GPC/SEC analysis of polymer branching

Introduction

Polymer branching can have an enormous effect on the physical properties of the final synthesized polymer.  Branching can disrupt crystallinity, increase chain entanglements, increase single chain density, and affect a variety of polymer properties such as glass transition temperature, modulus, toughness, etc.  Understanding polymer branching and its effects on a polymer’s end-use properties can help avoid failures and improve specifications.  This Application Note describes the GPC (or equivalently SEC) analysis of branching in polyvinylpyrrolidone (PVP).  PVP is a polymer used in applications as a binder, as a thickening agent, in adhesives and in coatings and films.  These uses necessitate the manufacturer to have an understanding of the polymer structure to know how the final product will perform.          

Experimental

Three PVP samples were analyzed for this Application Note.  As received, this set of samples was believed to contain one linear reference and two branched samples.  These polymers will be referred to herein as samples Linear, Branched 1, and Branched 2.  The samples were dissolved and analyzed in a blended mobile phase of 50:50 methanol and water with 0.05M sodium sulfate.    

A Malvern OMNISEC RESOLVE and REVEAL triple detection GPC system was used to analyze the three polymers (Figure 1).  The system is comprised of a refractive index (RI) detector to measure the concentration at each elution time, a viscometer detector to measure the intrinsic viscosity (IV) of the sample, and a light scattering detector to directly measure the molecular weight (MW).  The viscometer and light scattering detectors are crucial to determining polymer branching.    

Figure 1.  Malvern’s OMNISEC Triple Detection GPC/SEC System.

Chromatograms and Results

To start, let us look at the overlay of the RI chromatograms of these samples.  This is the classical way to compare samples using GPC/SEC.  Figure 2 shows the RI signal overlays for samples Linear (red), Branched 1 (purple), and Branched 2 (green). 

Figure 2.  Overlay of the RI chromatograms for the Linear (red), Branched 1 (purple), and Branched 2 (green) samples.

From inspection of Figure 2, samples Branched 1 and Branched 2 have larger chains in solution, as evidenced by the earlier elution volumes, and they have broader distributions as well.  However, this one detector does not tell us anything about the relative degree of branching in these samples.  We will need to look at the full triple detector suite to gather the information we are looking for.  Figures 3 - 5 show the triple detector chromatograms for samples Linear, Branched 1, and Branched 2, respectively.  In these chromatograms, the RI detector signal is shown in red, the viscometer signal is shown in blue, and the right angle light scattering signal is shown in green.  The calculated MW at each slice of the chromatogram is shown with the gold curve and the calculated IV at each slice of the chromatogram is shown with the light blue curve.

Figure 3.  Triple detector chromatogram of sample Linear.

Figure 4.  Triple detector chromatogram of sample Branched 1.

Figure 5.  Triple detector chromatogram of sample Branched 2.

The differences seen in Figure 2 are exacerbated in Figures 3 – 5 with the addition of the viscometer and light scattering detectors.  Sample Linear shows a fairly standard, Gaussian peak shape in all the detectors in the chromatogram.  Sample Branched 1 has a much more complex shape with a peak shoulder eluting at earlier retention volumes (~14.5 mL) that has markedly higher viscometer and light scattering signals than the predominant peak in the RI signal (~17.5 mL).  Sample Branched 2 does not have the Gaussian shape of sample Linear, nor the noticeable shoulder at earlier retention volumes of sample Branched 1.  It does have a broad distribution as it elutes over more than 5 mL of retention volume, but nothing else immediately stands out from Figure 5.     

To look at branching in these samples, we plot the IV (light blue curves above) as a function of the MW (gold curves above).  This gives us the Mark-Houwink plot, and the overlay of the three samples is shown in Figure 6 below.    

Figure 6.  Mark-Houwink overlay plot of samples Linear (red), Branched 1 (purple), and Branched 2 (green).

For a given polymer chemistry (as we have in this sample set), there are three categories of samples that we can define using this plot:

• The first category is a linear sample.  This plot would appear with the highest IV at a given MW and a fixed slope over the entirety of the plot. 
• The next sample category is a random, long-chain branched sample.  This plot would appear as having a gentle curvature downward.  The IV of a sample in this category would be very close to the linear sample at low MWs but would be noticeably lower than the linear sample at higher MWs as the level of branching becomes more significant with increasing MW.
• The last sample category is a systematic, controlled, short-chain branched sample.  This plot would appear as a parallel line to the linear sample, but would have a lower IV at all MWs.  Without a linear reference, it would be nearly impossible to differentiate this sample category from the linear category. 

From inspection of Figure 6, sample Linear appears linear (as expected) and sample Branched 1 appears to be systemically branched (as expected).  However, sample Branched 2 overlays with sample Linear and therefore appears to fall into the linear category as well.  This last one is unexpected and very interesting and lets the polymer manufacturer know that the attempt to induce branching in that sample was unsuccessful.       

In addition to a comparison and categorization of polymer branching, the GPC/SEC analysis gives quantitative information about the molecular properties of these polymers.  These results are summarized in Table 1 below.  The data shown is the average and standard deviation of three injections of each sample.  Presented values include the moments of the MW distribution (M_n, M_w, and M_z), dispersity (M_w/M_n), intrinsic viscosity (IV), and hydrodynamic radius (R_h).

Table 1.  Calculated results for the PVP samples

Immediately apparent from this data is that these three samples are very different.  Sample Linear is much lower MW than the other two samples and also has a much narrower distribution (lower dispersity).  Despite the vast differences between sample Linear and sample Branched 2, the Mark-Houwink plots overlay on each other over a wide range (see Figure 6).          

Conclusion

GPC/SEC provides an effective method of determining the presence polymer branching.  By combining the data obtained from the viscometer detector (IV) and the light scattering detector (absolute MW), information regarding the amount and type of branching can be discerned.  In addition to branching information, GPC/SEC provides a wide variety of molecular properties including MW, dispersity, solution viscosity, and hydrodynamic size.  This collection of information gives a thorough portrayal of the analyzed polymer and provides numerous avenues for correlation to the end-use properties of the polymer.