Density is a decisive specification parameter for high-density polyethylene (HDPE), not only because ASTM D4976-21 and ISO 17855-1 classify polyethylene by density at 0.941 g/cm³, but also because density is a rapid surrogate for crystalline fraction and an indirect predictor of stiffness, shrinkage, creep resistance, and barrier behavior. When the same lot is tested in two laboratories, a difference of 0.002 g/cm³ can trigger acceptance or rejection of a purchase lot, even though the discrepancy may originate in specimen preparation, thermal history, calibration, environmental control, or formulation rather than in the resin itself. The two dominant method families for HDPE density are displacement according to ASTM D792-20 and ISO 1183-1:2019, and density gradient column according to ASTM D1505-18 and ISO 1183-2:2019. Under identical conditioning and calibration, these methods can agree closely; in routine commercial practice, they frequently do not because the controlling variables are not standardized to the same tolerance.

What Happens When Two Laboratories Mold the Same Pellet with Different Cooling Rates?

Measured density in semicrystalline HDPE is governed primarily by the volume fraction and thickness of crystalline lamellae. Commercial HDPE grades commonly fall between 0.945 and 0.965 g/cm³, but the pellet density is only the starting point. Compression molding according to ASTM D4703-16 or ISO 293 can produce a plaque whose density differs from the pellet because the melt is cooled from 160 °C to 180 °C under controlled heat removal. Two presses set to the same platen temperature may still cool at different rates if one has a chilled water inlet at 15 °C and the other at 25 °C, or if one press closes with a 10 s longer transfer time. The resulting density shift is not a laboratory mistake but a real difference in crystalline order. Using published two-phase densities of 1.000 g/cm³ for the crystalline phase and 0.855 g/cm³ for the amorphous phase, a density difference of 0.001 g/cm³ corresponds to approximately 0.7 percent crystallinity change. On an injection molding line with a clamp force of 1,800 kN, cavity temperature setpoints of 40 °C and 80 °C can yield different skin-core crystallinity profiles; the 40 °C mold freezes the surface earlier and suppresses spherulite growth, while the 80 °C mold permits slower cooling and higher crystalline order in the same wall section. Pellet production is also relevant: a 25 mm twin-screw extruder with an L/D of 40 and a strand bath temperature that fluctuates by 10 °C can create pellet density variation before the laboratory receives the sample. Published data for exact density shifts on individual molding machines is limited because cooling rate is non-uniform and depends on part thickness, mold steel, and coolant turbulence.

Density gradient columns built to ASTM D1505-18 and ISO 1183-2:2019 are not self-calibrating; their accuracy depends on the stability of the liquid column, the certified density of glass floats, and the temperature uniformity of the jacket. For HDPE, the working liquid is frequently a water-isopropanol or water-ethanol mixture, chosen because its density span can bracket 0.941 g/cm³ to 0.965 g/cm³. The top of the column can lose the more volatile alcohol during an extended shift, causing a slow upward drift in float positions and therefore a biased reading if the laboratory does not recalibrate. Temperature stratification of 0.2 °C across the column height can also shift the density of the liquid mixture enough to move the apparent equilibrium height, because organic-water mixtures have larger thermal expansion than pure water near room temperature. A bubble of 0.5 mm diameter attached to the specimen during immersion adds buoyant volume and can move the specimen several millimetres up the column, which appears as a density change of multiple units in the third decimal place. The method requires the specimen to be completely wetted and free of trapped air; operators who wipe a specimen with alcohol before insertion can alter surface tension and create local density disturbances. Calibration floats supplied with certificates at a fixed temperature must be equilibrated to the column temperature, not used directly from a cold or warm storage cabinet. A freshly filled column also requires adequate time for diffusion to create a linear gradient; if a laboratory calibrates before the gradient stabilizes, the float positions change during the measurement shift and the density readout is not traceable to the calibrated floats. The column must be covered to reduce evaporation and protected from vibration, because mechanical disturbances create convection cells that destroy the linear profile. Some laboratories use isopropanol-based gradients; others use ethanol-based gradients. Although both can bracket HDPE density, the surface tension and evaporation rate differ enough that a calibration valid for one system is not transferable to the other. Two laboratories can operate the same method designation and still differ by more than 0.0005 g/cm³ when one uses fresh solvent, fresh floats, and stable jackets, and the other uses an aged column that has not been checked for linearity.

Immersion, Pyknometer, and Density Gradient Column Divergence

ASTM D792-20 and ISO 1183-1:2019 determine density by Archimedes displacement. Because HDPE has a density below water, a sinker is often required; the specimen is attached to or placed below a sinker of calibrated volume. The method is sensitive to balance repeatability, sinker volume correction, liquid temperature, and the complete removal of bubbles. A balance with 0.1 mg resolution is adequate only if the specimen mass is large enough to keep the combined uncertainty below the required tolerance. The immersion liquid must be measured for density at test temperature; using a handbook value for water without correcting for 23 °C and dissolved air introduces a small but avoidable bias. Water wetting of HDPE is poor because the surface energy is low. Without a wetting agent, bubbles adhere strongly to the surface, and the specimen must be degassed by brushing or vacuum. ISO 1183-1:2019 permits a wetting agent or a different liquid, but the liquid density must be redetermined after addition. The sinker volume must be corrected for thermal expansion; a stainless steel sinker of 10 cm³ volume changes volume negligibly compared with the required precision, but the suspension wire entering the liquid surface introduces meniscus force and should be thin and clean. If the same wire is wiped between weighings, the liquid level and wetting state change and repeatability degrades. In contrast, ASTM D1505-18 and ISO 1183-2:2019 do not require a volume measurement of the specimen; the specimen floats at the height where its density matches the column. For injection molded plaques, the immersion method tests the entire cross-section and averages the low-density skin and higher-density core, while a density gradient specimen cut from the surface may be dominated by skin material. A plaque with internal shrinkage voids behaves differently in the two methods: in displacement, sealed internal voids that do not fill with water add buoyant volume and lower the apparent density; in a density gradient column, the same void may also alter buoyancy if the specimen is not internally wetted. Laboratories that cut small specimens with a hot knife or razor can create stress whitening near the cut surface, which changes local density and wetting. The result is that the two method families can produce a spread of 0.002 g/cm³ to 0.004 g/cm³ on the same plaque when the skin-core difference and void population are large, even if each instrument is within its own calibration tolerance.

The following method-control comparison identifies the variables most likely to create a discrepancy when HDPE is tested in different laboratories.

When Carbon Black, Filler, or Moisture Alters Apparent Density

Formulation changes density more directly than thermal history. Carbon black used in HDPE pipe and film has a density near 1.8 g/cm³ to 2.1 g/cm³, and talc or calcium carbonate fillers are in the range of 2.6 g/cm³ to 2.8 g/cm³. A 2 wt% carbon black addition to a 0.950 g/cm³ natural resin increases the compound density on the order of 0.008 g/cm³ to 0.010 g/cm³ when the carbon black is well dispersed; the actual value depends on the grade, aggregate porosity, and whether the laboratory tests the final compound or the natural base resin. If one laboratory reports density for an unfilled natural pellet and the other reports a black pigmented pellet, the difference is not an interlaboratory error but a material difference embedded in the sample chain. Moisture uptake in HDPE itself is low, but carbon black and mineral fillers can adsorb water; conditioning according to ASTM D618-21 or ISO 291 at 23 °C and 50% relative humidity establishes a reproducible equilibrium state. A specimen pulled from a high-humidity warehouse and tested immediately may carry surface moisture or condensation, which changes wetting behavior in a density gradient column and can add mass in an immersion test. Temperature correction is a separate issue: because HDPE has a volumetric expansion coefficient near 2 × 10⁻⁴ K⁻¹, a difference of 5 °C between laboratories can shift the measured density by approximately 0.001 g/cm³. This is comparable to the crystallinity-driven spread described earlier and is corrected only when the test temperature and conditioning environment are recorded.

When an incoming HDPE lot is held for release and two certificates of analysis differ by 0.002 g/cm³, the first review should not compare only the reported averages but also the metadata for each measurement: specimen preparation standard, conditioning standard, method designation, liquid temperature, column calibration date, and whether the specimen was natural or pigmented. A compression-molded plaque conditioned for 40 h at 23 °C according to ASTM D618-21 is not technically equivalent to a pellet tested by ASTM D1505-18 minutes after arrival. Similarly, an ISO 1183-1:2019 immersion result at 23.0 °C and a density gradient result at the same nominal temperature can differ if one laboratory used a wetting agent and the other did not. Resolution of density conflicts requires isolating these preparative variables before assigning a material nonconformity.