Gradient Analysis of Polymer Blends, Copolymers, and Additives with ELSD and PDA Detection

  • Introduction
  • Experimental
  • Results & Discussion
  • Summary


In recent years there has been increased interest in using gradient HPLC techniques, such as Gradient Polymer Elution Chromatography (GPEC), with polymers for determining the compositional drift of copolymers, the composition of polymer blends, or for the analysis of polymer additives. Depending upon the gradient conditions and columns selected for analysis, separations may be obtained dependent on molecular weight or based upon precipitation, or adsorption mechanisms. The use of an Evaporative Light Scattering Detector (ELSD) allows one to perform solvent gradients with a universal mass detector and observe both UV absorbing and non-UV absorbing polymer samples without baseline disturbances from the solvent gradient. The addition of a Photodiode Array Detector (PDA) allows for compositional analysis across the molecular weight distribution of many copolymers, can be useful for the identification of components in a polymer blend, and also is invaluable for the quantitation of polymer additives and other small molecules in traditional reverse phase separations.

This section demonstrates the advantages of gradient analysis of polymers as compared to results obtainable with Gel Permeation Chromatography. The instrumentation used to carry out this work is described and examples of this technique for the analysis of polymer blends are shown. The effects of column functionality and solvent composition on the separation of polystyrene standards and samples is described and the best conditions observed are used to analyze various copolymers for monomer composition. Finally, the traditional use of gradient separations with the same instrumentation for the analysis of several types of polymer additives is also shown.


The most common chromatographic method for the analysis of polymers is Gel Permeation Chromatography (GPC) where the separation is based upon the size of the polymer sample in solution, or the hydrodynamic volume of the polymer solution.Figure 1 shows the chromatograms obtained using GPC for a polystyrene sample, polystyrene-acrylonitrile copolymer (25% acrylonitrile) and a polystyrene-butadiene rubber (50% styrene) analyzed separately.Even though the samples are of different molecular weight, the hydrodynamic volumes are similar enough that the polymer peaks are observed at nearly the same retention time.The chromatograms obtained for the GPC analysis of a blend of approximately the same concentration of each of the polystyrene, polystyrene-acrylonitrile, and the polystyrene-butadiene samples are also shown in Figure 1. This chromatogram shows no separation of the three different polymers and demonstrates the impracticality of GPC for the analysis of most polymer blends. 

However, when this same polymer blend is analyzed in a gradient mode, the three components can easily be baseline resolved as demonstrated in Figure 2 which shows the overlay of two replicate injections of the polymer blend run on a prototype divinylbenzene-vinylpyrolidone column with a gradient from 100% Acetonitrile (ACN) to 100% Tetrahydrofuran (THF) over 20 minutes.

Using this technique, the samples are dissolved in THF and then injected into the chromatographic system running 100% ACN. The polymers in the blend are insoluble in acetonitrile and precipitate onto the column. As the gradient proceeds, the polymers in the blend are redissolved according to their solubilities and are eluted from the column as well resolved peaks. This mechanism is similar to Gradient Polymer Elution Chromatography (GPEC). Other gradient methods for the analysis of polymers have been described in the literature which are performed under conditions where the polymers remain in solution and are separated by an adsorption mechanism, but these are generally for polar polymers that are soluble in alcohols or ketones run on bare silica columns and are not discussed here.


All gradient work was carried out using the following system configuration unless otherwise noted.

Waters Alliance 2690 Separations Module with column heater at 30 ºC
Detector 1:
Waters 996 Photodiode Array Detector
Detector 2: 
Alltech Model 500 ELSD with LTA Adapter
(Drift Tube at 40º C, 1.75 Liters/min Nitrogen)
Data System:
Waters Millennium 32 Chromatography Manager
As listed in Figures, 30 ºC
Flow Rate:
10 - 25 µl injections of 0.2 - 0.5% samples
Linear gradient, conditions and mobile phases as listed in Figures.


The most commonly used detector for GPC is the Refractive Index (RI) detector; however, the sensitivity of the RI to changes in mobile phase composition makes it unacceptable as a detector for Gradient Polymer Analysis. Figure 3 shows the chromatograms obtained for the 25 µl injection of a 0.5% solution of a styrene-acrylonitrile copolymer (25% Acrylonitrile) run on a prototype DVB/Vinylpyrolidone column with a gradient from 100% ACN to 100% THF in 20 minutes using a refractive index detector (RI), a photodiode array detector (PDA), and an evaporative light scattering detector (ELSD).

As soon as the mobile phase change from the gradient reaches the RI detector (~2.5 minutes) the RI signal goes offscale, completely overloading the detector. The chromatogram obtained from the PDA detector at 260nm (or any UV detector) demonstrates that UV detection is much better suited for gradient analysis than RI detection. The chromatogram does show baseline drift with the change in mobile phase but there is still good sensitivity for the polymer sample and the drift can easily be eliminated by baseline subtracting a blank gradient run. The third chromatogram in Figure 3, obtained using an ELSD, demonstrates the superior performance of the ELSD for gradient applications. The detector is essentially insensitive to the changes in the mobile phase composition since the solvents are evaporated prior to detection. This, combined with the excellent sensitivity for polymer samples, makes the ELSD the detector of choice for gradient analysis of polymers. By combining a PDA with the ELSD, one can detect and quantitate unknowns with the ELSD and use the PDA to determine peak purity, for the identification of unknowns through library matching, and for compositional analysis of copolymers.

Using this system, a wide variety of different types of polymers, polymer blends and copolymers can be analyzed. Figure 4 shows an overlay of chromatograms obtained for many types of polymers run on a Nova-Pak C18 Column with a 30 minute gradient from 100% ACN to 100% THF including polyvinylchloride, polymethylmethacrylate, polystyrene, polystyrene-butadiene block copolymer, polydimethylsiloxane, polystyrene-isoprene block copolymer, and butyl rubber.

When using this technique for the analysis of polymer blends or copolymers, it is necessary that the separation be independent of molecular weight so that the polymers are separated only by composition. Unfortunately, since this is primarily a precipitation/redissolution mechanism, some molecular weight dependence is inevitable, but it can be minimized through the judicious selection of columns, mobile phases, and gradient conditions.

Figure 5 shows an overlay of chromatograms obtained from a series of narrow polystyrene standards run on a SymmetryShield C8 Column (3.9 mm x 15 cm) with a gradient form 100% ACN to 100% THF in 10 min.

The standards from 43,900 to 2,890,000 MW elute in a band from approximately 9 to 9.5 minutes. The lower MW standards elute earlier, with many of the oligomers well resolved. These lower molecular weight standards are soluble or nearly soluble (9100 MW) in the starting conditions of the gradient (100% ACN) and are therefore separated by the traditional reverse phase mechanism. Figure 6 shows an overlay of the chromatograms obtained for the same standards run under identical conditions on a prototype DVB/vinylpyrolidone column (3.9 mm x 15 cm).

A similar pattern is observed with the 43,900 to 2,890,000 standards eluting in a slightly narrower band. The separation of the lower molecular weight standards is somewhat different; however, this is not surprising due to the different reverse phase characteristics of the two columns.

By changing to a Nova-Pak C18 Column (3.9 mm x 30 cm) and using a 30 min gradient, the chromatograms shown in Figure 7 were obtained. Using these conditions, the molecular weight dependence for the 43,900 MW and higher polystyrene standards is nearly eliminated. As expected the lower MW standards that are soluble in ACN are eluted earlier in the chromatogram, however, the low MW oligomers are being split into three peaks, indicating that they are being separated by their differing end groups.

The choice of mobile phase used as the non-solvent can have significant effects on the separations obtained from gradient analysis of polymers.

Figure 8 shows an overlay of chromatograms obtained for the same standards run on a Nova-Pak C18 Column (3.9 mm x 15 cm) with a linear gradient from 100% Methanol (MeOH) to 100% THF in 30 min. These results show a clear dependence on molecular weight from the well-resolved oligomers early in the chromatograms to the 8 million MW standard. This is undesirable for the purposes of copolymer or polymer blend analysis, as it would be difficult to determine whether differences in retention time were due to compositional differences or MW differences.

This non-solvent effect can also be seen when analyzing broad MW polymer samples.

Figure 9 shows the chromatograms obtained for NBS706 broad polystyrene standard run on a Nova-Pak C18 Column (3.9 mm x 15 cm) with a 30 min gradient using first ACN and then MeOH as the non-solvent with THF as the solvent for both injections. When using ACN as the non-solvent, a more desirable sharp peak is obtained whereas when MeOH is used as the non-solvent, a very broad peak is obtained. Our work has shown that for THF soluble polymers, the best separations were observed with the 100% ACN to 100% THF gradient. These conditions give a rugged method that can be used for a wide variety of polymer blends and copolymers.

Gradient analysis is a powerful tool for evaluating copolymer materials. A series of random styrene-butadiene rubbers (SBR) were run using this 100% ACN to 100% THF gradient on a prototype DVB/Vinylpyrolidone column (3.9 mm x 15 cm) in 20 min. Five different SBRs with composition ranging from 50% styrene to 5.2% styrene were injected along with a narrow polystyrene standard (355K MW) and a narrow polybutadiene standard (330K MW). An overlay of the resulting chromatograms is shown in Figure 10.

The different SBRs are easily separated by their relative amounts of styrene and butadiene. These SBRs were previously analyzed by traditional GPC to be sure that the molecular weights were high enough that molecular weight dependence would be insignificant, and the molecular weights were all found to be approximately 200,000 to 300,000 by relative calibration with polystyrene.

Using the gradient results, a calibration curve was constructed to determine % styrene vs retention time and is shown in Figure 11.

The plot exhibits a good correlation between % styrene and retention time so that this method could be used to determine the approximate composition of an unknown SBR. The UV data from the PDA could also be used to crosscheck the results from the ELSD.

In a similar manner, Figure 12 shows the chromatograms obtained for a series of block styrene-butadiene copolymers with a similar separation as the random SBRs.

The data is plotted in Figure 13 showing a calibration curve similar to the one obtained for the random SBRs. Using this gradient method, species with only slight differences in structure can easily be separated.

Figure 14 shows an overlay of individual injections of polymethylmethacrylate, polymethylmethacrylate, poly-n-butylmethacrylate, poly-n-hexylmethacrylate, and poly-laurelmethacrylate run on a Nova-Pak C18 Column (3.9 mm x 15 cm) with a gradient of 100% ACN to 100% THF in 30 minutes. The chromatograms show excellent separation between each component in the homologous series of methacrylates and could easily be resolved with a faster gradient.

The chromatogram in Figure 15 shows the separation of the same methacrylates injected as a mixture and run under identical conditions demonstrating an identical separation when the components are run in a mixture.

This same method using identical conditions also has utility for analyzing lower molecular weight compounds. Figure 16 shows an overlay of chromatograms for two low molecular weight waxes. The two waxes are well separated and slight differences between the oligomer ratios can be observed.

Low molecular weight polymer additives can be analyzed with this method by the traditional reverse phase mechanism. Many types of polymer additives will be shown using the following conditions that were chosen to be compatible with a mass spectrometer:

Waters Alliance 2690 Separations Module with column heater at 30º C
Detector 1:
Waters 996 Photodiode Array Detector
Detector 2:
Alltech Model 500 ELSD with LTA Adapter
(Drift Tube at 40º C, 1.75 Liters/min Nitrogen)
Data System:
Waters Millennium32 Chromatography Manager
Symmetry C8, 2.1mm x 15cm, 30º C
Flow Rate:
0.29 mL/min
 Linear Ternary Gradient, 30 mins; 70/10/20 to 1/79/20 H2O/ACN/THF

Figure 17 shows the separation of Tinuvin 440, Tinuvin 900, and Tinuvin 328 that are UV stabilizers commonly used in polyolefin resins. Even though these compounds are difficult to extract from polyolefin resins with good recovery, once extracted they can be analyzed easily with good sensitivity using this method.

Several different types of phthalate plasticizers are separated in Figure 18. Phthalates, which are commonly used as plasticizers in PVC resin, have come under scrutiny recently as possible carcinogens. Phthalates, particularly diethylhexylphthalate (DEHP), are used routinely in medical devices such as catheters and IV bags and in children's toys possibly exposing patients and children to high levels of this suspected carcinogen. This method is a simple means for analyzing these phthalate compounds.

Figure 19 shows the chromatograms for the slip agents oleamide and erucamide and the antistat stearic acid. These compounds, which have very little UV absorbance, exhibit poor sensitivity with UV detection but can easily be detected with the Evaporative Light Scattering Detector.

Figure 20 shows the separation of Irganox 1076 and Irgafos 168 that are two antioxidants commonly used in polyolefins and other polymers. Irganox 1076 is a hindered amine and Irgafos 168 is a phosphite ester that degrades easily. The chromatogram shows two peaks for Irgafos 168. The second peak is the main Irgafos 168 peak while the first peak is actually the oxidized Irgafos 168 impurity that was present in the sample. This method is not meant to be an optimized method but only a general method for use with a wide variety of additives.

Figure 21 shows 12 overlays of a separation of 10 common antioxidants run using a modified version of the approved ASTM method for the analysis of additives in polyolefins. The column, mobile phases, flowrate, and gradient conditions were optimized to obtain the shortest analysis time and maximum sensitivity allowing for the analysis of these 10 antioxidants in less than 10 minutes.

The method utilizes both a mobile phase gradient and a flow rate gradient resulting in an extremely reproducible and sensitive method. The analytes were detected with a PDA at 230 nm which besides giving excellent sensitivity, also allows for peak identification using the library matching capabilities of the photodiode array detector. The instrument and conditions used to carry out this separation are shown in Figure 22.


The use of gradient methods for the analysis of polymers allows for separations that are essentially independent of molecular weight. Individual polymers in blends having the same molecular weight distribution can easily be separated and copolymers can be separated by their monomer ratios. Using the same instrumentation, mostcommon polymer additives may also be analyzed. The Evaporative Light Scattering Detector is a universal detector which is unaffected by changes in mobile phase gradient composition and the Photodiode Array Detector allows for positive identification of many compounds and compositional analysis of copolymers. These gradient methods are highly reproducible techniques and are extremely well suited for deformulation applications.

< PreviousNext >

Contact Waters

Local Offices