LOW ION CONTENT, NANOPARTICLE-CONTAINING RESIN SYSTEMS

FIELD

[0001] The present disclosure relates to resin systems containing a curable resin and surface-modified nanoparticles. The resin systems have low ion contents, particularly, low contents of alkali metals and alkaline earth metals. The present disclosure also relates to curable resin systems (e.g., gel coats) containing treated, surface-modified nanoparticles that have low ion contents.

SUMMARY

[0002] Briefly, in one aspect, the present disclosure provides a resin system comprising a curable resin and at least 10 weight percent surface-modified silica nanoparticles. The resin system comprises no greater than 200 parts per million by weight alkali metal and alkaline earth metal ions based on the total weight of the silica nanoparticles and curable resin. In some embodiments, the resin system comprises at least 20 weight percent, or even at least 40 weight percent surface-modified silica nanoparticles. In some embodiments, the resin system comprises no greater than 100 parts per million by weight alkali metal and alkaline earth metal ions based on the total weight of the silica nanoparticles and curable resin.

[0003] In some embodiments, the curable resin comprises a vinyl ester, an unsaturated polyester resin, or a (meth)acrylate. In some embodiments, the resin system further comprises a reactive diluent.

[0004] In some embodiments, surface-modified silica nanoparticles comprise dual-ion exchanged, surface-modified silica nanoparticle. In some embodiments, the surface- modified silica nanoparticles are end-capped.

[0005] In another aspect, the present disclosure provides a resin system that comprises at least 20 weight percent surface-modified silica nanoparticles having an average particle size of greater than 50 nm dispersed in a curable resin; wherein a 425 micron thick sample of the resin system has a haze of no greater than 3% as measured by the Optical Property

Test Method. In some embodiments, the 425 micron thick sample of the resin system has a clarity of at least 99.5 as measured by the Optical Property Test Method.

[0006] In another aspect, the present disclosure provides a resin system that comprises at least 20 weight percent end-capped, surface-modified silica nanoparticles having an average particle size of greater than 50 nm dispersed in a curable resin system, wherein a 1.5 millimeter thick cured sample of the resin system has a clarity of greater than 50, as measured by the Optical Property Test Method. In some embodiments, the curable resin comprises an unsaturated polyester/styrene resin system. In some embodiments, the 1.5 millimeter thick cured sample has a clarity of greater than 90, as measured by the Optical Property Test Method. In some embodiments, the 1.5 millimeter thick cured sample has a transmission of greater than 60%, as measured by the Optical Property Test Method. In some embodiments, the 1.5 millimeter thick cured sample has a haze of no greater than 30%, as measured by the Optical Property Test Method.

[0007] In yet another aspect, the present disclosure provides a gel coat comprising one of the resin system according to an embodiment of the present disclosure.

[0008] In a further aspect, the present disclosure provides a method of preparing a curable resin system. The method comprises providing a silica nanoparticle sol comprising no greater than 200 micrograms of sodium per gram of silica as measured by IC Procedure #1; covalently bonding a surface-treatment agent to a surface of the silica nanoparticles to form surface-treated silica nanoparticles; and combining the surface- treated nanoparticles with a curable resin. In some embodiments, the method further comprises ionically binding an end-capping agent to the surface of the surface-modified nanoparticles. Curable resin systems made by such methods are also disclosed.

[0009] The above summary of the present disclosure is not intended to describe each embodiment of the present invention. The details of one or more embodiments of the invention are also set forth in the description below. Other features, objects, and advantages of the invention will be apparent from the description and from the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. 1 illustrates the increase in viscosity of a resin system containing dual-ion- exchanged, surface-modified nanoparticles.

[0011] FIG. 2 illustrates the effect of end-capping with DMAEMA on the viscosity of a resin system containing dual-ion-exchanged, surface-modified nanoparticles.

[0012] FIG. 3 illustrates the effect of end-capping with HMDS on the viscosity of a resin system containing dual-ion-exchanged, surface-modified nanoparticles.

[0013] FIG. 4A is a TEM image of the nanoparticle-containing resin system of EX-3. [0014] FIG. 4B is a TEM image of the nanoparticle-containing resin system of EX-4. [0015] FIG. 4C is a TEM image of the nanoparticle-containing resin system of EX-5.

DETAILED DESCRIPTION

[0016] As used herein, "agglomerated" is descriptive of a weak association of primary particles usually held together by charge or polarity. Agglomerated particles can typically be broken down into smaller entities by, for example, shearing forces encountered during dispersion of the agglomerated particles in a liquid.

[0017] In general, "aggregated" and "aggregates" are descriptive of a strong association of primary particles often bound together by, for example, residual chemical treatment, covalent chemical bonds, or ionic chemical bonds. Further breakdown of the aggregates into smaller entities is very difficult to achieve. Typically, aggregated particles are not broken down into smaller entities by, for example, shearing forces encountered during dispersion of the aggregated particles in a liquid.

[0018] Generally, curable resin systems (e.g., crosslinkable resin systems) are used in a wide variety of applications, e.g., as a protective layer (e.g., gel coats) and as the impregnation resin in composites (e.g., fibrous composites). Resin systems are often selected based on the desired mechanical properties of the final product including, e.g., hardness, toughness, fracture resistance, and the like. In some applications, the optical appearance of the finished product may be important such that properties like clarity and haze must be considered. In addition, process conditions may lead to preferred ranges of properties affecting processability such as viscosity. Finally, the desired end use of the product often leads to additional requirements, e.g., erosion resistance or anti-blistering.

[0019] Surface-modified nanoparticles have been added to curable resins to achieve cured resin systems exhibiting improved mechanical properties. However, the present

inventors have determined that nanoparticles and the impurities and by-products of nanoparticle processing can have a detrimental effect on other properties of the cured resin system. The present inventors have also discovered that steps taken to improve one property of a curable resin system comprising surface-modified nanoparticles can lead to undesirable changes in other properties.

[0020] Generally, any known curable resin may be used in the various embodiments of the present disclosure. In some embodiments, the curable resin may be an ethylenically- unsaturated curable resin. For example, in some embodiments, an unsaturated polyester resin may be used. In some embodiments, the unsaturated polyester resin is the condensation product of one or more carboxylic acids or derivatives thereof (e.g., anhydrides and esters) with one or more alcohols (e.g., polyhydric alcohols).

[0021] In other embodiments, vinyl ester resins are used. As used herein, the term "vinyl ester" refers to the reaction product of epoxy resins with ethylenically-unsaturated monocarboxylic acids. Exemplary epoxy resins include bisphenol A digycidal ether (e.g., EPON 828, available from Hexion Specialty Chemicals, Columbus, Ohio). Exemplary monocarboxylic acids include acrylic acid and methacrylic acid. Although such reaction products are acrylic or methacrylic esters, the term "vinyl ester" is used consistently in the gel coat industry. (See, e.g., Handbook of Thermoset Plastics (Second Edition), William Andrew Publishing, page 122 (1998).)

[0022] In still other embodiments, (meth)acrylate resins, including, e.g., urethane (meth)acrylates, polyethyleneglycol (multi)(meth)acrylates, and epoxy (multi)(meth)acrylates may be used. As used herein, the term (meth)acrylate refers to an acrylate and/or a methacrylate, i.e., ethyl (meth)acrylate refers to ethyl acrylate and/or ethyl methacrylate.

[0023] Depending on the selection of the curable resin, in some embodiments, the resin system may also include a reactive diluent. Exemplary reactive diluents include styrene, alpha-methylstyrene, vinyl toluene, divinylbenzene, triallyl cyanurate, methyl methacrylate, diallyl phthalate, ethylene glycol dimethacrylate, hydroxyethyl methacrylate, hydroxyethyl acrylate, and other mono- and multi-functional (meth)acrylates.

[0024] Curable resins are sometimes used to form gel coats for industrial applications. Generally a "gel coat" is a material used to provide a high quality finish on the visible surface of a fiber-reinforced composite material. The most common gel coats are made from curable resins such as those based on epoxy, unsaturated polyester diluted with styrene, or vinyl ester diluted with styrene resin chemistry. The curable resin is then formulated into a gel coat by adding additional components such as catalysts, air release agents, leveling agents, surfactants, thixotropic agents, wetting agents and pigments.

[0025] The formulated gel coat may be applied to a mold in the liquid state and cured to form crosslinked polymers that can be subsequently backed up with composite polymer matrices (e.g., unsaturated polyester resin, vinyl ester resin, or epoxy resin with glass and/or carbon fibers). A thixotropic additive may be included in the gel coat formulation to assist the gel coat's tenacity to vertical portions of the mold while it cures.

[0026] The manufactured component, when sufficiently cured and removed from the mold, presents the gel coat layer as the exposed, visible surface. This gel coat layer is usually pigmented to provide a colored, glossy surface which improves the aesthetic appearance of the article, such as a counter made with cultured marble.

[0027] Many marine craft are manufactured using composite materials. The outer layer is often a gel coat, typically 0.5 mm to 0.8 mm in thickness. In some embodiments, gel coats as thick as 1.5 mm have been used. Gel coats are typically designed to be durable and to provide resistance to ultraviolet degradation and hydrolysis.

[0028] Suitable curable resin chemistries for the manufacture of gel coats vary, but the most commonly encountered are unsaturated polyesters diluted with styrene, vinyl esters diluted with styrene, or epoxies. Within each of these categories, the resin chemistries are further subdivided. Viscosities of the curable resins used to make gel coats are typically 0.2 to 0.4 Pa ^» s at 25 ⁰ C and a shear rate of 1.0 (I/seconds).

[0029] Generally, "surface modified nanoparticles" comprise surface treatment agents attached to the surface of a core. As used herein, the term "silica nanoparticle" refers to a nanoparticle having a silica surface. This includes nanoparticles that are substantially, entirely silica, as well nanoparticles comprising other inorganic (e.g., metal oxide) or organic cores having a silica surface. In some embodiments, the core comprises a metal oxide. Any known metal oxide may be used. Exemplary metal oxides include silica,

titania, alumina, zirconia, vanadia, chromia, antimony oxide, tin oxide, zinc oxide, ceria, and mixtures thereof. In some embodiments, the core comprises a non-metal oxide.

[0030] Generally, surface treatment agents for silica nanoparticles are organic species having a first functional group capable of covalently chemically attaching to the surface of a nanoparticle, wherein the attached surface treatment agent alters one or more properties of the nanoparticle. In some embodiments, surface treatment agents have no more than three functional groups for attaching to the core. In some embodiments, the surface treatment agents have a low molecular weight, e.g. a weight average molecular weight less than 1000 gm/mole.

[0031] In some embodiments, the surface treatment agent further includes one or more additional functional groups providing one or more additional desired properties. For example, in some embodiments, an additional functional group may be selected to provide a desired degree of compatibility between the surface modified nanoparticles and one or more of the additional constituents of the resin system, e.g., one or more of the curable resins and/or reactive diluents. In some embodiments, an additional functional group may be selected to modify the rheology of the resin system, e.g., to increase or decrease the viscosity, or to provide non-Newtonian rheological behavior, e.g., thixotropy (shear- thinning).

[0032] In some embodiments, the surface-modified nanoparticles are reactive; that is, at least one of the surface treatment agents used to surface modify the nanoparticles of the present disclosure may include a second functional group capable of reacting with one or more of the curable resin(s) and/or one or more of the reactive diluent(s) of the resin system.

[0033] Particle size measurements can be based on, e.g., transmission electron microscopy (TEM). In some embodiments, the surface-modified nanoparticles have a primary particle size (as determined by TEM) of between about 5 nanometers to about 500 nanometers, and in some embodiments from about 5 nanometers to about 250 nanometers, and even in some embodiments from about 50 nanometers to about 200 nanometers. In some embodiments, the cores have an average diameter of at least about 5 nanometers, in some embodiments, at least about 10 nanometers, at least about 25 nanometers, at least about 50 nanometers, and in some embodiments, at least about 75 nanometers. In some

embodiments the cores have an average diameter of no greater than about 500 nanometers, no greater than about 250 nanometers, and in some embodiments no greater than about 150 nanometers.

[0034] In some embodiments, silica nanoparticles can have a particle size of ranging from about 5 to about 150 nm. Commercially available silicas include those available from Nalco Chemical Company, Naperville, Illinois (for example, NALCO 1040, 1042, 1050, 1060, 2327 and 2329); Nissan Chemical America Company, Houston, Texas (e.g., SNOWTEX-ZL, -OL, -O, -N, -C, -20L, -40, and -50); and Admatechs Co., Ltd., Japan (for example, SX009-MIE, SX009-MIF, SC1050-MJM, and SC1050-MLV).

[0035] In some embodiments, the core is substantially spherical. In some embodiments, the cores are relatively uniform in primary particle size. In some embodiments, the cores have a narrow particle size distribution. In some embodiments, the core is substantially fully condensed. In some embodiments, the core is amorphous. In some embodiments, the core is isotropic. In some embodiments, the core is at least partially crystalline. In some embodiments, the core is substantially crystalline. In some embodiments, the particles are substantially non-agglomerated. In some embodiments, the particles are substantially non-aggregated in contrast to, for example, fumed or pyrogenic silica.

[0036] Examples

[0037] The following examples illustrate some of the problems that can arise when surface-modified nanoparticles are used in a curable resin system and methods of further treating the nanoparticles. These examples also illustrate the beneficial and detrimental effects associated with various treatments of the surface-modified nanoparticles.

Table 1 : Description of materials used in the preparation of examples.

[0038] Comparative Example 1. CE-I was prepared as follows.

[0039] NP-2329 colloidal silica sol (1600 grams (g)) was added to a large jar and stirred, l-methoxy-2-propanol (1800 g), A- 174 (11.14 g), and SILQUEST A-1230 (22.45 g) were mixed together and then added to the stirring colloidal silica. The resulting colloidal silica sol was then heated for sixteen hours at 80 degrees Celsius ( ⁰ C) to produce a sol of surface-modified silica nanoparticles.

[0040] Solvents were removed from the sol using rotary evaporation to achieve a concentrated solution containing 70% by weight (70 wt.%) silica. The concentrated solution was then dried according to the procedures described in the U.S. Pat. No. 5,980,697 (KoIb et al.) and U.S. Pat. No. 5,694,701 (Huelsman, et. al.) with a dispersion coating thickness of about 0.25 mm (10 mils) and a residence time of 1.1 minutes (heating platen temperature 107 ⁰ C, and condensing platen temperature 21 ⁰ C) to yield a fine, free-

flowing white powder of surface-modified nanoparticles. The dried, surface-modified nanoparticles (233.12 g) were then combined with 316.9 g of HK-I and 1.27 g of 4- hydroxy-TEMPO and mixed with a DISPERMAT laboratory dissolver (BYK-Gardner, Columbia, MD) for ten minutes. The nanoparticle-containing resin system was then milled.

[0041] Milling Process. The dry nanoparticles were first mixed into the resin with a Cowles-blade-type, pneumatically-driven mixing blade. The resin was placed into a wide- mouth jar and the dry nanoparticle powder was added to the resin while stirring. The combination was then mixed more thoroughly with the DISPERMAT dissolver for approximately ten minutes until a uniform consistency was achieved.

[0042] After the preliminary mixing was complete, the nanoparticle/resin mixture was pumped through a peristaltic pump (MASTERFLEX LS, Cole Partner, Vernon Hills, Illinois), and into a MINICER mill (0.15 liter laboratory-sized MINIZETA horizontal mill, Netszch Fine Particle Technology, Exton, PA) equipped with ceramic internal components. The flowrate through the peristaltic pump was 150 ml/min. The milling media was 0.5 millimeter diameter yttria-stabilized zirconia milling media (Torayceram AGB-K-0.5, Toray International America, New York, NY). The mill run rate was set to 4320 rpm.

[0043] Comparative Example 2. CE-2 was prepared as follows. NP -2329 colloidal silica sol (1625 g) was added to a large jar and stirred. AMBERLITE IR-120plus(H) cation exchange resin (11.35 g) was added to the stirring colloidal silica sol and stirred for 30 minutes until the pH was 3.0. The cation exchange resin was filtered out and 1185 g of l-methoxy-2-propanol was added to the stirring colloidal silica sol. Ammonium hydroxide (30%) was added to the stirring colloidal silica sol until the pH was 9.5. The concentration of silica was 22.58 wt.%.

[0044] The resulting cation-exchanged colloidal silica sol (2720.4 g) was added to a large jar and stirred, l-methoxy-2-propanol (615 g), A- 174 (10.52 g), and SILQUEST A- 1230 (21.20 g) were mixed together and then added to the stirring colloidal silica and the mixture was then heated for 16 hours at 80 ⁰ C to produce a surface-modified silica sol. Solvents were removed from the sol using rotary evaporation to achieve a concentrated solution containing 70% by weight (70 wt.%) silica. The concentrated solution was then

dried according to the procedures described in the U.S. Pat. No. 5,980,697 (KoIb et al.) and U.S. Pat. No. 5,694,701 (Huelsman, et al.) with a dispersion coating thickness of about 0.18 mm (7 mils) and a residence time of 0.94 minutes (heating platen temperature 107 ⁰ C, and condensing platen temperature 21 ⁰ C) to yield a fine, free-flowing white powder.

[0045] The dried, surface-modified nanoparticles (233.12 g) were then combined with 316.9 g of HK-I and 1.27 g of 4- hydroxy-TEMPO and mixed with a DISPERMAT dissolver for ten minutes. The nanoparticle-containing resin system was then milled as described for Comparative Example 1.

[0046] Example 1. EX-I was prepared as follows.

[0047] NP-2329 colloidal silica sol (1625 g) was added to a large jar and stirred. DOWEX MONOSPHERE 550A(OH) anion exchange resin (88 g) was added to the jar and stirred for 30 minutes until the pH was 9.5. The anion exchange beads were filtered out. Next, 140 g of DOWEX MONOSPHERE 650C(H) cation exchange resin was added to the jar and stirred for 30 minutes until the pH was 3.0. The cation exchange beads were filtered out and 600 g of l-methoxy-2-propanol was added to the jar while stirring. Ammonium hydroxide (30%) was slowly added to the stirring colloidal silica sol until the pH was 9.5. The concentration of silica was 25.71 wt.%.

[0048] The resulting anion/cation-exchanged colloidal silica sol (2003.1 g) was added to a large jar and stirred, l-methoxy-2-propanol (1200 g), A-174 (8.82 g), and SILQUEST A-1230 (17.78 g) were mixed together and then added to the stirring colloidal silica and the mixture was then heated for 16 hours at 80 ⁰ C. Solvents were removed from the sol using rotary evaporation to achieve a concentrated solution containing 70% by weight (70 wt.%) silica. The concentrated solution was then dried according to the procedures described in the U.S. Pat. No. 5,980,697 (KoIb et al.) and U.S. Pat. No. 5,694,701 (Huelsman, et al.) with a dispersion coating thickness of about 0.25 mm (10 mils) and a residence time of 1.1 minutes (heating platen temperature 107 ⁰ C, and condensing platen temperature 21 ⁰ C) to yield a fine, free-flowing white powder..

[0049] The dried, surface-modified nanoparticles (233.12 g) were then combined with 316.9 g of HK-I and 1.27 g of 4- hydroxy-TEMPO and mixed with a DISPERMAT dissolver for ten minutes. The nanoparticle-containing resin system was then milled as

described for Comparative Example 1 , except the flow rate through the peristaltic pump was 200 ml/minute.

[0050] The milled, nanoparticle-containing, curable resin systems of CE-I, CE-2, and EX-I were used to prepare flat, gel-coated, 3 -ply, fiber-reinforced composite test panels according to Panel Preparation Procedure 1. A reference panel (REF-I) was also prepared using the unsaturated polyester/styrene resin system (GDW 8082 from HK Research Corp.) without nanoparticles according to Panel Preparation 1.

[0051] Panel Preparation Procedure 1. Each milled, nanoparticle-containing, curable resin system ("gel coat") was promoted with 1.00 wt% cobalt solution (JK 8033, available from HK Research (Hickory, NC)) and was initiated with 1.25 wt.% methyl ethyl ketone peroxide (MEKP B0410 46-702, available from HK Research (Hickory, NC )). The MEKP was mixed into the gel coat using a SPEEDMIXER DAC 600 FVZ available from FlackTek, Inc. (Landrum, SC) for 30 seconds at 2000 rpm. A 30.5 cm (12 inch) wide adjustable gap knife coater was set to a gap of 1.1 millimeters (0.045 inch) and used to coat 50 to 80 grams of the initiated gel coat across a mold-released glass plate.

[0052] After the gel coat layer was allowed to cure for one hour at room temperature, a 3 -ply panel was laid up on the cured gel coat, as follows. Fiberglass chopped strand mat weighing (0.92 kg/square meter (3 oz/square foot); Item 50219, CSM 3 oz x 38 inch, available from Fiberglass Warehouse, DMC, La Mesa, CA) was cut into 15 cm by 15 cm (6 inch by 6 inch) squares. Three plies of the chopped strand mat were weighed and this mass was doubled to calculate the amount of vinyl ester (ASHLAND 922-L25, available from Express Composites, Inc., Minneapolis, MN) to be used, resulting in a composite resin:fiber ratio of 2: 1. The calculated amount of vinyl ester resin was weighed out, and initiated with 1.25 wt.% methyl ethyl ketone peroxide (LUPEROX DDM-9, available from Sigma- Aldrich, Milwaukee, WI). One-third of the vinyl ester resin was poured evenly onto the cured gel coat layer. The first ply of chopped strand mat was placed on top of the vinyl ester resin and allowed to soak up the resin. A ridged roller was used to roll out the vinyl ester resin and remove air bubbles. An additional one -third of the vinyl ester resin was poured on top of the first ply, spread out, and covered with a second ply of chopped strand mat. This middle ply was rolled out, and then the last one-third of the

vinyl ester resin was poured on top of the middle ply and spread out. Finally, the third ply of chopped strand mat was laid up and rolled out.

[0053] This 3-ply construction was allowed to cure in a vented hood for 24 hours at room temperature. Then, while still on the glass, it was put into a 70 ⁰ C (158 ⁰ F) oven for 4 hours. After removal from the oven, the glass and panel are allowed to cool for one hour before the panel is removed from the glass, exposing the gel coat layer. The panel was cut into as many 7 cm by 7 cm (2.75 inch x 2.75 inch) square test specimens as possible.

[0054] Example 10. EX-10 was prepared as follows.

[0055] The NP-LMS colloidal silica sol (1201.6 g) was added to a large jar and stirred. l-methoxy-2-propanol (704 g), A- 174 (16.58 g), and SILQUEST A-1230 (33.46 g) were mixed together and then added to the stirring colloidal silica and the mixture was then heated for 16 hours at 80 ⁰ C.

[0056] The surface modified silica sol was dried to a powder in an oven at 80 ⁰ C. Next, 475.1 g of the dried powder and H lO g acetone were mixed in ajar and high shear mixed for 10 minutes using a SILVERSON 4LR high shear mixer. HK-3 curable resin (615.0 g), 4-hydroxy-TEMPO (2.5 g) and 2-(Dimethyl amino)ethyl methacrylate (DMAEMA) (3.5 g) were added to the surface modified silica dispersed in acetone. Acetone was removed via rotary evaporation. Gas chromatography confirmed no acetone remained in the sample. MMA (28.41 g) and styrene (114.3 g) were added back to the sample. The sample contained approximately 40 wt.% Siθ2-

[0057] The nanoparticle-containing, curable resin system of EX-10 was used to prepare flat, gel-coated, 3-ply, fiber-reinforced composite test panels according to Panel Preparation Procedure 2.

[0058] Panel Preparation Procedure 2.

[0059] The nanoparticle-containing curable resin system ("gel coat") was promoted with 1.00 wt% cobalt solution (JK 8033, available from HK Research (Hickory, NC)) and was initiated with 2.0 wt.% methyl ethyl ketone peroxide (MEKP B0410 46-702, available from HK Research (Hickory, NC )). The MEKP was mixed into the curable resin system using a SPEEDMIXER DAC 600 FVZ available from FlackTek, Inc. (Landrum, SC) for 15 seconds at 2000 rpm. A 30.5 cm (12 inch) wide adjustable gap knife coater was set to

a gap of 0.89 millimeters (0.035 inch) and used to coat 50 to 80 grams of the initiated gel coat across a mold-released glass plate.

[0060] After the gel coat layer was allowed to cure for one hour at room temperature, a 3 -ply panel was laid up upon the cured gel coat. Fiberglass chopped strand mat weighing 0.46 kg/square meter (1.5 oz/square foot) (CSM 1.5 oz x 38 inch, available from Fiberglass Warehouse, DMC, La Mesa, CA) was cut into 15 cm by 15 cm (6 inch by 6 inch) squares. Three plies of the chopped strand mat were weighed and this mass was doubled to calculate the amount of vinyl ester (HK Vinyl Ester R0904, available from HK Research (Hickory, NC)) to be used, resulting in a composite resin:fiber ratio of 2: 1. The calculated amount of vinyl ester resin was weighed out, initiated with 2.0 wt.% cumyl peroxide (NOROX CHM-50, available from HK Research (Hickory, NC)). One-third of the vinyl ester resin was poured evenly onto the cured gel coat layer. The first ply of chopped strand mat was placed on top of the vinyl ester resin and allowed to soak up the resin. A ridged roller was used to roll out the vinyl ester resin and remove air bubbles. An additional one-third of the vinyl ester resin was poured on top of the first ply, spread out, and covered with a second ply of chopped strand mat. This middle ply was rolled out, and then the last one-third of the vinyl ester resin was poured on top of the middle ply and spread out. Finally, the third ply of chopped strand mat was laid up and rolled out.

[0061] This 3 -ply construction was allowed to cure in a vented hood for 24 hours at room temperature. Then, while still on the glass, it was put into a 70 ⁰ C (158 ⁰ F) oven for 4 hours. After removal from the oven, the glass and panel are allowed to cool for one hour before the panel is removed from the glass, exposing the gel coat layer. The panel was cut into as many 7 cm by 7 cm (2.75 inch x 2.75 inch) square test specimens as possible.

[0062] The panels prepared from curable resin systems CE-I, CE-2, EX-I, EX-10, and REF-I were evaluated according to Boiling Blister Test A.

[0063] Boiling Blister Test A. A 5-liter, 3-necked flask was cut in half at the equator and an 8.9 cm (3.5 inch) wide cylindrical band was inserted. Portholes were cut into this band, each with a glass extension tube attached. These 7.6 cm (3 inch) diameter ports extended out from the flask about 7.6 cm (3 inches) and had a flange at the end. This flange accepted a 7.6 cm (3 inch) diameter by 6.4 mm (0.25 inch) thick O-ring onto which

a flat test specimen can be clamped with a spring-loaded U-clamp. The apparatus had four ports spaced evenly around the equator of the flask.

[0064] The flask was placed in a heating mantle controlled by a Vari-AC set at 100% (of 120 VAC). One neck of the flask contained a glass stopper and was used for pouring water into the flask. The center neck contained a condenser which was cooled by circulating cold water and returned vaporized water to the flask. The third neck contained a mercury thermometer inserted into the water.

[0065] Boiling chips were put into the bottom of the flask. Deionized water was preheated in a microwave oven and poured into the flask. Enough preheated water was added to bring the water level up to the bottom of the ports. When the water started to boil, the test specimens, which have been carefully weighed, were clamped onto the ends of the ports with the cured experimental resin layer side toward the water. More hot deionized water was added until the ports were completely filled.

[0066] When all the water returned to a boil, the clock was started. After 24 hours, the water was carefully drained to a level just below the bottom of the ports. The test specimens were removed, weighed, and observed. Any unusual surface effects, blisters, color changes, transparency changes, etc. were noted and described. A qualitative ranking system was applied to describe the severity of surface changes.

[0067] For some specimens, if no changes were observed after 24 hours, the test was continued for up to 100 hours. In order to extend the testing after the 24 test, the test specimens were re-clamped to the ends of the ports, the flask refilled with pre-heated deionized water, and timing restarted when the water began to boil. All water was replaced after 100 hours of testing.

Table 2: Boiling Blister Test A results.

* Commercially available low metal sol.

[0068] Although the presence of cations has been associated with the formation of blisters, cation-exchange alone was not an effective approach to solving the blistering problem. Rather, the present inventors discovered that anion exchange followed by cation exchange (i.e., "dual-ion-exchange") was required. Alternatively, low metal content sols could be used. The anion, cation, and total ion contents of the surface-modified nanoparticle sols that had been subjected to various ion exchange procedures were evaluated as follows.

[0069] IC Procedure #1. Ultra-pure (18 mega ohm) water was added to samples of various nanoparticle sols to obtain 2x, 10x, 5Ox, and 10Ox dilutions that were centrifuged at 10,000 rpm for 45 minutes to remove the solids. The remaining supernatant was analyzed by ion-exchange chromatography (IC). Aliquots (25 microliter) of each dilution were injected into a DIONEX ICS-3000 dual channel ion chromatograph. CS12A and CG12A columns, isocratic 20 mM methanesulfonic acid eluent, a CSRS suppressor, and conductivity detection were used for cation chromatography. The system was standardized with the DIONEX 6-cation standard (in dilutions). AS 18 and AGl 8 columns, gradient KOH eluent, an ASRS suppressor, and conductivity detection were used for anion chromatography. The system was standardized with the DIONEX 6-cation and 7-anion standards (in dilutions). Data were obtained as microgram per milliliter. Conversion of these data to parts per million (PPM) was done by dividing the reported data by percent solids and density of the original solution. The reported numbers are micrograms of ion per gram of Siθ2-

Table 3 : Ion content of silica sols determined by the IC Procedure #1.

[0070] Surface-modified nanoparticle sols that had been subjected to various ion exchange procedures were added to a curable acrylate resin. The optical properties of these surface-modified nanoparticle-containing resin systems were evaluated.

[0071] Comparative Example 3. CE-3 was prepared as follows. The surface-modified silica sol of CE-I (50.16 g) was combined with polyethylene glycol acrylate (12 g), and A- hydroxy-TEMPO (0.05 g) in a round bottom flask. Water and alcohol were removed with rotary evaporation until less than 1% water/methoxypropanol as measured by gas chromatography remained in the surface-modified nanoparticle-containing resin system.

[0072] Comparative Example 4. CE-4 was prepared as follows. The cation-exchanged, surface-modified silica sol of CE-2 (43.86 g) was combined with polyethylene glycol acrylate (12 g), and 4- hydroxy-TEMPO (0.05 g) in a round bottom flask. Water and alcohol were removed with rotary evaporation until less than 1% water/methoxypropanol as measured by gas chromatography remained in the surface-modified nanoparticle- containing resin system.

[0073] Example 2. EX-2 was prepared as follows. The dual-ion-exchanged, surface- modified silica sol of EX-I (42.22 g) was combined with polyethylene glycol acrylate (12 g), and 4- hydroxy-TEMPO (0.05 g) in a round bottom flask. Water and alcohol were removed with rotary evaporation until less than 1 wt.% of water/methoxypropanol as measured by gas chromatography remained in the surface-modified nanoparticle- containing resin system.

[0074] Optical Property Test Method. Various uncured, nanoparticle-containing resins were tested for optical transmission, clarity, and haze using a BYK GARDNER HAZE- GARD PLUS (catalog no. 4723, supplied by BYK Gardner, Silver Spring, Maryland). The transmission, clarity, and haze levels were defined according to ASTM-D 1003-00, titled "Standard Test Method for Haze and Luminous Transmittance for Transparent Plastics". The instrument was referenced against air during the measurements. Light transmission (T) measurements are provided as a percentage of transmitted light based on the incident light. Haze is the scattering of light by a specimen responsible for the

reduction in contrast of objects viewed through it. Haze, H, is presented as the percentage of transmitted light that is scattered so that its direction deviates more than a specified angle from the direction of the incident beam. Clarity is evaluated using a ring detector and comparing the small-angle scattered light component to the specularly transmitted component.

[0075] For the transmission, haze, and clarity measurements of the uncured resins, a spacer was mounted between two pieces of fused silica glass (0.3 cm (1/8 in.) thick, 6.4 cm (2.5 in.) wide, 5 cm (2 in.) tall), such that the spacer was outside the optical measurement area and created a gap of 425 microns (μm) (0.01675 in.) into which individual samples of nanoparticle-containing resins were placed. Clamps, also mounted outside the measurement area, were used to hold the glass pieces tightly to the spacer and ensure that the gap spacing was restricted to the thickness of the spacer. Between 10 and 15 individual measurements of the transmission, haze, and clarity were taken on each of the liquid resin samples. The average percent transmission, haze, and clarity are reported. For reference, the percent transmission, haze and clarity were also measured on a sample of the polyethylene glycol acrylate liquid resin without nanoparticles (REF. 2).

Table 4: Optical properties of uncured nanoparticle-containing resins.

* The resin system of REF. 2 did not include surface-modified nanoparticles. [0076] Rheological Properties

[0077] Example 3. EX-3 was prepared as follows. NP-2329 (3406 g) was added to a large jar and stirred. DOWEX MONOSPHERE 550A(OH) anion exchange resin (184 g)was added to the jar and stirred for 30 minutes until the pH was 9.5. The anion exchange beads were filtered out. Next, 257.9 g of DOWEX MONOSPHERE 650C(H) cation exchange resin was added to the jar and stirred for 30 minutes until the pH was 3.0. The cation exchange beads were filtered out and ammonium hydroxide (30%) was added while slowly stirring until the pH was 9.5. The ion exchanged sol contained 35.6 wt.% silica. This process was repeated to double the batch size available for the next step.

[0078] The resulting dual-ion-exchanged colloidal silica sol (1493.5 g) was added to a large jar and stirred, l-methoxy-2-propanol (1680.2 g), A- 174 (9.10 g), and SILQUEST A-1230 (18.34 g) were mixed together and then added to the stirring colloidal silica and the mixture was heated for 16 hours at 80 ⁰ C. This process was repeated five times to increase the batch size available for the next step. Solvents were removed from the sol using rotary evaporation to achieve a concentrated solution containing 70% by weight (70 wt.%) silica. The concentrated solution was then dried according to the procedures described in the U.S. Pat. No. 5,980,697 (KoIb et al.) and U.S. Pat. No. 5,694,701 (Huelsman, et al.) with a dispersion coating thickness of about 0.25 mm (10 mils) and a residence time of 1.1 minutes (heating platen temperature 107 ⁰ C, and condensing platen temperature 21 ⁰ C) to yield a fine, free-flowing white powder.

[0079] The dried, dual-ion-exchanged, surface-modified nanoparticles (509.18 g) were then combined with 690.8 g of HK-2 and mixed with a DISPERMAT dissolver for ten minutes. The nanoparticle-containing resin system was then milled as described for Comparative Example 1.

[0080] The rheology of CE-I , which contained surface-modified nanoparticles that were not ion exchanged, was compared to the rheology of EX-3, which contained dual- ion-exchanged, surface-modified nanoparticles. The shear-dependent viscosity was evaluated by the Rheology Test Method.

[0081] Rheology Test Method. The rheo logical properties of a sample were measured at 25 ⁰ C using an ARES rheometer (available from TA Instruments, New Castle, Delaware). The upper fixture was a cone with an angle of 0.099 radians and a diameter of 25 mm. The lower fixture was a flat plate with a diameter of 38 mm. With the fixtures mounted in the instrument, the torque and normal force displays were electronically zeroed. The two fixtures were brought into contact, the gap was electronically zeroed, and the fixtures were then separated by approximately 50 cm to enable sample loading.

[0082] A suitable amount of resin was placed in the center of the lower plate. Care was taken to avoid entrainment of air in the resin. The upper fixture (cone) was lowered into the resin until the gap was filled with resin. With the motor off, the lower fixture was slowly turned by hand while the upper fixture was lowered until a gap of 0.0465 mm was

obtained. During this gap-setting process, excess resin was gently scraped off the lower plate as it squeezed out.

[0083] With the oven temperature set to 25 ⁰ C, the oven was closed and the temperature was allowed to equilibrate. A 25 ⁰ C steady rate sweep test was employed from 100 to 0.1 reciprocal seconds, logarithmically, 5 points per decade. The measured viscosity was reported in Pascal ^» seconds (Pa ^» s).

[0084] As shown in FIG. 1, the dual-ion-exchanged nanoparticles of EX- 3 resulted in significantly higher viscosities across the range of shear rates relative to the untreated nanoparticles of CE-I. In some applications, this increase in viscosity may be tolerable or even desirable. However, in some applications, the increase in viscosity may be undesirable, e.g., lower viscosities may be required when spraying coatings (e.g., gel coats) or when infusing fibrous substrates to produce composites. Just as decreasing the salt content of the nanoparticles produced an increase in viscosity of nanoparticle/resin system, the effect would be reversible in the event of addition of salt to a low salt content resin/nanoparticle system.

[0085] End-Capping

[0086] As discussed above, the present inventors discovered that the dual-ion-exchange protocol resulted in an unexpected increase in viscosity of a resin containing such nanoparticles. The present inventors explored whether free silanol groups (Si-OH) remained at the surface of the nanoparticle. Such Si-OH groups may contribute to the observed increase in viscosity due to, e.g., hydrogen bonding either particle to particle bonding or particle to resin bonding. In particular, the present inventors experimented with the addition of low-molecular weight compounds in attempt to displace protons from the silanol groups and associate these compounds with the remaining surface Si-O- groups. This process is referred to herein as "end-capping," and the low molecular weight compounds are referred to as "end-capping agents."

[0087] Some end-capping agents associate with the nanoparticles by ionically-bonding with Si-O- groups at the surface of the nanoparticle. 2-(Dimethyl amino)ethyl methacrylate (DMAEMA) is one example of a low molecular weight, ionically-bonding, end-capping agent. Other ionically bonding end-capping agents include other amines such as triethylamine, and pyridines (e.g., vinyl pyridine).

[0088] Some end-capping agents associate more strongly with the nanoparticles by covalently bonding with the Si-O- groups at the surface of the nanoparticles. Hexamethyl disilazane (HMDS) is one example of a low molecular weight, covalently-bonding, end- capping agent.

[0089] In some embodiments, it may be desirable to minimize or eliminate migration of an ionically-bonded end-capping agent, particularly low molecular weight end-capping agents. In some embodiments, the end-capping agent may include a functional group capable of reacting with another component of the resin system, for example the curable resin or a reactive diluent. The reactive functional group may be selected based on the characteristic of the resin into which the end-capped, surface-modified nanoparticles are to be added.

[0090] Example 4. EX-4 was prepared as follows. Dried, surface-modified nanoparticles were prepared as described above for Example 3. These dried surface- modified nanoparticles (509.18 g) were combined with 690.8 g of HK-2, and 2.26 grams of DMAEMA and mixed with a DISPERMAT dissolver for ten minutes. The nanoparticle-containing resin system was then milled as described for Comparative Example 1.

[0091] Example 5. EX-5 was prepared in the same manner as EX-4, except that 4.53 grams of DMAEMA were used.

[0092] The shear-dependent viscosity of EX-4 and EX-5 were evaluated by using the Rheology Test Method. As shown in FIG. 2, the end-capping samples showed a reduction in viscosity across the range of shear rates when using dual ion exchanged nanoparticles, as compared to the untreated sample of EX-3.

[0093] As shown in FIGS. 1 and 2, the use of end-capped, surface-modified nanoparticles significantly reduced the viscosity of a resin system comprising a curable resin having moieties that are capable of hydrogen bonding. The increase in viscosity associated with the hydrogen-bonding effect of non-end-capped, surface-modified nanoparticles is expected to increase with an increase in any one of (a) the number of moieties capable of hydrogen bonding present in the curable resin, (b) the polarity of such moieties, and (c) the molecular weight of the curable resin.

[0094] Samples for TEM investigation were ultramicrotomed at room temperature to a section thickness of 85 nm. Images were collected using a HITACHI H-9000 transmission electron microscope.

[0095] Figure 4 A is a TEM image of the dual-ion exchanged nanoparticle-containing resin system of EX-3, which was not end-capped. Figure 4B is a TEM image of the dual- ion exchanged nanoparticle-containing resin system of EX-4, which was end-capped with 0.03 millimoles (mmol) DMAEMA per gram of silica. Figure 4C is TEM image of the dual-ion exchanged nanoparticle-containing resin system of EX-5, which was end-capped with 0.06 mmol DMAEMA per gram of silica. All images are at 10,000X magnification. The improvement in particle dispersion resulting from the use of an end-capping agent is apparent from a comparison FIG. 4B (EX-4) and FIG. 4C (EX-5) to FIG. 4A (EX-3). Specifically, FIGS. 4B and 4C illustrate increasingly well-dispersed, non-aggregated, non- agglomerated nanoparticles as the concentration of DMAEMA was increased.

[0096] The optical properties of approximately 1.5 mm thick cured samples of these end-capped nanoparticle-containing resin systems were measured according to the Optical Properties Test Method, except that no spacer or glass set-up was required to measure the properties of the cured samples. Between 5 and 7 individual measurements of transmission, haze and clarity were taken for each cured resin sample, and the average is reported in Table 5.

[0097] Two reference examples (REF-3 and REF-4) were also prepared using the unsaturated polyester/styrene resin system (GDW 8082 from HK Research Corp.) without nanoparticles. In REF-3, the resin sample was promoted with 1.00 wt% cobalt solution (JK 8033, available from HK Research (Hickory, NC)) and was initiated with 1.25 wt.% methyl ethyl ketone peroxide (MEKP B0410 46-702, available from HK Research (Hickory, NC )). The MEKP was mixed into the gel coat using a SPEEDMIXER DAC 600 FVZ available from FlackTek, Inc. (Landrum, SC) for 30 seconds at 2000 rpm. The liquid sample was poured between two pieces of mold-release-coated glass held apart by a spacer. The resin-glass-spacer construction was allowed to cure in a vented hood for 24 hours at room temperature. Then, the resin-glass-spacer construction was put into a 70 ⁰ C (158 ⁰ F) oven for 4 hours. After removal from the oven, the glass and sample were allowed to cool for one hour before the cured sample was removed from the glass. The

resin sample of REF-4 was pre -heated for 10 minutes at 80 ⁰ C prior to being prepared according to the process used to prepare REF-3. The resulting values of Transmission, Haze and Clarity are also shown in Table 5. Table 5: Effects of end-ca in on o tical ro erties avera e of 5-7 sam les .

[0098] Example 6. EX-6 was prepared as follows. Dual-ion-exchanged nanoparticles were prepared as follows. NP -2329 (3235 g) was added to a large jar and stirred. DOWEX MONOSPHERE 550A(OH) anion exchange resin (175 g) was added to the jar and stirred for 30 minutes until the pH was 9.5. The anion exchange beads were filtered out. Next, 203.5 g of DOWEX MONOSPHERE 650C(H) cation exchange resin was added to the jar and stirred for 15 minutes until the pH was 3.0. The cation exchange beads were filtered out and ammonium hydroxide (30%) was added while slowly stirring until the pH was 9.5. The dual-ion-exchanged sol contained 35.4 wt.% silica. This process was repeated to double the batch size available for the next step.

[0099] The resulting dual-ion-exchanged colloidal silica sol (1493.5 g) was added to a large jar and stirred, l-methoxy-2-propanol (1680.19 g), A- 174 (9.10 g), and SILQUEST A-1230 (18.34 g) were mixed together and then added to the stirring colloidal silica and the mixture was heated for 16 hours at 80 ⁰ C. This process was repeated five times to increase the batch size available for the next step. Solvents were removed from the sol using rotary evaporation to achieve a concentrated solution containing 70.4% by weight. A portion of the concentrated silica sol (461.7 g) and HMDS (52.46 g) were mixed together for 1 hour. The HMDS-treated, concentrated solution was then dried according to the procedures described in the U.S. Pat. No. 5,980,697 (KoIb et al.) and U.S. Pat. No. 5,694,701 (Huelsman, et al.) with a dispersion coating thickness of about 0.25 mm (10 mils) and a residence time of 1.1 minutes (heating platen temperature 107 ⁰ C, and condensing platen temperature 21 ⁰ C) to yield a fine, free-flowing white powder.

[0100] The dried, HMDS-treated, surface-modified nanoparticles (250.44 g) were then combined with 336.03 g of HK-2 and 4- hydroxy-TEMPO (1.3 g) and mixed with a DISPERMAT dissolver for ten minutes. The nanoparticle-containing resin system was then milled as described for Comparative Example 1.

[0101] As shown in FIG. 3, end-capping with HMDS, a covalently-bonding agent (Ex. 6), resulted in a lower viscosity at all shear rates relative to the untreated sample (EX-3), which was not end-capped. However, the viscosity reduction was less than that observed for the ionically-end-capped sample (EX-5).

[0102] Low ion content silica sols.

[0103] Example 7. EX-7 was prepared as follows. First, 1205.5 g of NP-136 low ion content silica sol were added to a large jar and stirred. 837 g l-methoxy-2-propanol, 7.73 g A- 174 and 6.18 g phenyl trimethoxy silane were mixed together and then added to the stirring colloidal silica and the mixture was then heated for 16 hours at 80 ⁰ C. Next, 562.95 g of this mixture, 86.20 g of HK-3, 0.34 g of 4-hydroxy-TEMPO, and 0.4 g DMAEMA were added to a round bottom flask. Water and alcohol were removed via rotary evaporation. Gas chromatography confirmed no methoxy-propanol remained in the sample.

[0104] Finally, 3.9 g MMA and 14 g of styrene were added to this sample and 72.76 g of the resulting mixture were combined with 27.32 g of HK-3 and mixed together to form EX-7. Thermogravinometric analysis confirmed 39.5% SiO2.

[0105] Example 8. EX-8 was prepared as follows. First, 1202.4 g NP-135 were added to a large jar and stirred. Next, 736.8 g l-methoxy-2-propanol, 9.40 g A- 174 and 7.52 g phenyl trimethoxy silane were mixed together and then added to the stirring colloidal silica and the mixture was then heated for 16 hours at 80 ⁰ C. Next, 440.35 g of this mixture, 86.20 g of HK-3, 0.34 g of 4-hydroxy-TEMPO, and 0.4 g DMAEMA were added to a round bottom flask. Water and alcohol were removed via rotary evaporation. Gas chromatography confirmed no methoxy-propanol remained in the sample.

[0106] Finally, 3.9 g MMA and 19.1 g of styrene were added to the sample and 72.73 g of the resulting mixture and 27.32 g of HK-3 were mixed together to form EX-8. Thermogravinometric analysis confirmed 40.2 wt.% silica.

[0107] Example 9. EX-9 was prepared as follows. First, 1294.0 NP-MJM, 841.7g of HK-3 and 3.4 g of 4-hydroxy-TEMPO were added to a round bottom flask. Methyl-ethyl ketone was removed via rotary evaporation. Gas chromatography confirmed no solvent remained in the sample. Then, 38.5 g MMA and 107.7 g styrene were added back to the sample to produce EX-9. The sample contained approximately 40 wt.% silica.

[0108] The ion content of resin systems prepared using low and high ion content silica sols were measured according to IC Procedure #2, as follows.

[0109] IC Procedure #2. Extractable anions: The sample was weighed (50-100 mg) into a small glass vial and dissolved in 200 mL dry acetonitrile. 1 mL ultrapure MiIIiQ water was added and the vials vigorously shaken. The contents were transferred to 1.5 mL microcentrifuge tubes. The tubes were placed in a microcentrifuge at 10,000 rpm for 10 minutes to settle the particles. An aliquot of the supernatant was transferred to an autosampler vial and injected onto a DIONEX ICS-3000 dual-channel ion chromatograph with suppressed conductivity detection. A ramp/step hydroxide gradient profile was applied, and the analysis was performed on AG 18/AS 18 columns. Blanks were also carried through this procedure for comparison. The samples were extracted in triplicate. Extractable cations: The same procedure was followed as in the anion preparation. The ion chromatography was run with a CG12A and CS12A column set, and 2OmM MSA isocratic elution. Data were obtained as micrograms of ion per gram total weight silica and resin.

[0110] The ions listed in Table 6 were specifically tested using the method described above. It is probable that there are other ionic species present, but they were not analyzed in the IC method described. Therefore, the list of ions in Table 6 is not exclusive of other ions which could potentially be present in the resin system.

[0111] Panels were prepared from the resin systems of EX-7, EX- 8, and EX-9 according to Panel Preparation 2 and tested according to Boiling Blister Test B. Ion contents and boil resistance are summarized in Table 6.

[0112] Boiling Blister Test B.

[0113] Boiling Blister Test B is the same as Boiling Blister Test A with the exception that, the test was continued for up to 6 days. In order to extend the testing after the 24 test,

the test specimens were re-clamped to the ends of the ports, the flask refilled with preheated DI water, and timing restarted when the water began to boil. All water was replaced after 6 days of testing.

Table 6: Comparison of low and high ion content resin systems as measured by IC Procedure #2. (Ion content in micrograms of ion per gram of total silica and resin.)

[0114] The presence of cations, specifically alkali metals (e.g., sodium, potassium, and lithium) and alkaline earth metals (e.g., magnesium and calcium), would be expected to cause blistering in nanoparticle containing curable resin systems. However, the methods used to reduce sodium ion content (ion exchange or low ion content sols) would be expected to concurrently reduce the alkali and alkaline earth metal content. Thus, curable resin systems with 40 wt.% surface-treated nanoparticles should contain no greater than 200 ppm, in some embodiments, no greater than 100 ppm, of total alkali metals and alkaline earth metals based on the total weight of silica and resin. For example, as shown in Table 6, the curable resin systems with 40 wt.% surface-modified nanoparticles having a total alkali and alkaline earth metal ion content of less than 100 ppm by total weight of silica and resin, produced good boil performance.

[0115] Using the sodium content for the curable resin of REF-I in Table 6, we can predict the sodium ion content contribution from the resin only in a curable resin system with varying nanoparticle loading. When combined with the information from Table 3, the additional sodium contribution arising from the addition of nanoparticles to the curable resin system can also be calculated. For example, a curable resin system containing 10 wt.% of the low metal silica nanoparticles would contain 81 ppm of sodium ion originating from the resin (i.e., 90 ppm Na/gram resin* 0.9 wt fraction resin) and an additional 15 ppm sodium ion originating from the NP-LMS sol nanoparticles (151 ppm Na/gram silica * 0.1 wt fraction silica). This yields a total sodium ion content of 96 ppm per total weight silica and resin. Although the other alkali and alkaline earth metal ions are not reported in Table 3, it is reasonable to expect that there would be some additional ion content contributions from those other species in the curable resin system from EX-10 which passed the Boiling Blister Test A.

[0116] In some embodiments, it may be desirable to have a resin system comprising a curable resin and at least 10 wt.%, in some embodiments, at least 20 wt.%, or even at least 40 wt.% surface-modified silica nanoparticles; wherein the resin system comprises no greater than 200 parts per million, and in some embodiments, no greater than 100 parts per million, by weight alkali metal and alkaline earth metal ions based on the total weight of the silica nanoparticles and curable resin.

[0117] The treated, surface-modified nanoparticles of the present disclosure and the resin systems containing such nanoparticles may be used in a wide variety of applications, including as coatings and as the resin system of composite articles (e.g., glass and/or carbon fiber composites). In some embodiments, the resin systems of the present disclosure may be useful as gel coats, e.g., gel coats for the marine industry.

[0118] Various modifications and alterations of this invention will become apparent to those skilled in the art without departing from the scope and spirit of this invention.