Ethylene-Propylene-Diene-Monomer elastomers have uses in automotive parts, roofing and motor oil. Historically, the molecular weight has been measured using high temperature GPC. This FIPA method uses cyclohexane at a temperature of 60°C
Ethylene-Propylene-Diene-Monomer (EPDM) elastomers are random amorphous polymers with outstanding properties and broad applications. These materials are commonly found in automotive parts, single-ply roofing membranes, motor oil formulations etc. Historically, molecular weight distribution (MWD) characterization of EPDM has been carried out using high temperature gel permeation chromatography (HT-GPC). Common run conditions involve the use of trichlorobenzene (TCB) as the solvent and measurements at 140 °C.
The FIPA method described below provides an alternative chromatographic method that is much more user friendly. In this application, the EPDM samples are first dissolved in hot (60 to 80 °C) cyclohexane. The time and temperature required for this dissolution can vary depending on the molecular weight and the crystallinity of the individual samples. EPDM samples are not soluble in THF, even up to the boiling point of THF. However, injecting samples dissolved in cyclohexane at 60 °C directly in the GPC running THF worked for the FIPA system and excellent molecular results were obtained.
The front end of the chromatography system, a Viscotek GPCmax combination of degasser, pump, and autosampler was used. The detection was made on a Triple Detection Array (TDA) system. The detectors in series were the light scattering (LS), the differential refractometer (RI) and the viscometer detector (differential pressure,DP)
The eluent used was tetrahydrofuran (THF) at a flow rate of 1.0 mL/min. The separation of the polymer and the solvent was carried out using a ViscoGEL FIPA-100H column. The samples were diluted to around 2 mg/ml and a volume of 100 µL was injected.
Figure 1a): FIPAgram of the EPDM sample #2.
Figure 1b): FIPAgram of the EPDM sample #3.
Figure 1 shows two EPDM FIPA chromatograms (FIPAgrams).which are quite different in Mw values. For figure 1a, the weight average molecular weight was 119,000 g/mol while the sample in figure 1b had a molecular weight of 1,329,000 g/mol. Although they have an order of magnitude difference in molecular weight, the two samples elute between 1.5 and 2 mL. The fact that FIPA does not provide separation of the sample, but just separation of the solvent is shown in this figure. The cyclohexane in which the EPDM was dissolved in eluted between 3 and 4 mL.
Table 1: Weight average molecular weight, weight average intrinsic viscosity standard deviation for 4 repeat FIPA measurements each.
Table 1 shows the great versatility of FIPA to analyze very different molecular weight range samples. FIPA can only provide molecular weight and IV averages (i.e., no distribution information). It is clear here that the repeatability of these FIPA runs (4 injections for each sample) is excellent for both Mw and IVw.
The FIPA results were compared to GPC under similar conditions. Table 2 presents the results for the GPC and FIPA of the same three samples. All samples were run in triplicates. The averaged RSD for Mw is 0.41% for FIPA and 2.74% for GPC. The averaged RSD for IVW is 0.24% for FIPA and 1.85% for GPC. The precision of the GPC measurements is fairly typical and FIPA represents a simple, non user-dependant way to very large gains in precision.
Table 2: Weight average molecular weight, weight average intrinsic viscosity standard deviation for FIPA measurements run in triplicate.
|Sample ID||Mw (Da)||SD (Da)||%RSD||IVW (dL/g)||SD (dL/g)||%RSD|
However, the GPC data will provide additional information of the distribution of molecular parameters. This gain in precision is also accomplished in about 1/3 of the GPC analysis time. The FIPA data was gathered at the Viscotek Houston laboratory, while the GPC data was provided by the Bayer Laboratory at Sarnia, ON, Canada.
In conclusion, FIPA is a fast, precise and accurate analytical tool for routine process and quality control. It is an ideal rapid and automated alternative to GPC if there is no requirement for information on the sample distribution.