Moisture diffusion through vinyl ester nanocomposites made with montmorillonite clay.

INTRODUCTION

Glass fiber reinforced polymer (GFRP) composites employing thermosetting polymer matrices--epoxy, vinyl ester or unsaturated polyester--are beginning to be used in the construction and repair of bridges and other civil structures (for example see refs. 1 and 2). Reasons for this trend include a high strength-to-weight ratio and a high stiffness-to-weight ratio of GFRPs as compared to conventional materials such as steel and aluminum. Widespread utilization of GFRPs in construction has, however, been hindered by the lack of long-term durability and performance data on which to base design calculations, especially when it is realized that GFRP composites used in infrastructure applications are intended to have a service life in excess of 50 years. In this regard, it is found that atmospheric moisture can diffuse to the fiber-matrix interface and cause both delamination and fiber weakening (3). One way to reduce moisture ingress is to reduce the moisture diffusion coefficient, and this may be done by dispersing and exfoliating clay platelets in the matrix polymer, forming a nanocomposite. Verification of this idea is the main focus of the present research. Note that other properties of GFRPs are also likely to be improved by the introduction of small amounts of clay. In particular, the extent of long-term creep is expected to decrease while the elastic modulus is expected to increase (4).

MATERIALS USED

There are a large number of different clays, each having somewhat different mineralogy, geologic occurrence, technology and uses (5). The most common variety of clay for nanocomposite synthesis and also the one used here is montmorillonite (4). It is a swelling clay made up of layers formed by a condensation of two silicon-oxygen tetrahedral sheets and one aluminum or magnesium-oxygen-hydroxyl octahedral sheet (5); the layer thickness is 0.96 nm (4). The individual layers or sheets generally exist stacked together, but they can be separated by a variety of methods. However, since the clay is inherently hydrophilic, a surface treatment is needed when the clay is to be dispersed in an organic matrix. In the present instance, Cloisite 10A[R], a montmorillonite treated with benzyl (hydrogenated tallow alkyl) dimethyl quaternary ammonium chloride, was used as received from Southern Clay Products Inc. of Gonzales, TX. The organic content of Cloisite 10A[R] is [approximately equal to] 39%, as determined by loss of weight on ignition. In addition, vinyl monomer clay (VMC) was prepared by treating Cloisite Na[R] with vinyl benzyl trimethyl ammonium chloride (6); Cloisite Na[R] was also supplied by Southern Clay Products Inc., and the vinyl group was expected to participate in the resin curing reaction. VMC preparation involved dispersing 30 grams of Cloisite Na[R] in 800 ml of distilled water with constant stirring with a mechanical stirrer for 5 hours. Separately, 6.35 grams of vinyl benzyl trimethyl ammonium chloride (VBTMAC), obtained from Sigma Aldrich and corresponding to 100 meq/100g of clay, was dissolved in 80 ml of water. At the end of 5 hours, the VBTMAC solution was added drop by drop to the clay dispersion, while stirring was continued. The result was the formation of a white precipitate of the treated clay. The mixture was stirred for another 3 hours, and the organomontmorillonite filtered using a vacuum filter and dried in an oven at 100[degrees]C. It was later ground in a ball mill. The organic content of VMC was 22% by weight.

The polymer used was Derakane[TM] 411-350 epoxy vinyl ester resin obtained from the Dow Chemical Co., and it contained 45 wt% dissolved styrene. It was processed as recommended by the manufacturer. To allow for curing to occur at room temperature, the resin was mixed with 0.5% by weight of 6% cobalt naphthenate catalyst (Sigma Aldrich Co.). Additionally, 0.05% of 99% N, N dimethyl aniline (Lancaster Synthesis, Pelham, NH) was used as an accelerator, while 1.5% of methyl ethyl ketone peroxide (9% active oxygen) obtained through Sigma Aldrich Co. was the initiator.

SAMPLE PREPARATION

Transparent, neat resin coupons were cast by pouring the above reaction mixture into (DuPont) Teflon[R] molds. For the diffusion experiments, mold dimensions were 50 X 12.5 mm, with the thickness ranging from 0.2 to 0.6 mm. The polymer was allowed to cure at room temperature for 24 hours, and it was subsequently post cured in an oven for 3 hours at 90[degrees]C. In the case of vinyl ester resin-clay nanocomposites, the organically treated clay was added to the liquid resin and manually stirred. The mixture was then degassed in a vacuum oven to remove air bubbles. The catalyst, initiator and accelerator were then added, and the resin was poured into the molds and subjected to the same cure cycle as the neat resin. Nanocomposite samples were prepared using both Cloisite 10A[R] and vinyl monomer clay; the samples were transparent at low concentrations of clay, while at higher concentrations (5 wt%) they became translucent. To examine the influence of mixing procedure, some of the samples containing clay were ul trasonically mixed for 30 min.

MATERIAL CHARACTERIZATION

Nanocomposite structure was probed with the help of X-ray diffraction, transmission electron microscopy and differential scanning calorimetry. XRD studies were performed at Southern Clay Products using a Scintag XDS-2000 Diffractometer with a CuK[alpha] source. The slit settings on the tube side were 4 mm for scatter and 2 mm for divergence. The slit settings on the detector side were 0.5 mm for scatter and 0.2 mm for receiving. The results revealed the spacing between the clay platelets. TEM work was done on relatively thick nanocomposite samples in the Department of Pathology at West Virginia University employing a JEOL 1010 Transmission Electron Microscope with an accelerating voltage of 80 kV. Micrographs were taken on 100-nm-thick sections cut perpendicular to the length direction. A TA Instruments DSC with a 910S cell was used to obtain thermograms on the various samples to measure the glass transition temperature of the completely cured material; samples were scanned from 30[degrees]C to 180[degrees]C at a rate of 10[degrees]C/min. Tensile testing was also carried out to measure the stiffness and strength of the nanocomposites. These tests were performed using a 100 kN Instron machine model 8501 at a displacement rate of 0.254 mm/min. The strain was measured independently using a strain gauge affixed to the mid-point of the specimen. Samples were made to conform to ASTM D638 specifications. Impact tests were conducted on an Instron/Satec BLI Impact testing machine equipped with a 2.71 J (2 ft-lb) Izod pendulum. Samples were made as per ASTM D4812 specifications; these had a thickness of 5.7 mm as against the recommended value of 3.17 mm, owing to the brittleness of the vinyl ester resin.

DIFFUSION TESTS AND DATA ANALYSIS

Diffusion tests were performed by immersing the rectangular cross section samples, having a dry mass ranging from 120 to 400 mg, in distilled water at room temperature (25[degrees]C). The samples were stored in a controlled humidity chamber, and it is estimated that they contained 0.05 [+ or -] 0.005 wt% water when diffusion experiments were started. The samples were periodically removed, blotted dry with a lint-free tissue, weighed and re-immersed in the water. A typical experiment lasted ten days, and, on the first day, readings were taken as frequently as every 15 min. The balance used had an accuracy of 1 [mu]g, and 3 to 5 replicate runs were carried out for a given set of conditions. To compute the diffusion coefficient from data on mass gain as a function of time, Fickian theory was employed. Here, the process of one-dimensional, unsteady diffusion is governed by (7):

[partial]c/[partial]t = D[[partial].sup.2]c/[partial][x.sup.2] (1)

in which c is the concentration of the diffusing species (in this case water), t is time, x is the position in the diffusing direction and D is the diffusion coefficient. Here D is taken to be a constant, at a particular temperature. Upon solving Eq 1 with constant boundary conditions we get moisture uptake [M.sub.t] to be:

[M.sub.t]/[M.sub.[infinity]] = [1 - [[summation over ([infinity]/0)] 8/[(2n + 1).sup.2][[pi].sup.2] exp[-D[(2n + 1).sup.2][[pi].sup.2]t/4[l.sup.2]]] (2)

where [M.sub.[infinity]] is the equilibrium increase in sample mass and 2l is the sample thickness. At the initial stages of diffusion, the solution for Fick's Law at small times reduces to:

[M.sub.t]/[M.sub.[infinity]] = 4[(Dt/[pi]][(2l).sup.2]).sup.1/2] (3)

so that the diffusion coefficient can be computed from the initial slope of [M.sub.t]/[M.sub.[infinity]] versus [t.sup.1/2]/2l using short time water-uptake data (7).

RESULTS AND DISCUSSION

XRD and TEM Tests

Figure 1 shows XRD scans on Cloisite 10A[R] and also of nanocomposite samples containing different amounts of this organoclay. None of the XRD scans on the Cloisite 10A[R] nanocomposite samples shows the characteristic basal reflection (peak) of Cloisite 10A[R], which occurs at 1.9 nm. The absence of such a peak is indicative of either exfoliation or breakdown of the ordered structure of the clay, and this question is probed further with the help of low magnification (1500X) TEM micrographs of the nanocomposite samples. Figures 2 and 3, corresponding to 0.5 and 5 wt% Cloisite 10A[R], show that the clay is reasonably uniformly distributed and the platelets are randomly oriented; the same trend was observed for the 1 and 2.5 wt% samples also. This is what would be expected given the manner in which the samples were made. High magnification (> 100,000X) TEM pictures of the same samples, shown in Figs. 4 and 5, reveal the presence of aggregates, but these micrographs also show individual platelets in the form of dark lines. It appears that the clay exists as expanded aggregates made up of 2 to 10 platelets. The size of these aggregates and the number of platelets in them increase with the percentage of clay within the sample, as can be seen from Figs. 2 and 3. The distance between adjacent platelets in these aggregates was also measured, and It turned out to be in the range of 4.4-5.0 nm, corresponding to a d-spacing of 5.4-6.0 inn. Peaks corresponding to this d-spacing are then expected to appear at a 20 value around 1[degree]-2[degrees] in the XRD scan. These peaks are not likely to be visible because the diffractometer looses its sensitivity at very small incident angles. This may explain why there are no peaks in the XRD scan despite the presence of aggregates or stacks of silicate plates seen in Figs. 4 and 5.

Although Cloisite 10A [R] is not completely exfoliated, the available evidence suggests more than Just intercalation because the d-spacing is greater than 2-3 nm (8). Indeed, the TEM pictures show all possible platelet morphologies, namely exfoliated, intercalated and stacked structures within the sample. It must be noted that a significant expansion of clay platelets has been observed even with only mild mixing when in-situ polymerization is carried out (9). This suggests that the monomer wets the clay surface easily, and, in the present case, there is very good compatibility of the surface treated Cloisite 10A [R] with the vinyl ester resin.

The XRD graphs and TEM pictures of VMC nanocomposites reveal a completely different picture from that presented in the case of Cloisite 10A [R]. Figure 6 shows a low magnification (1500X) TEM picture of a 5 wt% VMC sample, and it is seen that the clay exists in the form of large tactoids and stacks whose size is as large as 6-8 [mu]m. This suggests that the mixing was not very effective in shearing the clay platelets apart. Upon examining the tactoids more closely at higher magnifications of up to 150,000x (see Fig. 7), no individual platelets could be discerned, and this means that the spacing between the platelets is very small. This conclusion was confirmed through XRD scans (Fig. 8) that showed peaks at a spacing of around 1.5 nm, which is only a slight increase from the 1.486 nm d-spacing of the clay itself. This says that the clay platelets are, at best, intercalated, and that the clay exists as tactoids. We expected that violent mechanical stirring or ultrasonication would help in better dispersing the clay, but, even with ultrasonic mixing, XRD (Fig. 9) revealed that the spacing of VMC was only marginally increased from 1.486 nm for the clay itself to 1.51 nm for the nanocomposites. In this regard, note that according to the literature, an initial d-spacing of 1.7 nm is essential for expansion and exfoliation of clay to occur subsequent to crosslinking of the polymer (10). We speculate that, in the present case, crosslinking reactions took place between unsaturation sites on the surfaces of adjacent clay platelets and vinyl groups on the chains of the matrix polymer; the unsaturated sites referred to are the vinyl groups on the VBTMAC surface treatment given to the clay. The result was a covalent bond connecting two adjacent clay platelets through the intercalated polymer chain, thus preventing an expansion of the gallery spacing. A schematic diagram of this process is shown in Fig. 10. A similar explanation involving polymer bridging of silicate sheets was given by Messersmith and Giannelis (11) to ex plain why there was little or no change in the basal spacing in epoxy nanocomposites when a number of bifunctional primary and secondary amine curing agents were used to cure the epoxy resin.

Diffusion Tests

Figure 11 shows the fractional increase in mass of vinyl ester samples plotted versus [t.sup.1/2]/2l upon immersing neat resin coupons in distilled water at 25[degrees]C. As expected, based on Eq 3, water uptake increases linearly at short times, but it levels off at large times as equilibrium is attained. Also, superposed data from different runs reveal that moisture diffusion follows the predictions of Eq 2. One can, therefore, compute a (constant) diffusion coefficient, and the excellent fit of Eq 2 to all the data is shown by the solid line in Fig. 11. The behavior of the nanocomposites containing low loading of clay was similar. However, at higher percentages of clay, all the samples still showed a linear initial uptake, but the time to reach equilibrium was longer than what was theoretically expected (see the 5 wt% organoclay data in Fig. 12). Diffusion coefficients calculated using Eq 3 are listed in Table 1. The diffusivity of water in the neat resin is 9.19 x [10.sup.-7] [mm.sup.2]/s, and this is comparable to the value reported by others (12). It is seen that the addition of montmorillonite has a very significant effect on moisture diffusivity that reduces to about a fourth of its value in samples containing 2.5 wt% organoclay; this is true for both kinds of surface treatment. Although the diffusion coefficient reduces rapidly and progressively, there are diminishing returns, and a plateau appears to be reached at about 5 wt% organoclay.

As opposed to the decrease in diffusivity with clay addition, the equilibrium moisture content of the nanocomposite sample was found to increase with an increase in the amount of clay. At low concentrations of organoclay, i.e., < 1 wt%, the increase in the equilibrium moisture content was fairly linear, but it tapered off at higher concentrations of clay. These results are also listed in Table 1, and they appear to depend slightly on the type of clay surface treatment. It should be noted here that the samples were not completely dry when the diffusion experiments were started. Hence the "actual" moisture content when equilibrium was attained will be higher than the values reported in Table 1 by 0.05 [+ or -] 0.005 wt%. The reason for the higher equilibrium water content with increasing amounts of clay is the natural tendency of the clay to adsorb water. Even though the clay was treated with the organic quaternary compound, the clay remained hydrophilic. This is shown in Table 2, which summarizes the extent o f water uptake by Cloisite NA[R], Cloisite 10A[R] and VMC in a humid environment of 75% RH at 25[degrees]C; the untreated clay adsorbs more than three times as much moisture as the treated clay. It should also be recognized that the amount of water adsorbed by the clay would depend on the total exposed surface area of the clay platelets. It is the existence of multi-layer stacks and aggregates at higher clay concentrations, which have an effective exposed surface area that is less than the total area of the exfoliated clay, that leads to the negative deviation from a linear relationship between the equffibrium water content and amount of added clay. In addition, it appeared that Cloisite 10A[R] nanocomposites adsorbed more water than the VMC nanocomposite samples, and a consequence of water adsorbing onto the clay particles was that the transparent samples became whitish and lost their transparency. Cloisite 10A[R] nanocomposite samples started to become opaque and lose their transparency at high clay concent rations. This was not the case with VMC nanocomposites as they retained their transparency even days after they had reached concentration equilibrium. The reduced moisture uptake in the VMC nanocomposites was not expected based on the data given in Table 2, and it may have happened because the VMC surface treatment has a reactive group that can take part in the curing reaction. The crosslinked polymer molecules around the clay particles may have limited the access of water to the clay surface and also occupied sites where moisture may have resided.

The tests on untreated clay (Cloisite Na[R]) yielded diffusion coefficients, which, under comparable conditions, were about twice as large as both VMC and Cloisite 10A[R] nanocomposites. This was not surprising, because of the expected poor distribution of Cloisite Na[R] in the vinyl ester. It was, however, odd that the diffusion coefficient did not differ significantly between Cloisite 10A[R] and vinyl monomer clay nanocomposites. Since Cloisite 10A[R] platelets are better separated and distributed than VMC platelets, it was expected that Cloisite 10A[R] nanocomposites would show a larger decrease in the moisture diffusion coefficient as compared to VMC nanocomposites. Thus, it must be true that the increased barrier property resulting from clay addition is not just due to the presence of a "tortuous path" created by the clay platelets that are locally oriented; it must also be a function of some other structural variable. We speculate that this additional variable is the ability of polymer chains to move ar ound the nanometer sized clay platelets since moisture diffusion necessitates a rearrangement of polymer chains to allow for the creation of a diffusion path. It is for this reason that the diffusion coefficient becomes a function of the polymer relaxation time (13). In the case of the VMC nanocomposite samples. the clay platelets are not exfoliated but the movement of the polymer chains is more restricted than in the case of Cloisite 10A[R] samples because of the reactive nature of the clay surface. This leads to longer relaxation times and a slower rearrangement of polymer chains in the former case, resulting in comparable diffusion coefficients in the two cases. Hindered chain mobility is also the factor likely for the longer than anticipated time needed to reach diffusion equilibrium, as seen in Fig. 12.

Diffusion and equilibrium moisture content data allow for the determination of the permeability [P.sub.c] (product of diffusivity with the solubility, expressed as weight of water per unit volume of polymer), and these results are plotted in Fig. 13 as relative permeability versus the clay content; the relative permeability is [P.sub.c] divided by [P.sub.p], the permeability of the neat resin. Also plotted in Fig. 13 is the theoretical curve (solid line) for relative permeability using the tortuous path model of Yano et al. (14) according to which:

[P.sub.c]/[P.sub.p] = 1/[1 + (L/2W)[V.sub.f] (4)

where L is the length of the clay platelets and W is their thickness. Also, [V.sub.f] is the volume fraction of clay in the sample. Similar theoretical calculations could not be done for the VMC nanocomposites because the model applies only to exfoliated systems. It is seen that permeability of the nanocomposite samples decreases with increasing clay content in spite of an increase in solubility or equilibrium moisture content. Also, the data appear to be well represented by the theory.

DSC Tests

The polymer molecules in the regions neighboring the clay particles are partially immobilized (15), and this restriction in the mobility of the polymer chains should be reflected in an increase in the polymer glass transition temperature, [T.sub.g]. Also, the smaller the individual clay particles and the larger their number or the greater the extent and amount of exfoliation, the higher should be the [T.sub.g]. This phenomenon was indeed observed in the results obtained with all the Cloisite 10A[R] samples, and these results are displayed in Fig. 14. It was found that the glass transition temperature increased as the percentage of clay In the sample was increased, and this is in accord with the findings of others (11). However, consistent with the diffusion results, the glass transition temperature appeared to level off as the clay loading was increased. DSC scans were also made on VMC samples to determine the [T.sub.g], and results are again shown in Fig. 14. The glass transition temperatures of the 0.5% an d 1% VMC nanocomposites were even higher than those of the Cloisite 10A[R] nanocomposites and were around 109.6[degrees]C and 112.6[degrees]C, respectively, and did not seem to vary much with clay content. These significantly higher [T.sub.g] values are again attributed to the reaction between the clay and the polymer chains, leading to greater restriction in the motion of the chains. The drop in [T.sub.g] at higher percentages of clay seems to indicate that VMC is not likely to further improve barrier properties at still higher clay contents.

Mechanical Property Tests

Figure 15 shows the measured Young's modulus for nanocomposites made with both Cloisite 10A[R] and VMC. The stiffness has increased by about 25% with organoclay addition. However, as in the case of moisture diffusivity and glass transition temperature, the rate of property improvement decreases with increasing clay content. This again may be due to the clay existing as larger particles at higher clay contents as seen in the TEM pictures. Note that Cloisite 10A[R] appears to be more effective in raising the Young's modulus at low clay concentrations, while, at the higher clay loading, both clays appear to perform nearly equally well.

Results for the tensile strength are given in Fig. 16. There is greater variability in these results as compared to the stiffness data, and this is not surprising since strength is a failure property. Although strength appears to marginally reduce with clay content for Cloisite 10A[R] nanocomposite samples, VMC nanocomposites show a slightly different trend. At low concentrations of VMC, the tensile strength appears to increase. We postulate that this increase in strength is due to the chemical bond formed between VMC's surface treatment and the vinyl ester chains during curing, which is more effective in transferring stress to the clay platelets than in the case of an unreactive clay like Cloisite 10A[R]. The subsequent slight drop in tensile strength at 1 wt% and 2.5 wt% clay contents may be due to lack of proper distribution of clay. The increase seen in tensile strength at higher percentages of clay may be due to the properties of clay starting to influence the overall property of the nanocomposite.

Finally, data for the impact strength are presented in Table 3. Although it might superficially appear that clay addition can increase the impact strength, up to a certain weight-percent of organoclay (say 1-2.5 wt%), the standard deviation of the measurements is unusually large. Hence it is difficult to reach a definitive conclusion. All that one can say is that the addition of clay does not adversely affect the impact strength of the vinyl ester resin. The fracture area needs to be examined closely in future studies to determine the cause of such variation in impact strength.

The above results are similar to those reported by other researchers. Kornmann et al. (15) in their study of unsaturated polyester-MMT nanocomposites reported a linear increase in tensile modulus with MMT content. The increase in modulus was as much as 23% for 5 wt% of MMT. On the other hand, the tensile strength was found to be relatively unchanged, but it decreased slightly at higher clay contents of 10 wt%. Lan et al. (16) have reported a 35% increase in tensile modulus with 1 wt% exfoliated montmorrilonite in Epon 828 epoxy resin. They also found that d-spacing of the clay in the nanocomposite affected the level of property increase. For a nanocomposite containing approximately 2 wt% montmorillonite with a d-spacing of 1.66 nm, the tensile modulus and strength were found to be slightly lower than the corresponding values for the neat resin. Also Okada and Usuki (17) in their study on nylon 6 nanocomposites reported no change in tensile strength and modulus for situations where the d-spacing did not chang e after incorporation of the clay; for exfoliated nanocomposites the modulus increased from 1.1 GPa to 2.1 GPa for 4.2 wt% MMT.

CONCLUSIONS

The use of montmorillonite having two different surface treatments resulted in two different nanocomposites. VMC led to an intercalated structure with the clay existing as large tactoids with a small increase in d-spacing, while Cloisite 10A [R] gave a mixed morphology containing intercalated and partially exfoliated platelets. Diffusion tests conducted by immersing the nanocomposites in water showed that montmorillonite was effective in reducing the diffusion coefficient and permeability of water through the vinyl ester resin. Surprisingly, the reduction in the diffusion coefficient was only marginally more for the Cloisite 10A [R] than for VMC, which did not seem to disperse well according to TEM evidence. It is expected that better exfoliation of clay would give still better nanocomposite properties. As expected, the reduction in the diffusion coefficient was accompanied by an increase in the polymer glass transition temperature and an increase in the elastic modulus. However, the equilibrium moisture con tent of the vinyl ester increased upon clay addition. It remains to be seen whether this will help or hurt the durability of GRFP composites made with a clay-containing vinyl ester matrix. These durability tests are currently under way.

[FIGURE 1 OMITTED]

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

[FIGURE 14 OMITTED]

[FIGURE 15 OMITTED]

[FIGURE 16 OMITTED]

Table 1

Measured Values of Moisture Diffusivity and Equilibrium Moisture Content
of Vinyl Ester Resin Samples.

                         Diffusion Coefficient,     Equilibrium
                             x [10.sup.+6]       Moisture Content,
System                        [mm.sup.2]/s              wt%

Neat resin                   0.919  (0.030)        0.434 (0.012)
0.5 wt% VMC                  0.534  (0.025)        0.691 (0.004)
1 wt% VMC                    0.368  (0.041)        0.862 (0.005)
2.5 wt% VMC                  0.238  (0.019)        1.06  (0.021)
5 wt% VMC                    0.198  (0.008)        1.166 (0.021)
0.5 wt% Cloisite 10A[R]      0.4512 (0.025)        0.827 (0.007)
1 wt% Cloisite 10A[R]        0.336  (0.026)        0.927 (0.011)
2.5 wt% Cloisite 10A[R]      0.239  (0.021)        1.108 (0.006)
5 wt% Cloisite 10A[R]        0.125  (0.015)        1.484 (0.011)

Note: Numbers in parentheses are Standard Deviations.
Table 2

Results of Water Uptake Tests on Cloisite Na[R], Cloisite 10A[R], and
VMC (75% RH/25[degrees]C).

Time, h      Water Content (weight percent increase)

          Cloisite Na[R]  Cloisite 10A[R]   VMC

   0            0               0           0
  24          26.02             7.08        9.06
  48          31.62             9.16       11.56
 264          46.87            15.15       17.04
 720          59.68            18.94       19.36
Table 3

Impact Strength (in J/m) of Vinyl Ester Nanocomposite Samples.

Weight Percent
Organoclay      Cloisite 10A[R]  VMC

Net resin       58.07 (7.75)     58.07 (7.75)
0.5             69.95 (20.19)    57.42 (27.38)
1.0             64.39 (22.46)    89.77 (30.38)
2.5             67.44 (9.29)     56.33 (13.8)
5.0             40.3  (6.82)     61.18 (22.51)

Note: Numbers in parentheses are Standard Deviations.

ACKNOWLEDGMENTS

The authors would like to acknowledge the funding for this research from the U.S. Department of Transportation and Federal Highway Administration under contract number DTFH61-C-00021. We also sincerely thank Ms. Marilyn Howton, Technician, Electron Microscopy Laboratory, Department of Pathology, West Virginia University, for all the help and assistance with the TEM.

REFERENCES

(1.) H. V. S. GangaRao and C. Craigo, IABSE-Structural Engineering International, 9, 286 (1999).

(2.) S. Kshirsagar, R. A. Lopez-Anido, and R. K. Gupta. ACI Materials J., 97, 703 (2000).

(3.) S. Kajorncheappunngam, R. K. Gupta, and H. V. S. GangaRao. 'Effect of Aging Environment on Degradation of Glass-Reinforced Epoxy," ASCE J. Composites for Construction, 6, 61 (2002).

(4.) C. Chen and D. Curltss. Proc. Int. SAMPE Symposium and Exhibition, 46, 362 (2001).

(5.) H. van Olphen. An Introduction to Clay Colloid Chemistry, 2nd Ed., John Wiley & Sons, New York (1977).

(6.) A. Akelah and A. Moet, J. Mat. Sci., 31, 3589(1996).

(7.) A. Kumar and R. K. Gupta. Fundamentals of Polymers, McGraw-Hill, New York (1998).

(8.) H. R. Dennis, D. L. Hunter, D. Chang, S. Kim, J. L. White, J. W. Cho, and D. R. Paul, Polymer, 42, 9513 (2001).

(9.) P. B. Messersmith and E. P. Giannelis, J. Polym. Sci: Part A: Polym. Chain., 33, 1047 (1995).

(10.) G. W. Beall, "Molecular Modeling of Nanocomposite Systems," technical paper, Nanocor Corp. (www.nanocor. com).

(11.) P. B. Messersmith and E. P. Giannelis, Chem. Mater., 6, 1719 (1994).

(12.) J. W. Chin, T. Nguyen, and K. Aouadi, J. Appl Polym. Sci., 71, 483 (1999).

(13.) A. R. Berens and H. B. Hopfenberg, J. Polym. Sci.: Polym. Phys. Ed., 17, 1757 (1979).

(14.) K. Yano, A. Usuki, A. Okada, T. Kurauchi, and O. Kamigaito. J. Polym. Sci.: Part A: Polym. Chem., 31. 2493 (1993).

(15.) X Kornmann, L. A. Berglund, J. Sterte, and E. P. Giannelis, Polym. Eng. Sci., 36, 1351 (1998).

(16.) T. Lan, P. D. Kaviratna, and T. J. Pinnavaia, Chem. Mater., 7, 2144 (1995).

(17.) A. Okada and A. Usuki, Mat. Sci. Eng., C3, 109 (1995).

Apoorva P. Shah and Rakesh K. Gupta *

* Corresponding author. E-mail: Rakesh. [e-mail].