Size Exclusion Chromatography (SEC) for Polymers: a practical guide

This article is a practical guide to Size Exclusion Chromatography / Gel Permeation Chromatography covering instrumentation, calibration strategies, the data that can be extracted from a properly configured system, and the method development steps required to generate reliable and reproducible data. It assumes that the reader is already familiar with the basic principles of the technique. If you are new to SEC, we recommend reading our Introduction to GPC/SEC first, where the fundamental concepts are explained.

Every SEC measurement starts with one physical reality: the column does not separate molecules by molecular weight (MW). It separates them by hydrodynamic volume (V_h), which is the effective volume a polymer coil occupies in solution. Therefore, molecular weight is not measured directly, it is inferred from hydrodynamic volume through a calibration strategy. Choosing the right calibration approach is central to obtaining reliable molecular weight data from an SEC measurement.

The SEC Separation Mechanism

SEC columns contain a porous stationary phase, which is typically made of cross-linked polymer beads with a carefully engineered distribution of pore sizes. When a dissolved polymer sample passes through these columns, large molecules (those with large hydrodynamic volumes) cannot access the smaller pores and are carried rapidly through the interstitial volume (V₀) between the particles, eluting first. Conversely, smaller molecules penetrate more of the total permeation volume (V_t), spending more time inside the stationary phase and eluting later.

This is the opposite of most other chromatographic techniques, where smaller analytes travel faster. In SEC, larger molecules elute first because the pore size distribution of the stationary phase controls which fraction of the total permeation volume is accessible to any given molecule.

SEC, GPC, GFC: Resolving the Terminology Confusion

Three terms describe what is mechanistically the same separation technique:

• SEC (Size Exclusion Chromatography): Mechanism-based term. Applies to any polymer-solvent system.
• GPC (Gel Permeation Chromatography): Used for synthetic polymers in organic solvent systems.
• GFC (Gel Filtration Chromatography): Describes the same technique run in aqueous mobile phases, used for water-soluble polymers, proteins, and biopolymers.

The mechanism – separating molecules by hydrodynamic volume using a porous stationary phase – is identical in all three cases.

Throughout this article, both SEC and GPC will be used interchangeably.

Read more: GPC vs SEC vs GFC: Understanding the Difference & When to Use Each Term

A GPC/SEC analysis follows a simple sequence: pump → injector → column(s) → detector(s). Understanding what each component does, and where it can introduce errors, is what allows a polymer characterization laboratory to make sound decisions about setup, maintenance, and troubleshooting.

1. Pump: the engine of the system

Solvent flow-rate stability directly determines retention time consistency. Any drift in pump delivery would directly translate into apparent shifts in molecular weight averages between injections. This is one of the reasons why pump precision is a non-negotiable specification in GPC/SEC instruments.

Retention time reproducibility is monitored in practice using a flow rate marker: a small, chemically inert molecule injected with each sample and whose elution time serves as an internal reference. Any shift in the marker’s elution time flags a change in effective flow rate.

2. Injector: the starting point of every analysis

The injector introduces a precise volume of polymer solution into the mobile phase stream. In automated GPC/SEC systems this is performed by an automated valve injector, which improves volume precision and run-to-run repeatability compared to manual injection.

Injection volume affects both peak shape and column loading, and should be optimized during method development. Overloading the column by injecting a volume too large or a solution too concentrated causes peak broadening and distorts molecular weight averages.

3. Columns: the core of the separation

Defining the Operating Window: Exclusion Limit and Permeation Limit

Every column system has two hard boundaries that define the useful analytical window:

• Exclusion limit: The molecular weight above which all species co-elute at the void volume V₀ (the interstitial volume between the beads). Molecules above this size are 100% excluded from all pores. They appear as a single peak at the same elution volume (Ve) regardless of their actual size, and no molecular weight information is recoverable from this region.
• Permeation limit (or total permeation limit): The molecular weight below which all species fully permeate all pores and co-elute at the total permeation volume. Again, no resolution is possible.

The useful fractionation window sits between these two limits, and column selection is an exercise in matching this window to the expected molecular weight range of your sample. For instance, a column with predominantly large pores will poorly resolve the low-MW end, where species fully permeate regardless of size; and a column whose pores are too small relative to the sample’s MW range will push the high-MW fraction above the exclusion limit, where it co-elutes at the void volume with no recoverable resolution.

Column Technology: Pore Size, Particle Size, and Stationary Phase Materials

Three different parameters govern column performance:

• Particle size (typically 5 – 20 µm for analytical columns) governs separation efficiency: the number of theoretical plates (N). Smaller particles generate higher plate counts and better resolution, but also higher back-pressure. This constrains the pump and column hardware specifications.
• Pore size distribution governs the fractionation window: which molecular weight range is resolved. Mixed-bed columns blend particles with different pore sizes to extend the analytical window across a broad MW range. This convenience comes with a trade-off, as resolution in any specific MW region is somewhat lower than a matched single-pore-size column would provide.
• Stationary phase materials: If the polymer interacts chemically with the stationary phase, then retention is no longer purely size-based and the data become meaningless. For high-temperature analysis (above 100 °C), thermal stability of the column is a critical selection criterion.

4. Detector Selection: Concentration-Sensitive Detectors, Light Scattering and Viscometric detection.

Detector choice determines what information can be extracted from the chromatogram. The table below summarizes the main detector types used in GPC/SEC:

Related article: Chemical Composition measurement in GPC/SEC analysis. Differentiating polyolefins with similar Molar Mass Distribution

Related article: How to choose the most appropriate combination of detectors for high temperature SEC/GPC analysis of polyolefins

Full SEC system and its components

Calibration is the intellectual core of SEC practice. Understanding the limitations of each type is key to obtaining reliable data.

Conventional Calibration: Narrow Standards and Their Limitations

The most widely used calibration approach uses a series of narrow-distribution standards to build a calibration curve of log(M_W) versus elution volume. This curve relates a measured elution volume to a peak molecular weight for the standard polymer under the specific operating conditions.

Typical standard materials are Polystyrene (PS) and polymethyl methacrylate (PMMA) in organic systems, and Dextran and Pullulan in aqueous systems.

When, for instance, Polystyrene calibration is applied to characterize polyethylene (PE) – a chemically and architecturally different polymer – the result is an apparent molecular weight, not a true molecular weight. The calibration curve maps the elution volume to a PS-equivalent MW, not the actual MW of the PE chain.

Conventional calibration is widely used due to its simplicity, but it assumes that the test material behaves similarly to the calibration standards in solution, and that differences in molecular architecture do not significantly affect hydrodynamic size.

There are established approaches to convert molecular weights relative to polystyrene (PS) standards into molecular weights relative to the polymer of interest. The Q-factor method is based on a vertical shift of the calibration curve, while the use of the Mark–Houwink coefficients of the polymer of interest provides an alternative approach.

Although less common nowadays, calibration methods based on broad standards are also available. In such cases, the reference materials are typically selected to match the chemistry of the samples being analyzed.

Universal Calibration and the Mark-Houwink Relationship

The solution to the limitations of conventional calibration was established by Benoit and co-workers in 1967 [1]. They observed that the product of intrinsic viscosity and molecular weight (proportional to the hydrodynamic volume) is invariant in SEC separation. For flexible-chain polymers – the vast majority of synthetic polymers – this holds regardless of polymer chemistry, chain architecture, or branching, provided the polymer is in the same solvent and at the same temperature.

This enables the application of universal calibration: a single calibration curve of log([η]·M) versus elution volume can be applied to any polymer.

The way to apply universal calibration in a GPC/SEC system is by combining a four-capillary, differential viscometer detector with a concentration detector (Refractive Index or Infrared). This way, intrinsic viscosity [η] is measured at every elution slice.

When analyzing polymers at high temperatures, it is important to consider that the hydrodynamic volume of a polymer coil at high temperature is not the same as it is at ambient temperature. Thermal expansion of the chain affects its solution conformation. This is one reason why Mark-Houwink parameters are temperature-specific, and why transferring ambient-temperature calibrations to HT-GPC conditions is not valid.

Absolute Methods: Light Scattering Detector

When a Light Scattering detector is combined with a concentration detector, the absolute weight-average molecular weight (M_w) at each elution slice is obtained directly from the excess Rayleigh scattering, without the need for a calibration curve [2]. MALS also yields the radius of gyration (R_g) as a function of MW, which is useful for long-chain branching analysis.

Absolute does not mean “assumption-free”. Reliable LS measurements require accurate dn/dc for the polymer-solvent system at the measurement temperature, accurate concentration at every slice of the chromatogram, and adequate signal-to-noise at the high- and low-MW tails (which are concentration-limited).

Related article: Calibration and Data Interpretation – Additional Guidance and Extended Best Practices

GPC/SEC characterization of polymers operates at two levels. The first level (molecular weight distribution, averages, and dispersity) is accessible from any properly calibrated SEC system with a concentration detector, and it answers the most fundamental question in polymer characterization: how large are the chains, and how broadly are they distributed? The second level requires a purpose-built multi-detector configuration, and it answers the question that MW data alone cannot: what is the architecture of those chains?

4.1 Molecular Weight Distribution, Averages (M_n, M_w, M_z), and Dispersity (Ð)

GPC/SEC does not return a single molecular weight value, instead, it returns a distribution, and from that distribution, a set of statistical averages can be calculated. Each average is a different mathematical moment of the distribution curve, weighted toward a different region of it.

The practical significance of each average follows directly from which part of the distribution it reflects.

• M_n is dominated by the low-MW tail, the fraction most affected by chain ends, that includes oligomers and any low-MW additives or degradation products. Low Mn correlates with brittleness, poor solvent resistance, and elevated extractables levels, which is why it appears in food-contact and medical-device specifications.
• M_w is the most commonly specified commercial parameter, correlating with stiffness and tensile strength. It is sensitive to the middle of the distribution, which is why two materials can share an identical Mw and still behave differently in processing if their high-MW tails differ.
• M_z and M_z+1 are weighted toward the highest-MW fraction of the distribution. These are the averages that govern melt elasticity, die swell, and sagging behaviour in film and pipe extrusion, all driven by the longest chains in the sample.
• Dispersity (Ð = M_w/M_n) describes the overall breadth of the distribution. A narrow Ð (close to 2 for single-site PE) indicates relatively uniform chain lengths; a broad Ð (5–20 for multiple site PE) indicates a wide spread, with corresponding effects on the width of the processing window and batch-to-batch property uniformity. The terms Polydispersity or Polydispersity index (PDI) are still widely used as synonyms of Ð even if deprecated by the IUPAC.

Read more: Molecular Weight Distribution in Polymers: Averages, Dispersity, and How to Measure Them

 
Molecular weight distribution describes chain size but does not provide information on chain architecture. Two polyethylene samples can exhibit identical molecular weight distributions and still behave very differently in processing and end use if one contains long-chain branching while the other does not, as branched molecules elute at higher elution volumes and therefore appear to have a lower apparent molecular weight under conventional calibration. Differences in comonomer distribution across the molecular weight range can also lead to significant variations in properties.

These are aspects that molecular weight distribution alone cannot resolve, and a multi-detector configuration is typically required to address them.

4.2 Branching and Copolymer Composition: What MWD Alone Cannot Tell You

For some commercial polyolefin grades, both long and short-chain branches are present in the same sample. However, branching is not a single phenomenon: long-chain branching and short-chain branching arise from different polymerization mechanisms, affect different material properties, and require different analytical approaches to be detected and quantified.

 
Long-chain branching (LCB)  refers to polymer branches of lengths that are a significant fraction of the length of the main polymer backbone. They arise from chain transfer to polymer during synthesis, or from specific catalytic mechanisms.

LCB impacts the rheological properties of polymers, such as the melt strength and stability, as well as the shear strength and elastic behavior in melt. Even low levels of LCB can produce significant changes in melt behavior.

A SEC system with a concentration detector (Infrared or Refractive Index) and additionally a viscometer or/and Light Scattering detector resolves LCB through the following logic: a branched chain occupies a smaller hydrodynamic volume than a linear chain at the same molecular weight. Therefore, LCB can be detected by comparing the size (through R_g) or compactness (through [h]) of the branched sample to that of the linear reference at the same molecular weight.

 
Three ways to detect Long-chain Branching:

• SEC-LS: A plot of Log Rg versus Log M is known as a Conformation Plot. Comparing the conformation plot of a branched polymer with that of a linear standard allows determination of g ([R_g]²branched / [R_g]²linear) at each molar mass slice.
• SEC-VISC: A plot of Log [η] versus Log M is known as Mark-Houwink Plot. Comparing the Mark-Houwink Plot of a branched polymer with that of a linear standard allows determination of g′ ([η]branched / [η]linear) at each molar mass slice.
• SEC-LS-VISC (known as Triple Detector SEC): With this configuration the most accurate Mark-Houwink Plot is measured by using absolute molar mass as the x-axis.

Both g and g’ are indexes sensitive to LCB: g=1 indicates lack of branching, and g<1 indicates presence of long-chain branches, the lower the value the higher the LCB content. The relationship between g and g′ is given by the viscosity shielding ratio: g′=g^ε. The value of ε is dependent on a number of factors, including solvent, temperature, and branching. ε has been found to generally fall in the range 0.5 to 1.5.

The parameters g and g′, both decrease with increasing degree of branching and can be used to calculate the number of branches per chain (Branching number: B_n) as a continuous function of the molar mass of the branched macromolecule, according to Flory’s theory. For polyolefins the branching points can be considered as trifunctional and in SEC monodisperse slices can be assumed and therefore the following equation is used to quantify B_n from g:

An additional useful parameter that can be calculated using the branching number is the branching frequency, λ, defined as the number of long-chain branches per 1000 repeat units, which can be calculated from B_n:

The Conformation Plot method is considered most accurate, based directly on Flory’s theory, although the Mark Houwink Plot has the advantage of higher precision, covering larger molar mass range (it can be measured to lower molar masses), and being less sensitive to non-ideal SEC behavior (as the late elution of highly branched high molar mass molecules which distort heavily the Conformation Plot). The drawback of the Mark-Houwink Plot method is that it requires assuming a value for the viscosity shielding ratio, ε, which may even depend on molar mass.

 
Related application note: Why Is My HDPE Failing In Production When MFI and Density Look Fine.

Short-chain branching (SCB) refers to a macromolecular property distribution where small side branches are attached to the main polymer backbone. SCB occurs frequently in synthetic polymers like polyolefins, and it results from the incorporation of α-olefin comonomers into the polyethylene backbone during copolymerization.

The most common comonomers in commercial grades are 1-butene, 1-hexene, and 1-octene, producing pendant branches of 2, 4, and 6 carbons respectively. Each comonomer insertion is one branch.

Comonomer content and its distribution across the MW range are among the most consequential parameters in polyolefin characterization. SCB content and distribution directly controls crystallinity and crystallization kinetics: higher comonomer incorporation disrupts chain packing, reducing crystallinity and producing softer, more flexible material. Comonomer content also governs seal initiation temperature in packaging films, and environmental stress crack resistance (ESCR), which determines long-term performance in pipe and container applications.

In polypropylene, the methyl substituent on every other backbone carbon is itself an intrinsic form of SCB, and its stereoregularity – isotactic, syndiotactic, or atactic – determines crystallinity and mechanical performance.

SCB along the molecular weight distribution

Bulk comonomer content – provided by NMR or FTIR – doesn’t tell the full story. What matters is where in the molecular weight distribution the SCB resides. For example, in Ziegler-Natta LLDPE, comonomer is typically incorporated into shorter chains (the low-MW fraction carries more SCB than the high-MW fraction). In metallocene (single site) LLDPE, comonomer distributes far more uniformly across the MW range. Two materials can have identical bulk comonomer content and identical density and yet perform completely differently in film or pipe service because of this difference in how SCB is distributed across the MWD.

This is the measurement that standard GPC/SEC detectors cannot make. Refractive Index, Light Scattering, and Viscometry all respond to concentration or to the molar mass or hydrodynamic volume of the chain, but none carries information about the existence of pendant groups. An Infrared (IR) detector, however, responds to specific C-H absorption bands. The ratio of the methyl (CH₃) band intensity to the methylene (CH₂) band intensity, measured continuously across the elution profile, provides a quantitative measure of comonomer content at each MW slice.

The infrared detector at the core of Polymer Char’s GPC-IR instrument is designed for this measurement. The IR collects two absorbance channels simultaneously: the CH₂ band sensitive mostly to the polymer backbone carbons, and the CH₃ band, carrying the comonomer content signal. The ratio of these two signals, calibrated against standards of known comonomer content, yields short-chain branching frequency in branches per 1000 carbons as a continuous function of molecular weight – the SCB distribution across the full MWD – from a single injection.

Related article: Short-chain Branching in polyolefins: what causes short-chain branches and how to study them

Sample Preparation for SEC analysis

Solvent and temperature selection is driven by polymer solubility. A polymer that is not fully dissolved cannot be characterized.

For polyolefins, complete dissolution requires high-boiling-point solvents – primarily TCB or o-DCB – at elevated temperatures of around 150–160 °C, with dissolution times between 1–3 hours depending on the type of resin.

Read more: Practical considerations in Gel Permeation Chromatography analysis of polyolefins

Finding the right combination between dissolution time and temperature is fundamental to minimizing thermal degradation. The key is to find the minimum temperature-time mix that achieves full dissolution. Checking that the polymer is fully dissolved can be done by visual inspection or – in a SEC system with calibrated concentration detection – via the mass recovery value (%MR). This value, besides providing additional sample information, will indicate whether the sample has been fully dissolved without having to rely on manual methods based on solution appearance.

Sample concentration is another variable that can influence dissolution time. Typical sample concentration for SEC analysis is 0.5–2.0 mg/mL, but this can be adjusted downward for samples that are difficult to dissolve. Reducing sample concentration is also good practice when analyzing ultra-high-MW samples to avoid obstructions.

Injecting a sample too concentrated can also cause column overloading, compressing the early-eluting (high-MW) portion of the chromatogram, artificially narrowing the apparent distribution and shifting Mw downward.

Shear degradation is a phenomenon in which high-MW polymer chains are degraded by shear forces, resulting in chain scission. In an SEC chromatogram this leads to a shift of the polymer peak towards greater elution volumes (lower molar masses). Shear degradation can happen during dissolution (caused by mechanical stirring), during polymer filtration (caused by low-porosity frit-filters), and during analysis (caused by small particle-sized columns or narrow tubing).

Filtration is required to protect the column from fillers, gels, or undissolved material. At high temperature, in-line filtration integrated in the SEC system is preferred over manual filtration, both to avoid sample precipitation and to minimize handling. Filtration of very high-MW samples can remove the highest-MW fraction. When doing this manually, always inspect the filtrate. When filtration is integrated in the GPC system, pressure sensors and mass recovery values can be indicators of material loss.

Related Poster: Improvements in Sample Preparation of Polyolefins to prevent Polymer Degradation prior to GPC/SEC and CEF analysis

Related Poster: GPC analysis at different flow rates to overcome shear degradation

Key Operating Parameters in a SEC analysis

Flow rate:  0.5–1.0 mL/min for standard analytical columns. Higher flow rates reduce resolution, increase back-pressure, and raise the risk of shear degradation for high-MW samples. Flow rate precision directly affects the results. A shift in actual flow rate relative to the calibrated value translates to a shift in elution volume and therefore an error in apparent molecular weight. Validate flow rate using a flow rate marker: a small molecule (heptane is common in organic systems) injected at known concentration. The marker’s elution time should be reproducible to within ±0.3% across runs.

Temperature: Must be uniform across the entire flow path – column compartment, injector, and detectors. Temperature gradients cause baseline drift and retention time irreproducibility. For high-temperature systems, allow sufficient thermal equilibration time before injecting samples, often 30–60 minutes after reaching set temperature.

Injection volume:  Typically 100–200 µL for standard analytical columns. Injection volume affects peak shape and column loading. Too large a volume causes overloading, which broadens peaks (band broadening (BB)) and skews molecular weight averages. The optimum injection volume depends on column capacity and sample concentration, and should be determined during method development rather than assumed from general guidelines.

GPC/SEC columns: As discussed in the Column section, the type of column should be chosen according to the molecular weight range of the polymer and the operating conditions. The number of columns used will impact MWD resolution and analysis time: the more columns, the higher the resolution and the longer the analysis time.

• Replacing the columns: Some indicators will determine when it’s time to change the columns
  • Plate count is significantly lower compared to its initial value.
  • Pressure increases at the time of sample injection.
  • Flow rate marker changes shape in the chromatogram (broadening, front/rear tailing).

Baseline and Integration Limits for consistent SEC results

The integration limits define what molecular weight range is included in the reported averages.

Poorly placed baseline and integration limits can inflate or truncate both the low-MW tail directly distorting Mn (highly sensitive to low-MW content) and the high-MW tail, distorting M_z (sensitive to the highest-MW species). Baseline placement decisions should be documented and applied consistently across a sample set.

Method Validation: Building a Working SEC Methodology in the Laboratory

Method validation is the systematic process of demonstrating that the complete analytical system – instrument, column set, mobile phase, sample preparation protocol, detectors, and data processing parameters – delivers accurate, reproducible results for the specific materials being characterized. The subsections below outline the key steps.

System suitability and instrument qualification 

Before any sample data can be trusted, the instrument must be shown to perform within its specification. System suitability checks should be run at the start of each analytical session and should include: pump flow rate accuracy and precision (verified using the flow rate marker), column back-pressure within the expected range for the operating temperature and flow rate, baseline stability after thermal equilibration, and detector noise and drift within defined limits. These are not one-time qualification checks, they are routine operational controls that catch instrument drift, column degradation, and solvent quality issues before they corrupt sample data.

Calibration verification 

The calibration must be verified against reference materials before any sample series is run. For conventional calibration, this means confirming that a reference standard of known distribution returns the expected M_w, M_n, and Ð within defined acceptance criteria. For universal calibration, the Mark-Houwink parameters for both the standard and the analyte polymer must be confirmed to be appropriate for the operating conditions (solvent and temperature). For LS-based methods, the dn/dc value must be validated for the specific polymer-solvent-temperature combination in use.

Calibration should be repeated whenever columns are replaced, mobile phase batches change, or the system is moved or serviced. The frequency of routine recalibration checks should be defined as part of the method and justified based on demonstrated system stability.

Precision and reproducibility

Method precision is established by repeat injections of the same sample under identical conditions (repeatability) and by measurements across different days, operators, or instruments (intermediate precision). For most polymer characterization applications, Mw repeatability of 2-5% RSD and Mn repeatability of 2-7% RSD across a minimum of five injections are reasonable targets, though acceptance criteria should be defined based on the end-use requirements of the data. Mz and Mz+1 are more sensitive to integration limits and high-MW tail noise and will show higher variability; this should be documented and understood before these averages are used in specifications.

Artifact detection and system checks 

A validated method includes defined checks for the most common sources of artifacts: column overloading (checked by verifying that MW averages are independent of injection concentration over the working range), shear degradation (checked by comparing results at different flow rates for high-MW samples), and baseline placement reproducibility (checked by applying consistent integration limits and documenting the sensitivity of reported averages to small changes in those limits).

Documentation and traceability 

A working methodology is only as reliable as its documentation. The method record should specify, at minimum: solvent, flow rate, temperature, column set and serial numbers, injection volume and concentration range, calibration standards and their lot numbers, detector configuration and signal assignments, integration parameters, and acceptance criteria for all system suitability checks. Any deviation from these parameters should be recorded. Molecular weight data are only reproducible when the full analytical context is reproducible.

Related article: High Temperature GPC (HT-GPC): A Complete Guide

Click here to read more