|
Initially the water penetration at the surface is very quick and we observe a logarithmic rate of increase. However, the rate of water absorption
gradually decreases until it reaches the maximum water gain value or saturation point of the test coupons. If the resin matrix is inert to water, we have reached an equilibrium; but if
the polymer can be chemically broken down and solubilized, the weight of the composite will decrease as portions of the heavier matrix are broken down, leach out, and be replaced with
water. We know we have achieved a good gel coat/skin coat barrier when the water weight gain reaches a maximum value and stays there indefinitely.
  ASTM D570 is a useful screening tool to determine the relative
merits of polyester resins used for improving the composite structure. For this test, clear castings made from of a variety of unsaturated polyester resins were scanned at room
temperature, 150 ºF (66ºC), and 200º F (93 ºC) for water absorption rate, approximate saturation point, and percent solubles. A summary is presented in Table 1. The vinyl ester
resin, which has low water absorption and minimal solu b les, demonstrates the excellent physical properties Table 2 and superior fatigue resistance making it an excellent resin choice for marine and
swimming pool composites. These outstanding qualities further show that a vinyl ester is an excellent candidate for a barrier skin coat (labeled as Layer B in Figure 1). Figures 3 and 4 compare the vinyl ester resin
water absorption over time and temperature to a neopentyl glycol-based gel coat and Figure 5 illustrates comparisons  to other polyester resins, including a typical orthophthalic polyester laminating resin. The vinyl
ester resins will reach a point of saturation and maintain it while the less capable orthophthalic resins continue to absorb water and degrade.
We wanted to remove two sources of stress from the boil test so the accelerated test
conditions would more closely relate to actual use. To remove the stress induced by the clamps, the test panel surfaces were gel coated and its edges were sealed with
vinyl ester resin, which were exposed during trimming, this protected the orthophthalic composite from water wicking into the laminate and the subsequent hydrolysis of the
polymer. Then, the entire coupon was immersed in water
Experimental
Our goal was to gather and analyze water diffusion data through several gel coat/skin coat barriers using
various back-up resin castings and laminates. To accelerate water contact exposure, the composites were
tested under ambient conditions and at 150 ºF (66 ºC). We used a modified ASTM D570 (Water Absorption of
Plastics) test procedure where the weight gain of a standard composite specimen is recorded in specific time increments.
The first step in the test panel construction was preparing the gel coat/skin coat barrier. The gel coat was
catalyzed with 2% of Lupersol DDM-9 methyl ethyl ketone peroxide (MEKP) and drawn down using a wet film
applicator on two glass plates. When a vinyl ester resin interlayer was used, the same technique is employed
but at an increased mil thickness setting. We used the pre-promoted, thixotropic CoREZYN( brand VE8115
vinyl ester resin catalyzed with 2% Witco Hipoint 90 MEKP. The resin/glass interlaminate barrier was
approximately 30 mils thick and constructed using a 2:1 ratio of resin to glass mat consisting of two 3/4-ounce layers of Owens Corning M722.
To prepare the all-resin back-up castings, two barrier-coated plates were placed in a mold with U-shaped metal
stop edges and rubber gaskets, which prevent leaking. Five, mil-thick strips of DuPont Mylar were added to
offset the thickness of the gel coat barriers and control the all-resin casting cores at 1/8-inch thickness. The
laminate cores consisted of a 2:1 resin to glass mat ratio and layed-up using seven and eight layers
respectively of 3/4 ounce mat on each plate. Before gelation occured, the plates were pressed together to fuse into a double mold surface laminate having a 1/4-inch core.
 The laminates and castings were post cured and cut into test coupons. Figure 6 is an
exploded view of the 51/2-inch by 51/2-inch test coupons. To evaluate the effects of the barrier on water absorption rates, we chose the CoREZYN brand 1063-40
orthophthalic all-resin casting core to assure a homogeneous substrate back-up. Several resins are studied in laminate cores to determine the influence each resin matrix
had on the blister resistance of the gel coated composite. The moisture content at ambient and 150 ºF (66 ºC) was monitored over time and related to blister formation due
to the differences in the barrier structure.
Observations/Results
A compilation of the mechanical and liquid characteristics of the four back-up resins in
the composites' cores are in Table 2. The physical properties of a composite
manufactured from any of these resins will have similar composite strength. Any differences noted in the blistering or relative absorption rates should then be
attributed to the chemical resistance or hydrolytic stability of the resin used in the laminate substrate underlying the barrier.
 The initial boil test data are discussed in Table 3, which shows
the water weight gained and time-to-blister for several different composite structures. The 212 (F (100 (C) test temperature used is well above the heat
distortion point of most orthophthalic resins. The results show relative differences and illustrate that vinyl ester resins perform well under the severe boil test conditions.
 The diffusivity of the various barriers is difficult to predict and is generally measured on the membrane material under actual conditions. Table 4 shows the experimental data derived for an orthophthalic resin casting core with
different barrier constructions, immersed at room temperature through 29 weeks. To more readily compare the initial water permeation rates, a logarithmic curve fit has been
computed, via the least squares method, for the data through seven weeks exposure. Several curves are plotted in Figure 7. The initial slope of the curve is denoted by the
regression coefficient "b" and the values indicate that the relative permeation rates at room temperature are essentially the same for all barriers except the interlaminate with a
CoREZYN VE 8115 vinyl ester skin coat. The measured rate of water gain was essentially the same (b = 0.16) for samples prepared without a gel coat, with 20 mils of
gel coat, with 35 mils of gel coat, and with 20 mils of gel coat plus 15 mils of an interlayer of vinyl ester skin coat. The coefficient of determination r², indicates we
have a good curve fit and can have confidence that the permeation rate is initially unaffected by the thickness of the gel coat or a 15 mil thick barrier coat. The resin used
to construct the laminate cores is the key variable controlling the rate of diffusion. This is illustrated by the data on the casting without a gel coated surface, which has
the same rate of diffusion as the gel coated castings.
 Coupons with the same barrier coat construction as used in the ambient testing
described above were immersed in a 150 ºF (66 ºC) water bath to accelerate the blistering process. Figure 7 plots these values and indicates that the initial slope is
similar for the castings without gel coat, with 35 mils of gel coat, and with 20 mils of gel coat plus the 15 mils of an interlayer vinyl ester skin coat. The slope rapidly changes
as indicated by the logarithmic curve fit contained in Table 5 where we observe that
the "b" regression coefficient (the slope), increases with gel coat thickness and the correlation of coefficient decreases. An examination of 0 mils, 20 mils, and 35  mils of gel coat after ten weeks of exposure at 150 ºF (66 ºC), as shown in Figure 9, reveals that very large blisters are forming behind
the gel coat film and above the orthophthalic resin casting matrix on the 0 mils and 20 mils panels. The panel with 35 mils
of gel coat showed less blistering than the panel with 20 mils of gel coat. No blisters were formed on the panel with the vinyl ester interlaminate Figure 10  or with the interlayer skin coated film. The clear casting without gel coat exhibited
small cracks and subsurface disruptions. The lowest casting absorption rate, "b", was the 30 mils of the interlaminate skin coat made with CoREZYN( brand VE 8115.
  The data generated on barrier coated laminates made with orthophthalic and
isophthalic backup laminating resins, and two different types of vinyl ester are compiled in Tables 6 and 7. Some of the
absorption data collected at ambient and 150 ºF (66 ºC), immersion temperatures are plotted in Figure 8. The room
temperature absorption rates, "b", are the highest for the orthophthalic and isophthalic resins; the non-gel coated,
non-skin coated orthophthalic falls in the middle; and the vinyl ester skin coat and laminate have the lowest absorption rates.  The slope values would
seem to indicate that the presence of a gel coat on the orthophthalic composite actually increases the permeation rate while the addition of the interlaminate skin coat
significantly decreases the absorption rate. No blistering was visible at 150 ºF (66 ºC) through 29 weeks, suggesting that the vinyl ester skin coat barrier removed the
reaction zone beneath the gel coat surface.
The data collected for the accelerated test were run at 150ºF (66ºC), and are compiled in
Table 7. Note that the calculated absorption rate, "b", is the lowest and the correlation coefficient, r², is the highest for the CoREZYN VE 8100 and VE 8440vinyl ester resins.
These two panels did not exhibit blisters through 20 weeks of exposure. The low absorption rate and high
correlation coefficient is attributable to vinyl ester resin and an indication of the excellent hydrolytic stability
of the composite. The composite with a skin coated interlaminate barrier made of CoREZYN VE 8110 vinyl
ester resin with the orthophthalic back-up laminate exhibited no blisters through 29 weeks but had a higher
absorption rate. The isophthalic back-up resin had approximately the same absorption rate as the vinyl ester skin coat but exhibited significant blistering after ten weeks of immersion.
 One panel of each construction type was removed at ten weeks and at 29 weeks and
dried in an oven. The results showing the weight loss due to hydrolysis and the leaching of water soluble compounds are compiled in Table 8. The incorporation of the
vinyl ester as a skin coat interlayer barrier or as the entire back-up laminate into the composite significantly decreased the amount of solubles and dramatically increased
the time-to-blister. The substantial reduction of blister formation by using CoREZYN brand vinyl  ester VE 8110 resin is strikingly reproduced in Figure 11. The comparison is made in Figure 11 of the "Improved Boat Hull" (the vinyl
ester interlaminate) to the "Conventional Boat Hull" after immersion for ten weeks at 150 ºF (66 ºC).
Conclusions
The polymer membranes studied were unsaturated polyester
barriers. One of the membranes was constructed of a premium neopentyl glycol get coat with a vinyl ester resin skin coat and the other was a premium neopentyl glycol gel
coat. They both had an orthophthalic laminating resin back-up. The permeation and diffusion rates of water
through the polymer membranes were measured, tabulated, and graphed. A vinyl ester barrier applied on an
orthophthalic laminating resin reinforced composite substantially reduced the blistering caused by water
permeation and matrix solubility for immersion at 150 ºF(66 ºC). Further evaluation of the coupons immersed at
room temperature reveals the degree of correlation between the elevated temperature and the room
temperature studies. The increased hydrophobicity of the gel coat/skin coat membrane, as compared to gel
coat alone, improved the composite's hydrolytic stability, thus reducing water absorption rates and increased
its blister resistance. The vinyl ester skin coat barrier, when applied to a boat hull or swimming pool, should result in increased product durability and life of the composite.
References
Revised 1/01 from A Study of Permeation Barriers to Prevent Blisters in Marine Composites and a Novel
Technique for Evaluating Blister Formation, by Paul Burrell, David Herzog and Terrance McCabe. The original work, including full references, may be obtained from Interplastic Corporation.
|
