Grażyna KAMIŃSKA-BACH, Ewa LANGER ? Institute for Engineering of Polymer Materials&Dyes Paint&Plastics Department in Gliwice, Poland

Abstract:
Knowledge of the interactions between the polymer and solvents used as test substances provides important information about their miscibility and the thermodynamic properties of solutions. Inverse gas chromatography is a valuable tool for the study of thermodynamic properties of polymers, pigments and fillers. In this study thermodynamic properties of epoxy resins were studied. The purpose was to examine the impact of cross-linking of epoxy resins, epoxy number of resins, method of column preparation and test temperatures on the Flory-Huggins interaction parameters of epoxy resins and the identification of changes in the solubility parameters of cross-linked and uncross-linked resins.

Please cite as: CHEMIK 2013, 67, 4, 301-308

Introduction
Epoxy resins constitute a group of synthetic oligomers in the form of viscous liquid or thermoplastic solids, depending on their molecular weight. Due to their characteristics and diversity, they can be used in many, even the most demanding applications. In order to obtain a stable end product, a resin is first converted to an infusible, insoluble product through cross-linking with different hardeners. As a result of cross-linking, epoxy plastics acquire such properties as high mechanical strength, good dielectric properties, resistance to water and chemicals [1, 2].

The thermodynamic properties of epoxy resins, as well as the kinetics of the cross-linking process have been studied by many researchers around the world [3÷9]. The surface and mechanical properties of epoxy resins change in the course of the cross-linking process. It was found that the intermolecular interactions and surface properties are dependent on temperature [5]. The surface tension of two epoxy systems based on diglycidyl ether of bisphenol A (DGEBA) and tetraglycidyl methylene dianiline (TGMDA) has also been determined [6], finding that it decreases with temperature following a linear regression, which was confirmed in liquid systems [7] and molten polymers [8]. Mezzenga et al. [9] predicted change of surface energetic properties of epoxy resins during the curing process using changes in the measured surface tension and changes in the solubility parameters calculated by means of the additive group contribution method. On the basis of the results it was stated that there is good agreement between predicted and experimental results of the estimation of the surface energetic properties. Moreover, the described method of forecasting can be used to estimate the total solid surface energy of the cured epoxy resin. In the literature to determine the thermodynamic parameters of epoxy resins are not used inverse gas chromatography technique, as described in this paper.

Knowledge of the interactions between the polymer and solvents used as test substances provides important information about their miscibility and the thermodynamic properties of solutions. Inverse gas chromatography is a valuable tool for the study of thermodynamic properties of polymers, pigments and fillers [10]. During the dosing of a test substance into a chromatography column packed with the solid being tested, it is evaporated and sent through the column by a carrier gas, as a result of which the equilibrium distribution of sample molecules between the gas and the stationary phase (at infinite dilution) is determined. The absolute retention volume, VN [cm3] [10÷14], describes this equilibrium distribution. It is defined as the volume of gas required to elute the sample from the column and can be calculated using the equation:

The specific retention volume of the sample, Vg° [cm3/g], i.e. the relationship between the retention volume and 1 g of the liquid phase or 1g of adsorbent at a temperature of 0°C is calculated from the formula:

Based on the experimental values of the specific retention volumes (Vg0) the Flory-Huggins test substance ? stationary substance interaction parameters can be calculated [2]:

The calculated values make it possible to determine the solubility parameter of the test substance using the following interdependency:

Based on previous studies [15] it was found that in the case of acrylic resins, the values of the Flory-Huggins parameters and the solubility parameters are influenced by the type of polymer and its properties, as well as the measurement conditions, such as how the column has been prepared, and the test temperature.

The purpose of this study was to examine the impact of the following factors:
? cross-linking of epoxy resins
? the epoxy resin number
? how the column has been prepared
? test temperature.
on the Flory-Huggins interaction parameters of epoxy resins and
the identification of changes in the solubility parameters of cross-linked
and non cross-linked resins.

Experimental

Materials
Three epoxy resins, polycondensation products of bisphenol A and epichlorohydrin, have been examined. Characteristics of the resins are shown in Table 1.

Triethylenetetramine (TETA) in a stoichiometric ratio was used as a hardener for the epoxy resins (one NH group with one epoxy group).

Table 2 shows the characteristics of the analytically pure substances used in the tests (molecular weight ? M, critical temperature ? Tc, critical pressure ? pc, density ? ?, solubility parameter ? ?).

Test methods
The Flory-Huggins parameters and solubility parameters for epoxy resins have been determined using the inverse gas chromatography method. The study was performed using a Pye Unicam PU 4500 gas chromatograph with a flame ionization detector (FID) coupled to a PM 8000 Philips recorder. Separation was performed in glass chromatographic columns with a length of 50 cm and internal diameter of 0.4 cm. The carrier gas was pure helium (5.0) at a flow rate of 50 cm3/min. Empty columns and glassware used in the preparation of the columns was washed using dimethylchlorosilane (DMCS) in order to eliminate any interactions between the glass and the test substances entered into the system.

The columns were prepared for testing in two ways. The first (I) involved placing the epoxy resin being tested which has been non cross-linked or to which a hardener has been added (at 10% by weight in relation to the packing) on a filling material constituting Chromosorb W-AW-DMCS, 80/80 mesh. The second method (II) involved packing the column with only the cross-linked or non crosslinked resin being tested.

The study was conducted at the following temperatures: 40°C (chromatographic columns prepared according to method II), and 40°C, 100°C and 120°C (columns prepared according to method I). The injector temperature was 100 °C (in the case of column temperature of 40°C), 120°C (in the case of column temperature of 100°C), and 130°C (in the case of column temperature of 120°C) respectively. The temperature of the detector in each case was 250°C. Before performing the analysis, the column was conditioned at measurement temperature for 2 hours.

After obtaining a stable baseline, the test substances, both polar and non-polar, were injected using a micro-syringe with a capacity of 1?l, which allowed for capturing the individual retention times.

Results and discussion
The ??1,2 Flory-Huggins parameter values reflect the energy of interaction between a test substance and the stationary phase (column filling). Low values indicate a strong interaction, while a high value ? lack of such interactions [16].

Based on the experimental values of the specific retention volumes (Vg0) in case of different temperatures, the following parameters were calculated:
? the test substance ? stationary substance ? Flory-Huggins interaction parameters for each cross-linked and non cross-linked epoxy resin
? ?2 solubility parameters for different cross-linked and non crosslinked epoxy resins.

Effect of cross-linking on Flory-Huggins parameters
The curing process of epoxy resin results from the reaction of epoxy groups present in resin molecules with active groups of the hardeners. As mentioned earlier, the cross-linking reaction is an important process in obtaining the proper application properties. In the present study, the effect of the cross-linking process on the Flory- Huggins parameters and solubility parameters of selected epoxy resins has been studied. FTIR spectrum of a selected epoxy resin before and after curing is shown in Figure 1.

The spectrum of the cured resin does not differ substantially from the spectrum of the uncured resin. The main difference is the disappearance of epoxy groups of bands at wave number 917 cm-1, and appearance of the carbonyl group bands at wave numbers around 1725 cm-1.

The first parameter considered was the impact of the cross-linking process of epoxy resin on its thermodynamic parameters. Figure 2 shows the Flory-Huggins parameters of interaction on E2 and E3 non cross-linked and cross-linked resins analyzed according to method II. Resin E1 is liquid resin, so it is not possible to analyse it according to method II.

Analysis of the results shows that the cross-linking causes a drop in the Flory-Huggins interaction parameters in the most of cases. For resin E3 decrease of values of Flory-Huggins parameters occurs for all investigated test substances, whereas the resin E2 values fall Flory- Huggins parameters does not apply to diethyl ether and acetone, which is an increase in value. The changes in the Flory-Huggins parameter by cross-linking effect is likely to change in the quality and content groups characteristic of epoxy resin. These are the most common epoxy groups, ether linkages, and olefin, hydroxyl groups and chlorohydrine groups. The cross-linking of epoxy resin proceeds with the participation of the epoxy groups. In the case of cross-linking with primary and secondary amines by the reaction of hydroxyl groups, new, and the lowering of the content of epoxy groups. During the curing process, there is a decline in the amount of functional groups that are responsible for the strongest interactions between the molecules of the resin, which fills the column, and the test substances injected into the column. The differences in the parameter values depend on the test substance. The parameters themselves have high values for both the resins (from ca. 3 to 7.5).

Impact of the epoxy number on the Flory-Huggins interaction parameters
Another item considered was the epoxy resin number and its impact on the ? ? 1,2 interaction parameters.

Figure 3 show the ?? 1,2 Flory-Huggins parameter values for the cross-linked and non cross-linked epoxy resins. The selected resins vary in molecular weight, and thus in the content of functional groups (like epoxy groups, ether linkages, and olefin, hydroxyl groups and chlorohydrine groups) responsible for the interaction between the resin and solvents used as test substances. In the case of non cross-linked resins, an increase in the Flory-Huggins parameters can be observed as the molecular weight of the resin increases. Higher molecular weight means fewer functional groups, and therefore weaker interactions and a higher Flory-Huggins parameter. The situation is different in the case of cross-linked resins, where the functional groups disappeared after reacting with the hardener. The ?? 1,2 parameters assume a much higher value, because the impact is smaller.

Impact of column preparation on the Flory-Huggins interaction parameters

By analyzing the parameter values of a selected E3 uncured and cured epoxy resin, which has been tested at 40°C, we can see that there are clear differences in values depending on how the column has been packed ? 100% resin, or resin embedded on a filling in an amount of 10%, as shown in Figure 4. In the case of 100% resin, these values are much higher than for the embedded resins. These differences are likely due to the fact that at 40°C, the resin was studied below the glass transition temperature, and therefore the test substance molecules can be adsorbed only on its surface, causing weaker interactions. The active surface (on which adsorption occurs) relative to the total weight is greater for the resin deposited on a filling in comparison with 100% resin, leading to stronger interactions.

Impact of temperature on the Flory-Huggins interaction parameter
Figures 5 and 6 show changes in the interaction parameters for E1 non cross-linked (Fig.5) and cross-linked (Fig. 6) resins at 40, 100 and 120°C.

It was observed that the test temperature has a significant impact on the interaction parameters. In both cases, the parameters decrease in value with the increase of column temperature. This is probably due to a more lability of the polymer chains at higher temperatures, and thus easier access to functional groups.

Epoxy resin solubility parameter determination results
Based on the interaction parameters obtained, the solubility parameters (?) for non cross-linked and cross-linked epoxy resins at test temperatures were determined. The linking process has an impact on the ? solubility values. In most cases, the solubility parameter values for all the resins decrease after cross-linking. The smallest differences were observed in the case of E3 epoxy resin, where the epoxy number has the lowest value, and therefore includes the least epoxy groups capable of reacting with the hardener.

For non cross-linked E1 resin, the solubility values at temperatures from 40 to 120°C were in the range from 20.7 to 24.2 MPa1/2, while after cross-linking from 17.0 to 23.7MPa1/2. E3 solubility parameters before cross-linking ranged from 22.4 to 24.7MPa1/2, after cross-linking from 17.9 to 24.8 MPa1/ 2.

Conclusions
Cross-linking of epoxy resins causes also changes in the Flory- Huggins interaction parameters as well as in solubility parameters of resins. The results show that the cross-linking as well as the increase of column temperature cause a drop in the Flory-Huggins interaction parameters. An increase in the solubility parameters can be observed as the molecular weight of the resin increases. Higher molecular weight means fewer functional groups, and therefore weaker interactions and a higher Flory-Huggins parameter values. There are also clear differences in values of thermodynamic parameters depending on how the column has been packed.

Literature
1. Scholten H.: Epoxy resins: production, properties, applications. Europ. Coat. J.1995, 6, 458-463.
2. Bryan E.: Chemistry and Technology of Epoxy Resins. Chapman & Hall, London, 1994.
3. Mezzenga R., Boogh L., M?nson J.-A.E.: Evaluation of solubility parameters during polymerisation of amine-cured epoxy resins. J. Polym. Sci. Part B. 2000, 38, 14, 1883-1892.
4. Mezzenga R., Luciani A., M?nson J.-A.E.: Phase separation and gelation of epoxy resin/hyperbranched polymer blends. Polym. Eng. Sci. 2002, 42, 2, 249-257.
5. Scatchard, G.: Equilibrium in Solutions: Surface and Colloid Chemistry. Harvard Univ. Press, Cambridge, 1976.
6. Page SA, Berg J.C., M?nson J.-A.E.: Characterization of epoxy resin surface energetics. J. Adhesion Sci. Technol. 2001, 15, 2, 153-170.
7. Wu S.: Polymer Interface and Adhesion. Marcel Dekker, New York, 1982.
8. Adamson A.W., Gast A.P.: Physical Chemistry of Surfaces. Wiley-Interscience, New York, 1997.
9. Page SA, Mezzenga R., Boogh L., Berg J.C., M?nson J.-A.E.: Surface energetics evolution during processing of epoxy resins. Journal of Colloid and Interface Science 2000, 222, 55-62.
10. Kamińska-Bach G., Langer E.: Zastosowanie techniki odwróconej chromatografii gazowej do charakteryzowania nielotnych składników wyrobów lakierowych. Farby i Lakiery 2008, 6, 24-28.
11. Zhao S., Zhang F., Zhang W., Shi B.: Determination of the solubility parameter of cellulose acrylate using inverse gas chromatography. Chinese Science Bulletin 2007, 52, 3051-3055.
12. Surana R., Randall L., Pyne A., Murti Vemuri N., Suryanarayanan R.: Determination of glass transition temperature and in situ study of the plasticizing effect of water by inverse gas chromatography. Pharm. Res. 2003, 20, 1647-1653.
13. Huang J.-C.: Anomalous solubility parameter and probe dependency of polymerpolymer interaction parameter in inverse gas chromatography. Eur. Polym. J. 2006, 42, 1000-1007.
14. Fernández-Sánchez E., Fernández-Torres A., Garcia-Dominguez J.A., L?pez de Blas E.: Thermodynamic characterization of Superox 20M by inverse gas chromatography. J. Chromatogr. A 1993, 655, 11-20.
15. Langer E., Kamińska-Bach G.: Wybrane parametry termodynamiczne polimerów akrylowych oznaczone metodą odwróconej chromatografii gazowej. Przem. Chem. 2011, 1, 132-134.
16. Fall J., Milczewska K., Voelkel A.: Characterisation of mineral oil-polyester, mineral oil-poly-?-olefin mixtures by inverse gas chromatography. J. Mater. Chem. 2001, 11, 1042-1046.

Translation into English by the Author

Grażyna KAMIŃSKA-BACH ? M.Sc., graduated from the Faculty of Chemistry at the Silesian University of Technology (1997). Since 1997, she has been working as an assistant at the Institute for Engineering of Polymer Materials and Dyes in Torun, in the Paints and Plastics Department in Gliwice, in the accredited laboratory. Research interests: gas chromatography, infrared spectrophotometry, the study of the physicochemical properties of coating materials.

LANGER Ewa ? Ph.D., (Eng), graduated from the Faculty of Chemistry at the Silesian University of Technology (2005). Since 2005, she has been working as an assistant at the Institute for Engineering of Polymer Materials and Dyes in Torun, in the Paints and Plastics Department in Gliwice, in the coatings technology laboratory. In 2012 she defended her PhD thesis at the Faculty of Chemistry, Silesian University of Technology in Gliwice. Researchinterests: self-stratifying coatings, the study of the physicochemical  properties of coating materials.
e-mail: ., phone: + 41,
fax: +-74,