Quantitative Characterization of Solid Epoxy Resins Using

Purchase temporary access to this content. ACS Members purchase additional access options · Ask your library to provide you and your colleagues site-w...
1 downloads 0 Views 2MB Size
Anal. Chem. 2009, 81, 4271–4279

Quantitative Characterization of Solid Epoxy Resins Using Comprehensive Two Dimensional Liquid Chromatography Coupled with Electrospray Ionization-Time of Flight Mass Spectrometry Samir Julka,† Hernan Cortes,*,† Robert Harfmann,† Bruce Bell,† Andreas Schweizer-Theobaldt,‡ Matthias Pursch,‡ Luigi Mondello,§ Shawn Maynard,| and David West| The Dow Chemical Company, Analytical Sciences, 1897 Building, Midland, Michigan 48667, Dow Deutschland Anlagengesellschaft mbH, Analytical Sciences, Rheinmunster, Germany, Dipartimento Farmaco-chimico, Facolta` di Farmacia, Universita` dei Messina, Viale Annunziata, I-98168 - Messina, Italy, and The Dow Chemical Company, Epoxy R&D, B-4810, Freeport, Texas 77541 A comprehensive multidimensional liquid chromatography system coupled to Electrospray Ionization-Mass Spectrometry (LCxLC-ESI-MS) was developed for detailed characterization and quantitation of solid epoxy resin components. The two orthogonal modes of separation selected were size exclusion chromatography (SEC) in the first dimension and liquid chromatography at critical conditions (LCCC) in the second dimension. Different components present in the solid epoxy resins were separated and quantitated for the first time based on the functional groups and molecular weight heterogeneity. Coupling LCxLC separations with mass spectrometry enabled the identification of components resolved in the two-dimensional space. Several different functional group families of compounds were separated and identified, including epoxy-epoxy and epoxy-r-glycol functional oligomers, and their individual molecular weight ranges were determined. Repeatability obtained ranged from 0.5% for the main product to 21% for oligomers at the 0.4% concentration level. Detailed characterization of synthetic polymers is critical for better understanding of structure-performance relationships. Performance properties are dependent on a number of variables, such as molecular weight distribution, molecular topology, functionality, and chemical composition. A synthetic polymer of specific interest is epoxy resin. To anticipate the types of molecular structures that can be expected in Solid Epoxy Resins (SER) it is necessary to understand some aspects of the manufacturing processes used to make liquid and solid epoxy resins. The following gives a basic description of these processes and suggests some of the molecules and functional groups that can be expected. * To whom correspondence should be addressed. E-mail: [email protected]. Phone: 989 631 8207. Fax 989 636 6432. † The Dow Chemical Company, Analytical Sciences. ‡ Dow Deutschland Anlagengesellschaft mbH, Analytical Sciences. § Universita` dei Messina. | The Dow Chemical Company, Epoxy R&D. 10.1021/ac9001047 CCC: $40.75  2009 American Chemical Society Published on Web 04/29/2009

Liquid Epoxy Resin. Liquid epoxy resin (LER) is made in a two step process. First, the dichlorohydrin of bisphenol A is formed by reacting excess epichlorohydrin and bisphenol A. Next the dichlorohydrin of bisphenol A is dehydrochlorinated by reaction with aqueous sodium hydroxide giving primarily the diglycidyl ether of bisphenol A (DGEBA) in addition to other lower concentration impurities. The impurities may have one, two, or three epoxy groups, and higher molecular weight oligomers of DGEBA are also found. Other species may have non-epoxy functional end groups. These end groups can be diol or contain chlorine, for example, that form from side reactions that occur in the epoxidation process. The main product DGEBA and the various impurities are known collectively as LER.1 Fusion Process for Solid Epoxy Resin. In the so-called Fusion process for SER, excess LER is reacted with molten bisphenol A in the presence of catalyst to produce a distribution of higher oligomers of LER with even repeat unit (n) values. The idealized structure and the fusion chemistry are shown in Figure 1. The oligomer distribution is controlled by the stoichiometry. LER that contain two epoxy groups (difunctional) are advanced with bisphenol A to produce the even oligomer pattern. Chains that are formed from the advancement of non-epoxy terminated chains, mono- and trifunctional chains, and branched chains can be found (as small components) between the even oligomers.1 Taffy Process for Solid Epoxy Resin. In the “Taffy” process excess epichlorohydrin is mixed with an aqueous solution of sodium bisphenate forming an epoxy oligomer distribution with both even and odd repeat units. As the reaction proceeds a white highly viscous epoxy resin separates from the solution (the resin became known as taffy because of its appearance). Because of the large amount of water present during the reaction and the poor transport of reactants into the resin phase there are two types of non-epoxy end groups formed: chlorohydrin and diol. The chlorohydrin is present because of incomplete dehydrochlorination (ring closure). The diol end group is formed either by hydrolysis of an epoxy end group, or from hydrolysis of epichlo(1) Pham, H. Q.; Marks, M. J. Epoxy Resins in Encyclopedia of Polymer Science and Technology; John Wiley & Sons Inc.: New York, 2004.

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4271

Figure 1. Idealized structures and Epoxy resin fusion chemistry.

rohydrin to glycidol and subsequent reaction with sodium bisphenate or an oligomer.1 Since the liquid epoxy resins used for the synthesis of solid epoxy resins (SER) often contain considerable amounts of byproduct like chlorohydrins or R-glycols as described above, a number of possible side reactions during advancement with bisphenol A can lead to considerable macromolecular heterogeneity in the final SER. The broad molecular weight and functional end-group distribution due to advancement of several impurities in combination with linear versus branched distributions makes the detailed characterization of solid epoxy resins extremely challenging,.2,3 Reversed-phase liquid chromatography has been shown to provide good separation of the low molecular weight macromolecules according to functionality.4-6 Analyses of solid epoxy resins using reversed-phase liquid chromatography and mass spectrometry have been reported.7 However, with only single-dimensional separations, the limited peak capacity achieved and the limited separation of higher molecular weight components yielded incomplete characterization. Information on the number and chemical structures of side products, how each of the side products advances during the polymerization process, and determination of the molecular weight ranges for each of the side products is not possible using single dimensional separations. (2) Adrian, J.; Braun, D.; Rode, K.; Pasch, H. Angew. Makromol. Chem. 1999, 267, 73–81. (3) Adrian, J.; Braun, D.; Pasch, H. Angew. Makromol. Chem. 1999, 267, 82– 88. (4) Gloeckner, G.; Van den Berg, J. H. M. J. Chromatogr. 1986, 352, 511– 522. (5) Gorbunov, A. A.; Solovyova, L. Y.; Skvortsov, A. M. Polymer 1998, 39, 697–702. (6) Pasch, H.; Trathnigg, B. HPLC of Polymers; Springer: Berlin, 1997. (7) Fuchslueger, U.; Rissler, K.; Stephan, H.; Grether, H. J.; Grasserbauer, M. J. Appl. Polym. Sci. 1999, 72, 913–925. (8) Stoll, D. R.; Li, X.; Wang, X.; Carr, P. W.; Porter, S. E. G.; Rutan, S. C. J.Chromatogr. A 2007, 1168, 3–43. (9) Dugo, P.; Cacciola, F.; Kumm, T.; Dugo, G.; Mondello, L. J. Chromatogr. A 2008, 1184, 353–368. (10) Guiochon, G.; Marchetti, N.; Mriziq, K.; Shalliker, R. A. J. Chromatogr. A 2008, 1189, 106–168.

4272

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

In the past few years, comprehensive two-dimensional liquid chromatography (LCxLC) has gained popularity for analyses of such complex mixtures.8-10 The separation power of LCxLC has been successfully demonstrated on a variety of polymer systems11-15 and in the characterization of samples of biological interest,16-19 triacylglycerides,20 and polyols.21 The higher peak capacities obtained, as well as the ability to utilize a variety of orthogonal separation mechanisms, can prove invaluable in understanding structure-performance relationships. The heterogeneity of epoxy resins by LCxLC analyses has been previously studied.3 However, the second dimension used in those studies had relatively long cycle times (5 min) limiting the ability to obtain the desired 3-4 first-dimension cuts per peak to prevent loss of information in the two-dimensional experiment.22 The different eluting components were subsequently identified by collection of fractions using preparative Liquid Chromatography at Critical Conditions (LCCC) and analysis of the fractions using off-line Matrix-Assisted Laser Desorption/Ionization-Time of Flight (MALDI-ToF) mass spectrometry. The present report summarizes a research study undertaken to develop the LCxLC and LCxLC-ESI-MS methodologies for detailed characterization of solid epoxy resin functionalities as a (11) Pasch, H.; Adler, M.; Rittig, F.; Becker, S. Macromol. Rapid Commun. 2005, 26, 438–444. (12) Pasch, H.; Adler, M.; Knecht, D.; Rittig, F.; Lange, R. Macromol. Symp. 2006, 231, 166–177. (13) Knecht, D.; Rittig, F.; Lange, R. F. M.; Pasch, H. J. Chromatogr. A 2006, 1130, 43–53. (14) Weidner, S.; Falkenhagen, J.; Krueger, R.; Just, U. Anal. Chem. 2007, 79, 4814–4819. (15) van der Horst, A.; Schoenmakers, P. J. Chromatogr. A 2003, 1000, 693– 709. (16) Bushey, M. M.; Jorgenson, J. W. Anal. Chem. 1990, 62, 161–167. (17) Wagner, K.; Miliotis, T.; Marko-Varga, G.; Bischoff, R. Anal. Chem. 2002, 74, 809–820. (18) Delahunty, C.; Yates, J. R. Methods 2005, 35, 248–255. (19) Kajdan, T.; Cortes, H.; Kuppannan, K.; Young, S. J. Chromatogr. A. 2008, 1189, 183–195. (20) Dugo, P.; Kumm, T.; Chiofalo, M. L.; Cotroneo, B. A.; Mondello, L. J. Sep. Sci. 2006, 29, 1146–1154. (21) Entelis, S. G.; Gorshkov, A. V. Adv. Polym. Sci. 1986, 76, 129–175.

Figure 2. Representative overlaid size exclusion chromatograms (Diode array detector: Extracted 228nm) of solid epoxy resin sample SER-1.

Figure 3. Representative overlaid chromatograms of liquid chromatography at critical conditions for solid epoxy resin sample SER-1 (U.V.Detector; 228nm).

function of molecular weight by coupling Size Exclusion Chromatography (SEC) and LCCC. LCCC was chosen as the second dimension separation rather than the first since relatively fast (90 s) separations were achieved in the LCCC mode. This allowed sufficient sampling of the first dimension separation with better reproducibility. The objectives of the study were to separate oligomers with different functional end groups, determine their molecular weight ranges, and identify and quantitate the separated components. EXPERIMENTAL SECTION All LCxLC experiments were performed on a Shimadzu Prominence LC20square Comprehensive LC system (Shimadzu

Scientific Instruments, Columbia, MD). The data were analyzed using the LCxLC Manager Version 1.0 software (Shimadzu). Ammonium formate was purchased from Sigma (St. Louis, MO). HPLC-grade methanol (MeOH) and dichloromethane (CH2Cl2) were purchased from J.T. Baker (Phillipsburg, NJ). Three different solid epoxy resin samples, poly(bisphenol A-coepichlorohydrin) glycidyl end-capped were used. 1,3-Bis(4-(2-(4(oxiran-2ylmethoxy)phenyl)propan-2-yl)phenoxy)propan-2-ol, designated SER-1 (Sigma-Aldrich catalog number 405450); DER 661 (22) Murphy, R. E.; Schure, M. R.; Foley, J. P. Anal. Chem. 1998, 70, 1585– 1594.

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4273

Figure 4. LCxLC plot of solid epoxy resin sample SER-1. The tentative assignments for the components are summarized in Table 2. Displayed x-axis numerical values are minutes divided by 10. Table 1. Precision Data for Injection Repeatability for Solid Epoxy Resin SER-1a peak ID

peak area %(Inj 1)

peak area %(Inj 2)

peak area %(Inj 3)

average

std. dev.

RSD (%)

A B D G H I J K L

0.780 6.97 68.7 3.57 0.927 0.302 1.09 2.69 14.9

0.657 6.62 68.3 3.64 1.05 0.429 1.05 2.79 15.4

0.601 6.36 69.0 3.60 1.03 0.45 1.04 2.88 15.1

0.68 6.7 68.7 3.60 1.00 0.40 1.06 2.8 15.1

0.09 0.3 0.4 0.04 0.07 0.08 0.03 0.1 0.3

13 4.6 0.51 0.98 6.6 21 2.5 3.4 1.7

a The peak identifications are for the components labeled and identified in Figure 4 and Table 2, respectively. Note that the component with Peak ID of “C” was not present in this sample. Also, components identified as “E” and “F”, were observed, but were at too low a concentration to reliably quantitate.

made via the process delineated in Figure 1 (The Dow Chemical Company, Freeport, TX); and SER-2, prepared by the well-known Taffy process as described in the introduction. All samples were prepared as 3.0 wt % solutions in 3.5% MeOH in CH2Cl2. First dimension conditions were as follows: Column was a 4.6 mm × 250 mm PLgel MIXED-E, 5 µm particle size (Polymer Laboratories, Varian, Amherst, MA). The mobile phase was 3.5% v/v MeOH in CH2Cl2 with flow rate of 0.100 mL/min for 15 min, followed by 0.012 mL/min from 15.01 to 160 min. The injection volume was 1 µL. The second dimension conditions were as follows: Column was a 2.1 mm × 100 mm Agilent Rx-SIL, 1.8-µm particle size (Agilent Technologies, Santa Clara, CA, USA). Mobile phase A and B were 1% v/v and 6% v/v MeOH in CH2Cl2, respectively. The LCCC conditions were optimized daily between 48-52 A:B ratios. The flow rate used was 2.1 mL/min with an injection volume of 18 µL via two parallel loops on a 10-port switching valve. The temperature of the entire system was maintained at 30 °C. The first and the second dimension analyses were optimized with a micro-Diode Array Detector (Extract: 228nm) and a UV Detector (228nm), respectively. Quantitation was performed by normalization of UV detector signals, since standards for components separated are not available. LCxLC-ESI-MS Analysis. The Shimadzu Prominence LC20square Comprehensive LC system was coupled to a LCTPremier TOF-MS system (Waters, Framingham, MA) equipped with a Z-spray source. After the LCxLC analysis, effluent from 4274

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

the UV detector was split using a graduated splitter such that approximately 700-800 µL/min of the split flow was directed toward the mass spectrometer. This split eluent stream was merged with a 100 µL/min solution of 0.03 M ammonium formate in MeOH. The post-addition of ammonium formate solution ensured sufficient ionization of eluting epoxy resin components. The samples were analyzed in positive ion mode in the W configuration under nominal mass conditions. The capillary voltage was maintained at 2.5 kV while the source and desolvation temperature were maintained at 110 and 250 °C, respectively. The scan range was between 250 to 2000 amu. The MCP was maintained at 2250 V and the scan speed was set at 0.5 s per scan. RESULTS AND DISCUSSION To characterize the different functional groups present in solid epoxy resins, SEC and LCCC were selected as first and second separation dimensions. Before carrying out the LCxLC analyses, the two dimensions were independently optimized. In LCxLC analyses the first dimension separation is typically performed on a narrow-bore column using low flow rates.23,24 Because of lack of commercially available micro SEC columns, the first dimension separation was performed on a conventional-size column. Initial experimental results revealed that chromatographic peak shapes did not noticeably deteriorate because of band-broadening at very low flow rates (12 µL/min). However, the use of the conventionalsize column with a flow rate of 12 µL/min resulted in a dwell time of over 50 min. To shorten the run time and optimize for the subsequent two-dimensional experiments, the flow rate was set at 0.100 mL/min for the first 15 min. At 15.01 min the flow was reduced to 12 µL/min. Changing the flow rate in this manner did not affect peak shapes but significantly reduced the analysis time. Under these conditions, solid epoxy resins separate on the basis of their hydrodynamic volume. Figure 2 represents an overlay of chromatograms from three successive analyses of the SER-1 solid epoxy resin. The results demonstrate excellent repeatability in the first dimension separation. During LCCC analyses, the critical conditions for chromatographic separations have been defined as the point where the distribution coefficient of a polymer is unity regardless of the (23) Cortes, H. J. Chromatogr. 1992, 626, 3–23. (24) Dugo, P.; Cacciola, F.; Kumm, T.; Dugo, G.; Mondello, L. J.Chromatogr. A 2008, 1184, 353–368.

Table 2. Summary of Mass Spectral Data and Tentative Assignments for the Major Components Observed in SECxLCCC-ESI-MS Analysis of Solid Epoxy Resinsa

molecular weight.25,26 For this condition, the Gibbs free energy (∆G) change of the process must be zero, so that enthalpic

(adsorption) and entropic (exclusion) interactions are balanced. When these conditions are achieved, functional groups will interact Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4275

Table 2. Continued

a

a, b, c correspond to molecular weight ranges for solid epoxy resins SER-1, DER 661, and SER-2, respectively.

exclusively with the stationary phase, where polymer molecules will elute based on “defects” in the molecule independent of molecular mass. Critical chromatography has proven useful in the characterization of a variety of polymers.26-28 Figure 3 represents an overlay of UV chromatograms from three successive analyses of the SER-1 epoxy resin components eluting under critical conditions. The overlay also demonstrates excellent repeatability obtained for the second dimension LCCC separation. Throughout the course of the study, the second dimension mobile phase ratio was adjusted to achieve critical separation conditions whenever fresh mobile phase was prepared. Operating conditions were set so that each second dimension analysis was completed in 1.5 min. On the basis of Figure 3, as many as 12 components are at least partially resolved in the epoxy resin mixture. The first and second dimension columns were connected through a 10-port switching valve. Sample loops on the 10-port valve captured all of the effluent from the first dimension analysis and sequentially injected each onto the second dimension (25) Gorshkov, A. V.; Verenich, S. S.; Evreinov, V. V.; Entelis, S. G. Chromatographia 1988, 26, 338–342. (26) Pasch, H. Adv. Polym. Sci. 2000, 150, 1–66. (27) Falkenhagen, J.; Much, H.; Stauf, W.; Muller, A. H. E. Macromolecules 2000, 33, 3687–3693. (28) Heinz, L. C.; Macko, T.; Pasch, H.; Weiser, M. S.; Mulllhaupt, R. Int. J. Polym. Anal. Charact. 2006, 11, 47–55.

4276

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

column. A first dimension flow rate of 12 µL/min and a modulation (valve-switching) time of 1.5 min resulted in a second dimension injection volume of 18 µL. A total of 107 first dimension “aliquots” were analyzed in the second dimension using the orthogonal separation mechanism. The resulting data allowed two-dimensional contour plots to be generated for each sample. Figure 4 represents the LCxLC contour plot of solid epoxy resin sample SER-1. On the basis of the orthogonal separation mechanisms chosen, components in the epoxy samples were first separated on the basis of their hydrodynamic volume (molecular weight). Then, coeluting components with the same hydrodynamic volume were further separated based on their functionalities. As observed in the LCxLC map, substantial overlap of eluting components between 50-110 min was observed in the first dimension analyses. However, with the availability of additional separation space in the second dimension, each of the co-eluting components was further resolved. Currently, quantitation of components separated in the LCxLC space is in its developmental stage.29 Quantitative information on composition of a representative solid epoxy resin sample, SER-1, that was separated into individual components in the 2-dimensional space was obtained. Three such replicate injections were analyzed. (29) Mondello, L.; Herrero, M.; Kumm, T.; Dugo, P.; Cortes, H.; Dugo, G. Anal. Chem. 2008, 80, 5418–5424.

Figure 5. (a) Total ion chromatogram (TIC) in the positive ion mode for LCxLC separation of solid epoxy resin sample SER-1 and (b) Expanded TIC depicting the separation in one of the 2nd dimension cuts (1.5 min window).

If reliable quantitative information is to be obtained, peak response and retention times must be reproducible in both analysis dimensions. Since a number of the components separated had not been previously identified, and standards were not available, quantitation was performed by normalization of their respective UV responses. Because each of the components originated from the same epoxy repeating unit, this is not believed to be a significant detriment. The components present in the LCxLC maps were quantitated using the LCxLC Manager version 1.0 software. Quantitation repeatability in the 2D separation space was deter-

mined using the integration function of the LCSolution software. Integration results from quantitation of components in the LCxLC separation are presented in Table 1. Excellent injection repeatability was demonstrated for the main component, 0.5%. The RSD for components at the percent level ranged from 0.98% to 6.6%. The two components present below 1% had RSD values of 13% and 21%, reflecting higher relative error and also reflecting the inability to specify the exact size and shape of the integration zones using the current beta version of the quantitation software. Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4277

Figure 6. Extracted ion chromatograms (XICs) for the main product, SER, depicting the representative mass range for oligomers.

Figure 7. Comprehensive LCxLC plots of (a) solid epoxy resin sample DER 661 (Fusion process) and (b) solid epoxy resin sample SER-2 (Taffy process). Displayed x-axis numerical values are minutes divided by 10. 4278

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

LCxLC-ESI-MS Analyses. To identify the different eluting components, the LCxLC instrument was coupled to the ESI-MS as detailed in the Experimental Section. The mass spectrometric analysis provides an additional complementary dimension to the LCxLC analyses. Figure 5 represents the total ion chromatogram for the SECxLCCC-ESI-MS separation of SER-1 sample. In Figure 5, an expanded total ion chromatogram with a representative separation in one of the second dimension cuts is depicted. Different scan rates were evaluated to achieve acceptable resolution for the total ion chromatograms. The scan rate for the mass spectral acquisition was optimized at 0.5 scan/sec. Acquiring at a faster scan rate did not result in any further improvement in the resolution. In comparison to Figure 3, a decrease in chromatographic resolution was observed in the total ion chromatogram for the 1.5 min second dimension run depicted in Figure 5b. This decrease in resolution can be attributed to band-broadening resulting from extra column dead volume introduced by the connections used to first split off effluent for optimum flow compatibility with the MS, and then to allow post-addition of ammonium formate to facilitate ionization. Because the MS system is used only for compound identification and molecular weight assignments, this decrease in resolution is not a significant detriment if interpretable spectra can be obtained, as was the case in this study. Quantitation and further interpretation were performed using UV detection, where resolution was not affected. The epoxy resin samples depicted a characteristic oligomeric series for the major components with a repeating unit of 284.1 amu. Since the separation in the LCCC mode takes place on the basis of end group functionality, there is limited overlap of other functionalities. This prevents any issues related to quenching of ionization because of co-eluting components. Figure 6 represents the extracted ion chromatograms for the different oligomers of the main product, in the mass range 352.5 amu (n ) 0) to 5188.5 amu (n ) 17). Note that this mass range could not be observed in single-dimensional LC-MS analysis because of the polydispersity and significant overlap of components from the side reactions. The other major functional group of interest is the hydrolysis product of the glycidyl-ether, the R-glycol group. It was clear that SECxLCCC provides sufficient separation to provide detailed information on the molecular weight distribution of the two main components. Moreover, it also allows for quantitation of these two main components of interest. Table 2 represents the summary of the different components which were identified for the solid epoxy resin samples and their respective molecular weight ranges. It should be noted that since MS cannot differentiate isomers with the same molecular weight, structures described should be considered as tentative assignments. As an example, compounds D and G are expected to be different geometric isomers (ortho, para functionalities with same molecular weight) of the main product, or contain primary versus secondary hydroxyls from opening of the epoxide ring during polymerization. Hence, they elute at different retention times in the LCCC mode, as seen in Figure 4.

Although the components TGE-1, [2-((4-(2-(4-(2-oxiran-2-ylmethoxy)-3-(4-(2-(4-(oxiran-2-ylmethoxy)phenyl)propan-2-yl)phenoxy)propoxy)phenyl)propan-2-yl)phenoxy)methyl]oxirane], (triglycidyl ether of Bisphenol-A), TGE-2, [2-((4-(2-(4-(oxiran-2ylmethoxy)-2-(2-(4-(oxiran-2-ylmethoxy)phenyl)propan-2yl)phenyl)propane-2-yl)phenoxy)methyl)oxirane], (triglycidyl ether of Trisphenol), and SER were not well-resolved in the contour plots, the addition of a third dimension in mass spectrometry allowed identification of these components. On the basis of the analyses, it was observed that the current separation strategy in the second dimension will not separate species more polar than the MHR [3-(4-(2-(4-(2-hydroxy-3-(4-(2-(-4-(oxiran-2-ylmethoxy)phenyl)propan-2-yl)phenoxy)propoxy)phenyl)propan-2-yl)phenoxy)propane-1,2-diol], (Monohydroxy resin, epoxy - R-glycol), oligomers. Since the amount of such components is expected to be very low, this restriction does not represent a major disadvantage. Panels a and b of Figure 7 represent the LCxLC maps of the Fusion process and Taffy process resins, respectively. On the basis of the contour plots, different side reactions and their respective molecular weight ranges can be observed. As evident, significantly more MHR is made in the sample produced via the Taffy process. A continuous series of MHRs were observed for the epoxy resin made via the Taffy process in contrast to only even numbered (n ) 0, 2, 4,...) MHRs observed in the epoxy resin made via the Fusion process. CONCLUSIONS LCxLC-ESI-MS analyses can provide unique insights and understanding of the composition of synthetic polymers as demonstrated in the present work for epoxy resins. The different LCxLC patterns clearly depict significant differences in functionality and molecular weights caused by differing polymerization conditions. The technology developed separates numerous species with different combinations of end-group functionalities, allowing quantitation and determination of their individual molecular weight ranges for the first time. On the basis of the results obtained, it is concluded that LCxLC-ESI-MS provides a powerful technique for comparison of epoxy resins developed using different polymerization conditions. This information is useful for verifying molecular architecture models for the production of solid epoxy resins. The technologies developed are expected to be applicable to a wide variety of synthetic polymers. ACKNOWLEDGMENT The authors would like to thank Dr. Masayuki Nishimura and the Shimadzu Corporation for their help, assistance, and use of the Shimadzu Prominence LC20square Comprehensive LC system.

Received for review January 15, 2009. Accepted March 29, 2009. AC9001047

Analytical Chemistry, Vol. 81, No. 11, June 1, 2009

4279