Bruce Curry

October 1-3, 2003

Technical Paper
Fire Retardant Polyester Resin Formulations

Abstract 230°C (446°F)] providing a blanket effect, which limits the

oxygen available for combustion. Additionally, metal
The original fire retardant polyester resin formula- hydrates such as ATH absorb heat from the combustion
tions incorporated bromine and chlorine (halogens) into the zone, which reduces the prospect of continued burning.
polymer. The disadvantages to these formulations were ex- Finally, ATH will produce a char during burning that
cessive toxic products and smoke given off on exposure to a results in further flame retardant protection and less smoke
flame. Alumina trihydrate (ATH) has successfully been added generation.
to non-halogenated polyester resins to resolve the toxicity
and smoke issues. Another fire retardant filler is ammonium
polyphosphate, which has been used for many years but
Unfortunately the high concentration of AHT det- not extensively in fiber reinforced composites. It is a fine-
rimentally affects the mechanical properties of composites. grained non-toxic white powder that uses the intumescent
Ammonium polyphosphate (APP) is an intumescent filler effect for flame retardancy in fire resistant coatings. It will
that is widely used in coatings for fire resistance. Other foam on exposure to heat or flames. The carbon foam layer
papers have been presented that describe the benefits of protects the composite through its heat-insulating effect
using filler blends of ATH and APP in composites. This and reduces further oxygen access. An APP supplier is sug-
paper describes the development of fire retardant formula- gesting that lower viscosities and improved fire resistance
tions with emphasis on non- halogenated filled systems with properties can be achieved by decreasing the ATH content
varying amounts of ATH and APP. Neat and filled viscosi- and substituting part of the removed ATH will APP. The
ties, cast resin mechanical properties are reviewed. Also focus of this paper is to confirm this suggestion.
included in this paper are flammability test results on com-
posites tested according to BSS 7239 toxicity test, ASTM E One significant disadvantage of using ATH as the
84, ASTM E 162, ASTM E 662, and ASTM E 1354. only filler is that high loadings were required to meet spe-
cific fire retardant requirements such as ASTM E 84 class I
Introduction ratings. Other disadvantages are that ATH does settle, rein-
forcements are harder to wet out and finished parts were
Originally fire retardant polyester resins were made with ha- more brittle.
logenated raw materials such as tetrabromophthalic anhy-
dride, tetrachlorophthalic anhydydide or chlorendic acid or Five different test methods were used to evaluate these new
anhydride. These products performed well until toxicity is- blends. ASTM E 84 “Surface Burning Characteristics of
sues became important. In order to satisfy the needs for low Building Materials” is a standard test for general fire resis-
toxicity products, non-halogenated general purpose unsat- tance and is widely recognized. Three tests that are often
urated polyester resins were introduced that were filled with specified for mass transit applications are ASTM E 162 “Sur-
products such as alumina trihydrate (ATH). This filler was face Flammability of Material Using a Radiant Heat Energy
the key component in meeting the fire retardant and toxicity Source,” ASTM E 662 “Specific Optical Density of Smoke
requirements. Generated by Solid Materials” and Boeing Support Stan-
dard (BSS 7239) toxicity test. A test that is being recognized
ATH products provide effective flame redundancy by increasingly as a more reproducible fire retardant test is
several routes. First, they dilute the amount of fuel ASTM E 1354 “Heat and Visible Smoke Release Rates for
available to sustain combustion during fire by replacing Materials and Products Using an Oxygen Consumption Calo-
part of the resin. Secondly, ATH contains 34.6% bound rimeter.”
water that is released at high temperatures [beginning at
 Technical Paper
A summary of these tests follows: standard tests. The ASTM E 162 (surface flammability of
materialsusing a radiant heat energy source), ASTM E 662
a) ASTM E 84- Laminates that are a minimum of 45.7 cm (18 (specific optical density of smoke generated by solid materi-
inches) wide by 7.3 meters (24 feet) long are compared in als) and the BSS 7239 toxic gas generation tests were run
burning characteristics to a mineral fiber cement board, because they are often specified for mass transport applica-
which is rated 0, and a red oak board that is rated 100. To tions. ASTM E 1354 (heat and visible smoke release rates
meet an ASTM E 84 class I rating, the laminate in ques- for materials and products using an oxygen consumption
tion must have an ASTM E 84 rating of 25 or less. calorimeter) was run because it is becoming more recognized
b) ASTM E 162- This test measures surface flammability as one of the most reproducible fire tests. It is also an excel-
employing a radiant heat source rather than a flame. lent screening test since a minimum amount of composite is
c) ASTM E 662- This test measures the optical density of required to run the test.
smoke using a radiant energy source. The smoke is mea- Liquid properties were tested with standard equipment and
sured in a flaming mode and a non-flaming mode. In methods that are typically used in the industry. The casting
order to obtain the flaming mode, a row of equidistant preparation is found in Figure a. Fire testing was run on
flamelets is placed across the lower edge of the speci- composites only.
men.
d) BSS 7239- This test is a Boeing Support Standard that
measures the amount of toxic gases that are generated Results
from a burning laminate in a controlled environment.
e) ASTM E 1354- 100 mm square coupons are burned while 1) Resin liquid properties of Firepel® K133-AAA-00 with
being subjected to an external heat flux that is set from 0- different ratios of ATH/APP fillers are found in Figure
100 kW/m2. Time to ignition, heat release, smoke obscu- b.
ration and mass loss are recorded. 2) The neat casting of Firepel® K133-AAA-00 was tested
for mechanical properties and those values are in Fig-
This paper compares the fire/smoke properties of three ure c.
filled resin systems utilizing ATH and blends of ATH/APP. 3) Composite mechanical test results of neat resin versus
60% ATH filled laminate values are in Figure d.
4) Fire Test Results (Firepel® K133-AAA-00 composites
Experimental with different ratios of ATH/APP fillers)
a) ASTM E 84 (surface burning characteristics of
A non-halogenated general purpose unsaturated building materials) results are in Figure e.
polyester resin Firepel® K133-AAA-00 containing both sty- b) ASTM E 162 (surface flammability of materials us-
rene and methyl methacrylate (MMA) was chosen. In order ing a radiant heat energy source) results are in Fig-
to minimize the number of variables, three types of compos- ure f.
ites were prepared keeping the total filler content constant. A c) ASTM E 662 (specific optical density of smoke
standard formula that required a 40/60 ratio of resin/ATH was generated by solid materials) results are in Figure
chosen as a control. Ammonium polyphosphate replaced g.
12.5% and 25% of the ATH in two different composites. d) BSS 7239 (toxic gas generation) results are in Fig-
Liquid resin properties both neat and filled, cast neat ure h.
mechanical values, and the fire tests are provided in tables. e) ASTM E 1354 (heat and visible smoke release rates
The resin made with 12.5% ammonium polyphosphate and for materials and products using an oxygen con-
47.5% ATH are coded APP12.5/ATH 47.5. The resin made with sumption calorimeter) is in Figure i.
25% ammonium polyphosphate and 35% ATH are coded
APP25/ATH 35. The resin was promoted with 0.6% cobalt 6%, Discussion
0.2% N, N-Dimethyl aniline (DMA) and 1.5% methyl ethyl
ketone peroxide (MEKP). Composites were made by wetting 1) The cast neat mechanical properties of a non-haloge-
one layer of 450 grams per square meter (1.5 oz) chopped nated general purpose polyester resin are typical of prop-
strand mat. erties that are required in the FRP industry.
2) A filled composite is more rigid as expected. The strength
The ASTM E 84 tunnel test (surface burning characteristics values are lower due to not being able to achieve as
of building materials) was run since it is one of the most high of glass content.
3) The 60% ATH and the APP25/ATH 35 met ASTM E 84
class I flame spread and smoke spread rating but the
 Technical Paper
APP12.5/ATH 47.5 did not; however, ASTM E 84 is an Generally the tests with APP were disappointing. In discus-
inconsistent test. It is interesting to note that the smoke sions with our APP supplier, indications are that too much
rating was higher with increasing amounts of APP. ATH was used, making it difficult for the APP to exfoliate
4) The ASTM E 162 time to ignition, heat evolved and time properly. High performance flame-retardants like APP rely
to specific distances indicated that the composite was on the presence of charable substrate (carbon based poly-
progressively less fire resistant with increasing quanti- mer) to create an intumescent foam.
ties of APP.
5) The maximum smoke ratings, according to the ASTM E
662 results, were progressively higher in values as the Conclusion
quantity of APP increased. The smoke density results
at the 1.5 and 4.0 minute intervals varied significantly. 1) With increasing concentration of APP, the viscosity in-
6 Based on the BSS 7239 test results, the 60% ATH com- creased which is a negative factor.
posite was the best performer. 2) APP did not improve the fire and smoke properties of a
7) ASTM E 1354 data was difficult to analyze for conclu- non-halogenated general purpose polyester resin where
sions. The APP12.5/ATH 47.5 composite had the best the total filler content was kept at 60%.
result for the time to sustained ignition. The best aver- 3) Since desired results were not obtained in this series of
age peak heat release was APP25/ATH 35 and the best testing, it is suggested that a formula containing APP
smoke rating was with 60% ATH. be repeated using less ATH. If positive results were
obtained, the use of APP in composites could be an
important raw material in fiberglass reinforced compos-
ites.
 Technical Paper
Figure B
RESIN LIQUID PROPERTIES

PROPERTY Firepel ® K133-AAA-00

NEAT
FIREPEL ® APP12.5/ APP25/
60%ATH
K133-AAA- ATH 47.5 ATH 35
00
C Brookfield Viscosity Spindle #1 @ 50 rpm,
0.018 (18) 0.1 (100) 0.12 (120) 0.185 (185)
al second (cps)

Figure C
CAST MECHANICAL PROPERTIES

PROPERTY NEAT Firepel ® K133-AAA-00

Tensile Strength, Mpa (psi) 68 (9,900)
Tensile Modulus, Gpa (psi) 3.4 (490,000)
Elongation, % 2.4
Flexural Strength, Mpa (psi) 106 (15,400)
Flexural Modulus, Gpa (psi) 3.8 (550,000)
Heat Distortion Temperature, °C (°F) 85 (185)
Barcol Hardness (934) 45

Figure D

GLASS FIBER COMPOSITE MECHANICAL PROPERTIES

PROPERTY Firepel ® K133-AAA-00

NEAT* 60%ATH*
Tensile Strength, Mpa (psi) 107(15,600) 7.9(11,500)
Tensile Modulus, Gpa (psi) 90(1,300,000) 137(2,000,000)
Elongation, % 1.4 1.3

Flexural Strength, Mpa (psi) 165(24,000) 134(19,500)

Flexural Modulus, Gpa (psi) 7.6(1,100,000) 11.7(1,700,000)

Barcol Hardness (934) 45 62

Glass Content, % 30 16

* Composite made with 1 ply of 450 grams per square meter mat (1.5 ounce per square foot)
 Technical Paper
Figure E
ASTM E 84 RESULTS
SURFACE BURNING CHARACTERISTICS OF BUILDING MATERIALS
USING NON HALOGENATED GENERAL PURPOSE POLYESTER RESIN
COMPOSITE
(# of samples-1)

TEST RESULTS
Firepel ® K133-AAA-00
APP12.5/ APP25/
60%ATH
ATH 47.5 ATH 35
Flame Spread 25 40 25
Smoke Development 65 160 250

Figure F
ASTM E 162 RESULTS
SURFACE FLAMMABILITY OF MATERIALS USING A RADIANT
ENERGY SOURCE USING NON HALOGENATED GENERAL PURPOSE
POLYESTER RESIN COMPOSITE
(# of samples-4)
TEST RESULTS
Firepel ® K133-AAA-00
APP12.5/ APP25/
60%ATH
ATH 47.5 ATH 35
Time to ignition, seconds 71 51 44
% Standard deviation 14 4.8 7.9
Baseline temperature, °C 190 185 185
Maximum temperature, °C 245 276 291
% Standard deviation 1.6 3.2 4.8
Temperature Rise, °C 55 91 106
% Standard deviation 7.4 9.8 13
Time to 7.6 cm (3 inches), minutes 2.7 2.0 2.0
Standard deviation 30 8.2 5.7

Time to 15.2 cm (6 inches), minutes 4.7 2.9 3

% Standard deviation 32 8.8 4.3
Time to 22.8 cm (9 inches), minutes 8.3 4.5 5.0
% Standard deviation 24 3.0 3.8
Heat evolution factor, Q 8.8 14.4 17
% Standard deviation 8 9.8 13
Flame spread factor, Fs 2.2 3.0 3.0
% Standard deviation 18 4.9 5.2
 Figure G
Technical Paper
ASTM E 662 RESULTS
SPECFIC OPTICAL DENSITY OF SMOKE GENERATED USING NON
HALOGENATED GENERAL PURPOSE POLYESTER RESIN COMPOSITE
(# of samples-3)

TEST RESULTS
Firepel ® K133-AAA-00
APP12.5/ APP25/
60%ATH
ATH 47.5 ATH 35
NON FLAMING EXPOSURE
Maximum smoke density 40 89 134
% Standard deviation 17 85 6.5
Smoke density after 1.5 minutes 0.2 0.14 0.03
% Standard deviation 60 82 173
Smoke density after 4.0 minutes 2.0 0.47 1.0
% Standard deviation 40 84 120
FLAMING EXPOSURE
Maximum smoke density 61 106 163
% Standard deviation 28 18 20
Smoke density after 1.5 minutes 1.3 0.43 2.6
% Standard deviation 18 97 91
Smoke density after 4.0 minutes 16 8.5 23
% Standard deviation 41 37 10

Figure H
TOXIC GAS GENERATION BSS 7239 FLAMING MODE RESULTS
USING NON HALOGENATED GENERAL PURPOSE POLYESTER RESIN
COMPOSITE
(# of samples-1)
RESULTS
TEST
Firepel ® K133-AAA-00

60%ATH APP12.5/ATH 47.5 APP25/ATH 35

Hydrogen Cyanide (HCN),
<2 4 7
ppm
Carbon Monoxide, (CO),
67 275 290
ppm
Nitrous Oxides, (NOx), ppm <2 35 60

Sulfur Dioxide, (SO2), ppm <1 ND ND

 Figure I
Technical Paper
ASTM E 1354 RESULTS
OXYGEN CONSUMPTION CALORIMETER USING NON-HALOGENATED GENERAL
PURPOSE POLYESTER RESIN COMPOSITE

TEST RESULTS
Firepel ® K133-AAA-00
APP 12.5/ APP
ATH 47.5 25/
60%ATH
ATH 35
Time to sustained ignition, seconds 63 76 59
% Standard deviation 5 11 3.9
Average peak heat release rate, kw/m2 341 358 314
% Standard deviation 6.7 12 2.9
Average heat release rate after 60 seconds, kw/m2 235 258 262
% Standard deviation 11 3.0 2.5
Average heat release rate after 180 seconds, kw/m2 138 143 171
% Standard deviation 3.4 3.7 3.5
Average heat release rate after 300 seconds, kw/m2 104 113 126
% Standard deviation 2.7 3.8 5.1
Total heat release, mj/m2 27 27 33
% Standard deviation 5.2 5.0 3.9
Average effective heat of combustion, mj/kg 17 17 21
% Standard deviation 4.7 8.8 1.8
Smoke obscuration, average specific extinction area, m2/kg 433 498 642
% Standard deviation 0.8 6.9 4.3
Mass Loss, % 50.6 44.8 42
% Standard deviation 0.8 3.7 1.4