Importance of Alkyds
Ever since alkyd resins were first introduced some thirty-five years ago, they have enjoyed a consistent annual growth, with current production now running well over one-half billion pounds. Today alkyds outrank all other synthetic coating resins in importance, accounting for approximately half of all resins used by the paint industry, which approaches a size of two billion dollars annually in the United States.
The alkyd reaction is concerned to be the most versatile resin-forming reaction known. No other resin lends itself to greater internal variation or to more useful modification by physical or chemical blending with other polymers. Polymers commonly used to modify alkyd resins are listed in Table 1.
The purpose of this book is to elucidate the theories and calculations pertaining to this outstanding resin.
Chemical Nature of an Alkyd Resin
Alkyd resins, or alkyds, are tough resinous products formed by reacting polybasic organic acids with polyhydric alcohols. Broadly speaking, this type of esterification reaction produces compounds of the general class of polyesters. The key feature that distinguishes alkyds from other polyesters is the presence of monoacid (commonly fatty acid) as a major part of its composition. Typical alkyd structures are shown in Fig. 1.
Theoretically, any polyacid or polyol should lend itself to the manufacture of alkyds. However, from the standpoints of processing, paint performance, and pricing, only a relatively few raw materials (see listing in Table 2) have found commercial acceptance for trade sales and industrial applications.
An alkyd is classed as a polymer (a huge molecule) formed by chemical synthesis from many smaller molecules. The process whereby small unattached molecules are joined together by chemical reaction to form a tight network of interconnected molecules is called polymerization. Since the alkyd reaction usually releases a simple by-proudct molecule (commonly water) during the molecular tie-up, the process can be thought of as a compacting or condensing action. For this reason, the preparation of an alkyd is referred to as a condensation reaction and the end alkyd product as a condensation polymer.
To effect a chemical junction between two molecules, it is necessary that they be mutually chemically reactive. Tabulated in Fig. 1 are typical chemical reactions that can take place between two molecular species to effect a bonding or chemical linkage between them. The reaction between a carboxyl group (-COOH) and an hydroxyl group (-OH) is termed esterification. This reaction is basic to alkyd preparation.
If there is only one reactive site on each molecular species, that is, if they are both monofunctional, it is apparent that polymerisation cannot occur, for even at 100 per cent reaction, no more than double molecules can ever be formed. Let A and B be reactive sites on two monofunctional molecular species:
In fact, if either species of molecule has but one reaction site (the other may have two or more sites), true polymerization again fails to occur, for the most that can be expected is a saturation of the polyfunctional molecule with attachments of the monofunctional molecule and this is the maximum size of built-up molecule that can ever form under these conditions.
It is true that the resulting polyfunctional molecule may be quite large, but it can never progress to the stage of infinite interconnections which is the hallmark of a true polymeric structures.
However, if neither of the two reacting molecular species is mono-functional (they may be di-, tri-, tetra-, or of still higher polyfunctionalities), then infinite networks of tied-together molecules become possible, giving polymers of huge dimensions. This concept of functionality underlies the design of alkyd polymers suitable for paint vehicles.
At the very outset, it must be pointed out that the alkyd chemist is constantly faced with a dilemma in formulating alkyd resins. For an alkyd to be outstanding in performance, it must be processed to as high a molecular weight as possible. At the same time, the molecular weight must not be allowed to become excessive, or the alkyd vehicle will get out of control during processing (convert to an intractable gel) or exhibit an instability on shelf storage. Hence the alkyd formulator must at all times avoid the design of either unduly small or unduly large polymers. This is quite a trick. In general, an alkyd is formulated to a point just short of gelation at 100 per cent reaction. This criterion for proper alkyd design is used throughout this textbook.
Linear Versus Branch Molecules
A linear molecule is defined here as a difunctional molecule (two reactive sites). A branch molecule is defined as a tri-, tetra-, or higher functionality molecule (three, four, or more reactive sites).
Linear molecules react to form molecular chains of infinite extent; ring formation can also occur. Branch molecules also react to form molecular chains but, in addition, branching connections are formed which tie the chains together to give a three-dimensional polymeric network (Fig. 2). This is referred to as cross-linking.
Polymers of the linear type, with no branching (no cross-linking), are normally thermoplastic, which means that they are fusible and can be forced to take a new shape under heat and pressure-they are moldable. Polymers of the cross-linked type are normally thermosetting, which means that they are infusible and cannot be forced to take a new shape under heat and pressure-they are not remoldable. The rigidity and resistance to heat distortion that characterize a cross-linked polymeric structure can be understood by visualizing the many points of tie-in or cross-linking which restrict the movement of branched molecular chains, giving them a reinforced structure. However, the development of excessive cross-linking at the expense of chain extension leads to brittleness.
Tabulation of Alkyd Calculations
Alkyd calculations are greatly facilitated and better understood by using a tabular form for setting up and recording the data. For example, alkyd raw materials can be listed in a left-hand column, with headings for proportions and constants of these materials in a row across the top.
In solving a given problem, the initial or given data are first entered in the table at appropriate points. Then as computed values are obtained, they are filled in. This manner of recording the calculations is straightforward, information is posted in a systematic fashion, and a visual reference system is always available for checking. This recording technique is amply illustrated by examples throughout the text.
The Basic Chemistry of Unsaturated Polyesters
Unsaturated polyesters are the product of a condensation reaction between difunctional acids and alcohols one of which (generally the acid) contributes olefinic unsaturation. This polymer is dissolved in styrene or other monomeric material containing vinyl unsaturation. With heat and/or free radical initiation, the polyester and reactive diluent crosslink into a solid, non-melting network. The photos show a simplified model of the basic materials required to form a thermoset polyester.
With this picture in mind, the methods of varying the polyester to tailor it to specific application requirements are readily apparent. The principal polyester variations effected by constituent changes are the frequency of cross linking sites (crosslink density), the degree of steric protection afforded to vulnerable functional groups and the rotational freedom within the polymer chains.
Additional effects on properties are produced by polyester molecular size - a joint function of the ratio of ingredients and processing variations - and by coreactant selection and concentration.
Even this brief overview of polyesters indicates that several routes are usually available to achieve a particular property. A goal of this brochure is to suggest useful paths of investigation to resin formulators seeking cost effective resins to meet specific and use requirements. While the particular emphasis will be on the role of ingredients in polyesters based on isophthalic acid, many of the trends observed in our laboratory can be extrapolated to other types of unsaturated polyesters.
Processing Affects Polyester Properties
The brief description of processing included here is intended only as a basis for the discussion of processing's influence on properties. A more complete review of Amoco's recommendations for unsaturated isopolyester processing is available in other publications.
Fig. 1. Molecular Weight Increase Occurs As End Groups Disappear During Esterification
Equipment for processing good quality unsaturated polyesters includes: a heated kettle with agitator, temperature measurement devices, reactant addition and sampling ports; an overhead system with efficient fractionating and total condensers; and apparatus for safely diluting the polyester with monomer.
Good quality unsaturated polyesters are processed by a two stage reaction in which aromatic and saturated acids are reacted with all the glycol until at least one functional group of all diacid has reacted. The resin mixture is then cooled, maleic anhydride added and processing continued to final properties as determined by test methods, such as acid number or viscosity, that reflect polymer size.
Esterifying the slower reacting acids with a substantial glycol excess produces low molecular weight, hydroxyl terminated oligomers that react in the second stage to more evenly distribute unsaturated functionality throughout the polymer.
The esterification reaction is accelerated by efficiently removing the water of reaction (hence the need for an efficient partial condenser), higher than atmospheric pressure and/or certain catalysts.
The polyester chain grows as acid and hydroxyl groups combine and release water. Unreacted acid and hydroxyl groups left in the reaction mixture can be monitored by conventional wet chemical techniques to indicate the course of the reaction. Figure 1 illustrates the growth of polymer size and increase in viscosity as the available end groups are consumed. Because virtually all isopolyesters are formulated with a hydroxyl excess, acid number is commonly used to quickly indicate the remaining level of reactive material. Excessive residual carboxyl functionality contributes to viscosity drift and greater vulnerability to chemical attack in end use applications. Therefore, the preferred method for controlling polyester molecular weight is to adjust initial hydroxyl excess, rather than prematurely end the esterification reaction.
Evaluating the Unsaturated Polyester
A variety of tests are available for determining the identity and properties of polyester resins as solutions and as cured solids. The concern of the resin formulator is to use evaluation techniques that are useful for determining resin uniformity, can be quality control tests during manufacture, can predict performance in actual use conditions and are in-expensive and convenient to perform. Resin users need tests showing handling and quality characteristics.
Tests of the wet or uncured resin and knowledge of the resin formulation and ingredients can tell the molecular weight, level of unsaturation, probable thickening rate and cure response. Conventional tests of uncured resin and their implication for the resin formulator and user are shown in Table 1.
The focus of this brochure is on unsaturated resins themselves; however, in most applications the resins are combined with fibrous reinforcement or fillers. The user of the final product is most concerned with the properties of the composite which are affected by both the resin and the reinforcement.
Certain properties are dominated by reinforcement and the contribution of the cured resin is masked. Figure 2 illustrates the correlation of cured resin and composite tensile elongation for most of the resins studied in Amoco's Technical Service Laboratory. Up to about 2.5 percent the tensile elongation of the composite is influenced by that of the resin. Thereafter it is determined by that of the reinforcement.
Other composite properties, such as flexural strength (stress perpendicular to the orientation of most of the reinforcement), are largely determined by the resin. Corrosion resistance is among the properties that relate to the interaction between reinforcement and resin and that is best determined by testing a laminate.
Fig. 2. Reinforcement Masks Resin Tensile Elongation Properties
To most clearly focus on resin contribution to laminate performance and avoid the scatter of test results inevitable with composite testing, most properties discussed in ths brochure are based on testing of clear or unreinforced castings of resins. Laminate testing is used only for corrosion resistance and flexural fatgque evaluations that are functions of resin/reinforcement interaction.
Resin Properties are Intertwined
Change in one property usually causes a change in some other property. The properties most desired by the end user are frequently inferred from several measurable variables.
For example, toughness, a most desirable property, is classically considered as the area under the tensile stress/strain curve. Figure 3 depicts a hypothetical charting of destructive tensile stress of a clear resin casting using an Instron Tester. The area of the roughly triangular figure formed under the curve is proportional to the toughness of the tested resin. This property will obviously be affected by changes in either ultimate tensile elongation or ultimate tensile strength. If each component of a multivariate property, such as toughness changes slightly in the same direction, the cumulative change is much greater than each individual change. For instance, a 30 percent increase in both tensile elongation and tensile strength is a 70 percent increase in toughness.
Fig. 3. Tensile Stress/Strain Area Defines Toughness
Similarly, a difference in destructive stress testing of laminates can reflect a more substantial difference in non-destructive, cylic stress. Table 2 shows reinforcement partially masking the flexural strength advantage inherent in an isophthalic resin. When the laminate is subjected to flexural fatigue load testing (see Figure 22 for more detail of test and resins), the resin difference is again apparent.
Flexibility, which underlies many physical properties of resins, is usually closely correlated with tensile elongation. Figures 4 through 6 show the relationship between tensile elongation and flexural strength, tensile strength, stiffness (flexural and tensile modulus) and heat distortion temperature for polyesters studied in Amoco's laboratories. The graphs indicate relatively invariable relationships between some properties heat distortion resistance and high elongation are not compatible combinations of properties in the same class of resin. On the other hand, toughness can be increased in resins with low flexibility by increasing tensile elongation, while in very flexible resins toughness will be improved by reducing elongation.
Table 3 Summarizes the physical tests and instrumental analyses that indicate resin characteristics and usefulness for various applications.
Determination of Corrosion Resistance
Even in applications not normally referred to as corrosion resistant, the ability of reinforced polyester to resist conditions that would cause rusting or decay in other materials is a valued attribute. The vulnerability of a fiber glass laminate may be attack of the resin, the glass fibers or the interface between them. Despite normal fabrication of corrosion resistant laminates with an essentially fiber-free surface, liquids can permeate polyesters to some extent and contact the glass-resin interface. Consequently, Amoco's preference is for corrosion testing of complete laminates with edges protected from exposure.
Most of Amoco's testing has been conducted on laminates constructed and exposed in corrosive media by the procedures outlined in ASTM C581. Analysis of flexural properties and hardness at one, three, six and twelve months can be plotted on log-log graphs to project ten year performance. An example of such projections is shown in Figure 7. If the best straight line through the one year data indicates 50 percent or more retention of properties at ten years, the laminate is considered acceptable for commercial service in that media. Advantages of this test method are that its acceleration is through two-sided exposure, not heat which can distort results. The test evaluates the resistance of the total laminate and is reliable. In Amoco's tests property retentions have been reproducible to within 5 percent for flexural modulus and 10 percent for hardness and flexural strength.
Level of Unsaturation in Polyester
The essential ingredient for an unsaturated polyester is the carbon-carbon double bond or olefinic unsaturation that will subsequently crosslink with the reactive diluent. In virtually all commercial resins unsaturation is provided by maleic anhydride or fumaric acid. As shown by the molecular models, they are very similar in the esterified form, and their differences are minor compared with the effect of their use level in the polyester (Illustration 4).
The ratio of maleic or fumaric unsaturation to the total ingredients of the polyester is the primary determinant of reactive double bond frequency in the polymer. This frequency in turn determines the amount of crosslinking that can occur with a reactive diluent, such as styrene, and thus, strongly influences cured resin properties.
An unsaturated polyester could be made with only the unsaturated acid or anhydride and a glycol or oxide. As saturated acid replaces unsaturated, the frequency of double bonds in the polymer will decrease causing an increase in flexibility as reflected by tensile elongation (Figure 8). The same trend is displaced, but still true, for resins made with quite different glycols. Figure 9 shows the increased tensile elongation associated with higher aromatic acid content in resins made with diethylene glycol.
Heat distortion temperatures, as expected, decrease as crosslink density is reduced. However, strengths, as noted in the previous section, vary with change in elongation according to the resin flexibility. Thus the rigid, propylene glycol resin is strengthened (Figure 11) by increasing aromatic acid content, while the very flexible DEG resin exhibits lower strength as it is made more flexible (Figure 12) by increasing aromatic acid content.
These polyester double bonds react via a catalyzed free-radical process with the double bonds in styrene or other coreactant. The reaction of each double bond releases a discrete amount of heat energy. The total amount of heat released during crosslinking is indicated by the SPI Gel Test. Propagation time (a measure of the in-mold cure time) and peak exotherm as determined by the SPI Gel Test correlate well with the amount of unsaturation in the polyester, if all other factors are constant. Figures 13 and 14 illustrate these trends for a series of resins made with propylene glycol and crosslinked with 45 percent styrene. Generally, higher polyester unsaturation levels will result in faster cure and higher exotherm temperatures.
Other Effects of Crosslink Density Variation
Impact resistance is positively associated with flexibility. Thus, decreasing unsaturation will increase impact resistance, reducing molded part damage during fabrication and shipping.
Resins with higher unsaturation levels can tolerate more inert filler because the crosslink density remains sufficiently high to provide good strength as the resin portion of compound volume is reduced.
The higher heat distortion temperatures of more highly unsaturated resins allows service at higher temperatures.
Sources of Unsaturation
While theoretically a great variety of unsaturated difunctional acids and anhydrides could be used to provide the required double bonds in the polyester, virtually all commercial unsaturated polyesters incorporate maleic anhydride (cis configuration) or fumaric acid (trans configuration).
Maleic anhydride is usually less expensive than fumaric acid. It can be readily melted and handled as a liquid. The anhydride form reacts faster than the acid and one less mole of esterification water is released during processing.
Maleic acid readily isomerizes to form the more stable fumaric acid. During esterification the maleate structure can rearrange into the fumarate configuration. The rate of isomerization is apparently dependent on the type of glycol. Table 6 shows isomerization rates reported in the literature. The implication of these rates is that fumaric acid has no advantage over maleic anhydride unless 100 percent of the trans configuration is required for some purpose.
The advantages of total trans configuration as reported in the literature are typical of a more linear and crystalline polymer: greater hardness; higher moduli or stiffness; lower elongation; higher heat distortion temperature; reduced gel and propagation times; higher exotherms. The trend of these differences is to make the polymer more rigid.
Effects Attributable to Phthalic Isomer Differences
The basic aromatic di-acid can form three sterically distinct isomers: orthophthalic acid, isophthalic acid and terephthalic acid. These isomers have the same chemical formula, but differ in the location of the acid groups on the aromatic ring. Orthophthalic acid, the only one of the three isomers capable of forming an anhydride, is normally used in that form.
Each phthalic isomer has particular advantages and liabilities. For most applications the balance of cost and performance will generally provide a clear-cut choice for one of the isomers.
To a great extent the application performance and property difference result from the physical and chemical differences of the isomers. Amoco's evaluations indicate that IPA offers substantially better properties than equivalent formulations made with phthalic anhydride. The properties of terephthalic resins are generally better than orthophthalic resins. The only property improvement consistently offered by TA is greater heat distortion resistance.
A resin manufacturer's cost analysis includes raw material costs, processing costs and ancillary costs associated with special handling of any product. Such an analysis is an overly simple view of the real cost and value to the user. Customer oriented cost analysis must include not only resin purchase price and fabrication costs, but life-cycle cost factors such as maintenance, useful life before replacement and performance in multiple environments. The manufacturer producing high quality resins can offer the savings of longer service life in more varied conditions.
The combination of better physical properties and superior corrosion resistance reported for isophthalic polyesters in this section can be exploited in several ways:
Differences in Material Handling and Processing
The ortho isomer is normally used as the anhydride which reacts faster initially and releases one, rather than two, moles of esterification water. Phthalic anhydride can be melted, providing some convenience for plants equipped to pump hot liquids to storage and process units.
Terephthalic acid is the slowest reacting of the three phthalic acids. Catalysts or pressure are required to esterify TA within a reasonable time period.
Isophthalic acid reacts more readily than TA, but its initial rate is slower than phthalic anhydride. Isopolyesterification can be catalyzed to provide reaction times approximating those of anhydride esterification. A summary of typical processing time with and without catalysts and pressure is shown in Table 8.
The K1 reported in Table 7 for orthophthalic acid is not indicative of the initial reaction rate of the anhydride. Note, however, the significantly lower K2 of ortho versus iso. The practical implication of this low K2 is a much slower reaction rate for the second acid group of orthophthalic based polyesters.
Consequently, it is much more difficult to obtain high polyester molecular weight with phthalic anhydride than with the iso or tere conformations. Efforts to push orthopolyesters to high molecular weight increase the risk of gelation and aggravate the sublimation rate of phthalic anhydride. These efforts also negate the inherent advantages of ortho resins - fast reaction times and light color.
The sublimation tendency of phthalic anhydride requires caution during processing and can indirectly influence cured properties. Sublimed phthalic anhydride fumes are flammable and can cause fractionator and condenser plugging. Condensed phthalic anhydride and low molecular weight phthalic esters can drip back into the resin kettle during final stages of processing and after the reaction is complete. These low molecular weight materials have relatively high water solubility, act as plasticizers of the cured resin and reduce corrosion resistance.
Because terephthalic acid require catalysis for efficient processing, the consequences of residual catalyst in cured resin must be evaluated in end use applications. Amoco's experience indicates that certain catalysts may be detrimental to corrosion resistance properties and unacceptable for food contact applications. TA resins cooked with catalysts generally have dark color and poor shelf life. Gel characteristics can also be affected.
Liquid Resin Properties
Equivalent formulations made with the different phthalic isomers will have somewhat different liquid properties resulting from the different target end properties. Resins based on isophthalic or terephthalic acid will normally be processed to higher molecular weights and will show higher viscosities and lower end group counts (acid numbers and hydroxyl numbers) than equivalent phthalic anhydride formulations. To obtain proper solution viscosity for particular end uses, such as spray application, somewhat more styrene dilution or relatively more glycol may be required with IPA than with phthalic anhydride. Stopping the isopolyesterification reaction at a lower molecular weight (higher acid number) is generally not recommended as many of the cured resin advantages will be lost.
Terephthalic polyesters made with primary glycols such as neopentyl have poor solubility in styrene compared with ortho and isopolyesters. Blends of glycols and highly branched or cyclic glycols have been suggested to improve TA resin solubility.
Factors Affecting Alkyd Production
Except to briefly the more important factors that affect the preparation of an alkyd, a minimum of discussion will be devoted to the chemistry involved in alkyd polymerization and the processing conditions controlling the alkyd reaction.
An alkyd raw material may be supplied in one or more grades of purity. Thus glycerol is supplied in several concentrations as listed in Table 1. The term glycerol refers to the pure chemical
Trace amounts of foreign material may affect the alkyd reaction. For example, synthetic glycerol under certain conditions may act different than glycerol from natural sources.
The alkyd raw material may be introduced as such or combined with some other material. Thus fatty acids may be added directly the alkyd reactor, or they may be added, combined with glycerol, as an oil molecule (whole oil or glyceride oil).
The advantages and disadvantages of using fatty acids as opposed to glyceride (whole oils in formulating alkyds are summarized in Table 2. A review of this table suggests that when pricing takes precedence over performance, whole oils will be selected for the alkyd raw material, especially for long oil alkyds intended for trade sales items. However, where performance is all important, fatty acids are referred, for they afford the alkyd chemist a much greater latitude in alkyd design with a significantly higher probability of meeting the requirements of a tough specification. Fatty acids are most frequently used in the manufacture of short to medium oil length alkyds intended for industrial applications.
In calculating a theoretical alkyd formulation, all ingredients are broken down into their uncombined form before any computation of a formulation is attempted.
Order of Addition
It could be theoretically argued that any given alkyd composition should eventually reach the same end equilibrium structure regardless of the order in which the reactants are charged. From a practical and point, however, the order of addition is vitally important. In the formulation systems to be discussed, optimum alkyd compositions one are computed. The question of how the alkyd polymers are rest united chemically to achieve this optimum composition is a matter of preparation technique. Chemists versed in alkyd technology fully recognize the importance of the sequence of addition of alkyd ingredients and use it as a powerful tool in building up desirable structures.
For example, in the preparation of alkyds from glyceride oils, the first stage of the reaction almost always consists of forming a monoglyceride structure by alcoholysis of the glyceride oil with added polyol. This is a necessary first step, for (a) alcoholysis converts the insoluble polyol and glyceride phases into a single homogeneous monoglyceride phase and (b) the monoglyceride in turn provides a solvent for the phthalic anhydride added for the next step, esterification of the monoglyceride with diacid to complete the alkyd reaction. In the preparation of alkyds from fatty acids, order of addition can again be important. For example, it is generally conceded that improved physical properties are obtained when the polymer structure is predominantly linear and of high molecular weight. To achieve this, one proposed method stipulates a step by step esterification of the fatty acid. In this procedure, the alkyd is cooked with only a proportion of the total fatty acid present (from 40 to 90 per cent). By withholding a part of this chain-terminating ingredient a linear structure is encouraged during the first part of the reaction. Later, the remaining fatty acid is added to complete the preparation. Alkyds cooked by this so-called high polymer technique are said to be more viscous and of lighter color than those cooked by the conventional method, which calls for addition of all the fatty acid at the very beginning of the cook. Furthermore, films of the high polymer alkyd are said to exhibit a faster dry, improved flexibility, better adhesion, and enhanced resistance to detergents and alkaline solutions. However, results reported from a recent investigation specifically set up to study this high polymer theory only partially confirmed the original claims. Thus, some of the experimental results were contradictory, although the general trend supported the contention that this technique upgrades an oil such as tall oil, with greatest improvement in a long oil alkyd.
Certain significant differences are also observed for alkyds that are identical from a chemical composition standpoint but that differ in properties and performance depending on whether they are made by the alcoholysis or the fatty acid procedure.
A plausible explanation for this difference is afforded by considering the differential rates of reaction between -OH and -COOH groups, depending on their specific location on the parent molecule. Table 3 lists the relative rates of reactivity for several -OH/-COOH pairs, listed in descending order of reactivity.
In the fatty acid preparation method, where there is a free-for-all competition among the -COOH groups (all added at the beginning). the fatty acid -COOH groups lag behind in joining primary -OH groups and hence must settle for connections with secondary -OH groups. In the monoglyceride method, where the competition is rigged, the fatty acid -COOH groups are deliberately reacted with the primary groups of glycerol before any phthalic anhydride is added, and the diacid -COOH groups are placed at a competitive disadvantage and are forced to settle for a reaction with leftover -OH groups. Reasonable chemical structures for these alkyd compositions as formed by the two cooking procedures are graphically shown in Fig. 1.
An inspection of the two reaction rates at the bottom of Table-3 (for reactions that are responsible for cleaning up residual acidity) shows that a reduction in acid number should be attained more rapidly at the end of the reaction by the fatty acid procedure.
Rate of Agitation
Through agitation is necessary (a) to provide an intimate mixing of any immiscible ingredients (say soya oil and glycerol during alcoholysis) and (b) to accelerate the alcoholysis, polymerization, and allied reactions. The rate of a chemical reaction is promoted by agitation of the reacting molecules, and failure to provide adequate agitation leads to abnormally long cooks and inferior alkyd resins.
Agitation is furnished by both mechanical elements (propellers, paddles, turbines) and by sparging devices (bubbling of an inert gas through the reaction mixture). Sparging is also highly effective in removing liberated reaction products such as water (removal of water is necessary to permit the condensation reaction to proceed) and contaminants such as air (the oxygen encourages color development by oxidation).
The diameter of the rotating mechanical element furnishing the mixing agitation is generally about one-third the diameter of the alkyd kettle and should be located well down in the reactor. The revolutions per minute (rpm) of the rotating element is generally adjusted to an optimum peripheral speed of 10 feet/second.
With either nitrogen or carbon dioxide (which yield substantially equivalent results), sparging effectiveness is dependent on the rate of blow and the fineness with which the gas is dispersed as it courses upward through the reaction mixture.
For alcoholysis, a blow rate between 0.01 and 0.02 ft3/min/gal is satisfactory. Sparging is continued throughout the reaction, including the upheat and any thinning.
For esterification, a rate of 0.01 to 0.04 ft3/min/gal is satisfactory with the faster blow rate applying to the beginning and the slower to the end.
Gas dispersion is conventionally accomplished by introducing the gas into the reaction mix through many fine holes (facing downward to ensure drainage) drilled in a perforated tubular ring assembly which is spread out over the bottom of the alkyd kettle.
Adequate mechanical agitation and sparging are mandatory for the preparation of quality alkyds.
Temperature of Reaction
The choice of a reaction temperature or temperatures is usually a compromise, namely, a temperature that is sufficiently high to permit the reaction to be carried out within a reasonable time period, yet not so elevated as to cause destructive decomposition, discoloration, and/or an excessive loss of volatile material through the stack.
For example, a normal esterifying temperature for preparing a soya alkyd (day from soya oil, glycerol, phthalic anhydride) is 450 F (232 C). Increasing the esterification temperature by 40 F to 490 F halves the processing time, but loss of volatile material then becomes excessive. Decreasing the esterification temperature by 40 F to 410 F doubles the processing time, which makes the cook uneconomical from an over-all cost standpoint.
The temperature of the reaction at different stages may also be influenced and possibly controlled by such factors as the melt point or volatility of one of the alkyd raw material ingredients. Table 4 which compares reaction conditions for phthalic anhydride and iso-phthalic acid is a case in point.
Fusion (Solventless) Cook Versus Solvent Cook
Whereas alcoholysis (to form a monoglyceride) is invariably carried out with no solvent present, alkyd esterification may be carried out either in the absence of solvent (fusion cook) or with solvent present (solvent cook).
The amount of solvent used in the solvent cook must of necessity be held to a relative small percentage of the total charge by volume (5 to 10 per cent). Preferably it should be selected to have a boiling point in a range which is 75 to 100 F less than the temperature at which the alky is to be refluxed. Use of higher percentages of solvent prevents the attainment of an esterification temperature. Moreover, owing to its slow evaporation rate (only a high boiling solvent is applicable to a solvent cook), a high percentage of the solvent in the final alkyd extends the alkyd dry time to an intolerable degree.
The solvent cook is claimed (a) to facilitate the removal of water and gaseous contaminants owing to the continuous sparging action of the solvent, which is condensed and continuously returned to the alkyd kettle; (b) to give better temperature and viscosity control, as the presence of the solvent provides a more mobile mixture; (c) to establish conditions for a more uniform alkyd which is freer from skinning and overpolymerized gels; and (d) to yield a cleaner kettle after the batch is completed caromatic solvent dissolves any phthalic anhydride deposited in the condenser and returns it to the reaction mix). Despite these advantages, the fusion cook is a popular processing method and is preferred to the solvent cook when working with isophathalic acid. The fusion cook also requires less initial outlay of equipment, as no water condenser is necessary, and is less expensive to operate, as no heat input is required for refluxing of solvent.
Solvent Selection for Alkyd Reduction
Long oil alkyds are generally reduced to shelf-storage or application consistency with aliphatic solvents, whereas short oil alkyds require aromatic solvent, with or without some admixture of alcohol or other polar sovent, as the thinning system.
Viscosity properties for typical long and short alkyds reduced with appropriate solvents are given in Table 5.
Proper choice of catalyst is vital for facilitating the progress of the alkyd reaction.
The conversion of a glyceride oil and a polyol to a monoglyceride (say soya oil and glycerol to form a soya monoglyceride) can be satisfactorily accomplished without a catalyst by carrying out the alcoholysis reaction at a sufficiently high temperature (550 F). How ever, it is more expedient to carry out the reaction at lower temperatures (450 - 480 F) by resorting to a suitable catalyst which greatly accelerates the rate of conversion (by 10 - 20 times) and markedly cuts down on volatile losses. As little catalyst as possible should be used, for it also promotes color development and detracts from the water and alkali resistance of the end alkyd product.
Lime, Ca(OH)2, and litharge, PbO, are common alcoholysis catalysts which are quite effective at levels of 0.05 to 0.10 per cent based on the oil weight. Barium catalysts are rather ineffective and sodium catalysts, although effective, impart undesirable properties to the alkyd composition, such as considerable color development and slower dry Lithium ricinoleate, recently introduced, is finding favor owing to its excellent catalytic activity, relative freedom from phthalic anhydride poisoning, and clarity of product formed; calcium and litharge cause precipitation as metallic phthalates.
Alkyds Formulations Based on Theory
This chapter is devoted exclusively to the design of alkyds from theoretical considerations. Of the four theoretical systems discussed, one in particular has been amplified in Chapter 6 to give a highly practical formulating system for routinely designing and checking alkyd formulations.
The four formulating systems to be discussed in detail are predicated on:
a. Fav, an average over-all functionality for the alkyd composition
b. p, the probability of a branch-to-branch connection between reacting molecules at gelation
c. AN, the acid number of the alkyd composition at its gel point
d. Mav, the average molecular weight of the alkyd at gelation.
For the sake of brevity, they will be referred to hereafter as the Fav, p, AN, and Mav formulation systems, respectively.
To make the discussion completely general, two molecular species arbitrarily designated A and B will serve as co-reactants. For practical application work, they will be identified with actual molecules as required. It is postulated that A can react with B to form a chemical link. Since most alkyd work deals with bonding through esterification, A can be thought of as an acid or carboxyl group (-COOH) and B as a basic or hydroxyl group (-OH). However, it should be understood that A and B groups are not necessarily limited to this specific interpretation.
By the use of a subscript, it is possible to designate the functionality of the molecule to which an A or B group is connected. Thus, A2 will indicate that this group is part of a difunctional A molecule (a molecule with two reactive A sites). Or B4 will indicate that this B group is part of a tetrafunctional B molecule (a molecule with four reactive B sites). The use of subscripts provides a useful notation for tagging the A and B groups during the calculation work, identifying their source.
So much for likeness and a common ground for all four systems. It is the derivation of the third basic equation (a much more complex affair) which is radically different for each system and which serves to differentiate among them.
Assume k functional groups react for every molecule that disappears (through merger with another molecule). Normally k is 2, but with highly functional systems more than two functional groups can react per lost molecule to give multiconnected mutual molecules.
At P per cent of reaction, (m0 - mP) molecules have been lost. Hence the corresponding number of functional groups that have been reacted is given by Eq. 5.
The extent of the reaction (extent of the condensation polymerization) P can then be expressed in terms of m0, mP, and Fav by Eq. 6. Note that the product m0Fav in this expression equals the total number of functional groups initially available for the condensation reaction.
Both these expressions are beautifully simple statements of the conditions for gel formation when the reactants are present in stoichiometric proportions.
For example, if both A and B molecules are difunctional and if they are present in stoichior etric proportions, the reaction proceeds to completion (P = 2/2 = 1 = 100 per cent), at which stage there are present a number of tremendous linear molecules which are inter-twined to form a gel-like composition.
However, if either the A or B group is present in excess (which is usually the case), a downward adjustment in functionality must be made for this group corresponding to its excess.
Problems Illustrating the Formulation of Alkyds Based on the Fav System
The following problems serve to illustrate the technique of assessing and formulating alkyd compositions based on the use of an average functionality.
The short OL dehydrated castor oil (DCO) alkyd given in Table 1 has been proposed as an experimental alkyd composition. Check the feasibility of its preparation.
This 101 per cent value indicates that the reaction can be carried to completion with safety (1 per cent margin to spare). However, actual preparation of this particular alkyd by the solvent method resulted in gelation at an acid value of 12. This premature gelation can undoubtedly be attributed to dimerization of the DCO FA (a side reaction taking place along with the main alkyd reaction) which effectively contributes to the over-all polymerization.
The next problem illustrates how information obtained from an initial alkyd cook that gels prematurely can be employed to adjust the formulation of a subsequent cook.
From the data developed in Problem 1, formulate an adjusted alkyd composition that will avoid the onset of premature gelation.
The acid number of the alkyd composition as charged to the alkyd kettle in Problem 1 was 372. This value is calculated as shown in Table 3 from the total acid equivalents and the charged weight.
The extent of the reaction which was actually reached in Problem 1 was then 96.7 per cent.
This is comparable to the predicted value of 101 per cent. The discrepancy between predicted and experimental values calls for an adjustment of k in Eq. 8 to make it conform to the experimental findings of the initial cook. This is accomplished by substituting the experimental value of P = 96.7 in the equation; Fav remains unchanged at its value of 1.98. From Eq. 8,
Future calculations will now be based on this experimental k value of 1.91. This is comparable to the alue of 2 initially assumed for k.
To prepare a nongelling variant of the DCO alkyd of Problem 6.1, let the glycerol content be arbitrarily increased from 2.25 to 2.38 moles. This furnishes a greater excess of -OH groups (lowers the average effective functionality of the alkyd system) and permits the reaction to proceed further before onset of gelation.
The -OH excess here is 7.14/(1.00 + 4.36) = 1.33, which lowers the effective glycerol functionality by the reciprocal of 1.33, or the factor 0.750 (= 1.00/1.33). The effective over-all functionality of the alkyd composition Fav is then 1.93.
The predicted extent of the reaction before gelation, using the experimentally determined value of 1.91 for k, is 0.993, or 99.3 per cent.
This percentage appears borderline, as a P of 100 per cent is an optimum goal. However, this alkyd was successfully prepared at 450 F and brought to an acid value of 8 (corresponding to a P of 97.7 per cent) without gelation. The alkyd had a Z2 - Z3 viscosity when reduced to 50 per cent solids content in xylene. Note that the value of 97.7 per cent lies on the safe side of the predicted value of 99.3 per cent for incipient gelation.
Short-cut Method for Computing a P Value
Before proceeding with further illustrative problems, it is instructive to consider a short-cut method for arriving at a P value.
To explain this method, let eD denote the sum of the equivalents for that group which is not present in excess (is deficient in amount for a stoichiometric reaction) at the beginning of the condensation polymerization.
For example, an alkyd is conventionally prepared by the reaction of -OH and -COOH groups, with the -OH groups present in excess. According to the above definition, eD denotes the sum of the acid equivalents in the reaction mix, since this is the group that is deficient in amount for a stoichiometric reaction.
In the Problem 2, in order to compensate for an excess of one group in obtaining an effective Fav, an excess factor was first calculated and the reciprocal of this was used in turn to adjust the functionality of the excess component. But it can be shown that the net effect of this procedure is merely to make the effective number of equivalents of the group in excess equal to the actual number of equivalents of the deficient group.
This can be checked by referring back to Problems 1 and 2 where adjustments were made in the functionality of the glycerol to take care of its excess. Thus in both problems, the number of effective glycerol equivalents and the actual acid equivalents are both equal to 5.36. Note that the effective glycerol equivalents still remain equal to 5.36 in Problem 2 even though the glycerol content was increased.
Once this 1 : 1 relationship between effective equivalents of the group in excess (say -OH) and actual equivalents of the deficient group (say -COOH) is established as valid, it becomes apparent that 2eD can be used as a replacement for e0 in calculating an effective Fav for the system.
Before a P value can be calculated, the soya oil must be broken down into its FA and glycerol components. This is done in Table 7; the charged composition is tabulated to the left and the breakdown composition to the right.
From these values of m0 = 0.561 and eD = 0.552, the percentage of reaction P at incipient gelation can be calculated from Eq. 12.
Practical Alkyd Formulations
Derivation and Application of a Unique Alkyd Constant for Routinely Assessing, Adjusting, and Designing Alkyd Compositions
In Table 47, it was noted that of the four theoretical formulation systems considered, the one based on an average alkyd functionality was of most universal applicability. This particular system will now be developed in a different and somewhat unique fashion, leading to the concept of an alkyd constant K. The derivation, as before will be based on Carother's classical theorem with the adaptation applying to the specific case of alkyds.
By the use of this alkyd constant, it is possible to routinely assess the feasibility of preparing an untested alkyd, adjust an improperly formulated alkyd to a corrected composition, or formulate an alkyd from scratch.
The derivation initially follows the line of reasoning developed under the Fav formulating system of the previous chapter leading to Eq. 1.
At this point, a different approach is taken in containing the derivation. Equation 1 as written above covers the case in which the alkyd components are present in stoichiometric proportions. But what about the case in which one of the reacting groups is present in excess? Thus alkyds are invariably formulated with excess hydroxyl groups. How does this affect the foregoing derivation?
The problem can be resolved by considering the total of those equivalents that have actually reacted when the condensation reaction is 100 per cent completed, that is, when all the acid groups have been related. This total must of necessity equal twice the number of acid equivalents initially present in the alkyd composition (one hydroxyl group is reacted for each acid group reacted).
As far as effective use of functional groups is concerned, only those that have reacted count. The hydroxyl groups present in excess remain unused and in effect reduce the effective functionality of the molecule to which they are attached. Hence for the case in which hydroxyl groups are present in excess, the effective number of equivalents for the alkyd composition is 2eA (where eA equals the total number of acid equivalents initially present in the alkyd mixture) and not the actual number, e0.
There are more rigorous ways of deriving this basic equation, but this is the simplest and most straightforward.
For an alkyd to be outstanding in performance, it must be processed to as high a molecular weight as possible. At the same time, the molecular weight must not be allowed to become excessive, or the alkyd vehicle will drift out of control during processing (will gel) or the alkyd will exhibit instability on shelf aging and prematurely convert to a gel in the paint can. Hence the design must avoid unduly small or unduly large alkyd polymers.
In general, an alkyd is formulated to reach a point just short of incipient gelation at 100 per cent of reaction. This is equivalent to saying that at gelation (incipient), P should equal 1.00. The importance of Eq. 2 in the design and assessment of alkyds now becomes manifest, for it provides an extremely simple criterion for formulating an alkyd to meet the conditions of incipient gelation at P = 1.00, namely, that m0/eA also equals 1.00.
It is now postulated that the ratio of total moles to total acid equivalent for any properly formulated alkyd is equal to unity. This is a theoretical constant. As will be shown, this alkyd constant of 1.00 will be slightly increased in practical formulations to ensure a measure of safety during processing and storage.
Practical Validity of the Alkyd Constant K
It would not be unnatural if the simplicity of the expression for the alkyd constant should engender some doubt concerning its ability to hold for any but a narrow range of alkyd compositions. Actually, the reverse is true and it is rather remarkable that the alkyd constant should find such universal applicability.
For example, the mean value for K for 24 alkyds randomly abstracted from three literature sources was found to be 1.022 + 0.023. The average deviation of 0.023 for this empirically derived K value attests to its constancy, but it fails by a small fractional amount to coincide with the proposed alkyd constant of 1.00. This slight discrepancy in value can be reconciled by the following argument: The constant 1.00 is a theoretically derived value for the ratio m0/eA. It pertains to the conditions of full completion of the condensation reaction at incipient gelation. However, for any practical alkyd cook, formulating this close to gelation is dangerous. Accordingly, experience dictates that a margin of safety be provided. This accounts for the difference between the theoretical constant of 1.00 and the experimentally determined average value of 1.022. The 0.022 discrepancy is a safety factor.
This immediately suggests that it might be more expedient to peg the alkyd constant at a practical value of 1.02 rather than the theoretical 1.00 value, which provides no built-in safety factor. However, although this is a good practical recommendation, it is felt that the base value of 1.00 should be retained as a foundation figure from which working alkyd constants can be built in turn.
For example, an inspection of the data for the 24 alkyds revealed that those alkyds based on phthalic anhydride tended to have low K values, whereas those based on isophthalic acid tended to have high K values. By grouping the alkyds by diacid type and calculating the mean alkyd constant for each group it was shown (a) that better constancy among the K values was obtained (average deviation smaller), and (b) that with isophthalic acid, a somewhat larger processing safety factor must be provided.
Hence the following approach is suggested for establishing an alkyd constant for formulating practical alkyd resins. Unity will be retained as a base theoretical figure for the alkyd constant. From this theoretical value of 1.00, practical or working alkyd constants will be derived that will contain built-in safety factors corresponding to the type of alkyd being processed.
Typical adjustments of the theoretical alkyd constant of 1.00 to obtain practical target K values for formulating experimental alkyd cooks are given in Table 8.
Use of the Alkyd Constant for Assessing the Feasibility of Preparing a Given Untested Alkyd Composition
The working alkyd constant K provides the alkyd chemist with a powerful tool for assessing the feasibility of preparing a given untested formulation. By comparing the computed K (= m0/eA) for the proposed formulation against the working constant for that type of alkyd, the alkyd chemist is in a position to gauge whether the preparation of the alkyd resin is feasible.
If the computed K value is less than the working alkyd constant, then gelation short of 100 per cent reaction must be anticipated; if greater, then unacceptable polymer formation (with concomitant unsatisfactory vehicle performance) must be expected. About 0.05 units deviation from the working K value is probably the maximum that can be tolerated without encountering certain gelation on the one hand or jeopardizing ultimate alkyd performance on the other hand.
To illustrate, assume that an untested phthalic anhydride alkyd formulation is submitted for preparation approval. The working constant for this type of alkyd is K = 1.01. A first step would consist in determining whether the m0/eA value falls in the range 1.01 ± 0.05. A value close to 1.01 would indicate a normal alkyd preparation with satisfactory ultimate performance assured. If the m0/eA value for the alkyd fell in the extremes of this range, some question of the feasibility of its preparation would be in order. However, if the m0/eA value fell outside the 1.01 ± 0.05 range, it is almost certain that its preparation is foredoomed to failure or that its ultimate performance properties will be unsatisfactory. The problem that follows illustrates the details of computing and using an m0/eA value for alkyd assessment.
A medium OL alkyd is submitted for preparation approval. It consists of 300 parts soya oil, 200 parts PA, and 100 parts glycerol (weight basis). Is this a feasible formulation?
Calculate the actual K value (m0/eA) for the given alkyd and compare it with the working alkyd constant (optimum target K value) of 1.01 that applies to PA alkyd compositions.
Comparison of the actual alkyd constant of 1.02 with the optimum target constant of 1.01 reveals the complete feasibility of proceeding with the preparation of this soya alkyd.
Use of the Alkyd Constant for Adjusting an Improperly Formulated Alkyd Composition
An alkyd constant can be used to adjust improperly formulated alkyd compositions by setting up successive paper modifications and checking them against an appropriate alkyd constant until a reasonable m0/eA value has been attained. With a little experience, it is entirely possible to correct an improperly formulated alkyd by a single formulation adjustment. The key to any adjustment is course, judicious alteration of the relative proportions of the alkyd components until their m0/eA ratio equals the alkyd constant that applies to them. The next problem illustrates the details of how such an adjustment is carried out.
An IPA alkyd is submitted for preparation which has the following weight composition: soya FA 40%, IPA 38%, benzoic acid 2%, glycerol 20%. Is this a feasible preparation? If not, correct the formulation as necessary to carry out its successful preparation.
Calculate the alkyd m0/eA ratio as it now stands.
A comparison of this computed K value of 0.975 with the working alkyd constant of 1.05 for an IPA alkyd reveals the certain gelation danger that can be expected during the final stages processing. Adjustment to a safer and more practical formulation is necessary.
One way of accomplishing this is to increase the percentage of monofunctional benzoic acid at the expense of difunctional IPA. This lowers the over-all functionality of the system and permits a more complete reaction before the onset of incipient gelation. Based more or less on experience, the benzoic acid content is accordingly raised to 8 per cent and the IPA content is lowered to 32 per cent. This gives the alkyd resin in Table 5.
A comparison of this computed K value of 1.04 for the adjusted formulation with the working alkyd constant for IPA alkyds of 1.05 indicates the feasibility of proceeding with the preparation of the corrected composition.
Use of the Alkyd Constant for Formulating Alkyd Compositions From Scratch
One obvious approach for setting up an alkyd formulation from scratch would be to start with an educated guess for the required composition and then, by successive approximations, arrive at an acceptable K for the final alkyd. However, it is possible to achieve the same result in a more orderly fashion. The systematic approach will be illustrated by considering a common case in which the alkyd is formulated from monoacid, diacid, and a polyol of functionality x.
Assessment of the Performance of Single and Multicoat Red Iron Oxide-Alkyd Paint Systems
The prime requirements of exterior protective coating are resistance to the changes in weather conditions, ability to withstand the attack of environmental pollutens and protection of the substrate from corrosion. The growing industrial development potentials demand paint systems which effectively protect steel structures and other installations from pollutants in the surrounding environment. For this purpose the substrates are given a number of coats of protective coatings comprising primer, intercoat and top coat. However, in practice, instances have been observed where paint systems which score well in the initial testing fail prematurely, thus making the protection of the installations more expensive. It is, therefore, essential to assess the performance of individual paints with respect to the location and the performance of the paint system in combination (primer + intercoat + top coat).
Lingberg1 critically examined test data on the performance of paint coatings under outdoor and laboratory conditions and suggested that for good correlation among the test data strict methods and precise measurement of the individual properties should be adopted. Kilcullen2 studied the relative importance of various factors related to the environment and to conditions of application procedures in assessing the performance of paint coatings on steel structures and recommended that adequate thickness of coating of the paint systems is essential for good protection. Cooling and Wilkinson3 are of the view that accelerated weathering test data can be used confidently to predict the performance of coatings provided that the natural effects are taken into account in the accelerated weathering devices. Ellinger4 made an attempt to correlate the test date on the performance of coatings obtained from various accelerated weathering devices and from the natural weathering. He also quoted the views of others on these aspects of the students.
"Several workers5" have studied various properties of coatings and tried to correlate their findings in terms of quantitative assessment of the performance of the coatings. In the present study an attempt has been made to study the changes in the properties of paints having different contents of synthetic red iron oxide pigment in linseed oil- penta - phthalate alkyd medium. The paints which contain 55% PVC of iron oxide are used as primers and the ones containing 25, 30 and 35% PVC's of the pigment are applied as to coats on the primer coat. The effect of a number of coats on the properties of paints as well as the changes in their properties when exposed outdoors are studied and the findings are reported here.
Materials: Two resins11 (i) 66% linseed oil penta phthalic alkyd and (ii) 52% linseed oil-glycerol-phthalic alkyd were used as binders. Synthetic red iron oxide pigment (density 5.12g/cc) and white spirit + xylene mixture (1:1 v/v) were used for preparing paints in these alkyds. The details about the paint formulations are given in Table 1.
Preparation of paints: In order to have particles of uniform size the pigments was seived through 300 mesh. The calculated amount of the binder, pigment and solvents of a particular mill base was mixed in the pot and left overnight. The mill base was ground in a steel Cowlishaw high speed planetary ball mill to a fineness of 7-8 Hegmann - gauge. After grinding and filtering the solvent content was adjusted to 21.0 ± 0.5% on weight of the paint. The paints were stored in air tight sample bottles at 26 ± 1°C. 0.5% lead napthenate and 0.05% cobalt napthenate on the weight of the binder content were added to the paints 24 hours before application. The following tests were conducted on the coatings, either on a metal substrate or as free paint films:
1. Resistance to scratch in kg load.
2. Tensile strength of free film coatings in kg/cm2.
3. Adhesion strength of coatings in kg/cm2.
4. Permeation of water vapour through free films in g/m2/h/mil.
5. Resistance to corrosion in Salt Spray Test.
Coatings on Tin Foil:
Paint coatings applied on tin foil were used for conducting tests in which free films were required. Coatings of a particular paint were applied on tin foil using a mechanically driven applicator with Bird Blades (Gardner Laboratories Inc: USA). They were left in the application room for 48 hours to air dry before handling them for any type of test. In the case of multicoat application, a time gap of 24 hours was allowed for between two coats. The amalgamation of the tin substrate was carried out to separate it from the coating for initial testing. The free films of coatings thus prepared were used for conducting tensile strength and water vapour permeability tests.
Coatings on the Metal Substrate:
The surface preparation of mild steel panels (150 x 100 x 2mm) and the painting of them was carried out according to standard procedures prescribed for resistance to corrosion tests.12 The tin plated mild steel test panels (150 × 50 × 1 mm) were prepared for the scratch hardness test by a similar procedure. For the measurement of adhesion strength of coatings, mild steel discs (dia 30 mm and thickness 2 mm) which had been abraded, degreased and dried before storing them in desiccators were used. The paint coating was applied on these discs using an ICI spin coater.13
Coatings for Outdoor Weathering
The coatings (single coat, two and three coat systems) on tin foil backed with glass plate were tied to the support with mastic tapes. The painted mild steel test panels were given a protective coating on their back and edges. The mild steel discs prepared for adhesion test were coated on their back with a strippable coating. The prepared test specimens were exposed to outdoor weathering facing South at 45° angle on racks four feet above the floor on the terrace of the laboratory building14.
Results and Discussion
The 48 hours air dried coatings on tin foil were left for 24 hours in a mercury bath to amalgamate the tin substrate. The underside of the free film was then allowed to dry for another 24 hours. Thus the films were air dried for about 120 hours before any one of the tests was conducted on the free films. The data obtained from the tests conducted on coatings air dried for one week are referred to as the initial readings, i.e., zero period of outdoor exposure in the table and figures.
Scratch Hardness of Coatings:
In this study the hardness of paint coatings air dried initially and of those weathered outdoor for various periods of time is determined by using an automatic scratch hardness tester (Research Equipment Ltd., U.K.). The results reported in Table 2 and plotted in Figures 1a and 1b indicate a gradual increase in scratch hardness of coatings with a PVC of up to 50-55%. The hardness of coatings having pigment content beyond 55% PVC decreases, however the decrease is not significant. The reason for the over pigmented coating having good hardness is that when a wet coating is applied onto a freshly prepared metallic test panel the surface forces attract the active groups present in the binder. Consequently, the pigmented binder develops strength at the coating - substrate interface.
The plots of scratch hardness data of coatings weathered outdoors illustrate a gradual increase in hardness at up to 50 days of exposure and then a decrease during the following prolonged period of weathering (Figures 1a and 1b). In general, paints formulated with linseed oil - penta - alkyd show greater scratch hardness in comparison to those formulated with linseed oil-glycerol-alkyd. This is because penta-erthiritol-alkyd develops better cross linking in the course of its preparation as well as having good interaction with pigment and solvent. This is due to the greater functionality of its polyol. Penta-erthiritol-alkyd also attains greater strength during auto-oxidation of the coating in comparison to coatings based on linseed oil-glycerol-alkyd. However, in the course of prolonged weathering the coatings based on linseed oil-glycerol-alkyd retain their hardness properties for a longer period when compared to linseed oil-penta-alkyd based coatings.
The information obtained from this test is quantitative and can be used in assessing the performance of coatings in actual service.
Tensile Strength of Free Films
Tensile strength is one of the important properties of maintenance coatings. It indicates the reinforcing effect of the pigment in the binder and also the cohesive strength of the coating. The tensile strength of free films of coatings dried or weathered outdoor was determined using an Instron tensile tester (Instron Ltd. UK). The results are reported in Table 3 and plotted in Figures 2a. 2b and 2c. The data of tests conducted on initially air dried free film of paints show that their tensile strength values increase with pigment content in the paint formulation. For example the tensile strength of the free film of paint containing 25% PVC of pigment in linseed oil-penta-alkyd is 71 kg/cm2 and that of the paint having 50% PVC is 117 kg/cm2. However, at higher PVCs the drop in tensile strength indicates that the amount of binder is not sufficient to hold the pigment compact. Hence a pigment content around 50% PVC may be considered to be critical pigment volume concentration (CPVC). Paint formulations in linseed oil-glycerol-alkyd also show a similar trend, however their coatings attain the maximum tensile strength at about 55% PVC.
The tensile strength of weathered coatings reaches maximum value after 40 to 50 days of exposure. For example, in Figure 2a paint number P1 attains the optimum strength of 153 kg/cm2 after 50 days of exposure whereas the initial tensile strength of its air dried coating was 71 kg/cm2. Among the linseed oil-penta-alkyd paints the one which contains 45% PVC retains good tensile strength (134.66 kg/cm2), even after 180 days of outdoor weathering. Paints P6 and P7, being pigmented, have low tensile strength in comparison to other paints.
In figure 2b, the tensile data plots for paints in linseed oil-glycerol-alkyd show that, among these paints the free film of paint number pv (50% PVC) attains the maximum tensile strength and retains it even after 120 days of exposure. As observed in earlier cases, the tensile strength of the free film of over pigmented paint P11 (60% PVC) is relatively low.
The changes in the properties of multicoat systems, i.e. primer + one top coat and primer + two top coats, were studied with the objective to assess their performance in service. The tensile strength data reported in Table 3 and plotted in Figures 2a, 2b and 2c show that the tensile strength (load/cm2) of multicoat system is less than the average of the strengths of single primer coat and single top coat. Even in the course of weathering the multicoats do not build up as much tensile strength as it is found in the case of single coats. The reason for this observation is that the curing of the top surface of the thin single coat by auto-oxidation is fast and it is also catalysed by metal ions at the metal-coating interface. Consequently, the single coat attains good strength within a short period of time whereas top coat applied onto the primer does not come in contact with the metal substrate and therefore does not get catalysed by the metal ions. Due to this fact the multicoat systems do not attain as much strength as the single coat does in a specific period of time. The coatings of single and multicoats exposed for natural weathering also exhibit this feature in their properties.
Mechanical Properties of Alkyd Resin Varnish Films and the Effect of Different Weathering Conditions on them
The durability of a surface coating depends on the physical characteristics of its films, i.e., flexibility, tensile and bursting strengths, impact resistance, resistance to the permeation of water vapour, ions and gases, and its adhesion to the substrate. Generally, coatings fail in service by cracking and flaking due to the mechanical breakdown of their films, indicating that, at the time of failure, the magnitude of the stresses present in the film exceeds that of the cohesive forces. These stresses are set up as a result of the differential dimensional changes in the film and in its substrate, and changes in the chemical structure of the film as a result of weathering. The stresses induced in the film and the substrate are communicated to and resolved in the plane of the interface. The stress concentration and chemical changes lead ultimately to the breakdown of the film.
The stresses induced (or developed) on ageing are opposed to the forces bonding the coating to the surface of the substrate. In the process of drying, paints and varnishes are converted from the liquid to the solid state by oxygen-induced cross-linking. Since, under such conditions the paint film dries from the outside, rigidity will first develop at the exposed surface. To avoid stress concentration in the film, which will bring about cracking and flaking, a reasonably high order of mechanical properties is required. Thus, the determination of the mechanical properties of surface coatings under different weathering conditions is important in order to find out whether they will perform satisfactorily.
A close examination of methods available for measuring mechanical properties of paint and varnish coatings, such as rocker hardness, impact, bend and scratch tests, shows some limitations particularly with regard to their time dependence and reproducibility. Since these tests are normally carried out on films coated on metal surfaces, the values obtained may be affected by the nature of and adhesion to the substrate. An alternative approach based on measurements made on supported films could thus be advantageous.
A considerable amount of work has been done on the stress-strain properties of paint and varnish films, and it has been found that tensile strength measurements are in many ways a fare more reliable guide to film strength than the other methods. Film strength is related to the degree of cross-linking in the film-forming material. Cross-linking imparts rigidiy, which shows itself in increased tensile strength and lower elongation. The stress-strain curves can be used not merely to define the ultimate tensile strength and elongation of the material, but also to define its toughness, flexibility and hardness. Toughness is best measured by the total work required to break the film, as indicated by the area under the load elongation curve. In agreement with the definition of flexibility as extensibility or the ability to undergo deformation, the ultimate elongation can be considered as a measure of flexibility. The yield point on the load elongation curve is suggested as a measure of hardness. The importance of physical specifications for film-forming materials has been emphasized in connection with the relationship between the physical properties and the general durability of a coating. The initial values of the physical properties of the coatings vary with the composition of the film-forming materials. The durability depends more on the rate of change of physical properties than on their initial values. Periodical determination of these properties during the course of ageing will show the extent of changes taking place in the film up to its ultimate failure. Such data will provide a classification of the film-forming materials with regard to mechanical properties.
Certain physical properties of varnish films on metal substrates have been measured by several workers at definite ageing intervals. The flexibility of the coating was determined by bending a metal panel coated with the material under test around a mandrel. It was found that exposure to continuous carbon are light, in absence of water, had little visible effect on the films, hence the time of exposure required to crack the varnish film on bending over a progressive oxidation of oil in the course of ageing generally increases the tensile strength and decreases the flexibility of the film.
The function of the oil component of a varnish is to give the film the necessary extensibility to withstand the tensions caused by expansion, contraction, bending, etc., which would otherwise make it crack. In the present study a number of alkyd resin varnishes of the following compositions were selected for determining the mechanical properties:
The work outlined here was undertaken to determine the relationship between the composition of alkyd resin varnishes and their mechanical properties i.e., tensile strength, elongation, modulus of rigidity and bursting strength. The studies also included the periodical determination of some of the above properties during the course of ageing by exposing the free films of the varnishes to
(i) Natural weathering, and (ii) Carbon arc lamp.
The alkyd resin varnishes were prepared as described in Appendix I.
The free films of these varnishes were prepared by an amalgamation technique. The thickness of the dry varnish film can be controlled by variation in the percentage of thinner in the varnish and by using strips of known thickness while applying the varnish film over the tin foil with the help of a film applicator. For proper comparison of data obtained by the measurement of various film properties, the thickness of the dried film was controlled at 60 ± 5µ as measured by dial gauge.
Tensile Strength and Per Cent Elongation
The electrically operated Gardner tensile strength and elongation apparatus was used for determining the tensile strength and percentage elongation of the varnish films. Films were cut into test pieces of 12 × 1 cm and fastened to the upper and lower clamps of the apparatus so that the length of the film in between two clamps was 10 cm. The percentage elongation was noted for every reading of the load indicated by the scale and the observations were continued until the film failed. At least six determinations were made for each of the varnish films under test. The load values, calculated in terms of kg cm-2, were plotted against percentage elongation (Figs. 2) and the area under each curve estimated for the determination of the toughness of the film. The load elongation curves of varnishes No. 2, 5, 6 and 12 which have high tensile strength are plotted separately in Fig. 3. The tensile strength, elongation and toughness data are given in Table 1.
Modulus of Rigidity
The modulus of rigidity was also measured using the film as a torsion pendulum and results are reported in Table 1. These data were not useful in interpreting change of film properties and so are not further discussed.
The apparatus described in an earlier communication19 was used for determining the bursting strength of the varnish film. Steadily increasing air pressure was applied to a known area of the film. The pressure was indicated by a mercury manometer and the pressure at which the film burst gave the bursting strength expressed in terms of the height of mercury in mm.
In order to study the effect of ageing on mechanical properties of alkyd resin varnishes, their free films, supported on glass plates by pasting at the ends with cellulose adhesive tape, were exposed to natural weathering on racks at 45° facing south.
The mechanical properties of the varnish films were measured after 15, 30, 50, 70 and 90 days' exposure. On completion of each exposure period, the films were removed from the exposure rack and kept in a room maintained at 25 ± 2°C for 48 hours and their tensile strength and percentage elongation were determined. The data are graphically represented in Figs. 7.
Exposure to Carbon Arc Lamp
To study the effect of UV radiation on the mechanical properties of the film of alkyd resin varnishes under accelerated conditions, their free films were exposed to a carbon arc lamp in the Marr Fastness-to-light apparatus. The determinations of tensile strength and percentage elongation of the films were made after exposure periods of 30, 60, 100, 150, 200 and 300 hours and are graphically represented in Figs. 8 to 11. All determinations were carried out at 25 ± 2° C.
Results and Discussion
Tensile strength and elongation
The load elongation curves for the first six films are of similar form and are almost linear up to a certain value of stress (Fig. 1). This observation indicates that under low values of stress the extension of these varnishes is directly proportional to the load. The curves for the second set of films show slight variations in their form. These curves show that the extension of these varnishes is proportional to the applied load in a lower range of stress (Fig. 2).
From the observations of stress-elongation for the two groups of varnishes it may be concluded that there is a certain analogy between the behaviour of a varnish film and a strip of metal when subjected to tensile stress. The degree of recovery of varnish film depends mainly on the nature and extent of cross-linking in the surface coating material. The films differ from metals in that they are much affected by the duration of the stress or the number of time the stress is applied, and hence the ratio between stress and elongation fluctuates considerably with such factors and also changes with the age of the film and the type of exposure to which it is subjected.
Bosch et al. have summarized the mechanical properties of paints and varnishes as follows: (1) low elongation and low tensile strength mean hard brittle films liable to early failure, (2) low elongation and high tensile strength signify hard, tough films that are resistant to abrasion, (3) high elongation and low tensile strength result in flexible, soft and plastic films, (4) when both elongation and tensile strengths are high the film is flexible and tough and the film will have the best mechanical resistance.
In the case of linseed oil-phthalic anhydride-glycerol alkyds, both tensile strength and elongation increase considerably when the oil length is reduced from 66 per cent to 55 per cent. But in the case of pentaerythritol alkyds, the tensile strength increases from 23.1 kg cm-2 to 105.4 kg cm-2 and the elongation decreases slightly on reducing the oil length. The alkyds based on DCO do not show the same changes in their mechanical properties when the oil length is reduced from 66 per cent to 55 per cent.
The determinations of tensile strength and percentage elongation show that the oil lengths of the alkyds affect their general properties. The long oil length alkyd films are initially soft and flexible and have low toughness and hardness due to the high percentage of oil. The medium oil length alkyds appear to have a just sufficient amount of oil to impart a desirable flexibility, toughness and hardness, as is evident from the high initial values of their mechanical properties.
The load elongation curves for the second set of six alkyds are plotted in Fig. 2. Varnish no. 11 is very similar to varnish no. 1, except for the method of processing, varnish no. 11 being made by the fatty acid-oil process21 and varnish no. 1 by the alcoholysis process. Thus they may be expected to have similar mechanical properties. Varnishes no. 7 to 10 are modifications of varnish no. 1 and varnish no. 12 is the modification of varnish no. 2 Varnish no. 10, in which the cardanol-hexamine condensate was cooked in, was found to be very much inferior to varnish no. 9 where the modifications was by physical mixture. Varnish no. 10 was found to be inferior to varnish no. 1 also. Varnish no. 8 showed appreciable improvements in its properties as it attained high tensile strength and percentage elongation presumably due to the styrene modification. The modified varnishes no. 7 and 9 in general showed improvement in their mechanical properties (Table 1). The modification of varnish no. 2 by partial replacement of phthalic anhydride with styrenated rosin (varnish no. 12) was found not to have any appreciable effect on its mechanical properties.
The varnish films having high tensile strength and elongation were found to have high toughness as indicated by the area under the load-elongation curves. Amongst the varnishes studied, no. 6 was found to have the highest toughness.
Like tensile strength, elongation and toughness measurements, the bursting strength of varnish films can also be taken as one of the measures for the determination of the performance of coating. The initial values of bursting strength are reported in Table 1. It has been found that increase of oil length of the alkyd results in a decrease of bursting strength. Modification of the alkyd by incorporation of maleic-anhydride (varnish no. 7), styrene (varnish no. 8), cardanol-hexamine condensate as a physical mixture (varnish no. 9), and styrenated rosin (varnish no. 12) led to an increase in bursting strength. The pentaerythritol alkyds were found to have the greatest bursting strengths.
Natural Weathering and Exposure to Carbon Arc Lamp
Interesting information has been obtained from the data on the changes of the mechanical properties of alkyd resin varnish films subjected to natural weathering conditions and exposed to the carbon arc lamp.
The tensile strengths of varnishes no. 1 and 2 increase during natural ageing up to an exposure period of 20 days; afterwards there was a decrease in these values. These observations show that, after attaining maximum mechanical strengths, the film becomes brittle in the course of natural weathering and starts deteriorating. The tensile strength and the percentage elongation of 66 per cent linseed oil glycerine alkyd (varnish no. 1) films increase with ageing. Generally, in the course of ageing, the tensile strength of a coating increases and elongation decreases, but here both increase, showing that during ageing the toughness of the coating increases.
There is an improvement in the tensile strength of the 55 per cent linseed oil-glycerol alkyd film on ageing, but the percentage elongation sharply decreases. The films of this alkyd remained tought than those of the longer oil length alkyds throughout the exposure period. The results of a whole, however, suggest a deterioration of propertiesas on exposure. It would see m preferable to use a long oil alkyd as a medium for outdoor exposure as its mechanical properties improve on ageing. The pentaerythritol linseed oil alkyds have been found to possess better mechanical properties than linseed-glycerol alkyds.
In various mechanical properties, the medium oil length DCO alkyd (varnish no. 4) remained very similar to the 66 per cent linseed oil-glycerol alkyd (varnish no. 1), but it did not show as much improvement during ageing. The 66 per cent DCO alkyd (varnish no. 3) was much inferior to both of the above varnishes. Thus it may be concluded that, as far as mechanical properties are concerned, there is no particular advantage in the use of DCO in place of linseed oil in medium and long oil length alkyds.
Alkyd varnishes no. 7 to 9, which were obtained by modification of varnish no. 1, showed some improvement in the initial values of their mechanical properties. It was found (Fig. 6) that all the varnishes improved in tensile strength on weathering up to a certain period of time, after which the films became brittle and failed. Among the varnishes studied, varnish no. 7 showed the best performance and did not fail even after 90 days exposure. Varnish behaved similarly to varnish no. 1. Varnishes no. 8 and 10 attained the maximum tensile strength after 30 days ageing and failed immediately thereafter. Varnish no. 9, however, which also attained maximum tensile strength after 30 days, did not fail immediately, but slowly deteriorated and failed only after 70 days. Varnish no. 2 behaved similarly to varnish no. 9, the corresponding periods for maximum tensile strength and failure being 70 and 90 days respectively. On the other hand, varnish no. 12, which was a modification of varnish no. 2, behaved similarly to varnishes no. 8 and 10, failing immediately after attaining its maximum tensile strength at 30 days.
With regard to the change in the values of elongation of the varnishes on ageing, no uniformity is found. Varnishes no. 1, 3, 4, 7 and 11 showed increase in elongation up to 30 days, after which the elongation value decreased, but the elongation values of the other varnishes decreased continuously during ageing.
The weather data obtained from the meteorological department for the exposure period under study are given in Appendix-II.
The free films were exposed in the fastness-to-light chamber around the carbon arc lamp with a temperature in the vicinity of the varnish films of about 85° C. Varnish no. 1 attained maximum tensile strength of 90.0 kg cm-2 in 300 hours of exposure and varnish no. 2 attained a constant value of 146.6 kg cm-2 after 150 hours with a tendency to slight decrease on further exposure, whereas the same varnishes attained maximum tensile strength of 40.70 and 92.5 kg cm-2 respectively in 70 days under natural weathering. These observations show that the maximum tensile strength attained by the varnish films in natural weathering was comparatively less than that attained when exposed to carbon arc lamp. Both long and medium oil length DCO alkyds showed a constant increase in their tensile strength, varnishes no.5 and 6 showed better improvement in their tensile strengths, both in this test and under natural weathering conditions. Varnishes no. 1 to 6 followed a similar pattern with regard to their change of elongation in both the tests (Figs. 5 and 9).
Modification of Alkyds
No resinous polymer lends itself to more useful modification by other resins, both physically and chemically, than the alkyd type. To these blends, the alkyd contributes the vitally important properties of flexibility, toughness, adhesion, and durability.
Paints based on physical mixtures of alkyds with other resins provide the broad foundation for the major portion of modern industrial coatings. The combination of alkyds with urea/formaldehyde and melamine/formaldehyde resins is basic in appliance and automotive finishes. The upgrading of nitrocellulose lacquers by alkyd modification has enabled this oldest synthetic polymer to remain vigorously competitive with the newer coatings. Physical admixture of alkyds with chlorinated products (chlorinated paraffins, chlorinated rubbers) provide heavy-duty coatings for concrete floors, swimming pools, and corrosive environments.
Since all these paint systems are primarily physical mixtures, computations are normally relatively simple and straightforward. Resin formulation is mainly a matter of evaluating the mixed resin systems in which the percentage of alkyd content is systematically varied within conventional limits to establish an optimum balance between the alkyd and the modifying resin to meet any given set of requirements.
For this reason, only the chemical modification of alkyds will be considered, since chemical blending calls for somewhat involved computations.
Whether the modification is physical or chemical, the properties and performance of the blended system will be a reflection of the resins that make it up. Thus, an alkyd will be styrene-like in proportion to the amount of styrene it contains; or it will take on the properties of a silicone resin in proportion to the amount of silicone intermediate introduced.
Alkyd Modification with Styrene
Styrenated alkyds can be prepared by two main routes-a prestyrenation technique wherein one of the raw materials is styrenated prior to the main alkyd reaction; or a poststyrenation technique wherein the alkyd is reacted with styrene after the main reaction has been completed.
Of these two preparation routes, the poststyrenation procedure is generally preferred, as it gives better processing control and superior performance properties of the product. This in no way disparages the utility of styrenated oils, as such, as useful vehicles in their own right.
Since even partially polymerized styrene is incompatible with drying oils and alkyds, the key objective in styrenation is to chemically tie in at least a portion of the styrene to the alkyd polymer before the styrene monomer has had a chance to polymerize with itself to form an incompatible styrene homopolymer.
Furthermore, if the alkyd polymer prior to styrenation borders on a supermolecular size (shown by a very low acid number), it is almost certain that gelation or can instability will result, for the styrenation process will inevitably build up the polymer size to an uncontrollable dimension (to a gel). These two dangers must be kept constantly in mind in formulating styrenated alkyds.
Use of Conjugated Acids in Formulating Styrenated Alkyds
DCO, tung, and oiticica oils are rich in conjugated fatty acids; hence these drying oils are the ones commonly used in formulating styrenated alkyds. Maleic anhydride is hardly permissible in this type of formulation, for this unsaturated diacid would preempt the bulk of the reactive conjugated oil sites during the alkyd reaction, leaving insufficient residual conjugated unsaturation behind for adequate poststyrenation.
In formulating this type of alkyd, there is frequently a tendency to be overgenerous in supplying conjugated double bonds to the reaction mixture. Such oversupply is not necessary and in fact dilution with soya oil down to a 3 soya oil to 1 DCO weight mixture has been proposed for routinely preparing phthalate alkyds suitable for poststyrenation. A working alkyd constant of 1.04 is suggested here rather than 1.01, to allow for the viscosity boost imparted by the styrene addition. Table 1 gives an alkyd composition suitable for poststyrenation.
In processing this alkyd, the reaction must be terminated short of an acid number of about 15 to avoid excessive viscosity on subsequent poststyrenation.
If higher percentage of conjugated acids is required in the base alkyd, a correspondingly higher margin of safety must be formulated. This can be conveniently done by raising the alkyd constant to a higher value. The all DCO alkyd submitted in Table 2 is suitable for experimental poststyrenation. The high working alkyd constant of 1.09 compensates for the high degree of conjugation. Again the alkyd reaction should be terminated shy of an acid number of 15 to allow proper poststyrenation.
Use of a Diacid with a Reactive Double Bond
(Other than MA) in formulating styrenated alkyds
Of the several diacids with reactive double bonds that have been evaluated for the formulation of poststyrenated alkyds, only a few appear to be suitable. Of these, the maleic adduct of cyclopentadiene has received the most publicity. Unfortunately, of the two suppliers of this material, one has withdrawn its product from the market (Carbic Anhydride), whereas the other is presently offering its product (Nadic Anhydride) at over $1.00 per pound, a price that is uneconomical in most alkyd formulations. Further consideration of this material does not appear justified.
Use of Maleic Anhydride in Formulating
In the section on conjugated acids, the fact was stressed that for poststyrenation, maleic anhydride should be avoided. Conversely, when maleic anhydride is made the basis for formulating an alkyd for poststyrenation, it must be used exclusively with nonconjugated acids (2 per cent conjugation is probably the maximum that can be tolerated in this system).
The role of maleic anhydride is quite different during the alkyd and styrenation reactions. During the alkyd condensation reaction (esterification), maleic anhydride functions mainly as a diacid, whereas during the styrenation reaction it functions strictly as a source of unsaturation.
The percentage of maleic anhydride in the alkyd formulation is quite critical. Too low a percentage fails to provide sufficient double bonds for a tie-in of the styrene monomer, leading to a cloudy, or worse, and incompatible system. Too high a percentage furnishes an overabundance of double bond sites, leading to an overpolymerized or highly cross-linked gel. A formulating guide has been proposed to avoid these dangerous extremes. It is postulated that conditions for the poststyrenation are optimum (for a phthalate alkyd) when the alkyd polymer has a maleic functionality of. This is equivalent to saying that when 1 out of 3 polymer molecules in the alkyd provides a maleic-contributed double bond for poststyrenation, a clear homogeneous styrenated alkyd can be obtained which will exhibit a satisfactory and stable viscosity.
The maleic functionality of an alkyd is computed and expressed as follows: Consider an alkyd prepared from monoacid (nonconjugated), maleic anhydride, other diacid (phthalic anhydride), and polyol.
The replacement of PA by MA to permit poststyrenation depends to a major extent on the acid number to which the alkyd is cooked. Let acid numbers of 10 and 20 be selected as target values for the computation work. Substitute appropriate values in Eq. 3.
Alkyd Modification with Rosin
Since rosin is generally abietic acid, it is treated simply as another acid available for the alkyd reaction. In moderate proportions, rosin renders alkyds more soluble in aliphatic solvents, inhibits the onset of gelation (permits the use of a lower alkyd constant), improves adhesion, enhances gloss, reduces any tendency towards wrinkling, and increase the resistance of alkyds to aqueous soap and alkali solutions. However, rosin detracts from color, flexibility, toughness, and over-all durability; hence it must be used judiciously.
Alkyd Modification with Phenolic Resins
The modification of alkyds with phenolic resins is based largely on practical experience. This is partly because the exact structure of most phenolic resins is not known with any degree of certainty, and partly because the chemical reaction of a phenolic resin with an alkyd is not fully understood. Hence, with an unsure starting point and an indeterminate chemical reaction, the formulation of phenolic modified alkyds does not presently lend itself to precise theoretical treatment.
Mot novalac-type phenol/formaldehyde resins and practically all rosin-modified phenolics are physically compatible with alkyds. Unfortunately their admixture with the alkyd generally results in an intolerable reduction in durability. Resoles prepared from unmodified phenols thermoset too rapidly to permit chemical reaction with alkyds. However, modified phenols, in which one or more of the reactive positions in the aromatic ring are blocked by alkyl groups (say p-tert-butylphenol) give phenolic resins that react readily with alkyds (as they do with varnish oils) to give satisfactory modified alkyds.
Usually, the modification is held to a low percentage, 5 per cent being most common and 20 per cent being extreme. Presumably the modified phenolic resin reacts with the unsaturation of the fatty acids present in the alkyd composition, but side reactions also probably take place and contribute to the ultimate chemical tie-up.
Within the percentage modification noted, substantial improvement in resistance to water, aqueous solutions (alkaline or acid), and hydrocarbons can be effected with no appreciable reduction in durability.
As with any modification in which molecular size is enlarged by polymerization, a viscosity increase must be anticipated and allowance made to maintain a manageable viscosity.
Alkyd Modification with Silicones
Silicone-alkyds have enjoyed wide acceptance, especially for heat-resistant coatings, ever since their first appearance in the journal literature in 1947. As a result, the technology of a silicon-alkyd preparation for high-temperature service (400-550 F) is now fairly well established. In general, the formulation of superior silicone alkyd calls for a compromise composition. Basically, the silicone contributes thermal stability, gloss and gloss retention, nonyellowing, and solvent and chemical resistance, whereas the alkyd contributes flexibility, impact resistance, and freedom from crazing.
A silicone content somewhat in excess of 50 per cent is generally accepted as achieving optimum over-all properties for high-temperature service. The choice of alkyd ingredients as well as the selection of the silicone intermediate for the copolymerization reaction markedly affect the end performance properties.
Dimethyltriphenyltrimethoxytrisiloxane (20 per cent methoxy content) having a molecular weight of 470 and an E of 155 provides good alkyd compatibility and superior flexibility, adhesion, and impact resistance, but the product suffers from reduced thermal stability and gloss retention.
The functional concepts that have been previously applied with considerable success to the design of alkyd compositions appear to apply one casually to silicone intermediates. Presumably silicone polymerization takes a somewhat different course, with ring formation and intramolecular condensation competing with the chain and the intermolecular cross-linking type of condensation normally associated with an alkyd reaction.
Before considering silicone copolymerization, it is expedient to develop some generalized equations relating the variables of a self-condensation reaction. Consider the case of a single reactant that self-polymerizes to form a homopolymer.
Then m0F0 equals the initial number of equivalents present in the reaction mix. Let self-polymerization take place in which a given fraction f of the reactive groups is consumed (reacts). Assume that one mole of reactant disappears (by merger with another molecule) for each two equivalents that react. This assumes that the condensation proceeds by linear chain formation and by intermolecular cross linking (excludes ring formation and intramolecular condensation). This is, of course, an idealized type of condensation, which applies remarkably well to alkyd heteropolymers but, as will be shown, not so well to silicones. However, this assumption still provides a criterion for estimating the deviation of the silicone polymerization from an academic ideal.
Alkyd Modification with Formaldehyde
The effect of adding formaldehyde to an alkyd composition (in the form of p-formaldehyde or formalin) has been found to result in the formation of cyclic or inner formals, which leads to a reduction in the functionality of a polyol, rather than in a cross-linking reaction between two different polyols, which leads to polymerization.
It is therefore postulated, for purposes of alkyd formulation, that for each formaldehyde molecule (CH2O) present in an alkyd composition, two -OH groups will be tied together on a single polyolmolecule. This formal formation is strong enough to resist deformalization at alkyd processing temperatures.
Copolymerization of Alkyd Silicons for Coatings
Alkyd-Silicone coatings are comparatively new, being first mentioned in the literature in 1947. Patterson reviewed the properties of alkyd-silicones and reported that they are intermediate between alkyd-melamine and pure silicone enamels in heat and alkali resistance, adhesion, hardness and toughness. He also reported that varnishes made by chemical cocondensation of alkyds and silicones are usually superior to those made from cold-blend mixtures of the two.
Practically the only details of the synthesis of alkyd-silicone varnishes which appeared in the literature until early in 1952 were those disclosed by Bowman in their British patent. They heated oil-modified alkyd resins with organosilanols in a solvent reflux process. The several patents issued since the early part of 1952 prepare alkyd-silicone varnishes either by the reaction of an alkyd resin having excess hydroxyl groups with an organoalkoxysilane or by reacting the silane with glycerol and then reacting this intermediate with an acidic compound or an acidic ester. The one exception found was in the patent of Millar who used a process similar to that of Bowman and Evans.
The varnishes were prepared in ordinary round-bottomed, three-necked flasks heated with an electric mantle. The reactions were run at 200°C, under a carbon dioxide atmosphere, and with agitation.
The standard method for making the varnishes was to weight the desired quantities of dibasic acid, fatty acid and glycerol into the flask and heat at the maximum rate to 200°C. Samples of the alkyd reactants were withdrawn at 1/2 hour intervals to determine the acid number. When the acid number dropped from approximately 300 at the beginning of the reaction to less than 10, the organoalkoxysilane was added. The two-phase mixture was then checked for clarity at 5-minute intervals. A clear homogeneous cold pill was usually obtained after 15 minutes. The reactions was continued at 200°C, until gelation was imminent. This was determined by the cessation of cavitation around the stirrer. At this point the reaction was stopped by reducing the resin with high flash naphtha to approximately 50% solids. After cooling, the solids content of the varnish was adjusted to 50%.
A simple enamel formulation of varnish and rutile titanium dioxide in the ratio of 1:1 on a solids basis was used. Enamels were satisfactorily prepared both on ball mills and roller mills. The majority of enamels was made on a laboratory three-roll mill because of the versatility and speed of the mill. High flash naphtha was added to the enamel to obtain a viscosity of 30 seconds as measured with a No. 4 Ford cup at 80° F. The finished enamel was then centrifuged in a cup centrifuge at 2500 r.p.m. to remove any oversize pigment particles.
The finished enamels were sprayed on S.A.E. 1010, 20-gage cold-rolled steel panels and plate glass panels. The steel panels were degreased and treated with metal Prep, a commercial phosphate solution for preparing steel surfaces for enameling. The glass plates were washed with acetone.
The enamels were sprayed onto the panels to obtain a dry film thickness of 1.9 ± 0.1 mils as measured with a magnetic film thickness gage. Commercial film thickness range from 1 to more than 2 mils according to the desired amount of hiding. This thickness was chosen for this work in order to get optimum gloss retention and excellent hiding (see Figure 3). The films were cured for 1/2 hour at 400°F., and this bake was considered the initial point in the testing.
The enamels were tested for effect of film thickness on enamel properties and for gloss and color retention, craze life, toluene resistance, alkali resistance, impact resistance, flexibility surface hardness, adhesion and general appearance.
In discussing experimental results, the formulation nomenclature used should be kept in mind. It was assumed that the resins were composed of glyceryl siloxane and alkyd resin. The composition of the resins was then defined in terms of the silicone content (the percentage by weight of glyceryl organosiloxane in the totally reacted resin) and the oil length of the alkyd (the percentage by weight of fatty acid triglyceride in the alkyd portion. This method proved much more useful for correlating the results than did an equivalency basis. An additional benefit was that it is an adaptation of alkyd terminology and is therefore familiar to the coatings industry.
This formulation gives 87.5 grams of glycerol phthalate, 87.5 grams of glycerol trilaurate, and 175 grams of glycerol phenyl polysiloxane. Theoretically, 15.5 grams of water and 42.5 grams of ethanol should be split out by condensation. These calculations are based on phenylethoxypolysiloxane having an equivalent weight of 204. This equivalent weight was based on the ethoxy content of the silicone-in this case an ethoxy to silicon ratio of 0.80. It was assumed that all the ethoxy groups were available for reaction with glycerol. Other resins were formulated by determining the desired amounts of the reacted glycerol phthalate, glycerol trilaurate, and glycerol phenylpolysiloxane and calculating the required amounts of reactants from the chemical equations of the reactions.
Development of Varnish Procedure
Several attempts were made to prepare varnishes according to the method of Bowman. This consists of heating by a solvent process, an oil-modified alkyd resin having an acid number of approximately 40 with organosilanols. Clear varnishes were obtained using as the silicone intermediates phenyl, amyl-, nonyl-, and ethyl-trichlorosilanes hydrolyzed to the silanols. However, enamels made from these varnishes had poor gloss and only fair color retention. It is believed that little copolymerization was obtained because of the great tendency of the silanols to condense to silicones and their small tendency to react with the alkyd resin.
Organotriethoxysilanes were also used as the silicone intermediate. These compounds had more of a tendency to react with excess alcohol in the alkyd portion but were not very satisfactory because they were relatively volatile and it was difficult to remove the water and ethanol of condensation without losing some of the silicone. The resulting products were not too satisfactory.
For these reasons organoethoxypolysiloxanes were used in most of the work done in this investigation. These were formed by partially hydrolyzing and condensing organotriethoxysilanes to form low molecular weight silicone polymers containing residual ethoxy groups capable of reacting with the hydroxyl groups of alkyd resins. These compounds are nonvolatile, require less excess glycerol based on organic acid content and when reacted with alkyd resins connect them to stable silicone nuclei. Most of the silicone intermediates used had ethoxy-to-silicon ratios of 0.8.
Effect of Order of Addition
Homogeneous varnishes were obtained either by cooking all the alkyd and silicone ingredients together throughout the reaction or by forming the alkyd resin first and then reacting this with the silicone. Distinct differences in properties resulted from the two methods of cooking.
In one case, the phthalic anhydride, lauric acid, phenyl-ethoxypolysiloxane and glycerol were loaded the reaction flask at room temperature and heated to 200°C under agitation and an inert atmosphere. This temperature was maintained until gelation was imminent. The reaction was then stopped by adding high flash naphtha to the resin. The varnish had a color of 2, a viscosity of A (Gardener Holt) and an acid number of 75.
In the second case, the phthalic anhydride, lauric acid and glycerol were placed in the reaction flask at room temperature, heated to 200°C under agitation and an inert atmosphere and held at this temperature until the acid number had decreased to 11. At this time the phenylethoxypolysiloxane was added, 200°C. was regained, and the batch was held at this temperature until gelation appeared to be imminent. The resin was then thinned with high Flash naphtha. This resin had a color of 6, a viscosity of Cl and an acid number of 7.
These varnishes were numbered 62 and 6, respectively. The total cooking time of No. 62 was 30 minutes, and the alkyd reaction of No. 6 was 135 minutes with the reaction continuing for 15 minutes after addition of phenylethoxypolysiloxane. Enamels were prepared from these varnishes and tested. The results are given in Table 1.
The difference in alkali resistance is to be expected from the acid numbers of the varnishes. The large difference in gloss is typical of enamels prepared from varnishes cooked by the two procedures. Enamels prepared from varnishes cooked according to the technique used for varnish No. 62 always chalked very badly. The tendency was much less pronounced when the alkyd resin was formed first and then reacted with the organoalkoxypolysiloxane. Apparently the varnish coating the surface layer of pigment particles decomposes, leaving a chalklike layer of dust on the surface of the enamel. One explanation for the increased decomposition of varnishes having high acid numbers is that the phthalic half ester can easily revert to the alcohol and phthalic anhydride under the influence of heat, while the fully esterified phthalate does not depolymerize as easily. Another factor contributing to the poor gloss retention is that varnish No. 62 is probably not copolymerized to the extent that varnish No. 6 is copolymerized. The water of esterification can hydrolyze the ethoxy groups of the siloxane to silanols which tend to condense to the silicone structure. As will be shown later, mixtures of alkyds and silicones tend to be inferior in gloss to copolymers.
The effect of temperature was investigated by cooking varnishes at 190°, 200°, and 230°C. Resins cooked at 190°C. tended to be darker than those cooked at the other temperatures because of the long reaction time-200 minutes at 190°C. compared to 118 and 36 minutes at 200°C and 230°C., respectively. The silicone reaction is so fast at 230°C. that it is difficult to control. Therefore, the reactions were run at 200°C. As a result of recent work, it is believed that the best technique is to cook the alkyd at approximately 200°C. and reduce the temperature to approximately 165°C. for the silicone reaction. It is possible to control the reaction's end point by viscosity measurements with this technique.
Evidence of Copolymerization
The statement that copolymerization occurred in the varnish reactions is based on the following evidence.
When the organoalkoxysilane is first added to the alkyd resin in the reaction vessel, the mixture is incompatible and samples of the mixture are definitely two phase. As the reaction proceeds, the reaction mass becomes progressively clearer until finally cold-pill samples are completely clear and homogeneous. Either solubility is increasing as reaction proceeds in a highly functional reaction mass or the resin is becoming homogeneous because of copolymerization. The latter is much more probable.
Another line of evidence is based on Flory's theory of gelation. This theory states that gelation occurs when a rigid lattice work formed by primary valence bonds extends throughout the reaction mass, immobilizing the mass and causing a large increase in viscosity. If an alkyd resin contains sufficient monobasic acid and excess glycerol to form only linear polymers, gelation will not occur. If organoalkoxysiloxane is added to this mass and gelation occurs, either the silicone itself or a copolymer of the alkyd and the silicone is responsible for gelation. If a copolymer is formed, it is reasonable to expect that the more highly the alkyd resin is reacted before addition of the silicone, the faster gelation will occur if copolymerization is taking place. Table 2 gives the reaction time for three different varnishes. The more highly reacted the alkyd resin was before addition of the silicone, the shorter the time for gelation. This strongly suggests that copolymerization was occurring.
A third line of evidence is a study of the possible reactions. Phenylethoxysiloxane was heated by itself and with dioctyl phthalate without evolution of ethanol or evidence of further polymerization of the siloxane. Therefore it is itself stable and stable in the presence of esters. However, when heated with glycerol, ethanol rapidly split out and gelation occured. Figure 1 shows the condensate collected versus cooking time for an alkyd-silicone resin. The alkyd portion of the resin was cooked for 108 minutes at 200°C. At this point, the acid number of the resin was 10 and the rate of removal of water was zero. It is probable that the alkyd resin mass consisted essentially of alkyd esters and unreacted glycerol hydroxyl groups. Phenylethoxy, siloxane was added to this system which decreased the temperature to 160°C. As the temperature began to rise, ethanol split out of the reaction mass at an increasing rate. The only explanation for this evolution of ethanol is that it was split out by the reaction of phenylethoxysiloxane and the unreacted hydroxyl groups in the alkyd resin. Therefore, copolymerization was occurring.
Decomposition of Varnishes
One method for determining the heat stability of a resin is to measure its per cent weight loss when heated at a certain temperature. The weight losses of the resins prepared in this work are shown in Table 3. The value reported are the average of three determinations made by spraying the resin solutions on standard panels at 1-mil dry film thickness and measuring the weight % loss versus time baked at 450° F.
The weight loss depends primarily on the silicone content of the varnishes. This is best illustrated by comparing the weight loss of varnishes 810, 14, 15, 1, and the Dow Corning silicone mixture of 40% DC802 and 60% DC804. These varnishes contain 0, 25, 35, 50, and 100% silicone, respectively. Other factors that determine the weight loss are the polyol (ethylene glycol varnishes lost far more weight than equivalent glycerol ones) and whether the alkyd-silicone is a mixture or copolymer. Varnishes 13 and 20 compared to varnish 1 exemplify this statement. Apparently varnishes containing phenyl and dimethyl ethoxypolysiloxane are not as stable as those containing phenylethoxypolysiloxane as shown in varnishes 16 and 1 in the table.
Decomposition products of resins heated to 475° F. were collected. The principal product collected was phthalic anhydride. The other product was a yellow oily residue. This was unsaturated and did not contain carbonyl groups. Further analysis was unsuccessful. A small amount of water and ethanol was also collected.
The physical properties of the varnishes are given in Table 3. The procedure used in cooking the varnishes gives low viscosities and low acid numbers. Plaskon ST856 must be decidedly different because of the viscosity. The colors of the varnishes were, in general, light but were darker when the oil length was longer. Varnishes with low oil lengths were slightly more viscous than the others.
It was found that the film thickness of the enamels exerted a significant influence on the properties of the enamels. The variation in yellowing, craze life, and gloss of the enamels with film thickness is shown in Figure 2. The large variations in properties make it necessary to control the thickness of the films of the test enamels as accurately as possible. Three panels of each enamel were coated with films 1.9 ± 0.1 mils thick, evaluated, and the average values obtained were reported.
Gloss and Gloss Retention. Gloss values were measured with a photo-volt gloss meter that measured 60° specular gloss. Practically all declines in gloss occurred during the first few hours of baking at 400° F. Enamels attained their ultimate gloss after 100 hours. Therefore, gloss values of the enamels after baking for 1/2 hour and for 100 hours represent initial and final gloss. These two values and the change in gloss are given in Table 4.
The two alkyd-silicone cold-blend mixtures reported (V-20 and V-25) had very poor gloss retention. This was probably due to increased decomposition of the mixtures in comparison with the polymers and to incomplete homogeneity of the resins.
The gloss rentention of the enamels became worse as the silicone content was decreased and the fatty acid content increased Fig. 3).
Craze Life. In this investigation, craze life is defined as the length of time in hours that an enamel can be baked at the destinated temperature without film failure by cracking, checking, crazing, or loss of film integrity in any way. Craze life is very greatly influenced by the baking temperature. All the enamels ested had craze lives in excess of 400 hours at 350° F. and less than 3 hours at 500° F.
Styrene Copolymers in Alkyd Resins
Styrene is becoming an increasingly important raw material for use in organic surface coatings. However, at the present time it is not used in coatings as monomer nor as polystyrene but rather as copolymers with such materials as butadiene, drying oils, and most recently with alkyd resins. The volatility of the monomer and lack of compatibility of the polymer are the principal deterrents from the use of these materials as such. The advantages obtained from the styrene-copolymerized oils and alkyds are faster drying, harder film, and better water and chemical resistance than can be obtained with the straight oils or alkyd resins. However, the copolymers of styrene and various materials retain some of the sensitivity of polystyrene to certain hydrocarbon solvents. The tremendous production capacity for styrene, resulting from its extensive use in the synthetic rubber program during the last war, its relatively low cost, and very high degree of purity make it of definite interest for the surface-coating industry.
One of the first methods proposed for the reaction of styrene with a drying oil was disclosed in a British patent in 1931. This describes the polymerization of an aqueous emulsion of styrene and tung oil with hydrogen peroxide as catalyst. The next development was the use of the solvent method for copolymerization of styrene and film-forming materials in inert solvents in 1934. This work was investigated further by Wakeford and Hewitt, Wakeford, Hewitt, and Armitage, and Wakeford, Hewitt, and Davidson from 1942 onwards in a number of British patents. The mechanism of copolymerization between styrene and various drying oils is described by Hewitt and Armitage and the effect of various solvents was studied by Armitage, Hewitt, and Sleightholme. They used as a standard formula 50 parts solvent, 25 parts oil, and 25 parts styrene without catalyst. They also applied the same method for styrenation of a prepared alkyd; however, the method requires about 30 hours for reaction.
Dunlap and Wakeford, Hewitt, and Armitage in 1945 investigated the mass method of copolymerizing styrene and various drying oils. The mass method is much faster than the solvent method but only limited amounts of styrene can be copolymerized and still maintain homogeneous products. In this country the Dow Chemical Company developed a system for obtaining homogeneous products by the mass method by replacing part of the styrene with a-methylstyrene. This combination produced homogeneous products with drying oils containing a conjugated system of unsaturation. With oils such as linseed and soya, it is recommended that blends be made with tung or dehydrated castor oil to introduce some conjugated unsaturation. However, the use of a-methylstyrene has been found to detract from the fast drying time and also to reduce the resistance to solvents and chemicals. It is obvious that it would be desirable to use the mass method because of its speed of reaction and to avoid the use of a-methylstyrene, if possible. The present work knows that this may be done by first styrenating the fatty acids allowed by esterification with phthalic anhydride and glycerol to produce a styrenated alkyd resin.
Styrenation of Fatty Acids
The copolymerization reaction between styrene and drying oil fatty acids depends on the type of the fatty acids used. With the conjugated fatty acids, like tung and oiticica fatty acids, the reaction is believed to be similar to that between styrene and butadiene in GR-S manufacture. In this case, styrene is joined to butadiene by 1, 4- and 1, 2-additions. With dehydrated castor oil fatty acid and isomerized linseed fatty acid (diene value 22 and 20, respectively) it is believed that there is some polystyrene formed along with the copolymer. Armitage, Hewitt, and Sleightholme have stated that polystyrene of high molecular weight was not formed, but they had not proved that polystyrene of low molecular weight was absent. There is every reason for believing that polystyrene of low molecular weight might be present.
With linseed fatty acid, where there is a nonconjugated double bond system, the copolymerization reaction does not take place to any appreciable extent. This can be seen from other similar systems such as vinyl chloride (nonconjugated) and styrene (conjugated) which does not copolymerized. In general, a monomer containing a conjugated double bond system will copolymerize with another molecule containing a conjugated double bond. However, Armitage have recently suggested that copolymerization with nonconjugated systems might take place by the shift hydrogen mechanism in special circumstances.
The fatty acids used were those commercially available; tung from Archer Daniels Midland Company and oiticica, dehydrated castor oil, and linseed from Woburn Degreasing Company. Styrene was obtained from the Dow Chemical Company and it was found from a few experiments that it was not necessary to remove the inhibitor. Benzoyl peroxide was used as catalyst.
The copolymerization reaction between styrene and tung oil fatty acid, oiticica oil fatty acid, dehydrated castor oil fatty acid, linseed oil fatty acid, or isomerized linseed oil fatty acid was carried out by the mass method. In all cases 3% benzoyl peroxide on the weight of styrene was used.
The fatty acid and styrene with catalyst were placed in a 4-necked flask and heated to 145° C. by an electric mantel. Through the central neck a stirrer with mercury seal was attached. In the three side necks were attached a thermometer, a condenser, and an arrangement for withdrawing samples. In the case of dehydrated castor oil, linseed, and isomerized linseed fatty acids a separatory funnel was attached in one of the side necks through a Y-bend connection, and styrene catalyst mixture was added slowly to the fatty acid in the reaction flask. Samples were withdrawn at various intervals of time and the per cent styrene which had reacted was determined by removing the unreacted styrene using vacuum distillation (6 to 8 mm., 40° C.).
The amount of styrene which had reacted with the fatty acid was calculated from the difference between the known percentage of styrene present originally and the percentage found after vacuum distillation. This value was checked from acid number determinations. This value was checked from acid number determinations. The ratio of actual acid number to the theoretical of 195 for fatty acid indicates the per cent fatty acid present.
The value for the per cent styrene in the product as determined by the vacuum distillation method checks rather closely with that obtained by the acid number method.
Copolymerization of styrene with tung, oiticica, and dehydrated castor oil fatty acids was carried out using various molal ratios of styrene. The rate of reaction increased with increased ratio of styrene and the values are plotted in Figure 1. The corresponding values for per cent styrene reacted and actual and theoretical acid numbers are given in Table 1.
Attempts were made to styrenate standard linseed fatty acids but the products obtained were heterogeneous which is related to the nonconjugated unsaturation in this fatty acid. The isomerized linseed acids shown in Table 1 have a diene value of 20 and consequently produced homogeneous products after styrenation.
The styrenated products containing less than about 70% styrene are clear viscous liquids, with increasing viscosity as the styrene content is increased. Above 70% styrene content the products are hard resins. These styrenated products are mixtures of mutually soluble materials which include the copolymer of styrene and fatty acid in the largest amount and considerably smaller proportions of polystyrene and free fatty acid.
Rate of Styrenation of Fatty Acids
The general average values for rate of styrenation of a series of experiments are plotted in Figure 1; a comparison of these curves shows the slowest reaction rate for tung fatty acid, a somewhat faster rate for oiticica, and the fastest rate for dehydrated castor oil. These differences in rate apply at all ratios of styrene to fatty acid but in each case a faster rate is obtained by increasing the amount of excess styrene. Fast reaction rates are always desirable in commercial operations and these may be obtained by using about 4-mole ratio or 2-mole ratio of excess styrene.
The saving in time would be offset somewhat by the added cost of removing a larger quantity of free styrene and by the larger reactor capacity which would be required.
The rates for the three fatty acids are in the reverse order from what might be expected, from a consideration of the fact that tung contains the greatest percentage of conjugated unsaturation and dehydrated castor oil contains the least. This reversal of the order of the rate of reaction may be due to a smooth copolymerization reaction with the tung acids and a relatively low molecular weight product. The ketonic group in the oiticica fatty acid may be expected to exert an accelerating effect on the copolymerization reaction, hence a faster rate than the tung acid. In the dehydrated castor oil acid reaction there is formed, most probably, some polystyrene in addition to the copolymer. Because the rate of reaction for the polymer is faster than that for the copolymer, the over-all ratio is higher as shown. Although the styrenated product from the dehydrated castor oil reaction was sufficiently homogeneous for normal use in alkyd resins it was not so smooth as the others, indicating some polystyrene present.
A marked difference in rate of reaction and type of product formed could be obtained by varying the manner in which the styrene was added to the dehydrated castor oil fatty acid. When the styrene and dehydrated castor oil fatty acid are heated together a rapid, exothermic reaction develops when the temperature reaches 120° C and the product is very viscous and turbid. However, if the styrene is added to the fatty acid at a slow rate the reaction is smooth and the product is clear and apparently homogeneous. This method was adopted in making the styrenated fatty acids for the alkyd resins described later because it yields a homogeneous product with a minimum of polystyrene and a maximum of the copolymer.
Studies on Blends of Polystyrene Glycol and Alkyds in Surface Coatings
Polystyrene as such is not compatible with oil modified glyceryl phthalate resins, commonly known as alkyds. But, if polystyrene is made reactive by introducing functional groups, it can either be chemically reacted or physically blended with alkyds in small amounts. Blending of polystyrene glycol with a drying oil alkyd, viz., linseed oil alkyd was investigated earlier' and significant improvements in the film properties of the alkyd were observed. The present work describes the study of film properties of polystyrene glycol blends with nigerseed oil and castor oil alkyds. Nigerseed and castor oils are of semi- and non-drying type oils respectively. Since semi-drying and non-drying oil alkyds do not air-dry, they are usually mixed with amino resins and stoved. Therefore, butylated urea formaldehyde resin has been mixed with the alkyds and the polystyrene glycolalkyd blends.
Styrene (inhibitor free), benzoyl peroxide, dioxane, potassium hydroxide, methanol (all, L.R. grade) were used for the preparation of polystyrene glycol which has further purified by precipitation method using benzene and methanol (both L.R. grade). Butanol, acetic anhydride, and pyridine (all A.R. grade) were used for the determination of hydroxyl value of polystyrene glycol.
Phthalic anhydride, glycerol (both L.R. grade), alkali refined castor and nigerseed oils were used in the preparation of alkyds.
Urea, formaldehyde (37 per cent) and n-butanol (all L.r. grade) were used for the preparation of butylated urea formaldehyde resin.
Benzene, xylene, methanol (all L.R. grade) were used as solvents.
Preparation of polystyrene glycol
Polystyrene glycol was prepared by free radical polymerisation of styrene using benzoyl peroxide as initiator and subsequent hydrolysis of benzoate end groups to hydroxyl groups. The prepared sample had a hydroxyl value of 289.
Preparation of alkyds
Castor oil alkyd of oil length 35 was prepared by direct heating of the mixture of alkali refined castor oil, glycerol and phthalic anhydride at 220-230°C. The heating was continued until an acid value of about 10 was achieved. Nigerseed oil alkyd3 of oil length 40 was prepared by the monoglyceride process in which litharge (0.1 per cent by weight of oil) was used as a catalyst. Phthalic anhydride was added at 180° after monoglyceride formation and then the temperature was raised to 230°C. The heating was continued until an acid value of about 10 was obtained.
Preparation of butylated urea formaldehyde resin
Butylated urea formaldehyde resin was prepared in the laboratory by reacting 1 mole urea with 3 moles formaldehyde (37 per cent solution) at 93° C for one hour at pH 7.5. PH was brought down to 5.5 by adding phosphoric acid and then 2 moles n-butanol was added. The mixture was heated with stirring and reaction was continued until the calculated quantity of water of reaction was collected through Dean and Stark apparatus. The resin so obtained was clear and water white. Solid content was adjusted to 60 per cent by adding more n-butanol.
Preparation of polystyrene glycol-alkyd blends
Blends of polystyrene glycol with nigerseed oil and castor oil alkyds were prepared by adding powdered polystyrene glycol to the alkyd with agitation at a temperature of 100-120°C. Stirring was continued for about 1 hour. The maximum quantity of polystyrene glycol that gave clear miture with each alkyd was determined. In addition to these samples, other samples having lower amounts (about half of the maximum amount found compatible) of polystyrene glycol were also prepared.
Curing of polystyrene glycol-alkyd blends and plain alkyds
Butylated urea formaldehyde resin was mixed with polystyrene glycol-alkyd blends and plain alkyds. Usually, 25 per cent amino resin is mixed with alkyds. Therefore, 25 per cent butylated urea formaldehyde resin was added to all the samples of polystyrene glycol-alkyd blends and plain alkyds. In addition to this, in order to minimise the use of amino resin, samples containing 10 per cent butylated urea formaldehyde resin were also prepared. Butylated urea formaldehyde resin solution was added to the alkyds and the polystyrene glycol-alkyd blends at room temperature (25° C) under continuous stirring until a clear mixture was obtained.
All the samples were thinned with xylol to brushable consistency and films were applied on 6in × 2in glass and tin panels. Films of nigerseed alkyd and its polystyrene glycol blends were based at 120° for 40 minutes while the films of castor alkyd and its polystyrene glycol blends were baked at 120°C for 50 minutes. All the baked films were hard, smooth and glossy.
Baked films of all samples were tested for scratch hardness, flexibility and adhesion, water, acid, alkali and solvent resistance.
Results and Discussion
The composition of various blends of polystyrene glycol and nigerseed oil and castor oil alkyd is shown in Table 1. The amount of urea formaldehyde resin mixed with the blends and plain alkyds is also shown in this table. The physical properties, viz., scratch hardness and flexibility of the baked films are shown in Table 2. Water (cold and boiling), acid (hydrochloric, sulphuric and nitric) and alkali (sodium carbonate) resistance are given in Table 3, 4 and 5 respectively.
Compatibility of polystyrene glycol with alkyds
It was observed that when a higher proportion of polystyrene glycol was incorporated into alkyds, a hazy mixture was obtained and polystyrene glycol separated out on standing. Therefore, the maximum amount of polystyrene glycol that gave a clear mixture with each alkyd was determined. It was found that 11 per cent and 12 per cent polystyrene glycol was compatible with nigerseed oil alkyd of oil length 40 and castor oil alkyd of oil length 35 respectively.
Scratch hardness test was carried out by using a mechanically operated Sheen scratch hardness tester in which a hardened needle loaded with one kilogram weight moves over the film. All the samples of the polystyrene glycol-alkyd blends and plain alkyds passed the test showing good scratch hardness.
Maximum scratch hardness was also determined in each case by placing increasing load over the hardened needle. Maximum scratch hardness (in grammes) of all samples is shown in Table 2. It is clear that scratch hardness of polystyrene glycol alkyd blends is more than that of plain alkyd and further, it increases with increasing amount of polystyrene glycol. It is so, because large numbers of cyclic (benzene) rings present in polystyrene glycol contribute towards the hardness.
Flexibility and adhesion
Flexibility and adhesion of the dried film was tested on tin panels. The test was carried out by bending the tin panel in ¼ in diameter mandrel. All the samples of blends of polystyrene glycol-alkyd and plain alkyds passed the test as no detachment of film from the substrate or crack in the film was observed. Thus, all the films had good flexibility and adhesion. Further, it confirmed that the amount of polystyrene glycol blended was compatible with alkyds.
Films were applied on 6in × 2in glass panels and baked as described earlier. The sides of the glass panels were protected by wax before performing this test. The panels were immersed in cold distilled water at room temperature (25° C) for 48 hours and were taken out. The dipped portion of the films was allowed to dry and examined for appearance, loos in gloss, change in colour, and other visible damages. It was found that all the samples were practically unaffected. Panels were further immersed in water and were examined at regular intervals of 5 days.
It was observed that after 10 days, films of both the alkyds showed slight loss in gloss while the samples of polystyrene glycol-alkyd blends were practically unaffected. After 15 days, the alkyds showed considerable loss in gloss and change in colour while the blends having maximum amount of polystyrene glycol were practically unaffected and samples of the blend having half of the maximum amount showed a slight loss in gloss.
After 20 days, the alkyd films cracked. The condition of the film of all blends containing maximum amount of polystyrene glycol were still unaffected, except that in the case of nigerseed oil alkyd blends containing smaller amounts of amino resin, a slight loss in gloss was observed. All blends containing half of the maximum amount of polystyrene glycol found compatible showed only a slight loss in gloss except that in the case of castor oil alkyd blend some change in colour was also noticed. It clearly indicates that the incorporation of polystyrene glycol into alkyd improves the water resistance of the latter and further, as the amount of polystyrene glycol increases, water resistance also increases.
The films were also tested for boiling water resistance in order to reconfirm the above results. The results from Table 3 show that after 2 hours the boiling water caused loss in gloss and change in colour of plain alkyds except that in the case of nigerseed oil alkyd containing lower amounts of amino resin, some cracks were also noticed. Blends containing the maximum amount of polystyrene glycol found compatible, showed slight loss in gloss except the castor oil alkyd blend having a higher amount of amino resin, where the film was practically unaffected. Nigerseed oil alkyd blends containing half of the maximum amount of polystyrene glycol showed loss in gloss and change in colour while castor oil alkyd blends showed only slight loss in gloss. Effect of boiling water on the films was more significant after 4 hours of immersion. Films of all the plain alkyds were partially removed. Films of nigerseed oil alkyd blends containing smaller amounts of polystyrene glycol showed cracks. Films of all other blends showed only loss in gloss and change in colour. These results again confirm that the addition of polystyrene glycol imparts more water resistance to the alkyds.
For this test also, the glass panels of all the samples were prepared as described above and were immersed in 2 per cent solutions of each of sulphric acid, hydrochloric acid and nitric acid separately at room temperature (25°C). Panels were taken out after 24 hours, washed in running fresh water, allowed to air dry for an hour, and check for appearance, loss in gloss, change in colour and for sign of disintegration. It was observed that there was no loss in gloss or change in colour in any of the film.
Panels were further immersed in acids and films were checked at a regular interval of 24 hours. Table 4 gives the results of acid resistance on 15 and 20 days of immersion in the acids. After 15 days of immersion in hydrochloric and sulphuric acids, films of all the blends were practically unaffected except the nigerseed oil alkyd blends containing half of the maximum amount of polystyrene glycol found compatible and 25 per cent amino resin where slight loss in gloss was observed.
After 20 days of immersion in hydrochloric and sulphuric acids, films of all castor oil alkyd blends and nigerseed oil alkyd blends containing maximum amount of polystyrene glycol were practically unaffected. Nigerseed oil alkyd blends containing half of the maximum amount of polystyrene glycol found compatible showed slight loss in gloss.
The effect of nitric acid on the films was more pronounced as none of the sample remained unaffected on 15 days of immersion except the castor oil alkyd containing maximum amount of polystyrene glycol found compatible and lower amounts (10 per cent) of amino resin. The result established that, (i) castor oil alkyd had better acid resistance than nigerseed oil alkyd, (ii) acid resistance of the alkyds improved with the increasing amounts of polystyrene glycol, and (iii) samples containing 10 per cent amino resins had better acid resistance than those containing 25 per cent amino resin.
For this test, the glass panels of all the samples were prepared as described above and were immersed in 2 per cent solution of each of sodium carbonate and sodium hydroxide separately at ambient temperature (25° C). Panels dipped in sodium carbonate solution were taken out at a regular interval of 5 days, washed in running water, dried and the film examined for any visible damages. Film in sodium hydroxide solution was checked at a regular interval of 2 hours. Table 5 shows the results of the alkali resistance.
Nigerseed oil alkyd containing a maximum amount of polystyrene glycol was practically unaffected until 20 days of immersion in sodium carbonate solution. Nigerseed oil alkyd containing half of the maximum amount of polystyrene glycol showed a slight loss in gloss on 20 days of immersion. The effect on plain nigerseed oil alkyd was more pronounced than that on its blends e.g., the film of plain alkyd containing higher amounts (25 per cent) of amino resin cracked and the film of alkyd containing lower amount (10 per cent) of amino resin show loss in gloss and change in colour.
Mechanical Properties of Modified Alkyd Resins
A number of investigators have reported on the stress-strain properties of paint films. It has been found that the tensile strength of paint and varnish films is a more reliable guide to their strength than other tests commonly used, such as rocker hardness, scratch hardness impact and bend tests. Higher degrees of cross-linking combined with more homogeneity in the polymer film impart increased rigidity, which increases with increase in tensile strength and with decrease in elongation. The load-elongation curve can also be used to define the toughness, hardness and flexibility of the film.
The function of oil in oil-modified alkyds is to give the film the necessary flexibility to withstand the tensions caused by expansion, contraction and bending of the substrate, which could lead to cracking and flaking in the film. The present work has undertaken to study the effect of the fatty acids of varying chain lengths and degrees of unsaturation present in refined sardine oil and upgraded sardine oil on the mechanical properties of their alkyd films by comparison with those of vegetable alkyd films.
Materials and Methods
Alkyds of 66 and 50% oil lengths were prepared by the usual alcoholysis method from refined sardine, upgraded sardine (prepared by directed interesterification), linseed, soybean, safflower and dehydrated castor (prepared by the method of Sivasamban et al) oils. Free films of the alkyd resins were obtained by the procedure described earlier.
The Gardner tensile strength and percentage elongation apparatus was used to determine the tensile strength and percentage elongation of each film. Films of 35 ± 2µ thickness were cut into strips 1 cm wide and 12 cm long. They were fastened to the upper and lower clamps of the apparatus. The length of the strip between the two clamps was maintained at 10 cm without any strain on the film. The set screw of the apparatus was locked tight and the motor was started. The percentage elongation was noted for every reading of the load shown on the scale. Readings were continued until the film was torn, after which the motor was immediately stopped. Such readings were calculated in kg/cm2 of cross-section, and load-elongation curves (Figures 1 and 2) were drawn to estimate the area under each curve, which indicates the toughness of the film.
A strip of film, 10 cm long and 1 cm wide, was used as a torsional pendulum to determine the modulus of rigidity. One end of the strip was clipped by means of adhesive tape to a rigid support and the other end to the centre of the glass rod of calculated moment of inertia. Oscillations were started by disturbing the rod from its equilibrium position by slightly rotating it.
Bursting strength of each alkyd film was determined using the apparatus described by Vittal Rao et al. The film was held between the two flat flanges of the apparatus. In the enclosed space, air was compressed at a steady rate to exert pressure on the exposed area of the film. The pressure at which the film burst was noted from the mercury manometer attached to the system. The bursting strength of each film is expressed in terms of the height of mercury in mm.
Results and Discussions
Load-elongation curves for 66 and 50% oil length alkyd films are shown in Figures 1 and 2 respectively. Films of long oil alkyds modified with refined sardine and dehydrated castor oils were very soft and tacky and hence their mechanical properties could not be determined. The strengths of alkyd films modified with these oils have been found, however, to have increased considerably as their oil lengths were reduced to 50%.
A striking differences can be observed between the load-elongation curves of sardine fish oil and vegetable oil modified alkyd films. The curves for sardine oil alkyd films, irrespective of the oil length of the alkyd and unsaturation in the oil used, level off near the break points and bend towards the elongation axis. On the contrary, the curves of both long and medium oil length alkyd films modified with vegetable oils except DCO are steeper and at the break point they are bent towards the load-axis. From the nature of these curves it appears that sardine oil alkyd films are inherently softer and more flexible than vegetable oil modified alkyd films.
Sookne and Harris have studied the effect of a sharp molecular weight distribution and blends of such fractions of differing molecular weight ranges on the mechanical properties. They found that the presence of low molecular weight substance in a high molecular weight material has a disproportionately deleterious effect on mechanical properties. It has been observed that the presence of as little as 10-15% of low molecular weight fraction acts adversely on all mechanical properties, such as tensile strength and folding strength. These findings may be used to explain the difference in nature of sardine and vegetable oil alkyd films. Refined sardine oil contains large proportions of less reactive saturated and mono-un-saturated fatty acid components, mostly of C14 to C18 chain lengths, along with highly reactive components such as pentaenoic and hexaenoic acids belonging to C20 to C22 series16. This complex mixture of fatty acid components of the oil results in the alkyd resin modified by the oil becoming a blend of polymer units of very wide range of molecular weights, containing a large proportion of smaller molecular weight units. Hence, in spite of higher unsaturation in refined sardine oil (I.V. = 154), its long oil alkyd film is softer than the films of corresponding alkyds from safflower (I.V. = 141) and soybean (I.V. = 136) oils (Table 1).
In upgraded sardine oil a large proportion of saturated acids has been removed and consequently its unsaturation has been increased (I.V. = 221)9. The film of long oil alkyd modified with this upgraded oil has improved considerably in tensile strength.
With regard to the tensile strength of the long oil alkyds, the linseed and soybean alkyd films give the highest values, followed by the films of alkyds based on upgraded sardine and safflower oils. The films of alkyds based on DCO and refined sardine oils, as mentioned earlier, were too soft to allow the tensile strength to be measured. But again, inspite of the high unsaturation of upgraded sardine oil (I.V. = 221), its long oil alkyd film is more flexible and less tough than that of the corresponding alkyd from linseed oil (I.V. = 180). Here, perhaps, the effect of highly crosslinked large molecular weight alkyd polymers on tensile strength of its film has been reduced because of the presence of the low molecular weight polymer units formed due to the presence of large proportion of monoene acids of C14 to C18 carbon chain in upgraded sardine oil.
Fatty acids in vegetable oils are more or less confined to C18 chain length containing a small proportion of saturates and monoenes, along with a little C16 acids. Therefore, the alkyds abtained from such oils are expected to contain a very small range of polymer units and to be more homogeneous compared to their counterparts from define oils. Thus, the homogenous polymer units in alkyds resins from vegetable oils impart higher tensile strengths and less flexibility to their free films as compared to the films of alkyds based on sardine oils which are far less homogenous with regard to their polymer units. The softness and flexibility of DCO alkyd films can be attributed to the presence of hydroxy acids in conjunction with conjugated diene and larger proportion of non-conjugated diene which may result in a mixture of polymer units ranging from smaller molecular weight to highly crosslinked ones in the alkyd resin.
In medium oil length alkyds, the film of upgraded sardine oil alkyd is found to be the toughest. Here, the increased amount of glyceryl phthalate alkyd molecules have perhaps compensated the adverse effect due to less reactive fatty acids present in the oil on the mechanical properties of its alkyd film.
According to the scheme of Bosch et al, films of medium oil alkyds modified both with sardine and vegetable oils (except DCO) can be classified as tough and flexible films compared to their long oil alkyd films, as they have higher values for tensile strength and percentage elongation (Table 1). The DCO medium oil length alkyd films can be grouped as softer and more flexible, as they have lower values both for tensile strength and percentage elongation. Linseed oil alkyd films are found to be superior in tensile strength in both sets of alkyds.
The values for toughness given in Table 1 more or less confirm the above conclusions. However, although the medium oil length linseed oil alkyd film has the highest tensile strength, in toughness it is less than the medium oil length upgraded sardine oil alkyd film, which has the maximum value. This is because in the case of high tensile strength with high per cent elongation.
Polyblends of Polystyrene Glycol and Alkyd in Surface Coatings
Styrenated alkyds enjoy a wide variety of application in surface coatings. They are prepared either by first styrenation of fatty acid (or oil or monoglyceride) and then preparing alkyds or by styrenation after the preparation of the alkyd (poststyrenation).
Styrene monomer is incorporated into alkyd in order to improve chemical of the latter. In surface coatings, this is general practice to blend polymers physically in order to obtain desired film properties (e.g., alkyd - amino) povided that the polymers are compatible. However, a polystyrenealkyd system as such is not compatible and a useful composition cannot be obtained by simply blending them together. But if polystyrene can be made reactive by introducing functional groups such as hydroxyl, chloro, carboxyl etc., it can either be chemically reacted with alkyd or physically blended with alkyds.
Therefore, in the present work, polystyrene glycol has been prepared and the properties of its physical blends with alkyd have been reported. The chemical reaction of polystyrene glycol and alkyd will be reported later.
Styrene monomer used for the preparation of polystyrene glycol was purified in the following manner.
The monomer was washed with 4 per cent sodium hydroxide solution three to four times followed by washing with distilled water till free from alkalies. It was dried over anhydrous sodium sulphate overnight and then decanted off and distilled under reduced pressure.
Benzoyl peroxide (L.R. grade) was used as catalyst in the synthesis of polystyrene. Alkali refined linseed oil, phthalic anhydride and glycerol were used for preparation of the alkyd. Butanol, acetic anhydride, pyridine (all A.R. grade) were used for determination of hydroxyl value. Xylene, dioxane, benzene and methanol (all L.R. grade) were used as solvents. Lead and cobalt naphthenates were used as driers for the alkyd.
Preparation of polystyrene glycol
Polystyrene was first synthesised by free radical polymerization mechanism in bulk using benzoyl peroxide. The polystyrene formed contained benzoate and groups, which were saponified by use of alcoholic potassium hydroxide solution. Potassium benzoate formed on heating was removed by filtration and polystyrene glycol was obtained from dioxane solution by precipitation into water. Polystyrene glycol so obtained is a creamish white solid having an inherent viscosity of 0.15.
Presence of hydroxyl group in the polystyrene chain was confirmed by determining hydroxyl value according to ISI specification NO. !S: 548: Part I. The hydroxyl value of the prepared sample of polystyrene glycol was 289.
Preparation of linseed alkyd
Linseed alkyd of 50 per cent oil length was prepared by the monoglyceride process in the following manner.
Oil and 25 per cent (by weight of oil) glycerol was taken together and heated to 180°C. Litharge (0.1 per cent by weight of oil) was added and temperature raised to 240°C and heated till monoglyceride formation took place (checked by solubility in methanol). The mixture was cooled to 180°C and phthalic anhydride and the remaining glycerol was added. Temperature raised to 240°C and heated for about 4 hours. The product had an acid value of 10.
Blending of polystyrene glycol and alkyd
Physical blends of polystyrene glycol and alkyd were prepared by adding powdered polystyrene glycol into linseed alkyd resin and stirring the mixture for about one hour. During stirring, temperature was maintained at 100°C.
It was observed that polystyrene glycol was not compatible with oil modified alkyd in all proportions. The maximum quantity of polystyrene glycol that gave a clear mixture with the linseed alkyd was 20 per cent. Beside this, other blends containing lower amounts e.g., 5, 10, 15 per cent of polystyrene glycol were also prepared. Film properties of all these four blends and plain alkyd have been studied.
All samples were tested according to Indian Standard Specification (ISS) No. IS: 101 - 1964 for drying characteristic, scratch hardness, flexibility and adhesion, water, acid and alkali resistance, and solvent resistance including benzene, toluene and xylene. The colour of all the 5 samples were determined by Lovibond tintometer using a 1in. cell. Adequate amounts of driers on the basis of oil content of the resin were incorporated into the samples and thinned with xylene to a brushable consistency. Films were applied on 6in × 20in. glass and tin panels.
Results and discussion
During free radical polymerization by benzoyl peroxide, both phenyl and benzoyloxy radical may attack the double bond of the monomer. However, in case of bulk polymerization of styrene, end groups are mostly benzoate and not phenyl groups. Further, it has been established that the termination of two growing radical chains occur predominantly by combination, i.e., the radical ends of two growing chains combine to give a single molecule with an initiator fragment (i.e. benzoate) on each end. Use of a high proportion of initiator gives a low molecular weight product.
Keeping all the above facts in mind, successful preparation of polystyrene was performed. Benzoyl peroxide was used to impart difunctionality in the macromolecule. Polystyrene so obtained has benzoate end groups. These end groups were saponified by using alcoholic potassium hydroxide into hydroxyl ends.
The presence of hydroxyl groups in the prepared sample was confirmed by determination of the hydroxyl value of the sample which comes to 289.
Compatibility of polystyrene glycol with alkyd
It was observed that when higher proportions of polystyrene glycol were incorporated into alkyds they gave hazy mixtures. Beside this, polystyrene glycol had a tendency to separate out from the mixture on standing, when it was added in excess.
The maximum amount of polystyrene glycol found compatible with alkyd was 20 per cent. Further, the molecular weight of polystyrene glycol would play an important role in deciding the ease and maximum amount compatible with alkyd.
Films of linseed alkyd and its polystyrene glycol blends were air dried. Surface dry, hard dry and tack free times are recorded in Table 1. It was observed that the addition of polystyrene glycol into alkyd considerably reduces the drying time of linseed alkyd. For example, surface drying time of linseed alkyd was 1 hour. A 5 per cent blend surface dried in 30 minutes while 20 per cent dried in 15 minutes. It clearly shows that the addition of a higher amount of polystyrene glycol into alkyd from 5 to 20 per cent (by weight of alkyd) reduces the surface drying time 30 minutes to 15 minutes.
The hard dry time of linseed alkyd was 7 hours while its polystyrene glycol blends were hard dried in 3 to 4 hours only.
Scratch hardness was measured by mechanically operated ‘sheen' scratch hardness tester in which a hardened needle loaded with one kilogram weight moves over the film. All the samples of the polyblend and linseed alkyd pass the test which shows that polystyrene glycol and alkyd have good scratch hardness.
Maximum scratch hardness was determined by placing an increasing load over the hardened needle. Maximum scratch hardness in gramms of polyblends and linseed alkyd are given in Table 2. From the table, it is clear that polyblends have better scratch hardness than linseed alkyd because polystyrene imparts hardness to the film.
Flexibility and adhesion
Flexibility of the dried film was tested on tin panels. When bending the tin panels through 180° C with a ¼in. mandrel, all the samples of alkyd and their blends with polystyrene glycol passed the test, i.e., no detachment of the film from substrate or crack in the film was observed. Thus, the film had good flexibility and adhesion. Further, it confirms that polystyrene glycol is compatible with alkyd.
All the samples were allowed to air dry in a horizontal position for 48 hours. The sides of the panels were protected by wax. The panels were immersed in distilled water at room temperature (30°C) for 48 hours. After this the panels were taken out and washed with distilled water and allowed to air dry. The dipped portion was examined after four hours for appearance, loss in gloss and hardness. It was observed that all the samples were unaffected. Panels were further immersed in water and examined at a regular interval of 2 days. It was found that films of plain linseed alkyd softened after 8 days while the samples of all blends were practically unaffected even after 18 days of immersion. It concluded that blend of polystyrene glycol and alkyd have better water resistance.
Gas Chromatographic Analysis of the Carboxylic Acid Components of Alkyd Resins
In spite of the fact that alkyd resins were introduced some 45 years ago they are still the most widely used synthetic paint resins. In Australia, for the year 1968-69, alkyd resins constituted two-thirds of the total paint resin production. This may be attributed to the low cost and the suitability of the alkyd resin for modification by physical or chemical blending.
Alkyd resins consist of a back-bone of an aromatic dicarboxylic acid (e.g., 0-phthalic acid) esterified with a polyhydric alcohol (e.g. glycerol) to which fatty acids are joined at the remaining hydroxyl sites. The performance of alkyds as paint resins is largely dependent upon the nature and the concentration of the unsaturated fatty acid esters present. Various commercially available vegetable oils such as linseed, soya bean, safflower and sunflower are used in alkyd manufacture (see Appendix B). These oils contain the fatty acids, palmitic, stearic, oleic, linoleic and linolenic in the form of triglycerides. The last three of these acids contain, in turn, one two and three olefinic double bonds (sites of unsaturation). On exposure to oxygen the molecular chains crosslink and this results in the "drying" of the paint film.
In the formulation of an alkyd, a number of factors need to be considered which determine the type and quantity of vegetable oil used. For example, the concentration of polyenoic acids in the alkyd determines the drying time. However, an increase in their concentration renders the resin more prone to yellowing. Further, the hardness of a paint film is affected by the proportion of unsaturated fatty acids.
On this basis alone, a knowledge of the fatty acid composition of an alkyd will enable, to some extent, a prediction of its performance.
Previous analytical work
Early methods of characterising vegetable oils largely relied on the determination of iodine values (which is a measure of unsaturation).
Although ultra-violet spectroscopy has been used in the determination of fatty acids in vegetable oils, the main instrumental method of fatty acid analysis has been gas chromatography (GLC). The methyl esters of the fatty acids have been found to be suitable for GLC analysis and the conversion of vegetable oils and alkyd resins to fatty acid methyl esters has been achieved by transesterification, and by saponification followed by methylation. Due to the speed and simplicity of the transesterification method compared to the saponification procedure, the former method is more appealing. A comparison of the two methods has shown very little difference in the distribution of fatty acid methyl esters. Transesterification of lipids with methanolic sodium methoxide has given a fatty-acid methyl ester yield of 96 per cent of the theoretical yield and the saponification method gave a comparable result.
Saponification has a further disadvantage in that it has been observed that after 1h of reaction of linseed oil with 0.5 M methanolic potassium hydroxide, at reflux, followed by methylation of the liberated acids, an extraneous peak appeared in the chromatogram. This peak was attributed to the isomerisation of the polyenoic acids during saponification.
The analysis of the carboxylic acid components in alkyd resins by GLC has been largely confined to a quantitative determination of the fatty acids and qualitative estimation of the dicarboxylic acids. In this application a number of reagents have been used for transesterification reactions including methanolic solutions of sodium methoxide, potassium methoxide, lithium methoxide, boron trifluoride, hydrogen chloride, and diazomethane.
In the present work the transesterification technique was evaluated as a quantitative method for the determination of carboxylic acid derivatives in alkyd resins. The use of transesterifying agents was restricted to methanolic solutions of hydrogen chloride, boron trifluoride and lithium methoxide.
The fatty acid methyl esters were chromatographed using both diethyleneglycol succinate (DEGS) (Fig. 1.) and the Carbowax columns (Fig. 2.) It was found that, in general, the DEGS column afforded the better resolution. However, resolution of dimethyl o-phthalate and methyl linoleate was poor on this column, and the Carbowax column was found to be superior in this case.
The transesterifying agents HCl/MeOH, BF3/MeOH and LiOMe/MeOH were equally useful in the determination of the fatty acid derived components of the linseed oil (Table 1), soya bean oil (Table 2) and tall oil (Table 3) alkyds which were studied. However, the acidic reagents (HCl/MeOH and BF3/MeOH) gave very low results for the determination of the concentration of the phthalate derived components. The basic reagent, LiOMe/MeOH gave a result for the phthalate component which was significantly lower than the concentration determined by classical methods, or specified for the alkyd resins (Table 4).
Generally, the fatty acid distribution of alkyds, varies significantly from the parent glyceride oils, (Table 5). The linoleic and linolenic acid content is generally lower in the alkyd than in the parent oils. This is particularly noticeable in the case of linseed and soya alkyds. In addition, variations occur within alkyds derived from the same oil. A comparison of a long oil length (70 per cent oil) linseed alkyd with a medium oil length linseed alkyd (52 per cent oil), illustrates this point (Table 6). The long oil alkyd was produced by the fatty acid/oil process and the medium oil alkyd by the alcoholysis process (Appendix B). The source of the variation may be due to either the parent glyceride oils (Table 7), or to the reaction conditions used in the two processes. The fatty acid/oil process involves a higher temperature and twice the reaction time required for the alcoholysis process. These more vigorous conditions may induce decomposition or polymerisation of the lindeic acid liholenic acids thus reducing their concentrations in the resulting alkyds. Table 6 shows that the long oil linseed alkyd contains a lower concentration of the two acids in question than does the medium oil alkyd.
The effect of reaction time on the methyl ester determination was also considered. A medium length linseed alkyd was reacted with the LiOMe/MeOH reagent for various lengths of time. The relative proportions of the methyl esters produced was found to vary with changes in reaction time. The largest variation occurred in the formation of dimethyl o-phthalate. The concentration of dimethyl o-phthalate in the product was determined with respect to the fatty acid methyl ester concentration and so the variation with time of the dimethyl o-phthalate concentration may be due to a difference in rate of transesterification between the aromatic ester and the fatty acid esters. The reaction time also affected the individual fatty acid methyl ester concentrations. This may again be due to a difference in reaction rate or perhaps side reactions occurred which reduced the concentration of the more unsaturated fatty acids (Tables 8 and 9). It was found that a reaction time of 0.25 hours at reflux was the most suitable for obtaining a reliable determination of the individual fatty acid ratios. The dimethyl o-phthalate determination by this method was not satisfactory but may be used as a guide to oil length.
Three reagents have been examined in the transesterification of alkyd resins, with a view to developing a method of determining the concentration of the carboxylic acid components present. A 0.5M methanolic solution of lithium methoxide was found to be superior to the other reagents used, which were a methanolic solution of boron trifluoride and a methanolic solution of hydrogen chloride.
The lithium methoxide reagent was used for determining the concentration of the various fatty acids in the alkyd. The vegetable oil used in the manufacture of the resin could thereby ascertained. The phthalate content of the alkyd could not be determined accurately by any of the three reagents; however, the lithium methoxide solution gave a value for the phthalate content which served as a guide to the oil length of the resin.
The concentrations of the various methyl esters were determined by means of gas-liquid chromatography using a diethyleneglycol succinate column and a Carbowax 20 M column.
Methods of Analysis of Alkyd Resins
A number of the general references review the methods of analyzing alkyd resins.
The compositional analysis of alkyds is complicated by the wide variety of ingredient combinations that may be encountered. Neither systematic nor standard methods have been devised, so the analyst must use his ingenuity and devise methods to fit the product.
Determination of Composition
Gas-liquid chromatography (GLC) has proved effective for the compositional analysis of alkyd resins. The following discussion will cover a few applications of the technique.
Dibasic Acids. Esposito are credited with the first systematic study pertaining to the carboxylic acids in alkyds. In order to provide a sample with a sufficiently high vapor pressure, a rapid transes sterification procedure is used to convert the acids to the methyl esters. Fatty acids from soybean, linseed, or tall oil can be identified in the presence of 0-phthalic, isophthalic, fumaric, maleic, itaconic, succinic, adipic, azelaic, sebacic, diglycolic, pelargonic, and benzoic acids.
The samples are transesterified with lithium methoxide in methanol, except when it is necessary to distinguish maleic from fumaric acid, in which case boron trifluoride is used instead of lithium methoxide. The methyl esters are separated on a two-part polyester-Carbowax column, and the relative retention times of the individual esters are compared to data obtained from known esters. Table 1 lists relative retention time data for methyl esters; Esposito it convenient to refer these esters to triacetin, which was given a value of 1. Values are also included for a silicone column.
Prepare 0.5 N lithium methoxide in methanol by adding pea-sized pieces of lithium to absolute methanol, chilled in an ice bath. Determine the normality by titration and add enough methanol to adjust the solution to 0.5 N.
Prepare the boron trifluoride reagent by bubbling the gas into absolute methanol until 1 ml of the reagent, diluted with methanol, requires 11-12 ml of 0.5 N potassium hydroxide in methanol.
Place about 0.3 g of resin in a 125-ml flask and add 15 ml of 0.5 N lithium methoxide. Attach an air condenser and boil the contents on a steam bath for 2 min. Remove, cool, and add 5 ml of 6 N aqueous sulfuric acid. Transfer to a separatory funnel and dilute to 50 ml with water. Extract with 35 ml of methylene chloride and then wash the methylene chloride layer with 15-ml portions of water. If in soluble methyl esters are present, add tetrahydrofuran dropwise until dissolved. Remove the solvent on a steam bath. If maleic or fumaric acid is found in the subsequent chromatogram, repeat the transesterification procedure with a boron trifluoride catalyst using 5 ml of the reagent and a 5-min reflux.
Chromatograph a 5-ml sample along with 0.2 ml of triacetin on a 6-ft polyester-Carbowax column, and repeat the separation on a 6-ft silicone grease column. Identify the methyl ester peaks by determining their retention times relative to triacetin and comparing them to the values in Table 4.
The use of columns of different polarity makes it possible to resolve all of the dibasic acids and fatty acids commonly found in alkyd resins. Even dimethyl fumarate and dimethyl maleate, prepared with the boron trifluoride catalyst, are separated. These acids probably rearrange with alkali in a methanol medium to give dimethyl methoxy succinate.
The choice of the polyester substrate used in the column is not critical since almost any polyester is satisfactory. There is also a wide selection available for the nonpolar column; Apiezon greases, fluorosilicone, and polyamides. Figure 1 shows a chromatogram formed by the separation of the methyl esters of a typical alkyd prepared by transesterification with lithium methoxide. The chromatogram was run isothermally at 180°C on a 9ft × in. OD column of ethylene glycol succinate (EGS) on Gas-Chrom P with a 1-ml sample. The component acids were identified by comparing their retention times to known methyl esters.
Gas-liquid chromatography is ideal for the qualitative identifications of acids as it is remarkedly free of interferences, even from modifying resins. High-boiling solvents found in alkyd resin solutions sometimes obscure low-boiling methyl esters, but these can be removed by drying the sample before the ester interchange.
Rosin esters do not form methyl esters under the transesterification conditions used, but require strenuous saponification, isolation of the rosin acids by extraction, and methyl ester preparation with diazomethane.
The transesterification procedure does not lend itself to quantitative GLC, because the yields of methyl esters are different for individual acids. Furthermore, it appears that the detector response varies for different methyl esters, and that calibration with known mixtures is necessary. Nevertheless, studies are being made of the quantitative aspects, and single acids may be determined by the internal standard technique.
Esposito has determined phthalic acid quantitatively in alkyd resins as dimethyl phthalate through an internal standard, and recently, Haken described the determination of benzoic acid and p-tert-butyl benzoic acid. Since these acids form soluble potassium salts, a saponification mixture of the resin is evaporated to dryness, acidified, and extracted with ethyl ether. The free acids are methylated with diazomethane and the resulting methyl esters are examined by GLC.
Fig. 1. Chromatogram of the methyl esters of a typical alkyd.
The insoluble potassium salts isolated in the Kappelmeier procedure may be converted to methyl esters by the same procedure. However, a preferable route is to pass the salts through a column of cation exchange resin in hydrogen form, recover the free acid from the eluate, and methylate with diazomethane. Percival obtained dimethyle esters from polyesters by methanolysis of the latter with a 95:5 methanol-resin dilution, containing about 0.1 g of sodium methoxide. However, he extended the methanolysis to 18 hr and up to 42 hr for some resins to insure at least semiquantitative results for the combined dimethyl esters and glycols. A 12 ft × in. OD column, packed with 10 g of SF-96 silicone on 50 g of Fluoropak 80, is used with temperature programming from 110 to 180°C at 8°C/min. With suitable response factors and by working with known mixtures, the method should be useful for determining the identity of the dibasic acids as well as their ratio in an alkyd.
Maleic adducts are difficult to detect in alkyds and it is even more difficult to recover maleic acid from such mixtures. If the maleic anhydride has added to the oil to form an alkyl-succinic adduct, the fatty acids are recovered and their molecular weights determined by an acid value. Then the methyl esters of the maleinized acids are subjected to GLC, comparing the chromatogram with those obtained from similar samples of known composition.
Fatty Acids. One of the most important applications of GLC is in fatty acid analysis. The first paper on GLC by James and Martin in 1952 was concerned with fatty acids, and from then to 1958 more than fifty papers dealing with that subject were published. Polyester liquid phases, developed in 1958 permitted the separation of long-chain acids in respect to both chain length and degree of unsaturation, and in particular, the four C18 acids usually found in drying oils.
Apeizon or silicone greases (nonpolar phases) separate fatty acids by chain length only, unless used in extremely efficient packed columns or capillary columns. Silicone gum, SE-30, is less selective and the C18 acids will give but one peak. This is advantageous if separation into chain lengths is desired or if there is a question regarding the assignment of a peak.
With the unusual wealth of information and the almost universal availability of reliable instrumentation, analysis of fatty acids by GLC is almost routine. Recent developments, such as temperature programming, dual column operation, capillary columns, and more sensitive detectors, serve to make GLC more attractive and versatile, but most of the applications reported have been carried out with thermal conductivity detectors in conjunction with packed columns.
For fatty acid analysis, methyl esters are generally used. They can be prepared by several methods, starting with the resin as did Esposito or more likely, with the fatty acids isolated by the Kappelmeier separation. The transesterification technique is perfectly satisfactory for identifying the fatty acids, but the amount is best determined by an actual separation. The American Oil Chemists' Society (AOCS) Gas Chromatography Committee has studied esterification with methanol and sulfuric acid, with boron trifluoride and methanol with diazo-methane, and with 2,2-dimethoxypropane and has recommended the methanol-sulfuric acid system except with extremely small samples or with hydroxy or epoxy acids which should be methylated with diazomethane. Only a few minutes is involved in the preparation of methyl esters by either method.
Fatty acids may be run directly on columns containing polyesters with 1% of phosphoric acid added to the support for the purpose of inhibiting active sites and reducing tailing. The resolution is not so good as that obtained with methyl esters so that the extra methylation step is worth the effort.
AOCS Method Ce 1-62 has been studied collaboratively and is suitable for estimation of fatty acids derived from alkyds.
Methyl Ester Preparation. Methylate 2g, or less, of fatty acids by refluxing with 60 ml of 2% methanolic sulfuric acid for 1 hr. Cool, transfer to a separatory funnel, and add 100 ml of water. Extract twice with 50-ml portions of petroleum ether (bp 30-60°C). Wash the combined extracts with 20-ml portions of water until free of acids, dry with anhydrous sodium sulfate, and evaporate the solvent under a stream of nitrogen on the steam bath.
Apparatus. The gas chromatograph should have an injector port and a detector with an independent temperature control. Use a 4-10 ft × in. OD glass, stainless steel, aluminum, or copper column packed with 20% polydiethylene glycol succinate on 60-80 mesh, acid-washed Chromosorb P or W, and operate at a constant temperature between 190 and 210°C. The recorder should have 0-1 mV range with 1 sec full scale deflection. Use helium as the carrier gas.
Determination. Set the temperature of the injector 50°C above that of the column and the temperature of the detector 25°C higher than that of the injector. Condition the polyester column by holding at the operating temperature with the gas flowing until a steady base line is achieved. Following the manufacturer's directions, adjust the gas flow to permit elution of methyl stearate in 30 min or less; an adjustment of column length may also be necessary, but do not exceed 40-psig inlet pressure. Measure the gas flow periodically with a soap bubble meter but not during a run. Inject a 0.5-4 ml sample, adjusted so that the major peak is not attenuated more than eight times, and recore the chromatogram, attenuating as necessary. Mark the air peak as zero.
Calculations. Integrate the area of each peak by a convenient method. Identify the peaks by their relative positions on the chromatogram or by reference to a known mixture of methyl esters run under the same conditions. The esters appear in order of increasing number of carbon atoms and of increasing unsaturation for the same number of carbon atoms. Total the areas of all peaks and calculate the percentage of each. For ordinary work, the percentage area of a peak may be considered as the percentage of the corresponding component. For more accurate work, calibration factors should be determined and used to correct for non-linearity of detector response and for molecular weight differences. The known mixture used for such calibration should have a composition similar to that of the unknown.