Effects of sulphonic and phosphonic acrylic monomers on the crosslinking of acrylic latexes with cycloaliphatic epoxide

Progress in Organic Coatings 36 (1999) 21±33

Effects of sulphonic and phosphonic acrylic monomers on the crosslinking of acrylic latexes with cycloaliphatic epoxide Shaobing Wu1, Jon D. Jorgensen, Allen D. Skaja, Jonathan P. Williams, Mark D. Soucek* Department of Polymers and Coatings, North Dakota State University, Fargo, ND 58105, USA Received 15 July 1998; accepted 12 November 1998

Abstract Two model acrylic latexes were synthesized using methyl methacrylate (MMA) and butyl acrylate (BA) with methacrylic acid (MAA) or 2-hydroxyethyl methacrylate (HEMA). The MAA or HEMA were incorporated to provide the latexes with carboxyl and hydroxyl functionality, respectively. A cycloaliphatic diepoxide (3, 4-epoxycyclohexyl methyl-30 , 40 -epoxycyclo-hexane carboxylate) was used as a crosslinker for both the latexes. The crosslinking of the latexes with the diepoxide was catalyzed using a copolymerizable or free sulphonic or phosphonic acids. The copolymerizable acids (2-sulfoethyl methacrylate (SEM) and acid phosphoxyethyl methacrylate (PEM)) were added during the latex synthesis. The free acids (p-toluene sulphonic acid (TsOH) and phenylphosphonic acid (PPA)) were added into the latex emulsion shortly before crosslinking. The crosslinked coatings were evaluated in terms of water absorption, gel content, and pull-off adhesion. The crosslinking of the hydroxyl functional latex coatings with the cycloaliphatic epoxide required an acid catalyst, whereas the crosslinking of the carboxyl latex coatings did not need an acid catalyst. Sulphonic and phosphonic acrylic acid monomers not only functioned as catalysts for the crosslinking reactions, but also improved the adhesion and freeze±thaw stability of the latex coatings. In addition, neutralization of the acid catalysts led to reduction of the crosslinker hydrolysis, and consequently enhanced the overall coating properties. # 1999 Elsevier Science S.A. All rights reserved. Keywords: Crosslinking; Acrylic latex; Waterborne coatings; Cycloaliphatic epoxide; Catalyst; 2-Sulfoethyl methacrylate (SEM); Phosphoxyethyl methacrylate (PEM)

1. Introduction The use of water-borne coatings has been increasing steadily as the coatings industry strives to produce environmentally friendly coating products. Conventional latex coatings use thermoplastic polymers as the binders. Consequently, the ®lm formation of the latex coatings strictly relies on the coalescence of the latex particles to provide chemical and mechanical resistance [1]. Coalescence is a physical process of macromolecule entanglements, and it is not an adequate substitute for the chemical bonding which occurs during crosslinking [2]. Since chemical crosslinking can provide coatings with better chemical and physical resistance, crosslinking technology has been adopted into the latex coatings for higher performance applications [3,4].

*Corresponding author. Tel.: +1-701-231-8438; fax: +1-701-231-8439; e-mail: [email protected] 1 Presently address: Lilly Industries, High Point, North Carolina.

A number of crosslinkers and crosslinking technologies have been reported for latex coatings in recent years [3±8]. For the hydroxyl functional acrylates, amino resins or blocked isocyanates are the current crosslinking technology [5±8]. These systems, especially the amino resin crosslinked system, provide good cost-performance effectiveness. However, these coating systems present severe toxicity problems. In addition, a high baking temperature (150±2008C) is also required for the crosslinking. For the carboxyl functional latexes, typical crosslinkers are polycarbodiimides [9,10], polyaziridines [11,12], and polyoxazolines [13,14]. These coating systems, however, also present environment and public health related problems or unsatisfactory coating properties [15]. Cycloaliphatic epoxides differ from traditional diglycidylether of bisphenol-A (BPA) epoxy resins [16±20]. Cycloaliphatic epoxides exhibit higher reactivity under acidic conditions, excellent weatherability, and low resin viscosity. In addition, cycloaliphatic epoxides are more reactive towards hydroxyl and carboxyl containing com-

0300-9440/99/$ ± see front matter # 1999 Elsevier Science S.A. All rights reserved. PII: S 0 3 0 0 - 9 4 4 0 ( 9 8 ) 0 0 0 7 5 - 7

22

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

pounds. Consequently, a variety of hydroxyl or carboxyl functional compounds can be crosslinked with cycloaliphatic epoxides. In contrast, basic compounds such as amines react very sluggishly with cycloaliphatic epoxides [21]. However, amine compounds are very reactive towards glycidyl type epoxies [22]. Typically, latexes are neutralized to a basic condition. Consequently, when phenyl glycidyl ether type epoxies are used as latex crosslinkers, the basic condition can cause side-crosslinking reaction(s) of the crosslinkers with the neutralizer in the coatings reducing pot-life. For this reason, phenyl glycidyl ether type epoxies may be not suitable as crosslinkers for latex coatings. The crosslinking reactions of hydroxyl and carboxyl groups with cycloaliphatic epoxide have been studied previously [15,23±25]. The ring opening reactions of cycloaliphatic epoxide by hydroxyl and carboxyl groups are dependent on the pH, temperature, and catalysts [23±25]. As the pH decreases, the reaction rates increase. Generally, carboxyl groups are more reactive than hydroxyl groups. The reactions of the carboxyl groups with the epoxide also show lower activation energy than the hydroxyl groups. Consequently, the reactions of the carboxyl groups with cycloaliphatic epoxides are less dependent on the temperature. The overall observed reaction rate of hydroxyl groups with the epoxide is proportional to the concentration of the acid catalyst, and inversely proportional to the square root of the concentration of the base, suggesting a strong dependence on the pH of the reaction media. The reaction rate of carboxyl group with epoxide, however, is proportional to the concentration of the carboxylic acid group, and it does not depend on the catalyst or the base concentration indicating less dependence on the pH of the reaction media. Catalysts or bases used to control pH also in¯uence the reaction rates of cycloaliphatic epoxides with hydroxyl or carboxyl groups [26]. Amines, especially, tertiary amines are the typical catalysts for the curing of glycidyl type epoxies [27±29]. However, amines inhibit the reactions of cycloaliphatic epoxide with hydroxyl compounds [21,26]. For the reactions with carboxyl groups, amines speed up the reactions when the concentration is low. When the amine concentration increased to or greater than 0.1 M, the carboxylic reaction was also inhibited [26]. Strong acid catalysts, such as toluene sulphonic acid, were also shown to increase the hydrolysis and esteri®cation of the epoxide [26]. To develop a cycloaliphatic diepoxide crosslinkable acrylic latex coating, as shown in Scheme 1, we previously reported the synthesis and crosslinking of model carboxyl and hydroxyl functional latexes with cycloaliphatic epoxide [30]. The crosslinking reactions of the model latexes with the cycloaliphatic diepoxide are illustrated in Scheme 2 and Scheme 3. This paper was aimed at investigating the acid catalyst effects on the crosslinking of model hydroxyl and carboxyl functional acrylic latex coatings with a cycloaliphatic diepoxide. Both copolymerizable and free sulphonic or phosphonic acids were used as the catalysts. The crosslinking of the coatings with the diepoxide was evaluated in

Scheme 1. Proposed crosslinking reaction of acrylic latex coatings with cycloaliphatic diepoxide

terms of water absorption, gel content, and pull-off adhesion. In addition, the pot-life stability of cycloaliphatic epoxide crosslinker was investigated using the crosslinker emulsion as a function of pH. The results are reported herein. 2. Experimental All materials for the synthesis of cycloaliphatic diepoxide crosslinkable acrylic latexes were used without further

Scheme 2. Crosslinking reaction of hydroxyl functional latex with cycloaliphatic diepoxide

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

Scheme 3. Crosslinking reaction of carboxyl functional latex with cycloaliphatic diepoxide.

puri®cation. Methyl methacrylate (MMA), butyl acrylate (BA), methacrylic acid (MAA), 2-hydroxyethyl methacrylate (HEMA), ammonium persulfate, p-toluene sulphonic acid (p-TsOH), phenyl phosphonic acid (PPA), sodium bicarbonate, and ammonia were purchased from Aldrich. Defoamer Byk-035 was provided by BYK Chemie. 2-sulfoethyl methacrylate (SEM) was provided by Hampshire Chemical Corp. Phosphoxyethyl methacrylate (PEM), cycloaliphatic diepoxide (UVR-6105), and surfactants (Triton-200 and Tergital-XJ) were supplied by Union Carbide. The latexes were synthesized according to a method as previously described [30]. The synthesis was performed in a four neck ¯uted round bottom ¯ask (500 ml) equipped with a mechanical stirrer and an overhead condenser. The stirring rate was kept constant at 200 rev./min throughout the latex synthesis. The temperature of the ¯ask was controlled using a thermostated water bath. The glass transition temperature of the latex polymers was designed to be 08C, and the contents of the carboxyl or the hydroxyl functional monomer were designed to be at 6.0 and 9.1 wt.% providing an equivalent functionality. According to Fox equation, the corresponding recipes for the synthesis of the hydroxyl and carboxyl functional latexes are listed in Table 1. According the recipe, the monomers were mixed, and then slowly added into another ¯uted round bottom ¯ask containing a stirring water solution (0.2 g NaHCO3, 7.96 g Triton-200, and 96 g water) to prepare a monomer pre-

23

emulsion. After the monomer-emulsi®cation, the reactor was charged with 77 g water, 14 g of the pre-emulsion, 8.1 g of the initiator solution (0.77 g (NH4)2S2O8 in 35 g water), 0.2 g NaHCO3, and 0.08 g Triton-200 at 808C. The contents were then heated at 808C for 30 min. The preemulsion was then fed into the reactor over 3.0 h using a addition pump (ISMATEC REGLO-100), and the initiator solution was concurrently added to the reactor at a constant rate using a universal syringe pump (Valley Scienti®c, model 575). The polymerization was maintained at 808C under nitrogen atmosphere. After the addition of all the ingredients, the reaction mixture was heated for an additional 2 h at 858C to digest any residual monomers. Then, the reactor contents were cooled to 30±358C, and neutralized with ammonia to a pH of 8.5. After ®ltration of the contents through a 300 mesh screen to remove any residual coagulum, the ®nal latex was characterized in terms of glass transition temperature, particle size, and distribution [30]. The cycloaliphatic diepoxide (UVR-6105) was emulsi®ed using a mixture of anionic and non-ionic surfactants (Triton200 and Tergital-XJ). The anionic surfactant (3.21 g) and non-ionic surfactant (1.35 g) were dissolved in deionized water (55.0 g) containing 0.1 g defoamer (Byk-035). Then, the cycloaliphatic diepoxide (UVR 6105, 45.0 g) was emulsi®ed into the water solution at 608C at a high speed (2000± 3000 rev./min) in 30 min. After the emulsi®cation, a cycloaliphatic epoxide emulsion crosslinker containing 45 wt.% solids was afforded. A stoichiometric amount of the epoxide emulsion was then added into the latexes, respectively. All the coating formulations with the crosslinker were used 60 min after complete mixing. Sulphonic and phosphonic acid catalysts were used in this study. Both the strong acid catalysts were introduced into the latex formulation as free acids in the water phase, or tethered directly to the latex particles in the form of a functional acrylate. For the tethered catalysts, 2-sulfoethyl methacrylate and phosphoxyethyl methacrylate were copolymerized with the other acrylic monomers (MMA, BA, MAA, and HEMA) during the latex synthesis. The addition of the acid monomers was conducted according to a method described previously [31]. The free acids in the water phase were toluene sulphonic acid and phenyl phosphonic acid. They were added into the latex coating formulations before casting. The amount of these acids into the latexes was varied systematically from 0.6 to 1.5 wt.% based on the solids of the latexes. To study the pH effect on the crosslinking caused

Table 1 Recipes for the synthesis of the latexesa Latex

MMA

BA

HEMA

MAA

Ammonium persulfate

Triton-200

NaHCO3

HOb COOH

72.74 71.56

90.94 97.64

16.33 0.0

0.0 10.80

0.81 0.81

8.04 8.04

0.4 0.4

a b

All the ingredient quantities are listed in grams. The content of the hydroxyl and the carboxyl functional monomers of the two latexes are equivalent in moles (0.1256 moles).

24

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

by the addition of the acid catalysts, all the acids were neutralized or not neutralized by ammonia, and then added into the coating formulations, respectively. All the latex coatings were cast on aluminum or glass substrates using a drawdown bar (5 mils). After coalescence for 2 h at room temperature, the latex coatings were baked in a controlled oven (Blue M Electronic B-3898Q) with accuracy of 18C. For the carboxyl latex coatings, the crosslinking was performed at 1308C for 105 min. For the hydroxyl latex coatings, the crosslinking was carried-out at 1708C for 105 min. The water absorption (ASTM D3419), gel content (ASTM D2765), and the pull-off adhesion (ASTM D4541) of the crosslinked coatings were measured 24 h after baking. To measure the water absorption of a crosslinked coating, 0.2±0.3 g coating sample was boiled in 50 ml deionized water (1008C) in a one neck round bottom ¯ask (100 ml) equipped with a condenser for 24 h. Afterwards, the wet sample without free water was weighed. The water absorption was then calculated from the difference in the weights of the sample before and after soaking in the water according to Eq. (1): Water absorption ˆ

wtafter ÿ wtbefore  100% wtbefore

(1)

where wtafter is the weight of the sample after boiling, and wtbefore is the weight of the sample before boiling. To measure the gel content of a coating, 0.1±0.2 g sample was added into 100 ml solvent mixture (30 wt.% butanol, 30 wt.% butyl acetate, 40 wt.% MEK) in a 250 ml one neck round bottom ¯ask equipped with a condenser, and boiled for 12 h. After cooling, the content in the ¯ask was ®ltered through a 200 mesh screen (Te¯on) to retain any undissolved gel. The un-crosslinked polymer would dissolve in the solvent, and the crosslinked polymer would become gel. The retained gel was dried at room temperature for 2 h, and then dried in an oven at 1108C for 2 h. The gel content was calculated from the difference in the weights of the sample before and after boiling in the solvent according to Eq. (2) Gel content ˆ

wtafter  100% wtbefore

(2)

where wtafter is the dry weight of the sample after boiling, and wtbefore is the weight of the sample before boiling. The pull-off adhesion test was performed using a pull-off adhesion tester (Elcometer 106, Elcometer Instruments) with coatings on aluminum substrates. The surface of the coating was gently roughened using sand paper (200 mesh). A uniform layer of mixed two-component epoxy adhesive (Araldite) was applied on the coating, and subsequently glued to the metal dolt with a pressure. The glued sample was cured at room temperature for 4 h, and then cured in an oven at 608C for 24 h. The hydrolysis of the cycloaliphatic epoxide was studied using 1H-NMR. The disappearance of the cycloaliphatic epoxide as a function of pH was followed

based on the relative integration values of the oxirane protons before and after oxirane ring opening by water. 3. Results The objective of this study was to evaluate acid catalyst effects on the crosslinking of hydroxyl and carboxyl functional acrylic latexes with a cycloaliphatic diepoxide and the properties of the crosslinked coatings. Model compound study showed that the reaction rates of hydroxyl and carboxyl compounds with cycloaliphatic epoxide increased with the acidity of the reaction medium [23±25]. However, the reaction rate for hydroxyl compounds was more dependent on the acidity than carboxyl compounds. In addition, a model acrylic latex study also showed that the cycloaliphatic diepoxide crosslinked hydroxyl and carboxyl functional acrylic latexes. The crosslinking reactions (Scheme 2 and Scheme 3) were monitored via differential scanning calorimeter (DSC), Fourier-transform infrared (FTIR), and dynamic mechanical thermal analysis (DMTA) [30]. Spectroscopic data showed that carboxyl latexes were more reactive than hydroxyl latexes with cycloaliphatic epoxide. Model acrylic latexes were used to study acid catalyst effects on the crosslinking with cycloaliphatic diepoxide. A model hydroxyl functional acrylic latex was prepared, formulated, and crosslinked with cycloaliphatic diepoxide as a function of acid catalysts. Similarly, a model carboxyl functional acrylic latex was prepared, formulated, and crosslinked with cycloaliphatic diepoxide as a function of the acid catalysts. Strong acid catalysis was introduced for the crosslinking of both the hydroxyl and carboxyl model latexes with the cycloaliphatic diepoxide. Two types of strong acids, sulphonic and phosphonic, were investigated in accordance with previously reported model compound studies [32]. These strong acids were added into the latex formulations via tethering the strong acid catalysts directly to the latex particles. A methacrylate of sulphonic or phosphonic acid was copolymerized with latex monomers effectively bonding the strong acid catalyst to the latex particle as shown in Scheme 4. For comparison purposes, complimentary free strong acids, p-toluene sulphonic acid (p-TsOH) and phenyl phosphonic acid (PPA), were also used with the latex formulations which did not include tethered catalysts. The free acids were in the water phase as shown in Scheme 5. The effects of catalysts on the latex crosslinking were studied as a function of concentration, neutralization, and chemical crosslink bond formation (ester or ether). The effects of the acid catalysts on the crosslinking reactions were evaluated with respect to bulk coating properties of water absorption, gel content, and pull-off adhesion. The crosslinking bond formation for both the carboxyl and hydroxyl functionalized latexes with cycloaliphatic epoxide was previously reported. The formation of the ester and ether crosslinks were established using IR, 1H NMR, 13C NMR, and mass spectroscopy. The crosslinking temperature for the

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

25

Scheme 4. Carboxyl and hydroxyl functional acrylic latex particles with PEM and SEM acrylic acid monomers (the surfactant is not schematically included).

carboxyl and hydroxyl functionalized latexes were also determined using IR, DSC, and DMTA. In addition to the bulk ®lm properties, the hydrolytic stability of cycloaliphatic epoxide was investigated via 1H NMR. The hydrolytic stability of the cycloaliphatic diepoxide has great importance with respect to the pot-life of the ®nal latex coating formulations. The water absorption, gel content, and pull-off adhesion of the crosslinked hydroxyl functional latex coatings with

different catalysts were shown in Table 2. In comparison with the coatings with an acid catalyst, the coating crosslinked without an acid catalyst showed higher water absorption and lower gel content. The pull-off adhesion of the crosslinked coating without an acid catalyst, however, was higher than the coatings crosslinked with an acid catalyst. With an acid catalyst, the crosslinked coatings exhibited a signi®cant decrease in the water absorption and an increase in the gel content. In comparison, the coatings with a

Scheme 5. Carboxyl and hydroxyl functional acrylic latex particles with free PPA and p-TsOH acids (the surfactant is not schematically included).

26

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

Table 2 Catalyst effect on the crosslinking of hydroxyl (9.1 wt.%) latex coatingsa Catalyst

Amount (wt.%)

pH

Water uptake, (wt.%)

Gel content, (wt.%)

Adhesion, (lbs/inch2)

± SEM PEM TsOH PPA NH3 SEM ‡ NH3 PEM ‡ NH3 TsOH ‡ NH3 PPA ‡ NH3

± 0.60 0.65 0.53 0.49 ± 0.65 0.70 0.58 0.59

4.5 2.5 3.2 2.6 3.0 8.5 8.2 8.3 8.1 8.3

102.2 51.4 67.8 58.3 70.8 77.6 37.9 44.7 38.0 42.9

88.1 91.3 89.9 90.0 89.0 89.1 93.5 92.1 92.6 92.8

360 340 360 200 250 370 380 440 350 360

a

Cured at 1308C for 105 min.

copolymerizable acid (SEM or PEM) showed much higher adhesion than the coatings with the free acids (TsOH or PPA). However, both the types of the crosslinked coatings catalyzed with copolymerizable and free acids did not show much differences in the water absorption and gel content. In contrast, the crosslinked coatings with the sulphonic acids (SEM and TsOH) appeared to exhibit lower water absorption and higher gel content than phosphonic acids (PEM and PPA). However, the coatings with phosphonic acids exhibited the highest adhesion. Ammonia neutralized acids (Table 2) were also used as the catalysts to compare with the un-neutralized acids. Surprisingly, the coatings which were crosslinked with neutralized acids showed signi®cant improvement in the crosslinking of the latexes with the cycloaliphatic diepoxide. All the coatings exhibited signi®cantly lower water absorption, higher gel content, and higher pull-off adhesion than the coatings crosslinked with the un-neutralized acids. In addition, the crosslinked coatings catalyzed with neutralized sulphonic acids also showed lower water absorption and greater gel content than the coatings with neutralized phosphonic acids. In comparison with the sulphonic based catalysts, the coatings crosslinked with the phosphonic acids (PEM and PPA), especially with neutralized PEM, exhibited the highest adhesion. The crosslinking of the carboxyl functional acrylic latex coatings with cycloaliphatic diepoxide as a function of the

catalysts was also studied. Table 3 shows the water absorption, gel content, and pull-off adhesion of the crosslinked carboxyl functional latex coatings with the acid catalysts. In contrast with the hydroxyl latex coatings, without addition of an acid catalyst, the crosslinked coating showed lower water absorption, higher gel content, and lower adhesion than the coatings crosslinked with an acid catalyst. Similarly, after neutralization of the acid catalysts with ammonia, the crosslinked carboxyl latex coatings displayed signi®cantly lower water absorption, higher gel content, and higher adhesion than the coatings with non-neutralized acids. The neutralized carboxyl functional latex coating without addition of an external acid catalyst showed the lowest water absorption, highest gel content, and pull-off adhesion. Cycloaliphatic epoxides are more reactive in acidic conditions [23±25]. The hydrolytic stability of the cycloaliphatic diepoxide in water was affected by the acidity. Thus, the stability of the cycloaliphatic diepoxide against hydrolysis was studied using 1H-NMR as a function of the pH. Fig. 1 shows the 1H NMR spectrum of the cycloaliphatic diepoxide. The chemical shifts of proton a (Ha) of the cycloaliphatic diepoxide were between d3.047 and 3.131 ppm. The 1 H NMR spectrum of the hydrolyzed cycloaliphatic diepoxide is shown in Fig. 2. After hydrolysis, the chemical shifts of the original proton a (Ha) moved down®eld (d3.37± 3.63 ppm). Thus, the relative concentration of the cycloaliphatic diepoxide versus the hydrolyzed epoxide in the

Table 3 Catalyst effects on the crosslinking of carboxyl (6.0 wt.%) latex coatingsa Catalyst

Amount (wt.%)

pH

Water uptake (wt.%)

Gel Content (wt.%)

Adhesion (lbs/inch2)

COOH SEM PEM TsOH PPA COOH ‡ NH3 SEM ‡ NH3 PEM ‡ NH3 TsOH ‡ NH3 PPA ‡ NH3

6.0 0.60 0.65 0.53 0.49 ± 0.65 0.70 0.58 0.59

4.5 2.5 3.0 2.6 2.8 8.5 8.2 8.3 8.2 8.4

41.8 68.4 71.4 65.7 66.9 32.6 47.6 52.6 44.9 48.8

89.0 71.3 70.2 75.4 73.5 92.8 91.6 89.8 91.9 90.6

280 300 350 280 300 420 390 420 380 380

a

Cured at 1308C for 105 min.

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

Fig. 1. 1H-NMR spectrum of cycloaliphatic epoxide before hydrolysis.

Fig. 2. 1H-NMR spectrum of cycloaliphatic epoxide after hydrolysis.

27

28

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33 Table 5 Functionality effect on the crosslinking of phosphonic acrylic latex coatingsa Crosslinker

Functionality

pH

Water uptake (wt.%)

Gel content (wt.%)

No No No Yes Yes Yes

None Hydroxyl Carboxyl None Hydroxyl Carboxyl

8.4 8.2 8.4 8.3 8.3 8.3

170.7 210.1 266.1 164.5 44.7 52.6

7.3 5.7 5.9 8.9 92.1 89.8

a

Fig. 3. Disappearance of the cycloaliphatic diepoxide as a function of the pH and time.

emulsion system was calculated based on the integration values of protons Ha and Hc. The disappearance of the cycloaliphatic diepoxide in the emulsion as a function of the pH and time is shown in Fig. 3. The disappearance rate of the cycloaliphatic diepoxide increased as the pH of the emulsion decreased. To further highlight the effect of the hydroxyl or carboxyl functionality on the crosslinking of the latexes with the cycloaliphatic diepoxide, acrylic latexes without any hydroxyl or carboxyl functionality were synthesized with sulphonic or phosphonic acrylic monomer, and the resulting latexes were crosslinked with the cycloaliphatic diepoxide. Table 4 shows the water absorption and gel content of SEM functional acrylic latex coatings as a function of the crosslinker and the latex functionality (hydroxyl and carboxyl). All the latex coatings which were not crosslinked showed substantially higher water absorption and low gel content. Similarly, all the coatings of the latexes without any hydroxyl or carboxyl functionality, also showed extremely high water Table 4 Functionality effect on the crosslinking of sulphonic acrylic latex coatings cured at 1308C for 105 mina Crosslinker

Functionality

pH

Water uptake (wt.%)

Gel content (wt.%)

No No No Yes Yes Yes

None Hydroxyl Carboxyl None Hydroxyl Carboxyl

8.2 8.1 8.2 8.3 8.2 8.2

147.4 244.7 260.8 142.8 37.9 47.6

5.1 21.1 5.0 6.8 93.5 91.6

a All latexes contained 0.65 wt.% ammonia neutralized SEM as the catalyst.

All latexes contained 0.7 wt.% ammonia neutralized PEM as the catalyst

absorption and low gel content. Only the coatings of the latexes with hydroxyl or carboxyl functionality and crosslinked with the cycloaliphatic diepoxide exhibited the lowest water absorption and the highest gel content. The water absorption and gel content of the PEM functional acrylic latex coatings as a function of the crosslinker and functionality are listed in Table 5. Similar to the sulphonic acid results, all the latexes without any hydroxyl or carboxyl functionality showed substantially higher water absorptions and lower gel contents. All the functionalized coatings which were not crosslinked with the cycloaliphatic diepoxide also displayed high water absorptions and low gel contents. As before, only the coatings crosslinked with presence of both the crosslinker and the latex functionality (hydroxyl or carboxyl) showed the lowest water absorption and the highest gel content. The water absorption, gel content, and pull-off adhesion of the carboxyl functional acrylic latexes as a function of the curing time and catalysts are shown in Table 6. As expected, the water absorption of all the coatings systematically decreased with increasing curing time. As an overall trend, the gel content concomitantly increased. The pull-off adhesion also increased with the curing time, and appeared to reach maximum (410 lbs/inch2) at 105 min. Overall, the gel content and water absorption data was better without the strong acid catalysts than with the catalysts especially after 60 min of the cure time. However, the gel content and pulloff adhesion of the crosslinked coatings with an acid catalyst appeared to be higher than the crosslinked coating without an acid catalyst when the baking time was 30 min. The crosslinking of both carboxyl and hydroxyl functional latex coatings with the diepoxide was also evaluated as a function of the catalyst content. The water absorption of the hydroxyl and carboxyl latex coatings as a function of the SEM is shown in Fig. 4. For the hydroxyl functional latex coating, the water absorption decreased dramatically with addition of the catalyst reaching a minimum at 0.9 wt.% (SEM). Beyond that concentration (0.9 wt.%), the water absorption appeared to increase. In contrast, the water absorption of the carboxyl latex coating monotonously increased with increasing SEM. The gel content of the latex coatings as a function of the SEM is shown in Fig. 5. The gel content of the hydroxyl

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

29

Table 6 Catalyst effects on the crosslinking of carboxyl (6.0%) latex coatings Catalyst

Time (min)

Amount (wt.%)

Water uptake (wt.%)

Gel content (wt.%)

Adhesion (lbs/inch2)

COOH ‡ NH3 SEM ‡ NH3 TsOH ‡ NH3 COOH ‡ NH3 SEM ‡ NH3 TsOH ‡ NH3 COOH ‡ NH3 SEM ‡ NH3 TsOH ‡ NH3 COOH ‡ NH3 SEM ‡ NH3 TsOH ‡ NH3

30 30 30 60 60 60 105 105 105 165 165 165

± 0.65 0.58 ± 0.65 0.58 ± 0.65 0.58 ± 0.65 0.58

80.6 85.7 83.6 56.3 69.8 69.0 32.6 47.6 44.9 25.4 41.2 38.6

77.8 80.3 78.9 89.3 85.3 83.5 92.8 91.6 91.9 93.4 92.7 92.9

160 240 200 360 350 320 410 390 380 370 380 360

functional latex coating increased with increasing SEM, and reached a maximum value at 0.9 wt.% (SEM). When the addition of the SEM was greater than 0.9 wt.%, the gel content of the hydroxyl latex coating appeared to decrease. In contrast, the gel content of the carboxyl functional latex coating linearly decreased with increasing SEM. The pulloff adhesion of both the functional (hydroxyl and carboxyl) latexes as a function of SEM is depicted in Fig. 6. The pulloff adhesion of both the latex coatings appeared initially to increase with the addition of the SEM, however the results were within experimental error. Phase separation and relative viscosity were used to evaluate the latex stability increase of the latexes after freeze±thaw cycles. Table 7 show the freeze±thaw cycles without phase separation and the relative viscosity of the

latexes as a function of the acid monomer and functionality. For both the hydroxyl and carboxyl functional latexes without the tethered strong acid catalyst (SEM or PEM), the phase separation of the latexes was observed at one or two freeze±thaw cycles. The viscosity of the latexes was also observed to increase dramatically after the cycles. In contrast, all the latexes with the copolymerizable acid monomers (SEM or PEM) showed greater freeze±thaw cycles without the phase separation. Unlike the non-tethered latexes, the viscosity of these latexes increased slowly after cycling. In comparison with the phosphonic acid monomer (PEM), all the latexes with the sulphonic acid monomer (SEM) exhibited greater freeze±thaw cycles without phase separation. The SEM based latexes also exhibited slower increases in the viscosity after each cycle. In addition, the

Fig. 4. Water absorption of the latex coatings as a function of the SEM content.

Fig. 5. Gel content of the latex coatings as a function of the SEM content.

30

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

Fig. 6. Pull-off adhesion of the latex coatings as a function of the SEM content.

hydroxyl functional latexes underwent more freeze±thaw cycles without phase separation than the carboxyl functional latexes. 4. Discussion The crosslinking of hydroxyl and carboxyl functional acrylic latexes with cycloaliphatic diepoxide was investigated using spectroscopic methods, and evaluated using water absorption, gel content, and pull-off adhesion of the crosslinked coatings. The water absorption, gel content, and pull-off adhesion were very sensitive to changes in the polarity and molecular structures. By crosslinking, the polar pendant hydroxyl and carboxyl groups of the latexes were converted into less polar ether and ester bonds via the Table 7 SEM effect on the freeze-thaw stability of the latexesa Functionality

Catalyst

FTSb

1cc

2c

3c

4c

5c

Hydroxyl Hydroxyl Hydroxyl Carboxyl Carboxyl Carboxyl

None SEM (0.9%) PEM (0.98%) None SEM (0.9%) PEM (0.98%)

2 >8 6 1 4 3

6 0 0 107 0 4

12 0 0 276 7 13

18 0 5 386 18 27

18 0 9 445 27 53

20 0 11 472 31 97

a

The contents of the hydroxyl and carboxyl monomers were 9.0 and 13.6 wt.%. b FTC means the freeze-thaw cycles without visional observed phase separation. c 1c means the relative viscosity increase after one freeze-thaw cycle.

crosslinking (Scheme 2 and Scheme 3). Consequently, the polarity of the coating system was reduced. The degree of the reduction in the polarity was dependent on the extent of the crosslinking or crosslink density. Because water absorption of a polymeric material is a result of association and interaction of the material with water, higher crosslinking and lower polarity of a coating system will give lower water absorption for the coating system. The gel is the portion of a material that did not dissolve in a solvent. Given a proper solvent, the content of the gel is primarily determined by the molecular weight and the geometry of the molecules. After crosslinking, the molecules of the coatings were crosslinked into a three dimensional network. The molecular weight of the coatings was increased, and the entropy of the molecules was reduced. Consequently, the material dissolved less in the solvent and resulted in a high gel content. Thus, the gel content is proportional to the crosslink density. Generally, epoxy ring opening crosslinking reactions also result in signi®cant improvement on the adhesion. Consequently, the pull-off adhesion was related with the crosslinking, and increased by the crosslinking. Therefore, water absorption, gel content, and pull-off adhesion were used as primary means in this paper to probe the catalyst effects of the crosslinking reactions. Although the hydroxyl latex coating crosslinked without addition of an acid catalyst showed the highest water absorption and the lowest gel content (Table 2), the data still suggested some crosslinking within the latex coating when compared with the coating without crosslinking [30]. The crosslinking of the coating might be catalyzed by any residual acid generated during the latex synthesis [33]. The high water absorption and low gel content of the coating without an acid catalyst could be attributed to the low ef®ciency of the crosslinking. Previously, it was shown that the pull-off adhesion of a coating was proportional to the ¯exibility of the coating [34±36]. The higher pull-off adhesion of the coating crosslinked without an acid catalyst may be attributed to un-crosslinked cycloaliphatic diepoxide which functioned as a plasticizer increasing the ¯exibility of the resultant coating. With an acid catalyst, the crosslinked coatings exhibited a signi®cant decrease in the water absorption and an increase in the gel content. This was due to improved crosslinking by the addition of the acids. In comparison, the coatings with a copolymerizable acid (SEM or PEM) showed much higher adhesion than the coatings with the free acids (TsOH or PPA). The copolymerizable acids were attached onto the acrylic polymer chains. Consequently, the attached acid groups may provide more intimate interaction between the polymer and the substrate increasing the interfacial adhesion of the coating with the substrate [37]. The water absorption and gel content of both the coatings using the copolymerizable and free acids were comparable. This suggests that a similar catalytic activity (SEM verse TsOH and PEM verse PPA) was principally due to the acidity (pH)

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

[38]. In contrast, the crosslinked coatings with the sulphonic acids appeared to show lower water absorption and higher gel content than with phosphonic acid based catalysts probably due to the stronger acidity (pH) [32]. The coatings with phosphonic acids showed the highest adhesion. This could be attributed to the fact that phosphonic acid groups are more effective as coupling groups with metal substrates which promote adhesion [37]. Surprisingly, when the acids were neutralized with ammonia, all the coatings showed lower water absorption, higher gel content, and higher pull-off adhesion. These results suggest that neutralization of the acids improved the crosslinking of the latexes with the cycloaliphatic diepoxide. Clearly, the ammonia neutralization of the acids did not cause any inhibition to the latex crosslinking with the diepoxide. First, perhaps the ammonia neutralized acids decomposed at the crosslinking temperature regenerating the original acids which were then available to catalyze the crosslinking reactions. Secondly, when the un-neutralized acids were used, the coating formulations were acidic (lower pH). However, when the neutralized acids were used, the coating formulations became basic. Thus, it was assumed that the acidic conditions may cause hydrolysis of the cycloaliphatic diepoxide, whereas the basic conditions (higher pH) may increase the stability of the epoxy crosslinker in the coating formulation. Consequently, the greater stability of the crosslinker may contribute to the overall improved crosslinking of the latex coatings with the cycloaliphatic epoxide. In addition, the crosslinked coatings catalyzed with neutralized sulphonic acids also showed lower water absorption and greater gel content than neutralized phosphonic acids. This may also be attributed to the regeneration of the acids during the crosslinking. Although a weaker acid, the ammonia salts of sulphonic acids may dissociate more readily than the phosphonic acid salts [23±25]. However, the crosslinked coatings with the neutralized phosphonic acids, especially the neutralized PEM, exhibited the highest adhesion due to improved coupling effect of PEM with the substrates [38]. This result is consistent with the coatings crosslinked with the non-neutralized phosphonic acids (PPA and PEM). Thus, acids accelerated the crosslinking of the hydroxyl functional latexes with the cycloaliphatic diepoxide, and neutralization of the acids also appeared to further improve the ®lm properties. The crosslinked carboxyl latex coatings with neutralized acids exhibited lower water absorption, higher gel content, and adhesion than the coatings without the neutralization. Thus, neutralization of the latex and the catalyst also improved the crosslinking for the carboxyl functional latex coatings with the cycloaliphatic diepoxide. More interestingly, the neutralized carboxyl functional latex coating without using an external acid catalyst displayed the lowest water absorption, highest gel content and pull-off adhesion. Hence, the crosslinking of the carboxyl functional latexes with the cycloaliphatic diepoxide did not necessarily need

31

strong acid catalyst. From the data, and adhesion of the coatings with an neutralized acid, it was evident that acid catalysts did not effectively improve the crosslinking of the carboxyl functional latexes with the cycloaliphatic diepoxide, unlike the hydroxyl latex system. The addition of an acid catalyst in the carboxyl latex system resulted in enhancement of the water and solvent sensitivity of the crosslinked coatings. Previously, it was assumed that the stability of the cycloaliphatic diepoxide in¯uenced the crosslinking with functional groups (hydroxyl and carboxyl) and thus the overall coating properties. The disappearance rate of the cycloaliphatic diepoxide as a function of pH and time (Fig. 3) showed that acidic conditions (low pH) caused faster hydrolysis of the diepoxide crosslinker, and basic conditions (high pH) slowed down the hydrolysis increasing the stability against the hydrolysis. Therefore, the hydrolysis of the cycloaliphatic diepoxide crosslinker contributed to the higher water absorption and lower gel content, and adhesion for those coatings with non-neutralized acids. When the baking time was increased, there was more time available for the crosslinking of the latexes with the diepoxide. As a consequence, the water absorption of all the coatings decreased, and the gel content as well as the adhesion increased with the baking time (Table 6). In comparison, the higher water absorption, lower gel content, and adhesion of the coatings with an acid catalyst may be due to enhanced water and solvent sensitivities by the acid catalyst especially for those coatings which were crosslinked for more than 60 min. These results further suggest that the crosslinking of the carboxyl latexes with a cycloaliphatic diepoxide may not need an acid catalyst in terms of the data discussed. However, the higher gel content and pull-off adhesion of the crosslinked coating with an acid catalyst (neutralized SEM or TsOH) at 30 min seem to suggest that the acids accelerated the crosslinking. The higher water absorption, however, suggests that the acids did not necessarily improve the crosslinking at the condition. Thus, further study will be required to investigate these seemingly contradictory results. Upon freezing of a latex, the water in the latex starts icing excluding the latex particles from the water phase [39]. The latex particles become more and more concentrated. To a point at which there is no longer enough water to keep the latex particles being a continuous dispersed phase, and the repulsion force among particles become less than the force exerted by the icing expansion. Then, coagulation or coalescence of the latex particles occurs. Consequently, the latex become unstable after freezing, and may not reversibly restore to the original dispersed phase upon thawing. The repulsion force is one of the important factors in¯uencing the freeze±thaw stability of a latex coating. All the latexes with the copolymerizable acid monomers exhibited greater freeze±thaw cycles without the phase separation (Table 7). The tethered strong acid groups exerted higher electrical charge density than carboxylic acid groups [40], and

32

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33

increased the electrical repulsion among the latex particles. As a consequence, the greater freeze±thaw stability of the latexes with the tethered strong acid monomers was attributed to a higher electrical charge density on the latex particles. In comparison, the hydroxyl latexes were able to resist greater freeze±thaw cycles without the phase separation, and the viscosity increased slower than the carboxyl functional latexes. Although the electrical charges on the latex particles provided by the carboxylic groups could improve the stability against the particle coagulation, the hydration layer generated by the carboxylic groups on the latex particles increased the effective volume fraction of the latex particles in the systems. Consequently, the carboxyl functional latexes tended to be more easily destabilized with respect to the freeze±thaw cycles. It was evident that the crosslinking of the hydroxyl functional acrylic latexes with cycloaliphatic diepoxide required an acid catalyst. Although all the acids functioned as the catalysts for the crosslinking, the copolymerizable sulphonic acid (SEM) more effectively catalyzed the crosslinking reaction. In addition, the latexes with SEM also showed the strongest stability against freeze±thaw cycles. The adhesion of the coatings with SEM also was comparable to the PEM-based coatings. The crosslinking of the carboxyl functional latexes with the cycloaliphatic diepoxide did not required the addition of an acid catalyst. However, the carboxyl functional latexes with copolymerizable acid monomers, especially with SEM, showed greater stability against freeze±thaw cycles. Of the four acid catalysts, SEM provided the best balance of the overall coating properties for both the functional latex systems. For the hydroxyl functional latex coatings, the dramatic decrease in the water absorption, and signi®cant increases in the gel content and adhesion were evident when SEM was used as the catalyst. For the carboxyl functional latex coatings, a small addition of SEM improved the coating adhesion and freeze±thaw stability. The free acid, TsOH, catalyzed the crosslinking reactions of the latexes with the cycloaliphatic diepoxide. However, the free acid was unable to enhance the latex stability against freeze±thaw cycles. The pH of the coating system was an important factor affecting the crosslinking and coating properties. Neutralization of the acid catalysts with ammonia resulted in improved overall coating properties. 5. Conclusion Acid catalysts effected the ®lm properties of the carboxyl and hydroxyl functional latexes crosslinked with the cycloaliphatic diepoxide. An acid catalyst was essential to the crosslinking of hydroxyl functional latexes with the cycloaliphatic diepoxide. Without an acid catalyst, the coating properties were poor. However, the crosslinking of the carboxyl functional latex coatings with the cycloaliphatic

epoxide did not necessarily require an acid catalyst. The addition of an acid catalyst increased the water and solvent sensitivities. Sulphonic and phosphonic acrylic monomers copolymerized onto acrylic latex polymers, functioned as acid catalysts, and effectively catalyzed the crosslinking. The incorporation of the sulphonic and phosphonic acid monomers improved the adhesion and freeze±thaw stability of the latex coatings. Acidic conditions resulted in hydrolysis of the epoxy crosslinker decreasing the overall coating properties. However, basic conditions improved the stability against the epoxy hydrolysis, and consequently enhanced the overall coating properties. Acknowledgements The authors would like to thank the NSF-Coatings Research Center for the funding of this project and the Summer Undergraduate Research Program of the Polymers and Coatings Department at North Dakota State University.

References [1] B. Richey, T. Wood, Surf. Coat. Int. 1 (1994) 26. [2] Z.W. Wicks, F.N. Jones, S.P. Pappas, Organic Coatings Science and Technology, Wiley, New York, 1992, Chapter 3. [3] G. Bufkin, J.R. Grawe, J. Coat. Tech. 50(641) (1978) 41. [4] G. Pollano, Polym. Preprints (ACS) 38(1) (1997) 383. [5] T. Huang, F.N. Jones, Prog. Org. Coat. 28 (1996) 133. [6] Y. Inaba, E.S. Daniels, M.S. El-Aasser, J. Coat. Tech. 66(832) (1994) 63. [7] C.R. Hegedus, K.A. Kloiber, J. Coat. Technol. 68(860) (1996) 39. [8] J.L. Gardon, Polym. Mater. Sci. Eng. 76 (1997) 178. [9] W. Brown, Surf. Coat. Int. 6 (1995) 238. [10] A. Williams, A. Chem. Rev. 81 (1981) 589. [11] P. Moles, Polym. Paint Color J. 181(267) (1991) 4279. [12] G. Pollano, Polym. Mater. Sci. Eng. 77 (1997) 383. [13] G. Moor, D-W. Zhu, G. Clark, Surf. Coat. Int. 9 (1995) 376. [14] Nippon Shokubai Products Information Bulletin, Novel Waterborne Crosslinkers. [15] Wu, S., Soucek, M.D., J. Polym. Sci. A, submitted. [16] R.F. Eaton, D. Goldberg, B.D. Hanrahan, W.P. Mayer, D.F. Smith, T.C. Soon, Surf. Coat. Australia 29(5) (1992) 6. [17] Eaton, R.F., Hanrahan, B.D., Braddock, J.K., UV/EB Proc. RadTech Int. North America, NorthBrook, IL, 1990, p. 384. [18] D. Goldberg, R.F. Eaton, Mod. Paint Coat. 82 (1992) 36. [19] R.F. Eaton, K.T. Lamb, J. Coat. Technol. 68(860) (1996) 49. [20] R.F. Eaton, Polymer Preprints 38(2) (1997) 381. [21] H. Batzer, E. Nikles, Chimia (Aarau) 16 (1962) 57. [22] S. Sakai, T. Sugiyama, Y. Ishii, Kogyo Kazaku Zasshi 66 (1963) 355. [23] Wu, S., Soucek, M.D., Polymer (1998) in press. [24] S. Wu, M.D. Soucek, Polymer 39(15) (1998) 3583. [25] S. Wu, M.D. Soucek, J. Coat. Tech. 68(869) (1997) 43. [26] S. Wu, M.D. Soucek, Polymer Preprints (ACS) 38(1) (1997) 93. [27] H. Lee, K. Neville, Handbook of Epoxy Resins, McGraw-Hill, New York, 1967, Chapter 9. [28] B. Ellis, Chemistry and Technology of Epoxy Resins, Blackie Academic and Professional: New York, 1994, Chapter 2. [29] C.A. May, Y. Tanaka, Epoxy Resins Chemistry and Technology, Marcel Dekker: New York, 1973, Chapter 4. [30] Wu, S., Soucek, M.D., Polymer, submitted.

S. Wu et al. / Progress in Organic Coatings 36 (1999) 21±33 [31] T.L. Mill, R.H. Yocum, J. Paint. Technol. 39(512) (1967) 532. [32] M.D. Soucek, O.L. Abu-Shanab, C.D. Anderson, S. Wu, Macromol. Chem. Phys. 199 (1998) 1035. [33] H. Warson, The Application of Synthetic Resin Emulsions, Ernest Benn, London, 1972, Chapters 1±2. [34] Wu, S., Soucek, M.D., J. Coat. Technol. (1998) in press. [35] S. Wu, M.D. Soucek, Polymer Preprint 39(1) (1998) 541. [36] S. Wu, M.D. Soucek, Polymer Preprints 38(1) (1997) 492.

33

[37] Z.W. Wicks, F.N. Jones, S.P. Pappas, Organic Coatings Science and Technology, Vol. 2, Wiley, New York, 1992, Chapter 26. [38] J. March, Advanced Organic Chemistry, Wiley, New York, 1992, Chapter 8. [39] H. Warson, The Application of Synthetic Resin Emulsions, Ernest Benn, London, 1972, Chapter 3. [40] A.W. Adamson, A. Gast, Physical Chemistry of Surfaces, Wiley, New York, 1997, Chapter 5.