURES
Full-Size Specimens
Procedures for subjecting each of the 18
full-size specimens to uniform axial load
were based on applicable provisions of ANSI/
SPRI IA-1, Standard Field Test Procedure for
Determining the Uplift Resistance of Insulation
and Insulation Adhesive Combinations over
Various Substrates.2 Procedure IA-1 is commonly
used to determine the uplift resistance
of an installed roofing/insulation system in the
field, and it provides for a loading protocol of
incremental pressure increases and a dwell, or
holding period, at each load stage. Investigators
believed that this regimen would appropriately
provide for a combination of incremental and
sustained loading to effectively determine stiffness
response and any short-term, nonlinear
deformation tendencies. It is recognized that
this modest loading rate may produce slightly
lower ultimate load capacities compared with
a more rapid application of load (such as a load
application simulating wind gusts).
Figure 11. Schematic of existing steel reaction frame
for application of load to specimen.
Figure 12. Existing steel reaction frame for application of load to specimen.
Figure 13. Displacement transducer positioned at the edge of a test specimen to measure separation
and elongation of insulation during testing.
Load Cell
Reaction
Floor
Reaction
Beam
Upper Load
Frame
and Yoke
Test
Specimen
Lower Load
Frame
Tie-Down
Legs
Test
Specimen
Lower
Load
Frame
Tie-Down
Legs
Upper Load
Frame
and Yoke
32 • IIBEC Interface December 2022
The general testing sequence of each full-size
specimen consisted of the following:
1. Placing the adhered insulation specimen
onto the lower aluminum test frame
2. Lowering the upper aluminum test frame
onto the specimen
3. Securing the specimen to the upper and
lower load frames by hand-tightening the
wing nuts
4. Attaching the loading pin and pull-rod
assembly to the loading yoke of upper frame
5. Attaching the north, south, east, and west
edge displacement transducers
6. Attaching the two center-frame displacement
transducers
7. Zeroing the displacement transducers and
load cell and initiating data logging
8. Applying the preload/starting load of 30 lb/
ft2 plus tare weight
9. Maintaining the load for 1 minute
10. Incrementally increasing the load by 15 lb/
ft2 (240 lb)
11. Maintaining the load for 1 minute
12. Repeating incrementally increased loading
until failure
Figures 14, 15, and 16 present overall representative
views of test setups and in-progress testing.
Companion Specimens
Companion tests were performed on two
different small-scale insulation board specimen
configurations to provide supplemental
information about bond strength of polyisocyanurate
insulation layers as well as the tensile
strength of the low-rise foam adhesive.
1 × 1 ft Insulation Board Specimens
Nominal ⅜ in. threaded pull rods were
anchored to each plywood back to facilitate
load application. Specimens were installed in a
hydraulic test machine, which provided application
of a direct tensile load (Fig. 17). Load was
applied in a similar manner as was employed
for the full-size specimens: a preload of 30 lb/
ft2 was held for 1 minute, followed by incremental
loading of 15 lb/ft2 with 1-minute hold
periods. The load was incrementally increased
until failure occurred.
Puck Specimens
The nine aluminum puck specimens were
tested in tension in the hydraulic test machine
(Fig. 18). Load was applied in a similar manner
as for other tests: a preload of 30 lb/ft2 was held
for 1 minute, followed by incremental loading
of 15 lb/ft2 with 1-minute hold periods. The
load was incrementally increased until failure
occurred.
Figure 14. Positioning the upper aluminum load frame onto a test specimen.
Figure 15. Securing the upper load frame to a test specimen with a plate and wing nuts onto
threaded rods.
Figure 16. Application of axial load to a test specimen.
December 2022 IIBEC Interface • 33
TEST RESULTS
Full-Size Specimens
Direct tension testing of the 18 full-size specimens was performed
at our structural laboratories from October 22 to October
24, 2019. Portions of the testing were witnessed by Mark Langer,
representing MRCA. Specimens were tested at an age ranging
from four to six days from assembly. Each specimen was loaded
in 15 lb/ft2 (240 lb) increments followed by 1-minute hold periods
until separation failure occurred, in accordance with the previously
described test protocol. Total test time per specimen ranged from
approximately 27 to 67 minutes, depending on the magnitude of
the failure mode. Continuous readings of applied load, the four
displacement transducers at the specimen edges, and two at the
center were recorded for the duration of each test.
Direct tension strengths for each specimen were computed by
subtracting the tare weight of the upper load assembly (consisting
of the pull rod, shackles, the aluminum load frame, and a single
layer of foam/plywood) from the maximum measured test load
and dividing the result by the nominal cross-sectional area of the
specimen (16 ft2).
Average test strengths for Type A (polymer-coated glass fiber)
specimens for ribbon spacings of 6, 12, and 18 in. were 674, 497,
and 342 lb/ft2, respectively. Average test strengths for Type B (cellulosic
felt) specimens for ribbon spacings of 6, 12, and 18 in. were
614, 379, and 307 lb/ft2, respectively. Tables 2 and 3 summarize the
testing results. The relationship between measured failure loads
Figure 17. Testing of a 1 ft2 insulation board specimen with a
single ribbon of adhesive in the hydraulic test machine.
Figure 18. A 2-in.-diameter aluminum puck specimen
with ⅛-in.-thick low-rise foam adhesive being
tested in the hydraulic test machine.
Figure 19. Relationship
between measured failure
loads and adhesive spacing
for Type A (coated glass fiber
[blue]) and Type B (cellulosic
felt [yellow]) facers.
34 • IIBEC Interface December 2022
and adhesive spacing for Type A and B facers is
depicted in Fig. 19.
Separation Displacements
Displacements between the upper and lower
aluminum load frames were measured during
each test and represented the combined effects
of axial strain (stretching) of the two 2-in.-thick
insulation layers as well as separation and elongation
of the polyurethane foam adhesive during
loading. Figure 20 shows the plot of applied
load versus average edge displacement (P-∂ plot)
for specimen A-12-3. The curve is fairly linear
up to failure load, at which time displacements
rapidly increase due to large separation of the
insulation layers.
The P-∂ relationships provide two distinct
performance indicators for the tested tensile
specimens: (a) maximum displacement/separation
at failure and (b) stiffness. The data reveal
that the average edge displacements at failure for
specimens made with adhesive ribbon spacings
of 6, 12, and 18 in. were 0.16, 0.19, and 0.23 in.,
respectively. As expected, greater displacements
were achieved at higher ultimate loads; no significant
difference in maximum displacements
was noted between specimens with coated glass
Figure 20. Load displacement
plot measured for
specimen A-12-3.
2
2
6
A-6-1 4 719
A-6-2 5 680 674 7.1%
A-6-3 6 624
12
A-12-1 5 426
A-12-2 5 555 497 13.1%
A-12-3 4 509
18
A-18-1 4 348
A-18-2 5 329 342 3.4%
A-18-3 5 351
Specimen
2 2
6
B-6-1 5 6 558
B-6-2 6 6 643 614 7.9%
B-6-3 6 6 642
12
B-12-1 5 12 354
B-12-2 4 12 426 379 10.7%
B-12-3 5 12 357
18
B-18-1 4 18 274
B-18-2 5 18 301 307 11.9%
B-18-3 5 18 347
Table 2. Summary of test results for specimens with Type A (polymer-coated glass fiber) facers
Table 3. Summary of test results for specimens with Type B (cellulosic felt) facers
December 2022 IIBEC Interface • 35
fiber or organic facers. Table 4 summarizes the
measured maximum average displacements for
the test specimens.
Stiffness
The structural stiffness of a component is a
strong indicator of the overall performance of
an element. Stiffness is defined as the resistance
of a body to deflection or deformation from an
applied force—that is, elements with greater stiffness
will deflect or deform less than those with
lower fundamental stiffness properties. For the
adhered insulation test specimens, axial stiffness
was calculated for each specimen as the
ratio of applied tensile load to average separation
displacement between stages corresponding to
10% and 50% of ultimate loads, as shown schematically
in Fig. 20.
Computed axial stiffnesses for the six specimen
groups ranged from 40 to 65 ksi (Table
5a). The relationships of spacing versus stiffness
exhibit similar trends as that noted for ultimate
strength values, with coated glass fiber-faced
specimens having a somewhat higher (10% to
19%) stiffness than organic-faced specimens.
The higher stiffness values for tighter adhesive
ribbon spacing may be associated with the
greater amount of contact adhesive area for 6 in.
spacing as compared with 12 in. spacing. When
there is less spacing, forces are distributed over
more of the insulation surface, reducing overall
axial strain at comparable loads.
Post-test Observations
The failure mechanisms in the full-size
specimens were predominately delamination
and separation of the facers (either the coated
glass fiber or the cellulosic felt facers) and some
cohesive failure of the foam core. Cohesive failure
of the foam core was most prevalent in the
specimens with 6 and 12 in. ribbon spacings and
occurred to a lesser extent in the specimens with
18 in. ribbon spacing. For the specimens with 18
in. ribbon spacing, the primary failure plane was
delamination and separations from the insulation
facers. After testing, each specimen was
placed on a table for examination and documentation.
Representative examples are shown in Fig.
21, 22, and 23.
Companion Specimens
1 × 1 ft Square Specimens
Direct tension testing of the six 1 × 1 ft square
specimens (A1, A2, A3, B1, B2, and B3) was performed
after a five-day cure period. Each specimen
was loaded in 15 lb/ft2 (240 lb) increments followed
by 1-minute hold periods until separation failure
occurred. Test strengths were computed as the
maximum measured load divided by the nominal
area of the specimen (1 ft2). Average test strengths
were 673 lb/ft2 for Type A (polymer-coated glass
fiber) specimens and 484 lb/ft2 for Type B (cellulosic
felt) specimens. Table 5b summarizes the
testing results.
Testing of specimens FCA-1 (polymer-coated
glass fiber facer) and FCB-1 (cellulosic felt facer)
fabricated with full coverage of adhesive over the
entire 1 ft2 insulation surface had failure loads of
896 and 838 lb/ft2, respectively.
General post-test observations for the specimens
A1, A2, A3, B1, B2, and B3 included the
following:
• Coated glass fiber facers: The predominant failure
mechanism in the three specimens with a
single adhesive ribbon was delamination and
separation of the facers from the foam body of the
insulation boards in conjunction with varying
degrees of cohesive failure within the foam core.
• Cellulosic felt facers: The predominant failure
mechanism in the three specimens with a single
adhesive ribbon was delamination and separation
of the facers from the foam body of the
insulation.
General observations for the specimens FCA-1
and FCB-1 included the following:
• Coated glass fiber facers: The predominant
failure mechanism in the specimen with the
full-coverage adhesive layer was delamination
and separation of the facer from the foam body
of the insulation.
• Cellulosic felt facers: The failure mechanism in
the specimen with the full-coverage adhesive
layer was cohesive delamination and separation
within the facer.
Puck Specimens
Direct tension testing of the six aluminum puck
specimens (TBS-1 through TBS-6) that featured
the ⅛-in.-wide foam layer was performed after a
two- or five-day cure period. Each specimen was
loaded in 15 lb/ft2 (240 lb) increments followed
by 1-minute hold periods until separation failure
occurred. Test strengths were computed as the
Maximum average edge
6 0.21
A 12 0.20
18 0.17
6 0.15
B 12 0.18
18 0.24
Table 4. Maximum average edge displacements at failure load
2 2
A
A1 700
A2 660 673 3.4%
A3 660
B
B1 660
B2 435 485 32.2%
B3 360
6 65
A 12 50
18 44
6 58
B 12 42
18 40
Table 5a. Average axial stiffness for Type A and Type B specimens
Table 5b. Test strength results for 12 × 12 in. panels with single low-rise foam adhesive ribbon
36 • IIBEC Interface December 2022
maximum measured load divided by the nominal
area of the specimen (3.14 in.2). Average test
strengths were 1355 lb/ft2 for two-day-old specimens
and 3361 lb/ft2 for five-day-old specimens.
The data clearly show that for the configuration
tested, the adhesive gained appreciable
strength between curing ages of two to five
days for the nominal 70°F storage environment.
Table 6 summarizes the testing results.
Specimens ABS-1, ABS-2, and ABS-3 were
tested at an adhesive age of four days, and their
average measured strength was 295 psi (42,480
lb/ft2). These values are substantially greater than
for specimens fabricated with ⅛ in. of adhesive.
General post-test observations for specimens
TBS-1, TBS- 2, and TBS-3 (two-day-old
adhesive) revealed the failure mechanism to be
cohesive failure of the foam adhesive.
In testing of specimens ABS-1, ABS-2, and
ABS-3 with thin film adhesive (foam thickness
less than 1/16 in.), the failure mechanism was a
cohesive failure between the adhesive and the
aluminum puck.
SUMMARY AND FINDINGS
OF TEST PROGRAM
We have completed a research and testing
program for the MRCA to determine the effect
on bond capacity for various low-rise foam adhesive
ribbon spacings used to adhere layers of
polyisocyanurate roofing insulation boards. The
program evaluated both polymer-coated glass
fiber-faced and cellulosic felt-faced insulation
board specimens prepared with 6, 12, and 18
in. ribbon spacings. The work featured fabrication
of a custom test rig to accommodate and
apply direct tension loading of full-size (4 × 4 ft)
adhered insulation specimens, as well as testing
on small-scale companion specimens and tensile
tests of cured foam material.
Key findings of the test program include:
• Average measured direct tension strengths
for full-size polymer-coated glass fiber-faced
specimens tested at ribbon spacings of 6,
12, and 18 in. were 674, 497, and 342 lb/ft2,
respectively. Average test strengths for cellu-
Figure 21. Specimen A-6-2 shown in separated fashion, similar to an opened book. The plane
of failure was primarily within the foam core of the top facer of the insulation board (left), with
some delamination at the bottom glass fiber facer of the board (right).
Figure 22. In specimen A-18-2, the primary failure was separation of the top facer of the bottom
board from the foam core, with some cohesive bond separation within the foam core.
Figure 23. In specimen B-12-1, the failure planes were the bottom facer of the top board and
top facer of the bottom board, with cohesive separation within the organic facer along the foam
ribbons. The failure included some cohesive separation of the foam core.
2 2
2
TBS-1 36 11.4 1640
TBS-2 35 11.1 1601 1355 33.9%
TBS-3 18 5.7 825
5
TBS-4 80 25.5 3667
TBS-5 69 22.0 3163 3361 8.0%
TBS-6 71 22.6 3254
Table 6. Results for low-rise foam adhesive tensile test with ⅛-in.-thick foam beads
December 2022 IIBEC Interface • 37
losic felt-faced specimens for ribbon spacings
of 6, 12, and 18 in. were 614, 379, and 307 lb/
ft2, respectively. These test results were fairly
uniform based on computed coefficient of variation
of 7% to 13% for each facer type.
• The test data clearly show that direct tension
strength increased as adhesive foam ribbon
spacing decreased, for both the polymer-coated
glass fiber-faced and cellulosic felt-faced insulation
specimens. This strong correlation confirms
and quantifies industry knowledge and
practice that greater adhered insulation board
uplift resistance is achievable with closer foam
adhesive spacings.
• For the specific polyisocyanurate insulation and
foam adhesive used in this test series, boards
with polymer-coated glass fiber facers exhibited
approximately 9% to 24% greater strengths than
cellulosic felt-faced specimens at equal adhesive
spacings.
• Failure of the specimens primarily occurred as
separations and/or delaminations of the insulation
board facers from the foam body core along
the lines of the adhesive ribbons, with secondary
failures within the body of the insulation noted
in some instances.
• As expected, specimens with greater ultimate
strengths at decreased adhesive ribbon spacing
also exhibited higher axial tensile stiffness
characteristics. For a given uplift load, specimens
with closer ribbon spacings would have
less axial stretch or displacement than systems
with greater adhesive spacing.
• As expected, the measured ultimate strength
values of the full-size insulation-to-insulation
adhered specimens tested in this program were
found to be significantly greater than typical
uplift ratings for complete roof systems.
• Testing of the 1 × 1 ft adhered insulation companion
specimens, each with a single adhesive
ribbon, revealed that polymer-coated glass
fiber-faced specimens were 28% stronger than
those specimens with cellulosic felt facers, with
average strengths of 673 and 485 lb/ft2, respectively.
Companion specimens fabricated with
full coverage of adhesive over the entire 1 ft2
surface had substantially higher strengths of
833 lb/ft2 for the cellulosic and 896 lb/ ft2 for the
glass fiber. These comparative test results confirm
that, when greater amounts of adhesive are
present over a given area, larger portions of the
facer/insulation interfacial zone are mobilized
to transfer loads between adhered boards.
• In general, strengths derived from companion
tests are higher and not well correlated
with the results for the full-size specimens
with 12 in. adhesive spacings. Testing using
larger (4 × 4 ft) specimens is considered more
representative of real-world installations than
testing using smaller (1 or 2 ft2) specimen sizes
because the larger specimens have greater
surface areas.
• Tensile testing of the 2-in.-diameter pucks with
⅛-inch-thick foam adhesive indicated substantial
strength gains between test ages of two and
five days. Average measured strengths were
1355 lb/ft2 at two days and 3361 lb/ft2 at five
days. As expected with two-part polyurethanebased
adhesive, chemical cure and associated
strength development continues well after initial
setting, depending on environment conditions
at the time of dispensing and thereafter.
• In limited testing, pucks with adhesive of minimal
thickness (less than 1/16 in.) had substantially
greater adhesive tensile strength than those
specimens made with ⅛-in.-thick adhesive.
This behavior is not necessarily unexpected,
but it reinforces the value of restraining normal
foam expansion by appropriate ballasting to
minimize adhesive thicknesses (and also producing
wider contact zones of the ribbons) in
real-world installations.
The test values derived from this program are
specific to the types of insulation boards and foam
adhesive used. While overall general trends indicated
by test data may be similar and characteristic
of other manufacturers’ products, the limitations
of the values presented specifically in this report
should be noted.
RECOMMENDATIONS FOR
FURTHER TESTING AND RESEARCH
To help put these test results into the context
of industry practice and expectations, we
recommend additional testing that can narrow
other environmental, product, and installation
variables inherent to insulated roofing systems.
Such additional testing may include assessment
of insulation boards of varied thickness, as well
as insulation and foam adhesives from additional
product manufacturers; testing of installation
and curing in other environmental conditions;
and investigation of other factors that may
influence adhesion and overall strength in the
adhered foam interfacial regions. Panel orientations
and cure times may also be explored in
greater depth to obtain a better understanding
of short- and longer-term performance.
As the assessment of additional testing
is considered, we recommend that a careful
review of safety factors for adhesives in
insulation-to-insulation applications, as well
as overall roofing applications, be included.
Similarly, a review of ribbon spacings specified
in fully tested and rated assemblies for uplift
should be part of any further review.
REFERENCES
1. ASTM International. 2019. Standard Specification for Faced
Rigid Cellular Polyisocyanurate Thermal Insulation Board.
ASTM C1289-19. West Conshohocken, PA: ASTM
International. https://doi.org/10.1520/ C1289-19.
2. American National Standards Institute (ANSI) and
Single-ply Roofing Industry (SPRI). 2015. Standard Field
Test Procedure for Determining the Uplift Resistance of
Insulation and Insulation Adhesive Combinations over
Various Substrates. ANSI/SPRI IA-1 2015. Waltham,
MA: SPRI.
ABOUT THE AUTHORS
Randy Adams is chair
of the Midwest Roofing
Contractors Association
Technical and Research
Committee. He has
more than 40 years of
specialized experience in
the roofing industry and
is owner and president
of R. Adams Roofing
Inc., Environmental
Greenscapes Inc.
Richard S. Koziol is a
principal at Wiss,
Janney, Elstner Associates
Inc. He has more
than 35 years of specialized
experience in
investigating and testing
roofing systems. He
has developed and
designed repairs for
water infiltration and
condensation problems
in new and existing building enclosures. Koziol is
a graduate of the University of Illinois at Chicago
and a licensed architect in five states.
To help put these test results into the context of industry practice and expectations,
we recommend additional testing that can narrow other environmental, product,
and installation variables inherent to insulated roofing systems.
Randy Adams
Richard S. Koziol