By Esther López, Laura Santonja-Blasco, Felipe Cordero, Alberto Ortín (Polymer Char)

Why Is My HDPE Failing In Production When MFI and Density Look Fine?:

When the Polymer Datasheet doesn’t tell the whole story.

1. Introduction

HDPE bottle handle production is generally considered a highly stable process, particularly when the same resin supplier, grade, and operating conditions are maintained over long production periods. However, unexpected processing failures can still appear after the introduction of a new material lot, even when all incoming quality control parameters, including MFI, density, and average molecular weight, remain within specification. In the present case, a newly received HDPE lot generated recurrent cracking during drop-impact testing of the handles, while the previous lot processed under identical conditions showed no such behavior.

Although the failures were not observed in every produced part, the reject rate increased sufficiently to compromise process reliability and trigger a detailed material investigation.

Processing conditions were adjusted in an attempt to eliminate the defects, including modifications in melt temperature, but the failures persisted and additional issues such as bubble formation began to appear. Returning to the original processing conditions did not resolve the problem, and reject rates remained elevated. According to the supplier specifications, both lots were fully compliant: MFI values differed only marginally (6.7 vs. 6.9 g/10 min), remaining well within the accepted tolerance range. Conventional QC analysis therefore suggested equivalent materials, despite the fact that one lot consistently failed the 8 kg drop-impact test while the other showed stable mechanical performance.

This type of discrepancy is becoming increasingly common in polyolefin processing, particularly when materials from different production lots or alternative suppliers are introduced into established manufacturing operations.

In recent years, polyolefin converters have increasingly faced variability associated with supplier diversification, raw material shortages, and the qualification of alternative production sites. Geopolitical disruptions, trade realignments, and raw material shortages have forced large polyolefin converters to qualify new suppliers, or accept material from alternative production sites, in compressed timelines. A supplier relationship built over a decade, with its implicit quality consistency, gets replaced overnight by a new source that is ‘certified’ and ‘within specification’. The specification, however, was never designed to catch differences in molecular architecture. It was designed for a stable world where your trusted supplier’s material behaved predictably every time.

As a result, materials that formally satisfy conventional specifications do not necessarily guarantee identical processing behavior or final performance.

Conventional quality control approaches. The standard incoming QC toolkit for HDPE – MFI, density, and spectroscopic screening (FTIR, Raman, NIR) – seems, on the surface, like a reasonable answer. These techniques are fast, inexpensive, and widely implemented. MFI provides a single-point indication of melt flow behavior, while density is primarily associated with crystallinity and average short-chain branching content. FTIR, Raman, and NIR can detect contaminants, confirm polymer identity, and flag degradation products. For decades, this combination has been considered sufficient for routine lot release.

Analytical limitations of conventional methods. The main limitation of these conventional techniques is that they provide only average or indirect information about the material structure. Every one of these methods is structurally blind to molecular architecture, and architecture is precisely what governs real-world processing and performance.

MFI captures flow at a single, relatively low shear condition, while industrial processing can impose much higher shear rates [1,2]. Two materials with identical MFI values can differ dramatically in their molecular weight distribution (MWD), long-chain branching (LCB) content, and high-molecular-weight tail, all of which govern melt elasticity, strain hardening, and impact behavior. Long-Chain Branching in polyethylene is one of the key microstructures controlling processing and final properties [3], and MFI cannot see it. MFI is not sufficiently sensitive to detect variations in long-chain branching, molecular topology, or subtle changes in high-molecular-weight fractions that strongly influence melt behavior.

Density tells you about crystallinity and bulk Short-Chain Branching content. What it cannot tell you is anything about the topology of your high-molecular-weight chains, the ones that actually carry most of the load under impact. Two materials can have the same bulk density and completely different behavior under a drop-weight test.

FTIR, Raman, and NIR are genuine tools for compositional screening. They tell you what chemical groups are present. They will not tell you whether those chains are branched, how branching is distributed across the molecular weight range, or whether your melt will strain-harden appropriately during blow molding. Composition and architecture are not the same thing.

Triple-detection GPC as an advanced structural characterization approach. This is where Gel Permeation Chromatography with Triple Detection (TD GPC) comes in, and it is not a marginal improvement over conventional GPC. Unlike conventional single-detector GPC approaches, triple-detection systems provide simultaneous information on molecular weight, intrinsic viscosity, chain conformation, and branching architecture within a single analysis.

Polymer Char’s Triple detection GPC-IR system combines three complementary detectors in a single automated analysis:

• Infrared (IR) concentration detector: measures the full molecular weight distribution and chemical composition, including Short-Chain Branching (SCB/1000TC), across every elution fraction simultaneously [4].
• Differential viscometer: measures intrinsic viscosity across the MWD. Deviations from the Mark-Houwink linear reference behavior are direct evidence of Long-Chain Branching, something MFI cannot approach [3,5].
• Multi-Angle Light Scattering (MALS): measures absolute molecular weight without calibration standards, and provides radius of gyration for direct chain conformation analysis. Combined with viscometry, it delivers unambiguous LCBf quantification [5,6].

Together, these three detectors provide a complete molecular fingerprint. Not a proxy. Not a single-point flow number. A quantitative, reproducible, full-spectrum description of what your material is at the molecular level.

The objective of this work is straightforward: to show, using a real industrial case, that TD-GPC catches what conventional methods miss, and that what they miss is exactly what causes parts to fail.

2. Methodology

All measurements were performed in a Polymer Char high-temperature GPC-IR system configured with triple detection (TD-GPC). Analyses ran in 1,2,4-trichlorobenzene (TCB) at 160 °C under stabilized conditions, following standard high-temperature GPC protocols for polyolefins (ISO 16014-4, ASTM D6474). The three detection channels operated simultaneously on every injection:

• IR detector: concentration signal for conventional and absolute MWD calculation; simultaneous chemical composition measurement (CH₃/1000TC, SCB/1000TC) across the full molecular weight range.
• Differential viscometer (4-capillary bridge): Intrinsic Viscosity (IV) distribution across the MWD; input for universal calibration and Mark-Houwink analysis; primary channel for LCB detection via deviation from linear reference behavior.
• MALS detector: absolute molecular weight determination independent of calibration standards; radius of gyration (Rg) for chain conformation analysis; LCB frequency (LCBf) quantification via the Zimm-Stockmayer approach [5,6].

This configuration delivers a conventional, a universal, and an absolute reading from a single experiment, eliminating the need for multiple orthogonal measurements and providing internally consistent data on every sample.

3. Results and Discussion

3.1 Two lots, two identical datasheets, one problem

Two virgin HDPE samples from the same supplier and grade were submitted for analysis following an unexplained production failure. Both materials were intended for blown containers with injection-molded handles. One lot (HDPE 1) was associated with systematic handle failures in a standard drop-weight test, 8 kg mass, 15 cm drop height, simulating the weight of product inside the container. The other (HDPE 2) passed consistently.

3.2 Conventional data: the illusion of equivalence

Table 1 summarizes the conventional characterization results. The numbers are, frankly, unhelpful. MFI values of 6.7 and 6.9 g/10 min fall well within the same specification window. Average molecular weights from the conventional GPC-IR channel (Mw ≈ 76,000 g/mol for both) are nearly identical. Dispersity, intrinsic viscosity, and SCB values all fall within a range that would not trigger any standard flag. From a conventional QC standpoint, these materials are indistinguishable.

3.3 IR detector: ruling out chemistry as the culprit

Both resins show a broad molecular weight distribution consistent with Ziegler-Natta catalysis, and the MWD profiles are remarkably similar: no obvious shoulder, no anomalous high- or low-MW population that would immediately raise a flag. The IR detector picked up a subtle difference in SCB content: 0.9 vs. 1.8 per 1000 total carbons. The difference is real, but it falls within the expected range of lot-to-lot variation for the same grade and does not explain a brittle-vs.-ductile failure mode under impact.

What the IR result does here is close a door: the performance difference is not compositional. Both materials carry similar chemistry chain. The SCB distribution profile, methyl groups across the MWD, confirms this. The problem lies elsewhere, which means the investigation should focus on topology, not composition. In practice, this is a genuinely useful outcome: ruling out chemistry early narrows the diagnostic space considerably and points directly toward architecture as the suspect [4]

Figure 1. MW distribution by IR channel and SCB/1000TC profiles for HDPE 1 and HDPE 2, with HDPE 1 shown in green and HDPE 2 in blue.

3.4 Viscometry: where the two materials stop looking alike

Significant differences between both materials became evident after analysis of the viscometer response.

At equivalent elution volumes, HDPE 1 shows measurable deviation from the Mark-Houwink linear reference behavior for polyethylene. HDPE 2 tracks the linear reference closely. This deviation is the molecular signature of long-chain branching: branched chains occupy a more compact hydrodynamic volume than their linear counterparts at the same molecular weight, so they elute later in GPC and their intrinsic viscosity falls below the linear Mark-Houwink baseline [3,5].

The viscometer result does not yet give us a number, but it tells us clearly that these two materials carry different chain topologies. It is the point in the analysis where two lots that looked identical on the conventional GPC-IR channel visibly diverge, and the mechanism behind the production failure starts to take shape. Literature consistently confirms that even small differences in LCB frequency produce measurable rheological divergence in polyethylene systems [3,6].

Figure 2. Mark-Houwink plots for HDPE 1 and HDPE 2 – deviation from linear reference visible in HDPE 1.

3.5 Light scattering: putting a number on the architecture

MALS delivers the definitive answer. Because it measures absolute molecular weight directly, without calibration standards, it can calculate the branching contraction factor g’ and, from that, the Long-Chain branching frequency (LCBf) at each point across the molecular weight distribution.

The calculated LCBf values revealed substantial differences between both materials:

A tenfold difference in LCBf. From two materials that showed essentially the same MFI.

The implications are direct. LCB in polyethylene has a well-documented effect on viscoelastic behavior: even very low concentrations of Long-Chain Branches significantly increase melt strength and alter the relaxation spectrum [3]. HDPE 1, with its substantially higher LCBf, carries a more entangled chain topology that increases melt strength during processing, which sounds useful, but under the stress concentration at the handle weld line during impact, it correlates with reduced energy dissipation and earlier brittle failure. HDPE 2, with its more linear architecture, distributes the impact load more uniformly and holds [6,7].

From a structural standpoint, these are not the same material. The shared MFI value is a coincidence of bulk flow behavior, not evidence of molecular equivalence.

Despite sharing the same supplier origin, nominal grade, and nearly identical MFI values, the two HDPE materials exhibited an order-of-magnitude difference in long-chain branching frequency

Figure 3. Absolute MW distribution from MALS for HDPE 1 and HDPE 2.

4. Conclusions

Conventional analytical methods failed to identify the structural differences between both materials. MFI values and density measurements remained within specification, while infrared spectroscopy confirmed similar chemical composition in both samples. However, none of these techniques are capable of providing detailed information about molecular topology, branching architecture, or the structural heterogeneity that ultimately governs processing behavior and mechanical performance.

TD-GPC caught it in a single run: a tenfold difference in Long-Chain Branching frequency that conventional methods are structurally incapable of detecting.

For polymer converters navigating supplier diversification which, in 2025–2026, describes most of the industry, this gap is not academic. A lot that passes your incoming QC but carries ten times the LCB of your reference material will not behave like your reference material on the line. These structural differences may remain undetected during routine QC analysis and only become evident during processing or mechanical performance evaluation.

Polymer Char’s GPC-IR with triple detection provides:

• Full molecular weight distribution, not just an average
• Direct quantification of Long-chain branching, the single most impactful architectural variable for HDPE melt behavior and impact performance
• Absolute molecular weight, calibration-independent, from every single injection
• Chemical composition distribution across the full MWD in the same run

This is not a luxury measurement reserved for R&D. For anyone qualifying new polyolefin suppliers under real-world supply chain pressure, it is the only tool that tells you what you actually need to know, before you commit to production.

5. Acknowledgements

The authors thank the industrial partner for providing the HDPE samples and for granting permission to publish these results. Their willingness to share a real production challenge made it possible to demonstrate this approach under conditions that matter.

6. References

[1] ASTM International. Standard Test Method for Melt Flow Rates of Thermoplastics by Extrusion Plastometer (ASTM D1238). ASTM International, West Conshohocken, PA.

[2] ISO 1133-1:2011. Plastics — Determination of the melt mass-flow rate (MFR) and melt volume-flow rate (MVR) of thermoplastics.

[3] Subbaiah, Y., Shanmugam, S., & Hau, Q.E. (2024). Influence of high temperature GPC/SEC detection modes on the accurate molar mass estimation of slurry phase multimodal HDPE reactor materials. Journal of Separation Science, 47(2), 2300748.

[4] Polymer Char. (2019). Chemical Composition Analysis of Polyolefins by Multiple Detection GPC-IR5. Application Note. polymerchar.com

[5] Ortin, A., Monrabal, B., & Montesinos, J. (2013). Triple detection methods for HT-GPC analysis of polyolefins. Macromolecular Symposia, 330, 63–80.

[6] Wang, W.J., et al. (2004). Triple-detector GPC characterization and processing behavior of long-chain-branched polyethylene. Polymer, 45, 6495–6505.

[7] Pasch, H., & Eselem Bungu, P. (2023). Advanced characterization of polyolefins by multidimensional liquid chromatography. European Polymer Journal, 200, 112491.

[8] Polymer Char. (n.d.). How to Choose the Most Appropriate Detectors for GPC/SEC. polymerchar.com